PREFACE

In pursuance of the press restructuring approved by the Prime Minister and the

proposal made by the Ministry of Education and Training, the Ministry of Culture and

Information issued the Decision No 519/GP-BVHTT dated November 25th, 2002,

licensing the publication of the University of Danang journal entitled “Science and

Technology”.

The Press Bureau under the Ministry of Culture and Information issued the official

dispatch No 816/BC dated August 10th, 2006, permitting the University of Danang to

publish the “Science and Technology” every two months instead of every three months.

The National Scientific and Technological Information Centre under the Ministry

of Science and Technology issued the official dispatch No 44/TTKHCN-ISSN dated

February 6th, 2007, agreeing to grant an international standardized code, namely ISSN

1859-1531 to the UD journal “Science and Technology”.

The Press Bureau under the Ministry of Information and Communications issued

the official dispatch No 210/CBC of March 5th, 2008, allowing the UD journal “Science

and Technology” to be published in English and French in addition to the Vietnamese

publication.

On September 15th, 2011, the Ministry of Information and Communications issued

the Decision No 1487 /GP-BVHTT, granting an additional licence which permits the UD

journal “Science and Technology” to increase the publication period frequency from

every 2 months to every month and to raise the number of pages from 80 to 150 pages.

On January 7th, 2016, the Ministry of Information and Communications issued

Decision No 07 /GP-BVHTT, granting a print licence which permits the UD journal

“Science and Technology” to operate with a publication frequency of 15 issues per year

(of which 3 issues are published in English).

The UD journal “Science and Technology” has been established in view of

announcing and introducing scientific researches in the fields of teaching and training,

offering information on the domestic and foreign scientific research results that serve the

education in colleges and universities and propagating the State and Party’s guidelines

and policies in education and training as well as in scientific and technological researches.

Over the past 40 years, the UD journal “Science and Technology” has been

published as the continuity and development of the scientific information journals and

conference proceedings published at the University of Danang and its member

universities.

The journal’s editorial staff wishes to receive many-sided contributions of the

scientists and educators from within the university and from other colleges and

universities inside and outside the country so that the journal’s quality will be more and

more improved in the future.

Journal’s Editorial Staff

CONTENTS ISSN 1859-1531 – THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017

ENGINEERING AND TECHNOLOGY

Research on characteristics of drilling tool intergrating damping system using experimental modal analysis – EMA Ho Duong Dong 1

Dynamic behaviors of a single flexible link manipulator under different driving rules Bien Xuan Duong, My Anh Chu, Lac Van Duong 6

Assessment of the use of raw fly ash with high loss on ignition in concrete Ngo Si Huy, Huynh Trong Phuoc, Le Thi Thanh Tam 11

Analysis and treatment of trimethoprim residues in water by γ-irradiation Nam D. Le, Thang M. Ngo 15

An experimental study on properties of high-performance concrete using recycled aggregates Huynh Trong Phuoc, Nguyen Tien Dung, Ngo Si Huy, Vo Duy Hai 19

Design and control of an active prosthetic leg Hoang Trung Ngo, Thien Duc Ngo, Danh Ngoc Nguyen, Hoai Nam Le 24

Bi-level optimization model based evaluation of wholesale electricity price intervals considering the wind power uncertainty and elastic demand Pham Nang Van, Nguyen Duy Cung, Nguyen Duc Huy 30

An iterative subproblem method for thin shell finite element magnetic models Dang Quoc Vuong 35

NATURAL SCIENCES

The dependence of an acoustomagnetoelectic field on the temperature in a cylindrical quantum wire ALGaAs/GaAs Nguyen Van Hieu, Nguyen Quang Bau 40

SOCIAL SCIENCES

A study on syntactic, semantic and pragmatic features of hyperbolic expressions in CEO news items in English versus Vietnamese Dinh Thi Minh Hien, Chu Thi Mui 43

The impact of competition on credit risk: the case of Vietnam commercial banks Phan Tran Minh Hung, Phan Nguyen Bao Quynh 48

Insights into the implementation of English medium instruction at Vietnamese Universities: exploring the learning experiences of business students Le Thi Thuy Nhung 53

Work-integrated learning curriculum: an effective design to enhance graduate employment rate in Vietnam Quy H. Nguyen 58

Ideational meaning in IELTS sample essay introductions in light of systemic functional grammar Nguyen Thi Trung, Nguyen Thi Quynh Hoa 63

HUMAN SCIENCES

Correlation between self, peer and teacher assessment: a case study in a translation course Ho Si Thang Kiet 68

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 1

RESEARCH ON CHARACTERISTICS OF DRILLING TOOL INTERGRATING

DAMPING SYSTEM USING EXPERIMENTAL MODAL ANALYSIS - EMA

Ho Duong Dong

University of Science and Technology – The University of Danang; hddong@dut.udn.vn

Abstract - In recent years, the control of tool vibration has emerged as a critical area of scientific development. This is because tool vibration creates sound-noise and unwanted quality surface. Mathematically, the observed tool vibration is a ratio of the force acting on the tool to its dynamic stiffness. Because the dynamic stiffness is proportional to the tool damping, the tool vibration could be controlled by enhancing the damping ability of the tooling system. In this paper, the drilling tool integrating damping system is tested with Experimental Modal Analysis (EMA). The natural frequencies, the damping ratios, and mode shapes are detected. Through modal analysis testing the new damped drilling tool shows better damping ratio than the conventional tool. These tools are expected to have better performance than the conventional drilling tools as well.

Key words - cutting vibration; drilling tool; damping system; natural frequency; experimental modal analysis.

1. Introduction

In recent years cutting tools have continuously

developed to meet the great demands of improving work

accuracy and productivity. The cutting tools consist of

cutting inserts, attachment devices for the cutting inserts, the

proper tooling which has solid structure, and tool-holder.

The dynamic machining process can be represented as

a closed looped system comprising an elastic structure and

the metal cutting process. The elastic structure includes the

machine tool, the cutting tool and the work-holding fixture

whereas cutting process is defined by variables such as the

cutting tool geometry, cutting parameters.

Figure 1. The dynamic machining process

In Figure 1, F(t) is the instantaneous cutting force, F0(t)

is the cutting force nominal value, x(t) is the relative

displacement between cutting tool and workpiece, ∆d(t) is

the total deviation of the relative displacement x(t). P(t) and

Pd(t) are disturbances such as tool wear, thermal dilation of

the elastic structure, variation of the rigidity of the elastic

structure during a machining process, variation of cutting

parameters, etc. [1].

The mechanism of self-excited vibrations or chatter is

explained in Figure 2. As the dynamic cutting process

forms a closed looped system, any disturbances in the

system (i.e. vibrations which affect cutting forces) are fed

back into the system. Over time, the vibration amplitude

increases; and when the vibration frequency equals to one

of natural frequencies of the system, chatter occurs. The

variation in cutting force is due to the changes in chip

thickness, depth of cut, and cutting speed. These

parameters’ variations depend on the stability of machine

tool-workpiece system. The cutting process is considered

unstable if growing variations are generated; then the

cutting tool may either oscillate with increasing amplitude

or monotonically recede from the equilibrium position

until nonlinear or limiting restraints appear [2].

Figure 2. The closed - loop diagram

The variation of the shear angle and cutting forces

results from the unstable chip formation conditions (due to

the deflection of tool spindle or fixture). When the system

rigidity is low, force vibrations can cause excessive

vibration and lead to machine vibration. And the machine

vibration can cause additional force fluctuations. When the

dynamic cutting force is out of phase with the

instantaneous relative movement between the tool and the

workpiece, this leads to the development of self-excited

vibration [2]. Because the vibration reproduces itself in

subsequent revolution, it is called regenerative chatter.

When the chatter occurs without undulation, it is called

non-regenerative chatter. However, this type of chatter is

uncommon and has not been widely studied.

Figure 3. The SDOF System

V

F2F1

Yi

Yi-1

Yi-2

x(t)

ck

Mass - m

f(t)

2 Ho Duong Dong

In order to have a physical insight into the dynamic

behavior of vibrating system, a Single Degree of Freedom

(SDOF) Mass - Spring - Damping system is analyzed as

shown in Figure 3. Even though cutting tools are composed

of several components, it is possible to stimulate its

behavior under the influence of dynamic load by

considering it as an SDOF system.

The mechanical model in Figure 3 can represent cutting

process in which k, m, and c are the modal stiffness, mass

and damping values, respectively.

The equation of motion for the SDOF system shown is:

𝑚�̈�(𝑡) + 𝑐�̇�(𝑡) + 𝑘𝑥(𝑡) = 𝑓(𝑡) (1)

Where �̈�(𝑡), �̇�(𝑡), 𝑥(𝑡) are the acceleration, velocity,

and displacement respectively.

The natural angular frequency of free un-damped

oscillations, ωn (rad/sec), and the damping ratio, ξ, are

𝜔𝑛 = √𝑘

𝑚 and ξ =

𝑐

𝑐0=

𝑐

2√𝑘𝑚

The damping ratio ξ is the ratio of the actual damping c

to the critical damping cc, which is the smallest value of c for

which the free damped motion is non-oscillatory, that is, the

level of damping which would just prevent vibration [2].

When a system is excited by a specific sinusoidal force:

f(t) = Fejωt

in which F is the force amplitude, 𝑗 = √−1, t is time,

and ω is the exciting frequency in rad/sec.

Modal analysis is performed using the Fourier

transform X(ω) of the displacement x(t):

𝑋(𝜔) = ∫ 𝑥(𝑡)𝑒−𝑗𝜔𝑡

+∞

−∞

𝑑𝑡

The Fourier transforms of the time derivative of a

function can be determined by multiplying the Fourier

transform of the function by jω:

∫ �̇�(𝑡)𝑒−𝑗𝜔𝑡𝑑𝑡

+∞

−∞

= 𝑗𝜔𝑋(𝜔)

∫ �̈�(𝑡)𝑒−𝑗𝜔𝑡𝑑𝑡

+∞

−∞

= −𝜔2𝑋(𝜔)

taking the Fourier transform of both sides of Equation (1):

(-ω2m + jωc + k)X(ω) = F(ω)

F(ω) is the Fourier transform of f(t). The steady-state

response of this system is given by:

X(ω) = G(ω)F(ω)

Where 𝐺(𝜔) =𝑋(𝜔)

𝐹(𝜔)=

1

−𝜔2𝑚+𝑗𝜔𝑐+𝑘

G(ω), the frequency response function (FRF) of the

system, is the ratio of the complex amplitude of the

displacement (which is a harmonic motion with frequency

ω) to the magnitude F of the forcing function [2]. FRFs are

complex functions, with real and imaginary components.

They may also be represented in terms of magnitude and

phase.

In the paper, EMA is carried out to find out the natural

frequency of the tool. Determining the natural frequency of

the structure is important because it helps us to predict the

possibilities of when the resonance occurs. EMA is first

calculating the FRFs of the structure by artificially exciting

it with an impact hammer. The output is fixed, it means the

impact position is unchanged; FRFs are measured for

multiple inputs to form a single row of the FRF matrix. The

obtained data is then used to predict the frequency,

damping and mode shape.

Modal Analysis is a process to evaluate a mechanical

structure in terms of its natural characteristics which are

frequency, damping and mode shapes, i.e., its dynamic

properties. Experimental modal analysis comprises a set of

experimentally-based procedures which lead to the

construction of a mathematical model that can be used to

describe the dynamic behavior of the test object. This model

can be used on a variety of useful applications including:

- Visualization of the models of vibration of the test

structure for the purpose of gaining physical insight and an

understanding of the often complex dynamic properties of

real structures;

- Comparison of the actual (measured) vibration

behavior of a real structure with corresponding parameters

predicted from a theoretical model;

- Predicting the effects of making modifications, or

predicting what modifications to introduce to change the

structure’s behavior;

- Predicting the behavior of structures formed by

coupling two or more components together;

- Detecting damage or other changes in the integrity of

a structure during its service life;

- Identification of unpredictable parameters such as

damping, dynamic friction effects, excitation forces from

unknown sources, etc.

The basic sequence of steps in a model test is:

- Measurement of the test structure’s vibration response

to a controlled and known excitation;

- Analysis of the resulting response functions to

identify the underlying modal properties (natural

frequencies and mode shapes) of the test structure;

- Construction of a mathematical model from these

modal properties, suitable for the intended application.

2. Effect of damping, stiffness and mass to the dynamic

behavior of the cutting tool

Damping is the mechanism that converts vibration

energy into other forms of energy such as heat. There are

two sources of damping: internal damping and external

damping [1].

External damping is referred as dampers, whereas

internal damping is damped material and joint damping [6].

Traditionally, one can sacrifice the productivity by

reducing depth of cut, feed, etc. to minimize vibration and

avoid chatter. However, it is possible to achieve the same

level of system stability or even increase it by redesigning

the machine tools with high damping [1].

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 3

Damping defines the ability of a tool to dissipate energy

or absorb energy from the process to reduce vibration, prevent

chatter and absorb energy from structural modes excited by

the servos. According to [1], increased damping results in:

• More rapid decay of unforced vibrations.

• Faster decay of freely propagating structure-borne waves.

• Reduced amplitudes at resonance of structures subject

to steady periodic or random excitation with accompanied

reduction in stresses and increased fatigue life.

• Reduced rate of build-up of vibrations at resonance.

• Reduced amplitudes of self-excited vibrations, in

which the vibrating structure accepts energy from an

external source.

All materials available in nature dissipate some energy,

as evidenced by the fact that the amplitude of free

vibrations diminishes with time. The energy dissipation is

either due to the stress-strain hysteresis loop vibration

(material damping) or the friction (non-material damping)

[6]. The use of damped materials, such as viscoelastic, in

machining tool is quite common.

The effect of damping is illustrated in Figure 4. Two

levels of damping, c1 and c2 (c1>c2) are applied while

mass and stiffness are kept constant. It is seen that the

increased damping level affects the response at resonance.

The amplitude at resonance decreases with increasing

damping level.

Figure 4. Efect of damping change

The effect of mass changes is illustrated in Figure 5.

Mass m1 and m2 (m2 < m1) are applied while damping and

stiffness are kept constant. It can be seen that less mass

offers a higher natural frequency. It means that less mass

can improve the ability of machine tool to response to high

frequency input. Therefore, in machine design reducing

mass is one of the important objectives. There are two

approaches to reduce mass. One is selecting lighter

materials with higher performance; the other is structural

optimal design in which mass is reduced and the

requirement is met at the same time [3]. Even though there

are many new materials with very high stiffness-to-mass

ratio available in the market, the use of these materials in

machine tool structure is limited due to cost factor.

The new damped drilling tools have almost the same mass

to conventional tool used in the experiments. Therefore, we

do not expect them to have higher natural frequency.

Figure 5. Effect of masss change

The effect of stiffness changes for the forced excitation

case is shown in the Figure 6. It is seen that the reduction

in resonance amplitude is directly proportional to the

increase in stiffness of the system.

Stiffness is the ratio between force and the deformation

induced by this force. The structural stiffness of a machine

tool is one of the main criteria in the design of a precision

machine tool. Generally speaking, both the static and

dynamic stiffness of a machine tool structure determine the

machine’s accuracy and productivity. Therefore, high

stiffness is required both statically and dynamically, each

affecting different aspects of the machining process [3].

Figure 6. Effect of stiffness change

High static stiffness is required to minimize the

displacement between cutting tools and workpieces during

cutting. The need for high dynamic stiffness results mainly

from reducing self-excited and/or forced vibration.

However, there is a trade-off since high static stiffness

results often in a low damping system and therefore

reduces dynamic stiffness. And if stiffness is enhanced by

“beefing up” dimensions of the tool and/or of the spindle

and its bearing then masses, natural frequencies, damping,

and overall costs may significantly change. If stiffness is

changed by tightening structural joints, damping is usually

declining, thus defeating the purpose of reducing chatter

and forced vibrations [4].

3. Existing Damped Tools

As discussed, the mechanism of damping is to convert

energy from a vibrating system into other forms such as heat.

Basically, we can divide damped tools into two categories

Frequency (Hz)

Co

mp

lian

ce

k = const.m = const.

damping = c1

damping = c2 (c2>c1)

Frequency (Hz)

Co

mp

lian

ce

k = const.c = const.

mass = m1

mass = m2 (m2<m1)

Frequency (Hz)

Co

mp

lian

ce

m = const.c = const.

stiffness = s1

stiffness = s2 (s2>s1)

4 Ho Duong Dong

based on their damping mechanisms, active and passive.

The principle of active damping is to analyze in real

time the signal emitted during machining, recognize

instability (chatter) and compensate for it [5]. The cutting

process emits signals. The signals are sensed by sensors to

recognize chatter. When chatter is detected, cutting speed

is modified after that to avoid chatter.

The principle of passive damping is to enhance the

damping ability of the tool without actively compensating

for the upcoming vibrations [5]. Passive damped tools were

the subject for research years ago. Researches show that

while high stiffness of a cutting tool is an important

condition for successful performance, there are some

exceptions where reduction of tool stiffness is shown to be

beneficial. One example of damped turning tool is shown

in Figure 7. The tool is clamped in a fixture that is designed

to enhance tool stiffness in all directions except the radial

direction to workpiece. There is an elastomeric material

between tool and tool-holder. It is a thin-layered rubber-

metal laminate with steel interleaves, having a very high

stiffness in the normal directions to thin elastomeric layers

(compression), while very low stiffness is along the layers

(shear) [4]. The tool stiffness is intentionally reduced while

its damping is enhanced The result is that no chatter and

acceptable surface finish without using steady rests.

Figure 7. Damped drilling tool

Another application for enhancement of chatter

resistance by reducing stiffness of machine system is

presented by Rivin Eugene [4]. He proposed an elasto-

damping clamping device for drilling as shown in Figure

7. The sleeve (3) holds the drill tool (2). Cutting forces

shall push sleeve (3) back to shoulder (10) of spindle (1).

The elastic element (9) and damping element (5) connect

sleeve (3) and spindle (1) together. It is stated that the

vibrations of the drill are significantly reduced when the

system 9-5 is properly tuned. Damped tools designed with

viscoelastic materials are popular nowadays. There are

three different techniques in using viscoelastic materials:

as free-layer dampers, as constrained-layer dampers, and in

tuned viscoelastic dampers.

The two new damped drilling tools are added to

damping rings. These rings are made of viscoelastic

composite material and they are glued together on the tool.

More information about the structure of damped drilling

tools can be found in [5]. The new tools therefore are

expected to have higher damping ability than the

conventional tool. Theoretically they would absorb more

vibratory energy from the cutting process and the machined

surface quality would be improved.

The damped drilling tools and the conventional tool are

shown in Figure 8. It is important to note that the damped

tool’s length is greater than that of the conventional one

(248mm compared to 205mm) and it needs a special tool

holder. Each tool is tested two times with two different

clamp devices, the conventional clamp ETP Hydro-fix

NBB- 42/50-62 and the viscoelastic composite material

interface clamp ETP Hydro-fix NBC- 42/50-62.

Figure 8. Tools and clamping devices used in EMA

4. Experimental Modal Analysis (EMA)

Experimental modal analysis is the process of

determining the natural frequencies, damping ratios, and

mode shapes of a linear, time invariant system by

experimental approach. The result is a mathematical model

built to explain the dynamic behavior of the test object.

Figure 9. Measurement points for the EMA

The drilling tool, tool holder and clamping device were

mounted in the turret of a lathe SWEDTURN 300. The

machine engine is turned on to stimulate the same working

environment as in the machining test (to be presented in the

next section). EMA is then carried out with the hitting point

fixed at one position (point#3) whereas the accelerometer

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 5

moves from point #1 to point #4. All points are illustrated

in Figure 9. The overhang of the tools is 160mm.

As mentioned above about the effect of damping change,

increasing damping results in more rapid decay of unforced

vibrations, reducing amplitudes at resonance, etc. In Figure

4, the amplitude of resonance is decreased with increasing

damping level. On the other hand, the curve damping c2

does not have a sharp peak as curve damping c1 (c2 > c1).

Figure 10. EMA results. Red curve: Damped tool.

Green curve: Conventional tool

Figure 11. Modal analysis result. In red, the conventional tool

in conventional clamp. In green, the conventional tool in

damped clamp. In magenta, the damped tool in damped clamp.

In blue, the damped tool in conventional clamp

Figure 10 shows the synthesized compliance functions

when the drilling tools are mounted in the conventional

clamping device. It can be seen from the graph that at the

first natural frequency, both tools have the sharp peak, the

red curve (damped tool) has a slightly greater width at the

base than the green curve (conventional tool). At the

second mode, the red curve’s width is much greater and its

peak is not as sharp as the green one. It proves that the

damped drilling tools have a significantly higher damping

ratio than the conventional tool.

The result at that moment is satisfactory enough

because the damped tool is produced accurately. As

expected, the conventional tool has higher static stiffness

than the damped tool.

It is interesting to make a comparison of the different

clamping systems. Figure 11 shows that when mounted in

the damped clamp both tools improve their damping ratio

significantly. The green curve is larger than the red curve

at the second mode, and the magenta curve does not have

a sharp peak as the blue one.

5. Conclusion

Through modal analysis testing, the new damped

drilling tool shows better damping ratio than the

conventional tool. The damped clamp helps the

conventional tool improve the damping ratio significantly.

These tools are expected to have better performance than

the conventional drilling tools.

REFERENCES

[1] Amir Rashid, “On passive and active control of machining system

dynamics”, Doctoral Thesis, KTH Stockholm, Sweden, 2005.

[2] David A. Stephenson, “Metal Cutting Theory and Practice”,

Second edition, 2006.

[3] S.S. Dimov and Bertrand Fillon, “2nd International Conference on

Multi-Material Micro Manufacture”, 2006, pp.30-32.

[4] Rivin Eugene l, “Tooling structure: Interface between cutting edge

and machine tool”, Wayne State University, Detroit, USA.

[5] Lorenzo Daghini, “Theoretical and Experimental Study of Tooling

Systems – Passive control of machining vibration”, Licentiate Thesis, KTH Stockholm, Sweden, 2008.

[6] Ahid D. Nashif, David I. Jones, and John P. Henderson, “Vibration

Damping”, John Wiley & Sons Inc., 1985, pp. 26-33.

[7] S. Ema, H. Fujii, and E. Marui, Chatter Vibration in Drilling”,

Journal of Engineering for Industry, 1988.

(The Board of Editors received the paper on 29/8/2017, its review was completed on 27/9/2017)

0.00 3200.00Hz

-190.00

-90.00

dB

(m/N

)

0.00

1.00

Am

plit

ude

0.00 3200.00Hz

-190.00

-90.00

dB

(m/N

)

0.00

1.00

Am

plit

ude

Damped tool

Conventional tool

6 Bien Xuan Duong, My Anh Chu, Lac Van Duong

DYNAMIC BEHAVIORS OF A SINGLE FLEXIBLE LINK MANIPULATOR

UNDER DIFFERENT DRIVING RULES

Bien Xuan Duong1, My Anh Chu1, Lac Van Duong2 1Military Technical Academy; xuanbien82@yahoo.com

2Hanoi University of Science and Technology; duonglacbk@gmail.com

Abstract - In this article, nonlinear dynamic modeling and investigation into the effects of different driving rules on dynamic behavior of a single flexible link manipulator are presented. Dynamic equations are derived from finite element method based on Lagrange approach. The effects of elastic displacements on robot motion make dynamic modeling and control become complicated by highly nonlinear characteristics. Besides, most studies have not clearly determined driving rule or the reason why the driving rule is selected. The results have important meanings to select reasonable driving rule for single flexible link manipulator with payload or not. Dynamic behaviors of the system are simulated like a specific example on Matlab/Simulink software.

Key words - Flexible link manipulator; nonlinear dynamic; elastic displacements; driving rules; dynamic behaviors.

1. Introduction

The elastic displacements engendered by flexible

manipulator’s motion are the main reason for position

accuracy, velocity, stability and others which are hard to

determine exactly because of nonlinear characteristics of

model. A number of researchers tried to study on the

behavior of the flexible manipulator by using linearization

methods assumed as small deflections, small hub angle.

Tokhi [1] presented theoretical and experimental

investigations into the dynamic characterization of a single

flexible link manipulator system with dynamic model

developed using finite element method. Gavriloiu [2]

proved that linear control theory is not generally suitable

for manipulator system with flexible link. On the other

hand, many researchers focused on building control system

with optimal parameters. Based on the dynamic model,

Wang [3, 4] proposed a technique to investigate the

maximum load carrying capacity of a robot manipulator

given a dynamic robot trajectory. Korayem [5] studied on

the dynamic load carrying capacity of a two-link flexible

manipulator using finite element method and Pontryagin’s

minimum principle. Gee and Lee [6] discussed the

modeling and nonlinear dynamic equations of a one-

flexible link manipulator based on finite element method

in Lagrange approach with varied payload. Reddy [7]

presented the nonlinear modeling of a one-link flexible

manipulator based on AMM and LS-DYNA. The results of

both are compared. The numerical simulation shows that

the LS-DYNA model gives the smooth hub-angle profile.

In fact, the dynamic equations of systematic complex under

effect of hub inertial, mass of rotors, payload, structural

damping, coriolis forces and driving rules. It is noteworthy

that effects of driving rules on dynamic behavior of flexible

link manipulator are not much mentioned. Most studies

have not clearly determined driving rule or the reason why

any driving rule is selected. The dynamic characteristics of

model are not general.

In this work, the nonlinear dynamic equations of one-

link flexible manipulators which motion on horizontal

plane are presented by using finite element method in

Lagrange approach. The factors could be taken into

account as hub inertia, payload, structural damping,

coriolis and centrifugal in the dynamic equations. The

effect of Coulomb friction and gravity are ignored because

manipulator motions on horizontal plane. Dynamic

behaviors of single flexible link manipulators are studied

under different driving rules.

2. Dynamic modeling

2.1. Finite element method

The finite element method is a numerical method which

can be used for the accurate solution of complex

engineering problems. The FEM has also been utilized to

describe the behaviour of flexible manipulators [6]. The

flexible link is considered as an assemblage of a finite

number of small elements. The elements are assumed

interconnected at certain points, known as nodes. For each

finite element, the scalar kinetic and potential energy

functions are formulated as functions of the generalized

coordinate. The dynamic model of a one-link flexible

manipulator is depicted in Figure 1.

Figure 1. Schematic diagram of a one-link flexible manipulator

In Figure 1, the symbol q is the angle of rotation at the

hub. The link is assumed as Euler-Bernoulli’s beam and is

divided into n elements. Each element has two nodes. Each

node of element j has 2 elastic displacement variables

which are the flexural ( 2 1 2 1,j ju u ) and the slope

displacements ( 2 2 2,j ju u ). The coordinate system 1 1 1X O Y

is attached to link. Symbols ,E I and are mass density,

Young’s modulus, inertial moment of area of link and total

length of link ( L ), thickness ( h ), width ( b ) and cross

sectional area ( A ) of link. The symbol is the applied

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 7

torque at the joint and motor inertial moment is hI . The

vector from O to a point on j element in the coordinate

system is 01 jr . Symbols Pm and PJ are the mass and

inertial moment of payload on the end point of link. The

material of link is assumed homogeneity. Link is divided

into n elements, the length of any element ( jl ). The lengths

of elements are equal because the cross-sectional areas of

links are constant along length. Total elastic displacements

of j element with ( , )j jx y coordinate on 11 1X O Y [1].

j( , ) ( ) ( , )j j j j jx t x x tw N Q (1)

Where vector of shape function ( )j jN x and ( , )j jQ x t is

vector of elastic displacement of element j in FEM theory.

1 2 3 4( ) ( ) ( ) ( ) ( )j j j j j jx x x x x N (2)

2 1 2 2 1 2 2( , ) ( ) ( ) ( ) ( )T

j j j j j jx t u t u t u t u t Q (3)

Shape functions of j element ( )i jx are given by [4]

2 3 2 3

1 22 3 2

3 2 2( ) 1 ; ( )

j j j j

j j jjj j j

x x x xx x x

ll l l (4)

2 3 3 2

3 42 3 2

3 2( ) ; ( )

j j j j

j jjj j j

x x x xx x

ll l l (5)

Position vector 1 jr on 1 1 1X O Y and vector 0 jr of j

element on XOY are expressed as

1j

( 1)

( , )

j j

jj

j l x

w x t

r (6)

10 0 1j jr T r (7)

Where transformation matrix 10T form 1 1 1X O Y to

XOY is given by [1]

10

cos sin

sin cos

q q

q q

T (8)

The elastic kinetic energy of j element is determined as

2

0

0

1

2

jlj

j jA dxt

r

T (9)

If vector jQ is generalized vector of elastic

displacement of element and as defined

2 1 2 2 1 2 2( )T

j j j j jt q u u u u Q (10)

Then Eq. (6) can be rewritten as

1

2

Tj j j jT Q M Q (11)

And each element of inertial mass matrix jM can be

computed as

0 0

0( , )

( , 1,2,..,5)

j

Tl

j j

j jjm je

m e A dx

m e

r r

QM

Q (12)

Besides, the elastic potential energy of j element is

shown as jP and presented as

22

20

( , )1 1( ) ( )

2 2

jlj j T

j j j j j

j

x tEI dx t t

x

wP Q K Q (13)

The stiffness matrix jK is defined as [4]

2 2

3

2 2

0 0 0 0 0

0 12 6 -12 6

0 6 4 -6 2

0 -12 -6 12 -6

0 6 2 -6 4

j j

j j j j j

jj j

j j j j

l l

EIl l l l

ll l

l l l l

K (14)

Vector Q represents the generalized coordinate overall

system and is given as

1 1 2 2 1 2 2. .T

n nq u u u u Q (15)

The total kinetic and potential energy of elastic on link

can be written as

1

1( ) ( )

2

nT

dh j dh

j

t t

T T Q M Q (16)

1

1( ) ( )

2

nT

j

j

t t

P P Q KQ (17)

Where dhM and K are total inertial mass and

stiffness matrices. Each element of them is computed by

FEM theory [1].

2.2. Dynamic equations

Fundamentally, the method relies on the Lagrange

equations with Lagrange function L T P are given by

( ) ( )d

tdt

L LF

QQ (18)

Where T is the kinetic energy overall system. It is

given by

1

2

Tdh dc P T T T T Q MQ(t) (t) (19)

Kinetic energy of rotor dcT is determined as

1 1( ) ( ) ( )

2 2

Tdc h dcI q t t t T Q M Q (20)

Payload kinetic energy PT is computed as

1( ) ( )

2

TP Pt tT Q M Q (21)

8 Bien Xuan Duong, My Anh Chu, Lac Van Duong

Where matrices dcM and PM are determined from

variables Q vector, respectively. The total inertial mass

matrix M can be written as

dh dc P M M M M (22)

The total potential energy can be computed by

summing over all the potential energy of each element

following Eq. (17). Vector ( )tF is the external generalized

force and presented as

( ) 0 0 . . 0T

t F (23)

Thus, by implementation of Lagrange principle, the

nonlinear dynamic equations of the system are summarized

as follows

cMQ +CQ + DQ + F + KQ +G = F (24)

Where matrix cF is coulomb friction at the joints and G

represents forces from gravity. In this paper, coulomb

friction and gravity effects are ignored as the manipulator

movement is confined to the horizontal plane. The

centrifugal C and structural damping D terms are

calculated by using Eq. (25) and Eq. (26).

1( ( )

2

T T

CQ MQ Q MQ

Q (25)

D M K (26)

Where and are the damping ratios and are

determined by experiences [1]. Depending on the system,

it is possible to simplify Eq. (21). The variation in M

depends on the geometry and Q variables. This is different

in [1, 2] with assumption that the flexible manipulator

motions in small workspace, small angular of joints and

small elastic displacements. So, matrix M is constant. The

dynamic equations (24) become in compact from as

MQ+CQ+ DQ+KQ = F (27)

It must be further noticed that the first node of link is

coincided on the joint of the link. Thus, this node has elastic

displacements which are zero ( 1 2( ) 0; ( ) 0u t u t ). In this

work, as the link modeled by one element,

3 4

Tq u u Q is the generalized coordinate vector of the

system, where 3u and 4u are flexural and slope displacement

of the end point of link, respectively. The external generalized

force is ( ) 0 0T

t F . Matrices M and K can be

compacted by elimination 2nd, 3rd rows and 2nd, 3rd columns,

respectively. Matrices C and D are determined by using Eq.

(25, 26). The dynamic nonlinear equations of one-link flexible

manipulator can be derived as the following Eq. (27).

3. Dynamic behaviors of system

The initial values of dynamic model are given by [4]

2

11 4 3 10 2

5 2 5

1( ); 0.02( ); 0.003( ); . ( );

4.5 10 ( ); 2710( / ); 7.11 10 ( / );

5.86 10 ( . ); 0.1( ); 0.76; 5.6 10 ;h p

L m b m h m A b h m

I m kg m E N m

I kg m m kg

The driving rules are shown in Figure 2. They are

Bang-Bang [1], trapezoidal, triangular and sinusoidal,

respectively. The Bang-Bang and trapezoidal rule highly

consume energy. Minimum energy consumption of

triangular rule is the smallest. Torque impact of Sinusoidal

rule is smooth.

Figure 2. Bang-Bang, Trapezoidal, Triangular

and Sinusoidal driving rules

The time for torque transmission is 0,6( )s , simulation

time is 3( )s . The dynamic equations (27) solved on

MATLAB/SIMULINK software by using ODE45 function.

The Figure 3 is example with Bang-Bang driving rule.

Figure 3. System block diagram in SIMULINK

with Bang-Bang driving rule

The effects of different driving rules on dynamic

behavior of one-link flexible manipulators are show

through the values of joint variable, its velocity, elastic

displacements and their velocity at the end-point of link.

Two cases are presented to consider dynamic behavior.

The first case is given without payload and the other case

is opposite.

Case 1: One-link flexible manipulator without

payload ( 0( )Pm kg )

The Figure 4 shows the values of joint variable, flexural

and slope displacement. The maximums are 0.5( )rad ,

0.06( )m and 0.1( )rad with Bang-Bang driving rule,

respectively. The minimums are given by triangular

driving rule. The system appears vibrant and stable after

1.2(s) with 0.6(s) torque impact. Figure 5 shows the

velocity of joint variable, flexural and slope displacement.

Figure 6 shows that Bang-Bang driving rule creates

maximum distance on X and Y axis.

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 9

Case 2: One-link flexible manipulator with payload

( 0.1( )Pm kg )

In case 2, the dynamic behaviors of system are shown

more clearly than case 1. The system vibrates longer

especially with Sinusoidal driving rule. Results are shown

in Figure 7. The velocity in Bang-bang driving rule has

changed hugely in Figure 8. In general, it is easily

recognized that system with payload vibrates longer than

case which is without payload. The values of hub-angle and

distance on X, Y axis of end-effectors in case 2 are smaller

than case 1. Furthermore, velocity of joint variable

increases fast while elastic displacement values are not too

different from, other rules and dynamic behaviors of

system still are quite fully expressed.

Figure 4. Hub angle, flexural and slope displacement values ( 3 4, ,q u u )

Figure 5. Velocity of hub angle, flexural and slope displacement values ( 3 4, ,dq du du )

Figure 6. The X and Y coodinates of end-effector on XOY

Figure 7. Hub angle, flexural and slope displacement values

10 Bien Xuan Duong, My Anh Chu, Lac Van Duong

Figure 8. Velocity of hubangle, flexural and slope displacement values

4. Conclusion

In this paper, the nonlinear dynamic model of a one-

link flexible manipulator is presented with or without

payload. The motor inertial moment, centrifugal coriolis

and the damping are considered. The nonlinear dynamic

equations are clear and full by using FEM in Lagrange

approach. The effects of different driving rules on dynamic

behaviors of system are studied. Bang-Bang law can be a

reasonable driving rule which is used in most studies for

analyzing dynamic behaviors of flexible manipulators.

REFERENCES

[1] M O Tokhi, Z Mohamed and M H Shaheed, “Dynamic

characterization of a flexible manipulator system”. Robotica /Volume 19/Issue 05/pp 571 – 580, September 2001.

[2] Victor Gavriloiu, “Design of dynamic nonlinear control techniques for

flexible link manipulator”, Thesis submitted for the degree of master of

applied Science, Concordia University Montreal, Quebec, Canada (2005).

[3] L. T. Wang and B. Ravani, Dynamic load carrying capacity of

mechanical manipulators-part I: problem formulation, Journal of Dynamic system, measurement and control, March 1988, Vol. 110/47.

[4] L. T. Wang and B. Ravani, Dynamic load carrying capacity of

mechanical manipulators-part II: computational procedure and

applications, Journal of Dynamic system, measurement and control, March 1988, Vol. 110/53.

[5] Moharam H Korayem, Mostafa Nazemizadeh, and Hamed Rahimi

nahooji, Dynamic Load Carrying Capacity of Flexible Manipulators

Using Finite Element Method and Pontryagin’s Minimum Principle,

Journal of optimization in Industrial Engineering 12 (2013), 17-24.

[6] S. S. Ge, T. H. Lee and G. Zhu, A Nonlinear feedback controller for a single link flexible manipulator based on a finite element method,

Journal of robotics system 14(3), 165-178, 1997.

[7] M. Pala Prasad Reddy and Jeevamma Jacob, Accurate modeling and

nonlinear finite element analysis of a flexible link manipulator,

International Journal of Mechanical, Aerospace, Industrial and Mechatronics Engineering Vol.8, No.1, 2014.

(The Board of Editors received the paper on 18/4/2017, its review was completed on 24/6/2017)

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 11

ASSESSMENT OF THE USE OF RAW FLY ASH WITH HIGH LOSS ON

IGNITION IN CONCRETE

Ngo Si Huy1, Huynh Trong Phuoc2, Le Thi Thanh Tam1 1Hong Duc University; ngosihuy@hdu.edu.vn

2College of Rural Development, Can Tho University; htphuoc@ctu.edu.vn

Abstract - The use of fly ash to partially replace cement in concrete is one of the effective ways of reducing the negative effect on the environment due to the use and production of cement. However, properties of concrete are strongly associated with the quality of fly ash used. This paper investigates the applicable use of raw fly ash, which is taken from Nghi Son coal power plant, Vietnam with a high loss on ignition in concrete. There are two group mixtures with different water-to-binder ratios of 0.35 and 0.45. In each group, fly ash is used to replace 10%, 20%, and 30% cement. Test results indicate that raw fly ash can be used to replace up to 20% cement with improved compressive strength. Moreover, with increasing the fly ash content, workability of fresh concrete increases while its unit weight decreases. Furthermore, all tested concrete mixtures show the good durability performance with ultrasonic pulse velocity values of greater than 4100 m/s.

Key words - Raw fly ash; loss on ignition; slump; unit weight; compressive strength; ultrasonic pulse velocity; durability.

1. Introduction

Concrete is the most popular construction material in

the world. Annually, a large amount of concrete is used as

well as a large quantity of cement is consumed. The

production of ordinary Portland cement releases a large

amount of carbon dioxide (CO2) into the air and causes the

greenhouse effect as the global warming. On the other

hand, fly ash is an industrial by-product of thermal power

stations and is recognized as an environmental pollutant.

Recently, the utilization of fly ash as partial or full

replacement of ordinary Portland cement in concrete has

received much attention from researchers instead of

considering it as a waste material.

The properties of concrete with fly ash strongly depend

on the characteristics of fly ash used [1]. Some previous

studies have indicated that fly ash could be used to replace

10÷30% cement with improved compressive strength

[2-5]. Some other studies showed that fly ash could be used

to replace up to 40÷60% cement in concrete with better or

comparable compressive strength [6-9]. It is noted that fly

ash used in these previous studies were from different

sources with various qualities. Especially, the loss on

ignition of fly ashused in the studies [6-9] was ranged from

0.45% to 2.1%, satisfying the requirement of less than 6%

as required by ASTM C618 [10]. In other words, such fly

ash was selected with good quality so that it can be used to

replace up to 60% cement. It is also noted that the physical

and chemical properties of fly ash used in studies [2-5]

were not given. However, with the low applicable

replacement level for cement, it is proposed that their

quality is not as good as those used in studies [6-9].

The loss on ignition of fly ash is due to the loss amount

of carbon and sulphur at high temperature. The presence of

unburned (organic) carbon in fly ash affects the color of

concrete (may change to black or dark gray), increases the

water requirement, and reduces the efficiency of air

entraining admixtures in concrete [11-13]. Most of the

previous studies used fly ash as a selected material with the

loss on ignition satisfying the condition stipulated by

ASTM C618 [10]. So far, a study on the use of raw fly ash

with ahigh loss on ignition has been still limited. Thus, the

major objective of this study is to investigate the

applicability of the raw fly ash with a high loss on ignition

in concrete. Fly ash used in this research was taken from

Nghi Son thermal power plant, Viet Nam, with the loss on

ignition of 15.76%, which exceeds the ASTM C618 [10]

requirement.

2. Materials and experimental program

2.1. Materials

Table 1 shows the physical and chemical properties of

cement and fly ash (FA) used in this study. Ordinary Portland

cement was Nghi Son type-PC40 with a specific gravity of

3.12. Raw fly ash was taken from Nghi Son coal power plant,

containing a low amount of calcium. According to ASTM

C618 [10], it is classified as class-F. The possibility of using

the raw fly ash with a high loss on ignition value of 15.75%

in concrete is examined in this study.

Natural sand with fineness modulus of 2.87, density of

2.62 T/m3, dry rodded weight of 1.50 T/m3, moisture content

of 4.35%, and water absorption capacity of 1.08% was used

as the fine aggregate. The coarse aggregate used was crushed

stone with the nominal maximum size of 12.5 mm, density

of 2.69 T/m3, dry rodded weight of 1.39 T/m3, moisture

content of 0.25%, and water absorption capacity of 0.08%.

In order to reduce the amount of water for good concrete

quality and to increase the workability of fresh concrete, the

superplasticizer (SP) of Sikament R4 with a specific gravity

of 1.15 was used with a fixed dosage of 1% (by total amount

of binder materials) for all concrete mixtures.

Table 1. Physical and chemical properties of cement and FA

Items Cement FA

Physical properties Specific gravity 3.12 2.16

Chemical

composition (wt.%)

SiO2 22.38 48.38

Al2O3 5.31 20.42

Fe2O3 4.03 4.79

CaO 55.93 2.80

MgO 2.80 1.41

Others 4.45 4.28

Loss on ignition 1.98 15.76

2.2. Mixture proportions

Table 2 shows the mixture proportions for the

preparation of concrete samples used in this investigation.

12 Ngo Si Huy, Huynh Trong Phuoc, Le Thi Thanh Tam

There are eight mixtures that are divided into two groups

(M35 and M45) with different water-to-binder ratios of

0.35 and 0.45. Nomenclature of the mixtures is described

as follows: M35 and M45 denote the water-to-binder ratio

of 0.35 and 0.45, respectively; the number after them (0,

10, 20 and 30) are the percentages of fly ash replacing

cement. In each group, the control mixture was designed

without fly ash, while three others used fly ash to replace

cement by 10, 20, and 30% by weight. The purpose of these

design mixtures is to find out the optimal level of fly ash

replacement for cement in concrete. All of the mixtures

were designed in accordance with ACI 211.91 [14]. It is

noted that the amount of water was controlled lower than

180 kg/m3 in order to ensure the good quality of concrete.

Table 2. Concrete mixture proportions

Mixture Concrete ingredient proportions (kg/m3)

Cement FA Sand Stone Water SP

M35-0 514.1 0.0 985.2 751.1 175.5 5.1

M35-10 459.7 50.6 978.1 745.7 174.3 5.1

M35-20 405.1 101.5 970.9 740.3 173.0 5.0

M35-30 352.3 150.7 964.0 735.0 171.8 5.0

M45-0 400.4 0.0 1082.1 751.8 175.7 4.0

M45-10 358.3 39.8 1076.0 747.5 174.7 4.0

M45-20 316.7 79.2 1069.9 743.3 173.7 4.0

M45-30 275.6 118.1 1063.9 739.1 172.7 3.9

2.3. Specimens preparation and test programs

The experimental works were conducted in the

construction material laboratory of Hong Duc University.

After mixing, fresh concrete properties including slump

and unit weight were measured. Concrete specimens were

prepared in a cylinder with 10 cm in diameter and 20 cm in

height. After one day of casting, they were demolded and

immersed in water at a room temperature until the testing

age as shown in Figure 1. Compressive strength and

ultrasonic pulse velocity (UPV) were measured at 3, 7, 14,

28, 56, and 91 days. Figure 2 shows pictures of the

compressive strength test and ultrasonic pulse velocity test.

The values reported herein are the average value of three

specimens. Slump, compressive strength, and ultrasonic

pulse velocity tests were performed in accordance

withASTM C143 [15], ASTM C39 [16], and ASTM C597

[17], respectively.

Figure 1. Concrete specimens were cured in water

(a) (b)

Figure 2. a) Compressive strength test; b) Ultrasonic pulse

velocity test

3. Results and discussion

3.1. Fresh concrete properties

The slump and unit weight of all concrete mixtures are

given in Table 3. The slump values of M45 mixtures are

greater than those of corresponding M35 mixtures. It

means that higher water-to-binder ratio results in a higher

slump. It is clear that the workability of fresh concrete is

closely associated with the amount of water. Thus, the

amount of water increases as the increase of water-to-

binder ratio, leading to increasing concrete workability.

As shown in Table 3, for the same water-to-binder ratio,

the slump of fresh concrete increases with increasing the

replacement level of cement by fly ash. In general, the fly

ash particles are spherical while cement particles are

irregular polygonal [18]. The spherical shape of fly ash

contributes to reducing the friction of the aggregate-paste

interface, thus increases the workability of fresh concrete.

Moreover, the paste volume of fly ash is greater than that

of cement because the specific gravity of fly ash is lower

than that of cement (Table 1). The increase in the paste

volume leads to the increase of plasticity and cohesion,

then increases the workability of fresh concrete. Therefore,

with increasing fly ash content, the slump of concrete

increases. This finding is similar to the result from Naik

and Ramme [6].

On the other hand, for the same water-to-binder ratio, the

unit weight of fresh concrete decreases with increasing fly

ash content. This is also due to the low specific gravity of fly

ash as compared with that of ordinary Portland cement. With

the same amount, the volume of fly ash is more than that of

cement, therefore the unit weight of fresh concrete decreases

with increasing fly ash content in the concrete mixture.

Moreover, when water-to-binder ratio increases, the unit

weight of concrete decreases. Similarly, due to the highest

specific gravity of cement as compared with other

ingredients in concrete, the amount of cement increases with

the water-to-binder ratio, leading to an increase in mass of

concrete with the higher water-to-binder ratio.

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 13

Table 3. Fresh concrete properties

Mixture FA content

(%)

Slump

(mm)

Unit weight

(T/m3)

M35-0 0 0 2.58

M35-10 10 5 2.55

M35-20 20 10 2.51

M35-30 30 30 2.49

M45-0 0 30 2.53

M45-10 10 45 2.51

M45-20 20 55 2.48

M45-30 30 70 2.46

3.2. Compressive strength development

Compressive strength is an important property used to

evaluate the concrete quality. The compressive strength

development of concrete with different water-to-binder

ratios of 0.35 and 0.45 is presented in Figures 3 and 4,

respectively. The concrete mixtures with lower water-to-

binder ratio have higher compressive strength values than

those of corresponding mixtures with the higher water-to-

binder ratio. A low water-to-binder ratio is associated with

a high content of cement paste, hence a high value of

compressive strength is.

For the water-to-binder ratio of 0.35, the mixture with

10% fly ash (M35-10) shows the highest compressive

strength after 14 days. The mixtures with 20% and 30% fly

ash content exhibit lower compressive strength than

control mixture without fly ash (M35-0). Similarly, for the

water-to-binder ratio of 0.45, the mixtures with 10%

(M45-10) and 20% (M45-20) fly ash show higher

compressive strength than the control mixture (M45-0),

while mixture with 30% fly ash (M45-30) shows the lowest

compressive strength value. The replacement level of

cement by fly ash in this study is lower than that reported

from previous studies [6-9] because the fly ash used herein

is a raw material with low quality as compared with that of

previous studies [6-9]. This also means that even with a

high loss on ignition (15.76%), raw fly ash can be used to

replace up to 20% cement in concrete with improved

compressive strength. Although the compressive strength

of concrete mixtures with 30% fly ash is lower than those

of control mixture, the ultimate compressive strength of the

M35-30 and M45-30 mixtures are 35.1 MPa and 24.8 MPa,

respectively. It means that depending on the strength

requirement, the replacement level of cement by fly ash is

determined accordingly.

Figure 3. Compressive strength development of M35 mixtures

Figure 4. Compressive strength development of M45 mixtures

As can be seen from Figures 3 and 4, before 28-day

ages, the fly ash-free concrete mixture shows a relatively

high compressive strength value as compared with that of

fly ash concrete mixtures. Additionally, after 56 days, the fly

ash-free concrete shows a lower strength development rate

than the other mixtures containing fly ash. The low

compressive strength value at the early age and the

increased strength at the later age of the fly ash concrete

are associated with the slow and continuous pozzolanic

reaction of fly ash in concrete, which only starts

significantly after one or more weeks [19]. The major

products of cement hydration (see equation (1)) are

calcium silicate hydrate (C-S-H) gel and calcium

hydroxide (Ca(OH)2). While C-S-H is the main carrier of

strength in hardened concrete, Ca(OH)2 has a negative

effect on quality of hardened concrete because of its

solubility in water to form cavities and its low strength.

However, when fly ash is added to the mixture as a cement

substitution, Ca(OH)2 is transformed into the secondary

C-S-H gel as a result of the pozzolanic reaction (see equation

(2)). However, if the fly ash content is added over the

optimum value, such fly ash amount does not fully involve

in the chemical reaction process. In this case, it mainly acts

as fine aggregate in the mixture rather than a cementitious

additive. In other words, the fly ash is not used efficiently.

Cement hydration:

3 2 2 2( , )C S C S H O C S H Ca OH (1)

Pozzolanic reaction:

22Ca OH SiO C S H (2)

3.3. Ultrasonic pulse velocity (UPV)

To assess the uniformity and relative quality of concrete

as the presence of voids and cracks, the ultrasonic pulse

velocity (UPV) test is conducted in accordance with ASTM

C597 [17]. Generally, a high UPV value indicates good

quality concrete. As suggested by Malhotra [20], concrete has

good durability when it obtains UPV value of higher than

3660 m/s. Figures 5 and 6 show the UPV of concrete with

different water-to-binder ratios of 0.35 and 0.45, respectively.

As a result, all concrete mixtures show the good quality with

the UPV values of higher than 4100 m/s. In addition, before

91 days, the UPV values of the fly ash-free concrete mixtures

(M35-0 and M45-0) are higher than those of mixtures

containing fly ash. However, at 91 days, the UPV value of

14 Ngo Si Huy, Huynh Trong Phuoc, Le Thi Thanh Tam

the M35-10 mixture is close to that of M35-0 mixture.

Moreover, the increasing rate of the UPV values of the fly

ash concrete is higher than that of the fly ash-free concrete

regardless of the water-to-binder ratios of 0.35 or 0.45. This

phenomenon may be explained by the fact that the control

mixture contains a greater quantity of cement than fly ash

mixtures. Thus, due to the higher specific gravity of cement

in comparison with that of fly ash, the control mixture shows

a higher density and a faster cement hydration than the fly ash

mixtures at the early age. This leads to a higher UPV value

of control concrete as compared with that of the fly ash

concrete. However, after 56 days, with a great contribution of

pozzolanic reaction due to the presence of fly ash in concrete,

the inner structure of fly ash concrete is improved, indicated

by an increase of UPV values, and such concrete mixtures

exhibit good quality and durability.

Figure 5. Ultrasonic pulse velocity values of M35 mixtures

Figure 6. Ultrasonic pulse velocity values of M45 mixtures

4. Conclusions

This paper investigates the applicable use of raw fly ash

with a high loss on ignition in concrete. Based on the above

experimental results, the main conclusions are summarized

as follows.

1) The slump of concrete increases and the unit weight

of fresh concrete decreases with the increased level of fly

ash replacement for cement.

2) Raw fly ash can be used to replace up to 20% of

cement in concrete with the enhanced compressive

strength. Further increasing fly ash content results in a

reduction in compressive strength. Thus, the level of fly

ash replacement can be increased or decreased depending

on the requirement for compressive strength of concrete.

3) Fly ash concrete shows a low compressive strength

at an early age and a great strength improvement at the later

age. This is mainly due to the pozzolanic reaction of fly ash

in concrete.

4) All concrete mixtures with raw fly ash in this study

exhibit good durability with the ultrasonic pulse velocity of

higher than 4100 m/s.

Acknowledgments

The authors would like to thank Hong Duc University

for financial support. The experimental works were carried

out at the construction material laboratory of the Department

of Engineering and Technology, Hong Duc University.

REFERENCES

[1] A. Bilodeau, and M. Malhotra, “High-volume fly ash system:

Concrete solution for sustainable development”, ACI Material Journal, 2000, Vol. 97, No. 1, pp. 41-47.

[2] P. R. Wankhede, and V. A. Fulari, “Effect of fly ash on properties of

concrete”, International Journal of Emerging Technology and

Advanced Engineering, 2014, Vol. 4, No. 7, pp. 284-289.

[3] C. Marthong, and T. P. Agrawal, “Effect of fly ash additive on

concrete pro perties”, International Journal of Engineering Research and Applications, 2012, Vol. 2, No. 4, pp. 1986-1991.

[4] R. Bansal, V. Singh, and R. K. Pareek, “Effect on compressive

strength with partial replacement of fly ash”, International Journal

onEmmerging Technologies, 2015, Vol. 6, No. 1, pp. 1-6.

[5] H. A. Mohamed, “Effect of fly ash and silica fume on compressive

strength of self-compacting concrete under different curing conditions”, Ain Shams Engineering Journal, 2011, Vol. 2, pp. 79-86.

[6] T. R. Nail and B. W. Ramme, “Effect of high-lime fly ash content

on water demand, time of set, and compressive strength of concrete”,

ACI Materials Journal, 1990, Vol. 87, No. 6, pp. 619-626.

[7] Oner, S. Akyuz, and R. Yildiz, “An experimental study on strength

development of concrete containing fly ash and optimum usage of

fly ash in concrete”, Cement and Concrete Research, 2004, Vol. 35,

pp. 1165-1171.

[8] R. Siddique, “Performance characteristics of high-volume class F fly ash

concrete”, Cement and Concrete Research, 2003, Vol. 34, pp. 487-493.

[9] G. Li, and X. Zhao, “Properties of concrete incorporating fly ash and

ground granulated blast-furnace slag”, Cement and Concrete Composites, 2003, Vol. 25, pp. 293-299.

[10] ASTM C618, Standard specification for coal fly ash and raw or

calcined natural pozzolan for use in concrete, 2005.

[11] M. Mohebbi, F. Rajabipour, B.E. Scheetz, “Reliability of loss on

ignition test for determining the unburned carbon content in fly ash”,

World of Coal Ash Conference in Nasvhille, 2015.

[12] K. Wesch, Fly ash in concrete: Properties and performance. Report of Technical Committee 67-FAB use of fly ash in buildings, E&FN

Spon, London, 1991.

[13] J. Paya, J. Monzo, M.V. Borrachero, E. Perris, and F. Amahjour,

“Thermo gravimetric methods for determining carbon content in fly

ashes,” Cement and Concrete Research, 1998, Vol. 28, pp. 675-688.

[14] ACI 211.1, Standard practice for selecting proportions for normal, heavyweight, and mass concrete, 1991.

[15] ASTM C143, Standard test method for slump of hydraulic-cement

concrete, 2015.

[16] ASTM C39, Standard test method for compressive strength of

cylindrical concrete specimens, 2012.

[17] ASTM C597, Standard test method for pulse veloscity through

concrete, 2009.

[18] V. G. Papadakis, “Effect of fly ash on Portland cement systems –

Part I. Low-calcium fly ash”, Cement and Concrete Research, 1999, Vol. 29, No. 11, pp. 1727-1736.

[19] A. L. A. Fraay, J. M. Bijen, and Y. M. de Haan, “The reaction of fly

ash in concrete a critical examination”, Cement and Concrete

Research, 1989, Vol. 19, No. 2, pp. 235-246.

[20] V.M Malhotra, Testing hardened concrete: Nondestructure methods,

American Concrete Institute Monograph Series, 1976.

(The Board of Editors received the paper on 02/6/2017, its review was completed on 13/7/2017)

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 15

ANALYSIS AND TREATMENT OF TRIMETHOPRIM

RESIDUES IN WATER BY γ-IRRADIATION

Nam D. Le, Thang M. Ngo

Ho Chi Minh City University of Technology; namleduyhc06@gmail.com, nmthang@hcmut.edu.vn

Abstract - Trimethoprim (TMP) is widely applied in veterinary and also frequently prescribed together with sulfa-methoxazole (SMX) for human medicine. Therefore TMP residues accumulated in agricultural as well as municipal waste water further contaminate surface water. In this paper, the capability of HPLC/UV to detect and quantify TMP residues in water is thoroughly investigated, yielding LOD = 0.06 µM, LOQ = 0.2 µM and very good reproducible calibration line in the concentration range of 2 µM ÷ 100 µM. The resulting procedure is applied to evaluate the capability to treat TMP residues in water (init. 20 µM ÷ 140 µM) by gamma irradiation. Removal yields greater than 99 % are obtained using absorbed doses 0.3 ÷ 3.0 kGy, respectively. Based on the HPLC/UV chromatograms obtained, some aspects of the TMP radio-lytic products in the investigated samples are briefly discussed.

Key words - analysis; antibiotic residues; gamma-irradiation; trimethoprim; water treatment and reuse.

1. Introduction

Residues of pharmaceutical products, especially those of

antibiotics in natural aquifers have been detected worldwide

[1-4]. A joint research project showed that residues of

sulfamethoxazole (SMX) in surface waters in Vietnam are

higher compared with those in other countries [5]. As SMX

is frequently applied together with trimethoprim (TMP),

residues of the latter one in Vietnam’s surface water are

supposed at elevated levels, too.

Trimethoprim (TMP) with molecule formula

C14H18N4O3 and structure shown in Figure 1 is an antibiotic

against a broad spectrum of bacterial species and applied

both in veterinary as well as in human medicine (mostly in

combination with SMX).

Figure 1. Molecular structure of TMP

At ambient conditions (1 atm, 20oC), TMP’s solubility

in water is about 400 mg/L, which increases with

temperature and/or concentration of other organic solvents

in the order ethyl acetate, 2-propanol, acetonitrile, ethanol

[6]. Moreover, TMP is known to be persistent in

conventional wastewater treatment facilities. Therefore,

TMP could be involved, accumulated and transported in

the environment alongside the water streams.

Attention of several research groups has been focused

to treatment of TMP residues in water by diverse methods.

E.g. Electro-catalytic degradation on surfaces of carbon

electrodes doped by porphyrin manganese was

investigated and theoretically validated by means of

computational chemistry [7]. Sorption of TMP onto some

agricultural soil samples and its desorption by CaCl2

solutions or outflow from an wastewater treatment facility

was reported, demonstrating that the local soil- and aquifer

compositions play an important role in transport of TMP

[8]. Recently, peroxydisulphate initiated by heating to

temperatures 50 ÷ 65oC has been applied to activate TMP

removal, depending on the sample matrices – while natural

organic matters and bicarbonate ions suppress this process,

chloride ions accelerate it [9].

By means of photolysis and photo-catalysis, TMP in both

distilled water and sea water matrices is relatively stable

under natural light illumination. Although an intermediate

photolytic product is photosensitive and acted as auto-

catalyst, the sample DOC decreases very slowly. Using TiO2

increases the mineralization degrees of TMP in both

matrices, but the rates in sea water are substantially lower

because the inorganic components act as hydroxyl radical

scavengers [10]. Under similar illuminating conditions by

UV-A, UV-C and VUV, hydroxyl radicals play an important

role in the samples investigated, enabling up to ~ 73% the

total removal yields, while direct photolysis accounts for

about ~ 27% [11]. A somewhat more complicated situation

is TMP and SMX treatment in urine matrices due to diverse

effects of the matrix components [12, 13].

Gamma irradiation using 60Co sources is classified

among the advanced oxidation methods as it produces

hydroxyl radicals, too. It is applied mainly to discuss the

mechanism and intermediate products of TMP (init. 1 mM)

transformation by hydroxyl radicals [14]. In a more recent

paper, the TMP (init. 20 mg/l ~ 69 μM) removal is reported

but focuses on the effects of persulfate concentration

0.5 ÷ 2.0 mM and matrix pH 6,5 ÷ 8,5 [15].

The effect of initial concentration of TMP is not reported

in both these 2 publications. Moreover, the HPLC/UV

procedures seem very complicated and differ from each

other, causing confusion about the reported TMP removal

yields. Our paper first focuses on the HPLC/UV procedure

for TMP analysis and then on the TMP removal yields

depending on its initial concentration and the applied doses.

2. Materials and methods

TMP 99.0% purchased from Sigma-Adrich, formic

acid p.a. from Merck, Acetonitrile HPLC grade from

J. Baker and other chemicals of analytical grade are used

without further purification. Bi-distilled water is used for

preparing solutions.

A 1000 μM TMP stock solution is prepared by

dissolving 0.0726 g TMP in 250 ml bi-distilled water,

stored in dark at ~ 4oC and diluted accordingly to actual

samples (TMP conc. in μM: 140, 100, 70, 50, 30, 20, 10, 5,

2, 1, 0.5, and 0.2 μM, respectively) before use.

16 Nam D. Le, Thang M. Ngo

TMP concentrations are analysed using an HPLC

equipment typed Agilent 1290 infinity series, equipped with

an Agilent Eclipse Plus C18 guard column (1.8 µm x 2.1 mm

x 50 mm), an Agilent Poroshell 120 EC–C18 analytical

column (2.7 µm x 4.6 mm x 100 mm) and a diode array

detector (DAD). The column was let at room temperature,

the injection volume fixed at 10 μl, the wavelength set at

254, 265, 270 and 275 nm. First, various compositions of the

mobile phase are tested. Then the mobile phase flowrate, and

finally the calibration line is constructed as function of peak

volume vs. sample concentration.

Gamma irradiation experiments are conducted at Da

Lat Nuclear Research Institute as described in previous

paper [16], using a Gamma chamber 5000 (India) 60Co source with dose rate ~ 46.6 Gy/min. Briefly, 8 ml

sample (TMP init. conc. 20 μM ÷ 140 μM) is filled in

12 ml glass tube (Hach, USA), tightly closed and irradiated

to the pre-determined absorbed dose (0.3 kGy ÷ 3.0 kGy).

TMP concentrations before and after irradiation were

analyzed using the established procedure and constructed

calibration line. Each experiment was conducted in

triplicate to validate the experimental errors.

3. Results and discussion

3.1. HPLC/UV procedure

In order to avoid the pressure change, the isocratic

mode of mobile phase is applied throughout this work.

Fixing the flowrate at 1 ml/min., various mixing ratios of

bi-distilled water / acetonitrile (90% ÷ 10%), 10mM

phosphate buffer pH 3.5 / acetonitrile (10% ÷ 30%) do not

result in any peaks of TMP even at prolonged measuring

time, despite its success using gradient mode [e.g. 14,15].

However, mixtures of 0.1% formic acid / acetonitrile (70%

÷ 90%) work relatively well. Taking into account effects of

the mobile phase flowrate (1.0 ml/min ÷ 0.25 ml/min) onto

the retention time, peak area and – symmetry, the mobile

phase composition is chosen 0.1% formic acide /

acetonitrile = 82% / 18% (v/v) and its flowrate 0.5 ml/min.

TMP signals are the highest at wavelength 270 nm instead

of 275 nm as stated in [15]. Figure 2 and Figure 3 show

chromatograms of the most dilute TMP samples which are

measured in this work, and the constructed calibration line

using the HPLC parameters mentioned above. One can see

a linear relationship between the TMP peak areas and the

corresponding TMP concentrations up to 100 μM with

confidence coefficient of 0.9999.

The reproducibility of TMP retention times is pretty

good, e.g. (3.42 0.02) min.rt results from

12 measurements presented in Figure 2 (triplicate

measurement each sample). Certainly, this retention time

increases with increasing the 0.1% formic acid /

acetonitrile in the mobile phase, e.g. to ~ 90/10. However,

the peak shape and symmetry suffer a lot. The TMP peak

areas are well reproducible, too. The estimated relative

errors are within 3%, even for the most dilute sample (0.2

μM TMP). For the sake of our further application, 0.2 μM

TMP is considered the real limit of quantitation (LOQ) and

therefore 0.067 μM TMP comes out as the corresponding

limit of detection (LOD) of this analytical procedure.

It is worth noting this analytical procedure does not aim

to analyze TMP concentration in surface water samples,

which are at least about 100x lower than 0.2 μM [14,15].

In such cases, an additional pre-concentration step, e.g. by

solid phase extraction (SPE), is necessary. However, it

confirms that the calculated removal yields up to 99% even

from the initial concentration 20 μM TMP (see below) are

reliably determined.

Figure 2. Chromatograms of dilute TMP samples *

Figure 3. Calibration line for TMP analysis

* Mobile phase: 0.1% formic acid / acetonitrile = 82 / 18 (v/v),

0.5 ml/min. Detector wavelength 270 nm, injection volume 10 μl

3.2. TMP removal yields, -rates, and radio-lytic products

Based on the constructed calibration line in Figure 3, TMP

removal yields RD% are calculated according to the formula:

0 0

0 0

% D D

D

C C S SR

C S

Where symbols C and S refer to TMP concentrations

and peak areas, indexes 0 and D refer to samples before and

after absorbing dose D, respectively. The sample with

initial TMP concentration 140 μM is diluted before

analysis and the measured peak area is re-calculated.

Figure 4. shows the TMP removal yields and –rates

due to the absorbed doses. Most quantitative removal of

the initial TMP concentrations 20 μM, 50 μM, 70 μM,

100 μM, 140 μM is achieved at absorbed doses 0.3 kGy,

1.0 kGy, 1.5 kGy and 3.0 kGy, respectively.

As C0 = 70 μM is comparable with 20 mg/l (~ 69 μM) in

the literature [15], our determined dose for a practically

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 17

quantitative TMP removal is slightly higher (1.5 vs.

1 kGy). The origins of this difference might be, but not

limited to the difference in dose rates of the 60Co sources.

Except for the lowest conc. C0 = 20 μM, the experimental

results show that TMP removal rates fit well to kinetic

equations of pseudo-first order reactions, with the

reaction rate depending on the initial TMP concentration.

This finding is frequently reported in the literatures

[e.g. 16, 17]. In fact, from the theoretical point of view

reactions between substrate molecules – in this case TMP

molecules – and hydroxyl radicals in irradiated samples

are of second order [e.g. 14]. Anyway, these results

demonstrate the potential of γ-irradiation as an alternative

treatment method for TMP contaminating water. Even for

such a high level of contamination as ~ 140 μM TMP, an

absorbed dose just about 3.0 kGy is sufficient for its

almost quantitative removal.

Figure 4. TMP removal yields (upper) and –rates (lower) depending on its init. conc. and doses

Figure 5. Chromatograms of TMP samples depending on its initial concentrations and absorbed doses

Beside the removal yields, the toxicity or even identity

of the treatment products has recently become important

factors from both theoretical and practical points of view.

Normally, sophisticated equipment such as liquid

chromatography – time of flight mass spectrometry

(LC-TOF-MS) or conventional LC-MS are required

[e.g. 14-16]. However, the HPLC-UV chromatograms

reveal some characteristics of the treatment products which

18 Nam D. Le, Thang M. Ngo

absorb UV lights [e.g. 17]. Figure 5 shows up to 6 peaks of

TMP radio-lytic products which have retention times

shorter than TMP itself. According to the principle of

reverse-phase chromatography all these detected UV-

absorbing products have higher polarity than TMP. These

peaks gradually diminish with increasing the absorbed

dose, except for two with the shortest retention times,

suggesting that only the corresponding products are stable

under γ-irradiation. In addition, comparing chromatograms

on Figure 5 and Figure 2 would suggest that these

remaining 2 peaks represent the inorganic products. It is in

good accordance with the reported ~ 20% TMP

mineralized under comparable conditions [15]. However,

nothing more could be stated and, moreover, the number of

detected peaks are lower than the number of TMP radio-

lytic products reported in the literature [14, 15].

It is well known from the literature that in irradiated

aqueous samples, water is first radio-lysed to produce many

chemically active species including hydroxyl radical OH –

strong oxidant and ,aqua

H e – strong reductants, which

further attack the substrate molecules [e.g. 14, 15, 18].

Under the experimental conditions prevailing in this

work, mainly hydroxyl radical is responsible for TMP

radiolysis and it is believed to preferentially attack the

trimethoxybenzene moiety (TMB), as illustrated in Figure

6, resulting in up to 5 products. They all contain aromatic

rings [14, 15] and therefore should be able to absorb UV

radiations, too.

Figure 6. Preferential attack of OH to TMP

It is questionable whether some of the TMP radio-lytic

products identified by LC-MS in the literature could be

ascribed to the peaks mentioned above in the HPLC/UV

chromatograms. As indirect evidence, the octanol / water

distribution coefficients of the detected products and their

precursor – trimethropim – could be accessed by means of

computational chemistry and compared with each other [16].

4. Conclusions and outlooks

A suitable HPLC/UV procedure for rapid TMP analysis

in aqueous samples is described in details, which enables us

to analyze samples in the concentration range of 0.2μM ÷

100μM TMP. γ-irradiation proves to be an efficient

alternative method for treatment of TMP residues in water,

as an absorbed dose about of 0.3kGy should be sufficient to

quantitatively remove TMP residues at all contamination

levels typically found in wastewater effluents and natural

aquifers. Further investigation is required to identify/

quantify the radio-lytic products of TMP and/or to compare

the toxicity of samples before and after irradiation.

Acknowledgement

This research is funded by Ho Chi Minh City

University of Technology – VNU-HCM under grant

number Tc-KTHH-2017-04.

REFERENCES

[1] Sui Q., Cao X., Lu S., Zhao W., Qiu Z., Yu G., Occurrence, sources,

and fates of pharmaceu-ticals and personal care products in the groundwater: A review, Emerging Contaminants 1 (2015) 14-24.

[2] Schaefer M.M., Doyle L.A., Fleenor W.E., Johnson M.L., Fate and

Transport of Three Pharmaceuticals in the Sacramento–San Joaquin

Delta, San Francisco Estuary and Watershed Science 11 (2013) 1-13.

[3] Dinh Q.T., Moreau-Guigon E., Labadie P., Alliot F., Marie-Jeanne

T., Blanchard M., Joelle E., Chevreuil M., Fate of antibiotics from hospital and domestic sources in a sewage network, Sci. Total

Environ. 575 (2017) 758-766.

[4] Dinh Q.T., Moreau-Guigon E., Labadie P., Alliot F., Marie-Jeanne

T., Blanchard M., Chevreuil M., Occurrence of antibiotics in rural catchments, Chemosphere 168 (2017) 483-490.

[5] Shimizu A., Takada H., Koike T., Takeshita A., Saha M., Rinawati,

Nakada N., Murata A., Suzuki T., Suzuki S., Nguyen H.C., Bui C.T.,

Pham H.V., Siringan M.A., Kwan C., Zakaria M.P., Reungsang A.,

Ubiquitous occurrence of sulfonamides in tropical Asian waters, Science of the Total Environment 452-453 (2013)108-115.

[6] Yin D.P., Liu M.X., Fu H.L., Shu G., Zhou J.Y., Qing X.Y., Wu

W.B., Solubility of Trimethoprim in Selected Pure Solvents and

(Water + Ethanol/2-Propanol) Mixed-Solvent Systems, J. Chem. Eng. Data 61 (2016) 404-411.

[7] Rajith L., Jissy A.K., Kumar K.G., Datta A., Mechanistic Study for

the Facile Oxidation of Trimethoprim on a Manganese Porphyrin

Incorporated Glassy Carbon Electrode, J. Phys. Chem. C 115 (2011)

21858–21864.

[8] Zhang Y.L., Lin S.S., Dai C.M., Shi L., Zhou X.F., Sorption–desorption and transport of trime-thoprim and sulfonamide

antibiotics in agricultural soil: effect of soil type, dissolved organic

matter, and pH, Environ. Sci. Pollut. Res. 21 (2014) 5827-5835.

[9] Ji Y. Xie W., Fan Y., Shi Y., Kong D., Lu J. (2016), Degradation of

trimethoprim by thermo-activated persulfate oxidation: Reaction kinetics and transfor-mation mechanisms, Chem. Eng. J. 286 (2016) 16–24.

[10] Sirtori C., Aguera A., Gernjak W., Malato S. (2010), Effect of water

matrix composition on Trimethoprim solar photodegradation

kinetics and pathways, Water Research 44 (2010) 2735-2744.

[11] Kim H.Y., Kim T.H., Yu S., Photolytic degradation of sulfamethoxazole and trimethoprim using UV-A, UV-C and vacuum-

UV (VUV), J. Environ. Sci. & Health A 50 (2015) 292–300.

[12] Zhang R., Sun P., Boyer T.H., Zhao L., Huang C.H., Degradation of

Pharmaceuticals and Metabolite in Synthetic Human Urine by UV,

UV/H2O2, and UV / PDS, Environ. Sci. Technol. 49 (2015) 3056-3066.

[13] Zhang R., Yang Y., Huang C.H., Li N., Liu H., Zhao L., Sun P., UV/H2O2 and UV/PDS Treatment of Tri-methoprim and

Sulfamethoxazole in Synthetic Human Urine: Transformation

Products and Toxicity, Environ. Sci. Technol. 50 (2016) 2573–2583

[14] Luo X., Zheng Z., Greaves J., Cooper W.J., Song W., Trimethoprim:

Kinetic and mechanistic consideration in photochemical environmental fate and AOP treatment, Water Research 46 (2012) 1327-1336.

[15] Zhang Z., Yang Q., Wang J., Degradation of thrimethoprim by

gamma irradiation in the presence of persulfate, Rad. Physics Chem.

127 (2016) 85-91.

[16] Le D.N., Le T.T.T., Mai V.T.T., Huynh K.L., Ngo M.T.,

Transformation Products of Aqueous Sulfame-thoxazole by 60Co Gamma Irradiation - A Combined Computational and Experimental

Study, Proc. 5th World Conf. Appl. Sci. Eng. & Technol., (2016) 64-

69, ISBN 13:978-81-930222-2-1.

[17] Ngo M.T., Hoang M.N., Tran T.M.T., Radiolysis of 1-naphthol in aqueous solutions, J. Radioanal. Nucl. Chem. 286 (2010) 287-293.

[18] Wojnarovitz L., Takacs E., Wastewater treatment with ionizing

radiation, J. Radioanal. Nucl. Chem. 311 (2017) 973-981.

(The Board of Editors received the paper on 24/10/2017, its review was completed on 22/12/2017)

OH•OH•

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 19

AN EXPERIMENTAL STUDY ON PROPERTIES OF HIGH-PERFORMANCE

CONCRETE USING RECYCLED AGGREGATES

Huynh Trong Phuoc1, Nguyen Tien Dung2, Ngo Si Huy3, Vo Duy Hai4 1College of Rural Development, Can Tho University; htphuoc@ctu.edu.vn

2University of Science and Technology - The University of Danang; nguyentiendung261@gmail.com 3Hong Duc University; ngosihuy@hdu.edu.vn

4University of Technology and Education - The University of Danang; duyhai88@gmail.com

Abtract - This study evaluates the properties of high-performance concrete (HPC) using recycled coarse and fine aggregates sourced from Taiwan. Densified mixture design algorithm is used to calculate the mix proportion of the HPC. The HPC samples are prepared with a constant water-to-binder ratio of 0.35, using either natural fine aggregate (NFA) or recycled fine aggregate (RFA) combined with natural coarse aggregate (NCA) and recycled coarse aggregate (RCA). The effect of replacing 0%, 30%, and 100% NCA by RCA in each group mixture is studied. Both fresh and hardened properties of the HPC are evaluated. Test results show that the compressive strength values of the HPC range from 31.7 to 56.7 MPa. Moreover, all of the HPC samples produced in this study exhibit the excellent anti-erosion ability and durability performance with electrical surface resistivity and ultrasonic pulse velocity values of above 20 kΩ.cm and 3660 m/s, respectively.

Key words - High-performance concrete; recycled aggregates; engineering property; compressive strength; durability

1. Introduction

In recent years, the urbanization level in the developing

countries has been rapidly exploding. Consequently, a large

number of old infrastructures are demolished for new ones.

The demolition and the construction of new infrastructures

generate a lot of construction wastes such as demolished

concrete, brick, tile, wood, plastic, steel, etc. On the other

hand, the use of natural resources in some areas is over

exploited. The exploitation process and production of

construction materials such as aggregate and cement from

natural resources also generate a large amount of carbon

dioxide (CO2), causing greenhouse effect. For the sustainable

development, it is necessary to reduce CO2 and limit the use

of natural resources by recycling industrial by-products and

wastes such as silica fume, fly ash, slag, and recycled

aggregate from building rubbles. The turning rubble from

demolished buildings into fine and coarse aggregate used in

concrete reduces both the consumption of natural resources

and landfill for disposing of waste materials.

The performance of recycled aggregate concrete is

generally not as good as that of conventional concrete made

from natural aggregate [1–7]. It is noted that the properties

of concrete made from recycled aggregate strongly depend

on concrete mix proportion, sources, and quality of the

recycled aggregate [1–2]. Tabsh and Abdelfatah [1] stated

that the compressive and splitting tensile strength of recycled

coarse aggregate concrete were 10–25% lower than that of

the normal concrete. This was a similar trend when using

recycled concrete coarse aggregate from different sources

[2]. Etxeberria et al. [3] investigated the use of four different

recycled coarse aggregates by crushed concrete to replace

0%, 25%, 50%, and 100% natural coarse aggregate in

concrete mixtures. Their results showed that the

compressive strength of concrete made from 100% recycled

coarse aggregate was 20–25% lower than that of the

conventional concrete. Xiao et al. [4] indicated that

increasing the replacement level of recycled aggregate

resulted in a reduction in elastic modulus and compressive

strength of concrete. Casuccio et al. [5] studied the use of

recycled coarse aggregate obtained by crushing a normal

strength and high strength concrete. The compressive

strength and modulus of elasticity of the recycled coarse

aggregate concrete were about 1–15% and 13–18% lower

than those of the conventional concrete, respectively. Chen

et al. [6] recycled building rubble such as demolished

concrete, brick, and tile into coarse aggregate used in

concrete. The compressive strength of the concrete samples

prepared with such recycled coarse aggregates was 25–40%

below the strength of normal concrete. Furthermore, Khatib

[7] examined the use of recycled fine aggregate made from

demolished concrete and brick. Experimental results

indicated that the compressive strength of recycled fine

aggregate concrete was 10–30% lower than that of the

conventional concrete. Kou and Poon [8] investigated the

possible application of recycled concrete fine and coarse

aggregate in self-compacting concrete. Test results proved

the feasibility of utilization both recycled fine and coarse

aggregate in self-compacting concrete.

In order to enhance both physical and engineering

properties of the recycled aggregate concrete and reduce

the amount of cement used, the additive materials such as

silica fume, fly ash, slag, and metakaolin, which are the

industrial by-products, were added into concrete mixtures.

A previous study reported that the addition of 25% fly ash to

concrete mixture reduced the drying shrinkage, enhanced

the compressive strength, and increased the resistance to

chloride ion penetration of the concrete [9]. Kou et al. [10]

pointed out that the use of fly ash as cement substitution

reduced the compressive strength, tensile strength, and

elastic modulus, however, increased the resistance to

chloride ion penetration and decreased the drying shrinkage

and creep of recycled aggregate concrete. The use of a

combination of mineral admixtures such as fly ash and

silica fume [11], fly ash and slag [12], and silica fume, fly

ash, slag and metakaolin [13] enhanced both physical and

engineering properties of recycled aggregate concrete.

The previous studies have demonstrated the possibility of

recycling all building rubbles into the useful fine and coarse

aggregates used in concrete. The properties of concrete made

from such the recycled aggregates were normally lower than

properties of the conventional concrete; however, the concrete

20 Huynh Trong Phuoc, Nguyen Tien Dung, Ngo Si Huy, Vo Duy Hai

properties were improved with the inclusion of mineral

additives. In order to enhance not only the mechanical

properties but also the durability of concrete, Hwang and

Hung [14] have proposed the Densified Mixture Design

Algorithm (DMDA) method to design concrete mixture

proportion. This method has been considered as a green

design method for high-performance concrete (HPC) with

increasing the physical density and reducing the amount of

cement used. Thus, the primary objective of this study is to

evaluate the properties of HPC designed by the DMDA

method with the incorporation of recycled aggregates.

2. Materials and test methods

2.1. Materials

A mixture of type-I ordinary Portland cement (OPC),

class-F fly ash (FA), and ground granulated blast furnace slag

(GGBFS) are used as binder materials for producing concrete

samples. All of the binder materials used in the present study

are sourced from Taiwan with characteristics as presented in

Table 1. Table 2 shows the sieve analysis and fineness

modulus (FM) of both natural and recycled aggregates, while

their physical and mechanical properties are shown in Table

3. It is noted that the recycled fine and coarse aggregates are

provided by a local construction material company in

Taiwan, which was made from a mixture of demolished

crushed concrete, brick, and tile. As shown in Table 3, the

recycled aggregates have properties that are not as good as

those of natural aggregates. This is attributable to the high

porosity of recycled aggregates compared with natural

aggregates. In other words, the quality of the recycled

aggregates is not as good as that of the natural aggregates. In

order to reduce the amount of water, but keep the desired

workability of the fresh concrete mixture, type-G

superplasticizer (SP) with a specific gravity of 1.1 is used.

Table 1. Characteristics of binder materials

Materials OPC FA GGBFS

Specific gravity 3.15 2.29 2.90

Chemical

composition

(wt.%)

SiO2 20.0 64.0 35.6

Al2O3 4.2 22.1 11.2

Fe2O3 3.1 5.6 0.5

CaO 62.4 2.7 41.0

MgO 4.1 0.9 6.4

SO3 2.9 0.6 0.9

Others 1.4 1.1 2.1

Table 2. Sieve analysis and fineness modulus of the aggregates

Sieve size

(mm)

Percentage of

passing, (%) Sieve size

(mm)

Percentage of

passing, (%)

NFA RFA NCA RCA

4.75 (#4) 99.3 98.7 19 (3/4 in) 100 100

2.36 (#8) 76.1 77.8 12.5 (1/2 in) 75.0 94.1

1.18 (#16) 53.4 55.8 9.5 (3/8 in) 46.6 64.4

0.6 (#30) 38.2 36.9 4.75 (#4) 11.0 13.5

0.3 (#50) 23.8 20.5 2.36 (#8) 3.9 4.1

0.15 (#100) 14.3 11.7 FM 6.2 5.8

FM 3.1 3.0

Note: NFA = Natural fine aggregate; RFA = Recycled fine

aggregate; NCA = Natural coarse aggregate; RCA = Recycled

coarse aggregate; FM= Fineness modulus.

Table 3. Physical and mechanical properties of the aggregates

Physical properties NFA RFA NCA RCA

OD density (kg/m3) 2640 2540 2646 2589

Absorption capacity (%) 1.5 5.6 0.7 3.2

Crushing strength (MPa) NA NA 61.5 44.3

Note: OD = Oven dry condition; NA= Not available.

2.2. Mix proportions and test methods

In this study, all of the mix proportions are designed using

the DMDA method with the same water-to-binder ratio of

0.35. The design concept and the design procedures of

DMDA method were previously published by Hwang and

Hung [14]. In this method, concrete mix proportions are

divided into two phases as aggregate and paste. The aggregate

phase consists of coarse aggregate, fine aggregate, and fly ash,

which forms the major skeleton of the concrete structure. The

void among coarse particles is filled by fine aggregate and fly

ash to minimize the porosity. Meanwhile, the paste phase

includes cement, slag, water, and superplasticizer, which is for

lubricating and filling pores to achieve concrete workability.

The key point of DMDA method is to determine the optimized

amount of concrete ingredients by experimental work so that

concrete samples have the highest density as well as good

quality. It also means that the amount of each concrete

ingredient is different from a mixture to another mixture even

they have the same a water-to-binder ratio.

Table 4 shows the mix proportions for all concrete

mixtures, which includes two groups. The first group is

designed with natural fine aggregate, while the second group

was designed with recycled fine aggregate. In each group,

recycled coarse aggregate content is 0%, 30% and 100% of

the total coarse aggregate amount. The M1, M2, and M3

mixtures denote concrete designed with 100% natural fine

aggregate and 0%, 30%, and 100% recycled coarse aggregate,

respectively. Meanwhile, M4, M5, and M6 mixtures denote

concrete designed with 100% recycled fine aggregate and 0%,

30%, and 100% recycled coarse aggregate, respectively. It is

noted that the total binder amount increases with increasing

the recycled aggregates content. In details, the total binder

amount of group II mixtures is higher than those of

corresponding group I mixtures. Due to the high porosity of

recycled aggregate, all of the concrete mixtures reach the

highest density incorporating a high amount of binder.

Table 4. Mixture proportion for preparing concrete samples

Mixture Concrete ingredient proportions (kg/m3)

OPC Slag FA CAg FAg Water SP

Group

I

M1 220.2 55.1 117.1 786.8 1123.5 133.4 4.1

M2 322.6 80.6 105.2 453.2 1250.5 174.8 3.3

M3 363.4 90.8 60.2 813.7 879.3 177.5 2.4

Group

II

M4 293.6 73.4 190.2 665.4 894.4 191.1 4.1

M5 433.6 108.4 115.9 398.2 998.2 227.0 3.2

M6 363.1 90.8 121.0 763.1 775.1 197.5 3.5

Note: CAg = Coarse aggregate; FAg = Fine aggregate

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 21

2.3. Test programs

Cylinder concrete samples with 10 cm in diameter and

20 cm in height are prepared. The properties of fresh and

hardened concrete such as workability, compressive

strength, water absorption, thermal conductivity, ultrasonic

pulse velocity, and electrical surface resistivity are tested.

The concrete compressive strength is measured at 7, 14,

and 28 days, while other properties of the samples are

measured at 28 days. The values presented herein are the

average value of three concrete samples.

3. Results and discussion

3.1. Fresh concrete properties

The designed workability of the fresh HPC mixtures is

controlled in the range of 22 ± 2 cm by using various dosages

of SP. Thus, right after mixing, the fresh HPC mixtures are

checked for the slump, slump flow spread, and flow time

with the results as provided in Table 5. It is noted that the

workability of fresh concrete is tested in accordance with

ASTM C143. As a results, the slump, slump flow spread,

and flow time of all concrete mixtures are in the ranges of

21–23 cm, 32–36 cm, and 5–10 seconds, respectively. Thus,

all of these mixtures show good fresh properties.

Table 5. Properties of fresh concrete mixtures

Mixture Slump

(cm)

Slump

flow (cm)

Flow time

(sec.) % SP

Group

I

M1 21 36 10 1.09

M2 23 34 5 0.64

M3 21 32 7 0.45

Group

II

M4 23 35 10 0.86

M5 23 36 5 0.48

M6 22 32 7 0.64

3.2. Compressive strength development

Compressive strength is a very important property of

hardened concrete. Figures 1 and 2 show the compressive

strength development of the HPC samples in group I and

group II, respectively. It is observed that the compressive

strength of all concrete mixtures increases with curing ages,

with the strength values ranging from 31.7 to 56.7 MPa.

Figure 1. Compressive strength development of

the HPC samples in group I

Figure 2. Compressive strength development of

the HPC samples in group II

As shown, the compressive strength of M5 and M6

mixtures were lower than that of the M2 and M3 mixtures,

respectively. This phenomenon is due to the use 100%

recycled fine aggregate in group II mixtures as compared to

the use of 100% natural fine aggregate in group I mixtures.

However, the compressive strength of M4 mixture was higher

than that of M1 mixture because the total binder amount of

M4 mixture is 42% higher than that of M1 mixture.

For group I, the HPC mixture with 30% recycled coarse

aggregate content (M2) show the highest compressive

strength, followed by mixture with 100% recycled coarse

aggregate content (M3) and recycled coarse aggregate-free

mixture (M1). It is noted that the total amount of binder of

M2 and M3 mixtures is around 30% higher than that of

control mixture (M1), which is used to minimize the

porosity of recycled aggregate concrete structure. This is

mainly attributable to the 43% and 35% higher

compressive strength of M2 and M3 mixtures compared

with the M1 mixture, respectively. It means that using

DMDA method not only compensates the negative effects

of the recycled coarse aggregate on concrete strength but

also improves the strength of the recycled aggregate

concrete. This finding is in line with the experimental

results from a previous study [9]. With a similar amount of

binder, M2 mixture with the use of 70% recycled coarse

aggregate has the compressive strength value of around 6%

higher than that of the M3 mixture with the use of 100%

recycled coarse aggregate. It means that increasing the

replacement level of recycled coarse aggregate results in a

reduction in the concrete strength. This finding is in good

agreement with previous studies [1–7].

For group II, all of the HPC mixtures are designed with

100% recycled fine aggregate and using recycled coarse

aggregate to replace 0%, 30%, and 100% natural coarse

aggregate. Among three mixtures in group II, M4 mixture has

the highest compressive strength, followed by M5 and M6

mixtures. It reveals that compressive strength of the concrete

samples reduced with increasing the content of recycled

coarse aggregate. This phenomenon is associated with the

high porosity of recycled aggregate [1–7]. However, all

recycled aggregate concrete mixtures examined herein show

similar compressive strength value of around 44.5 MPa at 28

days. Similar to the group I mixtures, this result is based on

the fact that the total binder amount of M5 and M6 mixtures

22 Huynh Trong Phuoc, Nguyen Tien Dung, Ngo Si Huy, Vo Duy Hai

is higher than that of M4 mixture. This finding also proves

the feasibility of using DMDA method to improve the

properties of HPC made wfrom recycled aggregate.

3.3. Water absorption

Water absorption is an important parameter reflecting the

permeable property of concrete. Concrete with low water

absorption will have high resistance to the sulfate ions,

chloride ions, alkali ions, and other harmful substance,

which causes a chemical attack in concrete. The test result

of the water absorption of the hardened concrete mixtures at

28-day ages is presented in Table 6. Generally, group II

concrete mixtures show a higher water absorption rates than

group I concrete mixtures. This is because the 100%

recycled fine aggregate is used to replace natural fine

aggregate in group II mixtures. For group I, the M1 and M3

mixtures have a similar water absorption level, whereas M2

shows the lowest water absorption rate. This is mainly due

to the use of more binder amount and less recycled coarse

aggregate content in M2 mixture as compared to those in M3

mixture. For group II, the water absorption increases with

increasing the recycled coarse aggregate content.

Table 6. Properties of hardened concrete at 28-day age

Mixture

Hardened concrete properties

WA

(%)

TC

(W/mK) UPV (m/s)

ESR

(kΩ.cm)

M1 1.83 1.78 4481 32.1

M2 1.79 1.88 4741 36.9

M3 1.86 1.61 4452 32.3

M4 2.58 1.30 4420 30.5

M5 3.00 1.27 4370 35.0

M6 3.35 1.22 4297 38.9

Note: WA = Water absorption; TC = Thermal conductivity;

UPV = Ultrasonic pulse velocity; ESR = Electrical surface

resistivity

The water absorption strongly depends on the pore

structure and the cracks in concrete. Recycled fine and

coarse aggregates initially have higher porosity and water

absorption capacity than natural fine and coarse aggregates

(Table 3). Therefore, the concrete mixtures with higher

recycled aggregate content exhibited higher water absorption

level. However, the porosity of M2 and M3 mixtures is

reduced by increasing the amount of binder, thus these

mixtures have a good permeable quality as compared with

the control mixture (M1).

3.4. Thermal conductivity

The results of thermal conductivity measurement for all

HPC mixtures are shown in Table 6. In general, the thermal

conductivity of all the examined concrete mixtures ranges from

1.22 to 1.88 W/mK. As a result, group II concrete mixtures

have lower thermal conductivity values than that of group I

concrete mixtures. The thermal conductivity is associated with

the density of concrete and inversely related to water

absorption. The increasing porosity of concrete samples due to

the replacement of natural aggregate by recycled aggregate

reduces the thermal conductivity of the concrete.

In group I, the M2 mixture show the highest value of

thermal conductivity because of the highest amount of binder

used as above mentioned. In group II, the thermal conductivity

of the HPC decreases with the increase in recycled coarse

aggregate content. The high porosity of the recycled aggregate

as aforementioned is attributable to a lower thermal

conductivity. However, the difference of thermal conductivity

value among the mixtures in each group is insignificant due to

the effectiveness of DMDA mix design method.

3.5. Ultrasonic pulse velocity

The ultrasonic pulse velocity (UPV) test is often used

for assessing the uniformity and the relative quality of

concrete, which relates to the presence of voids and cracks

inside the concrete structure. The results of UPV test of the

HPC samples are presented in Table 6. As shown, all of the

concrete samples have UPV values of above 3660 m/s at

28-day ages, indicating a good durability as suggested by

Malhotra [15]. In a similar trend to thermal conductivity,

M2 mixture registers the highest UPV value, whereas M1

and M3 mixtures have similar UPV values. That is also due

to the less porosity, as mentioned above, of M2 mixture in

comparison with that of the other mixtures. With the use of

recycled fine aggregate to replace natural fine aggregate,

group II concrete mixtures exhibit lower UPV values than

group I concrete mixtures.

3.6. Electrical surface resistivity

The electrical surface resistivity (ESR) of concrete is an

important factor to evaluate the corrosion resistance of the

concrete. The test results for the ESR of the HPC samples

are shown in Table 6. As shown in the table, the ESR

values of all concrete mixtures range from 30.5 to

38.9 kΩ.cm. Abdefatah and Tabsh [16] previously reported

that concrete had excellent anti-erosion ability if its ESR

value was above 20 kΩ.cm. This means that all of the

recycled aggregate concrete mixtures used in this investigation

exhibit excellent electrical resistivity. In other words, all

concrete mixtures investigated herein demonstrate an

excellent anti-erosion ability.

4. Conclusions

Properties of the HPC using recycled aggregates are

evaluated in this study. The obtained results lead to the

following conclusions:

- With the use of various SP dosages, all of the fresh

concrete mixtures made from recycled aggregate exhibit a

good workability with a less amount of water.

- All of the HPC samples show the good performance

in terms of compressive strength, which ranges from

31.7 MPa to 56.7 MPa. The use of DMDA method to

design recycled aggregate concrete mixtures is found to

have a positive improvement of concrete strength.

- This study finds that increasing recycled aggregate

content leads to increasing the water absorption capacity

and electrical surface resistivity, but reducing the thermal

conductivity and ultrasonic pulse velocity of the HPC

samples. However, the addition of binder can compensate

for or even improve properties of recycled aggregate

concrete.

- All of the concrete samples examined in this study

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 23

exhibit the good durability performance and excellent anti-

erosion ability.

- The application of DMDA method for mix design of

HPC with recycled aggregate effectively enhances both

engineering properties and durability performance of the

concrete.

REFERENCES

[1] Tabsh S.W, Abdelfatah A.S. Influence of recycled concrete aggregates on strength properties of concrete. Construction and Building

Materials, 2009; 23: 1163–1167.

[2] Pedro D, Brito J.D, Evangelista L. Influence of the use of recycled concrete aggregates from different sources on structural concrete.

Construction and Building Materials, 2014; 71: 141–151.

[3] Etxeberria M, Vazquez E, Mari A, Barra M. Influence of amount of

recycled coarse aggregates and production process on properties of recycled aggregate concrete. Cement and Concrete Research, 2007;

37: 735–742.

[4] Xiao J, Li J, Zhang C. Mechanical properties of recycled aggregate concrete under uniaxial loading. Cement and Concrete Research,

2005; 35: 1187–1194.

[5] Casuccio M, Torrijos M.C, Giaccio G, Zerbino R. Failure mechanism

of recycled aggregate concrete. Construction and Building Materials, 2008; 22: 1150–1506.

[6] Chen H.J, Yen T, Chen K.H. Use of building rubbles as recycled

aggregates. Cement and Concrete Research, 2003; 33: 125–132.

[7] Khatib J.M. Properties of concrete incorporating fine recycled

aggregate. Cement and Concrete Research, 2005; 35: 763–769.

[8] Kou S.C, Poon C.S. Properties of self-compacting concrete prepared

with coarse and fine recycled concrete aggregates. Cement and

Concrete Composites, 2009; 31: 622–627.

[9] Kou S.C, Poon C.S, Chan D. Influence of fly ash as a cement addition on the hardened properties of recycled aggregate concrete.

Materials and Structures, 2008; 41: 1191–1201.

[10] Kou S.C, Poon C.S, Chan D. Influence of fly ash as cement replacement on the properties of recycled aggregate concrete.

Journal of Materials in Civil Engineering, 2007; 19: 709–717.

[11] Corinaldesi V, Moriconi G. Influence of mineral additions on the

performance of 100% recycled aggregate concrete. Construction and Building Materials, 2009; 23: 2869–2876.

[12] Ann K.Y, Moon H.Y, Kim Y.B, Ryou J. Durability of recycled

aggregate concrete using pozzolanic materials. Waste Management, 2008; 28: 993–999.

[13] Kou S.C, Poon C.S, Agrela F. Comparisons of natural and recycled

aggregate concretes prepared with the addition of different mineral

admixtures. Cement and Concrete Composites, 2011; 33: 788–795.

[14] Hwang C.L, Hung M.F. Durability design and performance of self-

consolidating lightweight concrete. Construction and Building

Materials, 2005; 19: 619–626.

[15] Abdefatah A.S, Tabsh S.W. Review of research on and implementation

of recycled concrete aggregate in the GCC. Advances in Civil

Engineering, 2011; doi:10.1155/2011/567924.

[16] Malhotra V.M. Testing hardened concrete: Nondestructure methods, American Concrete Institute Monograph Series, 1976.

(The Board of Editors received the paper on 27/8/2017, its review was completed on 02/10/2017)

24 Hoang Trung Ngo, Thien Duc Ngo, Danh Ngoc Nguyen, Hoai Nam Le

DESIGN AND CONTROL OF AN ACTIVE PROSTHETIC LEG

Hoang Trung Ngo1, Thien Duc Ngo2, Danh Ngoc Nguyen3, Hoai Nam Le3 1FPT Global Automation; hoangtrungngovan.td@gmail.com

2R&P Logistics Company Limited; ngothienduc94@gmail.com 3Univesity of Science and Technology, The University of Danang; ndngoc@dut.udn.vn, lehoainam@dut.udn.vn

Abstract - The purpose of this paper is to present the design, prototype and control of an active prosthetic leg. The research scope encompasses the mechanical design, the electronic design and the control system design. The mechanical design covers kinematic diagram design, dynamic system modelling and 3D design. Meanwhile, the electronic one includes PCB mainboard design and sensor system design. The control system design consists of system parameter estimation and self-tuning regulator adaptive controller design, creating the adaptive controller which plays a role as the brain of the prosthetic leg to control the whole system. The main objective of this work is to produce a very first prototype of an active prosthetic leg which mimics the movement, experimenting on a healthy man before being adopted on amputees. In addition, the prosthetic leg can be also served as a test platform for testing the control methods.

Key words - Active prosthesis; control design; adaptive control; leg trajectory; communication protocols.

1. Introduction

In order to improve the mobility of people with limb

amputation, intensive research activities on the prosthetic

field have been developed during the last decades. Various

prosthetic designs were introduced which can be divided

into two kinds of devices: passive and active. Passive

prosthetics require the user to move it with their own effort

[1]. This can make the disabled people feel uncomfortable.

In such a case, different kinds of active prosthetic with

energy sources and auxiliary sensors were developed and

commercialised [2]. The prototype SmartLeg was

introduced [3] which applies machine learning to provide

optimal gait to the user by active movement controlling in

knee and ankle joints through embedded hydraulic

actuators. [4] presents the mechanical design and controller

design of a semi-active above-knee. Somes different kinds

of control algorithm for active prosthetic leg were proposed

[5, 6, 7]. Current active prosthetics such as the Ottobock

C-Leg offer a high level of comfort for patients, but most

of them have a high price for most customers, especially

ones in Vietnam. The objective of this work is to develop

a flexible active above-knee prosthetic leg model with a

considerable lower cost. To this aim, a prototype with a

sophisticated mechanism of above-knee [8] was built with

3D printing technology. For the control purpose, a novel

leg’s trajectory construction method and an adequate

adaptive control design are proposed. Besides, a variety of

system modelling methods, electronic interface protocols,

PCB design as well as mechanical design were brought into

the leg. The developed model can be also served as a

platform for testing and applying the control methods,

notably for mechatronic students.

The remainder of this paper is organised as follows.

Section II is devoted to the description of the system

design. Section III describes the characteristics of the built

prototype. Some numerical results are also discussed here.

Finally, the conclusion drawn from this work and possible

ways for further studies are given.

2. System design

This section describes the overall structure of the

designed prosthetic leg and its control system.

2.1. General structure

Figure 1 presents the active prosthetic leg overall

structure which includes two principal parts: offline and

online process. The online process is the indispensable task

as in every active prosthetic. Firstly, the sensors acquire

information about the leg’s state, then convert it into the

data which is transfered to the set point generator. The

generator compares the data with the leg’s trajectory

dictionary, produces set points for above-knee and ankle

and then converts them into screw nut positions. Based on

these desired positions, the controller controls the

prosthetic system. Besides, different states of the leg would

have a different need of force because of various loads;

therefore, the controller must use the data from sensors to

decide how much energy should be used for the

contemporary state.

In addition, because the human leg has an untold

number of states, an offline process must be introduced to

undertake this crucial task, collecting the various states of

the human leg to conduct a leg’s trajectory. This process is

done using motion tracking sensor system attached to the

human leg to detect the angles of thigh, calf and foot.

Motion tracking

system

Data

processing

Leg s

trajectory

dictionary

Angle of

thigh, calf

and foot

Offline process

Prosthetic

system

Self-tuning PID

Controller

Set point

generatorSensors

Online process

Figure 1. The general structure of designed prosthetic leg

2.2. Mechanical design

In order to build a prosthetic leg, a kinematic diagram

was introduced. As illustrated in Figure 2, the prosthetic leg

can be divided into two main mechanisms: the above-knee

and the ankle. Besides, two axes were put into the diagram

in order to calculate the angles of thigh, calf and foot.

In the human leg, the flexible of above-knee comes from

ligaments, but the contact place of two bones and muscles are

responsible for the body’s load [9]. By the way of changing

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 25

the shape, the quadrilateral structure plays the same role as

ligaments in the above-knee mechanism. In addition, a cam

was integrated into the quadrilateral structure to act as the

contact place of two bones to strengthen the load bearing

capacity. Finally, the screw mechanism and timing belt with

motor operate like muscles to control the leg’s movement in

a considerably adequate way.

Figure 2. Kinematic diagram of the prosthetic leg

In terms of movement, the angle between the calf and

the foreside b of the quadrilateral structure varies in a small

range (less than 15 degrees). In terms of dimension, the calf

is far longer than b. From these points, without changing

the whole structure of kinematic diagram, the combination

structure of quadrilateral structure and calf can be

completely simplified to become a pivot with its heart

located in the intersection of the thigh and b (of the

quadrilateral structure).

Screw

nut 1F1

L1

x1

H1

E1

G1

θ1

E2

F2

G2

H2

Srew

nut 2

x2

L2

I2

90o

θ2

α2

(a) (b)

Figure 3. Brief kinematic diagram: (a) the lower leg part with

ankle and (b) the higher leg part with above-knee

In order to determine the leg’s movement, both

movement sensors and encoder sensors (which track the

movement by evaluating the position of screw nuts) can be

used. So, how to evaluate the exact movement of the leg by

encoders? In fact, encoders can only indirectly calculate

the position of screw nuts by determining the rotation of

motors. At the same time, we must have some functions to

transform the desired knee’s angle and ankle’s angle into

the desired position of screw nuts. Figure 3(a), 3(b) and the

following functions have met these urgent needs.

Table 1. The dimension of prosthetic leg’s prototype

Symbol Description

1 Angle between calf and foot

E1 Distance between the screw 1 and the pivot

F1 Distance between the pivot in E1 and ankle

G1 Distance between two pivots of foot

L1 Distance between the pivot and the screw nut 1

H1 Distance between the pivot in foot and E1

x1 Distance between the screw nut 1 and position switch

2 Angle between thigh and calf

2 Angle between E2 and F2

E2 Part of prosthetic thigh to fasten the pivot

F2 Part which sticks to the remnant human thigh’s part

G2 Distance between two pivots in the higher leg part

H2 Distance between the screw 2 and the pivot

L2 Distance between the pivot and the screw nut 2

I2 Distance between pivot in calf with H2

x2 Distance between the screw nut 2 with position switch

Figure 3(a) includes:

2 2 2 2

1 1 1 1

1

1 1

1

2.arccos 902. .

    90 180 1' '   80ang

F

le t

G H E

F G

cal high anf ls s g e

By using the data from movement tracking sensors, we

have the calf’s angle and the thigh’s angle. From (1), we

can completely calculate H1. x1 can be then calculated

using the following equation:

1 1 1x H L

Like the Figure 3(a), Figure 3(b) includes:

2 2 2 2 2 2 2

2 2 2 2 2 2 2

2

2 2 2 2

2    

arccos arccos2.

180  

. 2. .

' '  

K F E K G

angle cal

I H

K F K G

thig f anglh s s e

For more details, 2 2

2 2 2 2 2 22. . .cosK E F E F ,

and x2 can be then calculated by the following equation:

2 2 2x I L

x1 and x2 can be used for controlling the prosthetic leg’s

states. To mimic the movement of a real leg, a leg’s

trajectory or phases of human gait cycle [1,10] must be

built in the offline process.

The self-tuning PID controller is embedded in dsPIC33

will calculate and conduct voltage value to control the

motor’s rotation through the driver according to the position

of screw nuts. The encoder data feedback is used to ensure

these positions are accomplished. Apart from that, an RF

module is used to send data of prosthetic leg to another CPU

for visualizing and building the initial transfer functions.

26 Hoang Trung Ngo, Thien Duc Ngo, Danh Ngoc Nguyen, Hoai Nam Le

2.3. Control system design

Figure 4 describes the prosthetic leg’s system

architecture, including control algorithm and sensors. The

motion tracking sensor system measures the contemporary

state of the leg. Based on this state, the last nearest state of

leg and the leg’s trajectory dictionary, the setpoint

generator produces the desired calf’s angle and the foot’s

angle. These desired angles can be then converted into the

positions of screw nuts using the equations from (1) to (4).

Thigh s angle Foot s angle

Setpoint

generator

UART

Leg s trajectory

dictionary

Driver

Motor 1

Voltage

Motor 1

Voltage

Encoder 1 Encoder 2Screw nuts

position

Calf s angle

Motion tracking

sensor system

Online

estimation

PID

Controller

Comparator

Controller

design

Self-tuning

regulator

PWM

Figure 4. The system architecture of the prosthetic leg

The main objects in the prosthetic leg system are two

motors. With the purpose of motor modelling, a second

order transfer function is used [12]:

1 2

2

1 2

V

Y z b z bG z

U z z a z a

Y(z), U(z) are the Laplace transform of the velocity of

screw nuts y(k) and the PWM value u(k), respectively.

As mentioned above, in the process of human walking,

the load force varies depending on the different leg’s states.

In the other words, the parameters a1, a2, b1, b2 will

probably not be constant. Thus, with the purpose of

stabilising and controlling prosthetic leg, an estimation

parameter process must be introduced.

Set 1 2 1 2

T

k a k a k b k b k is

parameters vector of the above-knee (or ankle) at k time

and set 1 2 1 20 0 0 0 0

T

a a b b .

From (5):

1 2 1

2

. 1 . 2 . 1

. 2

y k a y k a y k b u k

b u k

Parameters vectors can be easily updated by using

recursive least squares filter [11,13]:

1 .k k L k k

. 1Tk y k k k

1 .

1 .T

P k kL k

k P k k

11TP k I L k k P k

With:

1 2

1 2

1 . 1 1 . 2

1 . 1 1 . 2

k a k y k a k y k

y k b k u k b k u k

1 2 1 2    T

k y k y k u k u k

For adapting the rapid dynamic of the system, a self-

tuning PID controller is adopted. At each instant, the set-

point generator uses the measurement of prosthetic leg and

the leg’s trajectory dictionary to determine the set-point

(angles of thigh, calf and foot). Using the estimated

prosthetic leg transfer function, a self-tuning regulator is

used to change the parameters of the PID controller. The

control system can be illustrated by Figure 4.

According to (5), open system transfer function

illustrates the relationship between the position of screw

nut (x1 and x2) and the PWM value:

1 2

2

1 2

. .1

P x

b z b TG z t

zz a z a

With: T – sample time, tx – dimension of the screw’s pitch.

To control the screws, a PI controller is clearly simple and

sufficient, with the controller’s transfer function [11, 13]:

. 1.

2 1

IC P

K T zG z K

z

The characteristic equation of the closed loop:

1 . 0C PG z G z

From (13), (14) and (15) we have:

3 2

1 1

2 1 2 1 1

2 2 2

(2. . ) 2 1 1

2 2. . .

. . 2. . 2. . . 0

P I P I

P I x

I P x

K T K z a K T a K z

a a K T a a K T t b z

a T K a K T t b

Open system (13) is a third order strictly proper transfer

function with a pole of z = 1. According to [13], the further

the pole of the system is from the imaginary axis, the

smaller the time constant and the faster the time responds

to the system. As a result, the system can be approximate

to the second order system. In this research, the quality

control includes:

Damping constant: with the ITAE standard, the second

order system should completely have a damping constant

ξ = 0.707.

Natural oscillation frequency: according to 5%

standard, with the average setting time of 1s:

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 27

3 31 4.243

. n

tn

Choose ωn = 5 (rad/s)

From (16) the characteristic equation is a third order

one. In addition, this equation must have two conjugate

poles corresponding to ξ = 0,707 and ωn = 5:

2 2

2 2

3 2

2 0

2 0.707 5 5 0

7.07 2 7.07 25 25 0

n nz a z z

z a z z

z a z a z a

From (16) and (17), we have:

1 12 1 . 1 .

7.07 2. .

P I

P I

a K T a K

a K T K

2 1 2 1 12 . . 2. . .

25 7.07 2. .

P I x

P I

a a K T a a K T t b

a K T K

2 2 22. . . . 2. . .

25. . 2  .  .  

P I x

P I

a K T a K T t b

a K T K

From 7,07×(18)–(19) and 25×(18)–(20), we eliminate

the variable a:

1 2

1 2 1

16.14 2 64.14 .

6.07 17.93 . 2. .

P

I x

a a K

T a a K T t b

1 2

1 2 2

50 2 403.5 .

       50 126.75 . 2. .

P

I x

a a K

T a a K T t b

Set:

1 2 1 2

1 2 1 2

16.14 2 64.14 6.07 17.93

50 2 403.5 50 126.75

a a T a a

a a T aA

a

1

2

2. .

2. .

x

x

T t b

TB

t b;

P

I

KX

K

We have the equation: .A X B (23)

Finally:

1.P

I

KX A B

K

In conclusion, corresponding to system, parameters

have been continuously estimated and the PI controller’s

parameters KP and KI are also updated by using the

equation (24).

3. Implementation and result

3.1. 3D prosthetic leg prototyping

The entire prosthetic leg is fabricated from plastic using

3D printing technology. A sophisticated flexible

mechanism which consists of a cam and a quadrilateral

structure was used for imitating the above-knee, called

“remotion knee” [8]. Figure 5 illustrates the 3D model in

CAD software, the real prototype and the real prototype

attached to a human body for operational testing.

(a) (b) (c)

Figure 5. (a) 3D model, (b) real prototype with motion tracking

sensors, (c) real prototype attached on human

The prototype hardware configuration is resumed in

Figure 4. The motors are high-torque DC brushed motors

with a peak operating speed of 468 RPM. To measure the

angle of the thigh, we use an InvenSense MPU6050 sensor

[14, 15] which contains a MEMS accelerometer and a

MEMS gyroscope in a single chip. The foot and calf angles

are deducted by measurements of encoder integrated into

two motors which move the foot and the calf joints. To

control the motors, to process the data from sensors and to

perform the self-tuning PID controller, a 16-bit PIC

microcontroller is used [16].

Table 2. Kinematic dimensions of the prosthetic leg

Symbol Dimension Symbol Dimension

E1 17,0 mm E2 70,0 mm

F1 136,0 mm F2 35,5 mm

G1 35,4 mm G2 200,2 mm

L1 94,4 mm H2 25,0 mm

2 90,0° L2 126,4 mm

The PCB was designed by Altium designer software,

including the main dsPIC33 microcontroller as in Figure 6.

Figure 6. The main controlling board of prosthetic leg

3.2. The leg’s trajectory

In order to obtain a leg’s trajectory dictionary, the

angles of thigh, calf and foot of a healthy typical adult were

collected and processed by a motion tracking sensor system

as in the offline process part in Figure 1.

There are some essential reasons why we use sensors to

track the angles of thigh, calf and foot instead of hip, knee,

ankle joint angles. Firstly, the sensors used for tracking hip,

knee and ankle joint angles are probably the contacting

driver

microcontroller

28 Hoang Trung Ngo, Thien Duc Ngo, Danh Ngoc Nguyen, Hoai Nam Le

ones (e.g. resistive sensors) or must be the absolute

encoders. The former type will be gradually eroded, while

the encoders are relatively hard to assembly into prosthetic

knee without bloating the knee’s volume. Besides, the

encoders with an adequate precision are relatively

expensive, and it is hard enough to measure the angle of

knee joint because of the complicated structure of

“remotion knee”. So, we should use a motion tracking

sensor system to track the thigh, calf and foot’s angles, then

convert it to the angle of hip, knee, ankle joint angles. After

that, we transform this relationship into the position of the

screw nuts. In order to ensure the accuracy of the

converting process, we utilise 3D simulation (SolidWorks)

before manufacturing using 3D printing technology.

Figure 8 shows how the data of leg’s movement was

collected (angles of the thigh, the calf and the foot) by a

motion tracking sensor system (Figure 7) using InvenSense

MPU6050 sensors and Arduino.

Figure 7. Motion tracking sensor system

MPU6050 MPU6050 MPU6050

Angle (thigh) Angle (calf) Angle (foot)

Arduino

Pro mini

Arduino

Pro mini

Arduino

Pro mini

I2C I2C I2C

Arduino

Pro mini

SOFT UART

SOFT UARTSOFT UART

CPU

(Arduino IDE software)

COM to USB

Dictionary

(.txt) Figure 8. Algorithm diagram of sensor system and

processing data

There are a few stages of investigating and

transforming the data to a dictionary have been using in

leg’s controller. Firstly, data was collected at least three

times each type of topographic. Next, Microsoft Excel was

used to illustrate data as a line graph showing the

relationship of three angles of leg in comparison with the

axes (thigh, calf, foot).

The obtained measurements of leg’s movement can be

represented by the angles’ relation graph (Figure 9). The

angles in this graph were adjusted to make the angles of

thigh, calf and foot have the same value of -90 degree as

the patient state’s standstill. The data analysis bases on the

different states of value in Figure 9 rather than the human

gait. As seen in Figure 9, we can divide the data into 4

phases based on the value of thigh’s angle and its velocity

as well as calf’s angle velocity. These 4 stages can be then

easily converted to a continuous function, much more easy

to implement to the microcontroller (MCU) than

electromyography method [5, 9], and still gives the leg a

relatively trusty trajectory.

This relationship can then be transformed to the

relationships of 1 and 2 (in Figure 2) with the thigh’s

angle and calf’s angle. Next, these correlations can be

converted to the positions of screw nuts, and the two

motors will do the rest of tasks to provide the prosthetic leg

with an adequate movement.

Figure 9. The relationship between the angles of thigh,

calf and foot

3.3. Initial transfer function of the system

At the initial state of the prosthetic leg, when the human

state is a standstill, we collect the data and estimate the

initial values for the systems:

Above-knee transfer function:

1 2

0.06068 0.072

0.3656 0.4994V

Y z zG z

U z z z

Ankle transfer function:

2 2

0.1318 0.1463

0.5266 0.2994V

Y z zG z

U z z z

These transfer functions play a role as initial values of

transfer functions of the prosthetic leg, and the parameters

will be continuously updated by using recursive least

squares filter in 2.3.

3.4. The performance of the self-tuning PI controller

Figure 10 illustrates the performance of the prosthetic

leg with sample time of 50ms. There are some disorders in

the first 5s of the operation process because of the leg

initializing.

With the average time instant as 1s, the performance of

the above-knee is adequate with the average static error of

only over 10%. The first state of disabled human’s legs is

standing, so the errors tend to be larger than 0.

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 29

The performance of ankle has a quite large average static

error (27%) but the error is considerably stable. The

performance is probably acceptable because the static error

must be accepted to maintain the steady movement of the

prosthetic leg, which is mainly used to walk rather than flaunt.

As we have mentioned, the static error comes from the

way we implement the algorithms into MCU, with some

anti-wind-up and stabilised filter to preserve the prosthetic

leg’s stabilisation. Besides, the adaptive self-tuning PI

controller is still not optimised, and the rounding of MCU in

calculation also plays some part in producing the static error.

Figure 10. The performance of the prosthetic leg

4. Conclusion

In this paper, a sophisticated and flexible active

prosthetic leg model was design and built. The difference

between leg’s movement and on leg’s trajectory is to stay

to an acceptable degree. Besides, the control theories are

successfully adapted to the prosthetic leg, revealing as

system modelling and self-tuning adaptive PID controller.

Future research will focus on the developing of a switch

mechanism for battery used and non-battery used operation

mode. Another perspective relates to expanding the

capability of the prosthetic leg to move on undulating

surface and stairs. Finally, many functions (e.g. accelerating

speed, medical tracking sensor…) can be integrated to the

prototype to enhance the comfort and safety of amputees.

REFERENCES

[1] Alzaydi, Ammar A., et al. "Active prosthetic knee fuzzy logic-PID

motion control, sensors and test platform design”, International

Journal of Scientific & Engineering Research 2.12 (2011): 1-17.

[2] Lim, James. "The mechanical design and analysis of an active

prosthetic knee”, (2008).

[3] Dedić, Remzo, and Haris Dindo. "SmartLeg: an intelligent active

robotic prosthesis for lower-limb amputees”, Information, Communication and Automation Technologies (ICAT), 2011 XXIII

International Symposium on. IEEE, 2011.

[4] Lambrecht, Bram GA, and Homayoon Kazerooni. "Design of a

semi-active knee prosthesis”, Robotics and Automation, 2009. ICRA'09. IEEE International Conference on. IEEE, 2009.

[5] Ashmi, M., S. Jayaraj, and K. S. Sivanandan. "Development of a

robust microcontroller based output feedback control for assistive

limb”, Biomedical Engineering and Sciences (IECBES), 2014 IEEE

Conference on. IEEE, 2014.

[6] Huff, Amanda M., Brian E. Lawson, and Michael Goldfarb. "A

running controller for a powered transfemoral prosthesis”, Engineering in Medicine and Biology Society (EMBC),

2012 Annual International Conference of the IEEE. IEEE, 2012.

[7] Azimi, Vahid, et al. "Robust composite adaptive transfemoral prosthesis

control with non-scalar boundary layer trajectories”, American Control

Conference (ACC), 2016. IEEE, 2016.

[8] Hamner, Samuel R., Vinesh G. Narayan, and Krista M. Donaldson.

"Designing for scale: Development of the remotion knee for global emerging markets", Annals of biomedical engineering 41.9 (2013):

1851-1859.

[9] Fanelli, Gregory C., ed. The multiple ligament injured knee: a practical

guide to management. Springer Science & Business Media, 2012.

[10] Borjian, Roozbeh. "Design, modeling, and control of an active

prosthetic knee”, (2008).

[11] Åström, Karl J., and Björn Wittenmark. Adaptive control. Courier

Corporation, 2013.

[12] Xue, Chi, Hui Zhu, and Biao Yu. "Modeling and Simulation of

Parameter Self-Tuning Fuzzy PID Controller for DC Motor Speed

Control System”, Applied Mechanics and Materials. Vol. 195. Trans Tech Publications, 2012.

[13] Nguyễn Thị Phương Hà, Huỳnh Thái Hoàng, “Lý thuyết điều khiển

tự động”, 2005.

[14] Rowberg, Jeff. "I2Cdevlib: MPU-6050 6-axis

accelerometer/gyroscope”, Publicación electrónica: http://www.

i2cdevlib. com/devices/mpu6050 Consultada 12.01 (2014).

[15] [InvenSense Inc., 2013] MPU-6000 and MPU-6050 Product

Specification Revision 3.4.

[16] [Microchip Technology Inc., 2012] dsPIC33FJ32MC202204 and

dsPIC33FJ16MC304 datasheet..

(The Board of Editors received the paper on 01/8/2017, its review was completed on 05/9/2017)

30 Pham Nang Van, Nguyen Duy Cung, Nguyen Duc Huy

BI-LEVEL OPTIMIZATION MODEL BASED EVALUATION OF WHOLESALE

ELECTRICITY PRICE INTERVALS CONSIDERING THE WIND POWER

UNCERTAINTY AND ELASTIC DEMAND

Pham Nang Van1, Nguyen Duy Cung2, Nguyen Duc Huy1 1Hanoi University of Science and Technology (HUST); van.phamnang@hust.edu.vn, huy.nguyenduc1@hust.edu.vn

2Student at Hanoi University of Science and Technology (HUST)

Abstract - In the electricity market operation, wholesale electricity prices or Locational Marginal Prices (LMP) vary according to electric demand (including consumption power consumed, bidding prices and the level of price-sensitivity) as well as the penetration level of the wind power. The variable domain identification of LMP plays a very important role for market participants to assess and mitigate the risk on account of the uncertainty of wind power output forecasting. Traditionally, the Monte Carlo simulation (MCS) method can be used in order to determine the variable intervals of LMP. However, in this paper, the authors deploy a bi-level optimization model to calculate the upper and lower bounds of LMP when considering the uncertainty of wind power and elastic demand. The objective function of the upper-level optimization problem is to maximize (or minimize) LMP at a node whereas the objective function of the lower-level optimization problems is to calculate the optimal power generation of the units participating in supplying the load.

Key words - Wholesale electricity market; mathematical program with equilibrium constraints (MPEC); mixed-integer linear programming (MILP); wind power uncertainty; elastic demand.

1. Introduction

Currently, many countries around the world, including

Vietnam, have been operating wholesale electricity markets.

In the wholesale electricity market, the market participants

are generation companies (GENCOS) and distribution

companies (DISCOS). The market operator collects

generating offers by producers, load bids by consumers and

clears the market by maximizing the social welfare [1]-[2].

The uncertainty from wind output has brought

unprecedented challenges to the optimal operation of the

electricity market. The power system operation has been

dealing with the uncertainty of load, different from load

uncertainty; however, wind output is characterized with

large uncertainties and low prediction precision [3]. On the

other hand, load demand has an intrinsic pattern and thus the

load prediction, especially, in short-term, has a significantly

high forecast accuracy [3]. Therefore, the optimal operation

and dispatching model considering stochastic wind power

output has been a hot topic for research.

Reference [4] studied the effect of wind integration and

wind uncertainty on power system reliability, using an ARMA

model to analyze short-term wind forecast. Reference [5]

studied the impact of stochastic wind power on the unit

commitment (UC) problem and constructed a UC stochastic

optimization problem with the objective to minimize the

expected operation cost. In reference [6], the influence of

distribution generation on a heavily loaded distribution system

with a wind forecast model based on statistics is tackled. A

mixed-integer stochastic optimization model is established in

[7] where the wind uncertainty is modeled with ARMA as

well as Latin hypercube sampling and a scenario reduction

method is adopted to simplify the computation.

The first step to investigate the effect of uncertainty is

to model the uncertain wind output by using a variety of

methods, for instance, probability distribution model [8],

fuzzy model [9] and interval number model [10]. In the

next steps, different optimization models are applied to

find the solution.

To make payments in the electricity market, locational

marginal prices (LMP) are calculated. The difference in

LMPs between two nodes of a branch depends upon the

congestion and losses on that branch [2]. The locational

marginal pricing methodology is widely used in electricity

markets to determine the electricity prices and to evaluate

the transmission congestion cost [11]-[12]. Step change

characterizes of LMP under system load variation has been

identified and discussed [13]. Moreover, the concept of

critical load level (CLL) is defined and employed for load

frequency control [13]. Based on a similar idea, the

investigation of the impact of variable wind power outputs

on LMPs must be worth launching. It is important to find a

method to efficiently obtain the wholesale electricity price

intervals under the variation of wind power output and

elastic demand.

This paper proposes an approach to determine the

intervals of LMP using a bi-level optimization model, which

is similar to the interval number-based optimization model

regarded as the optimization of optimization. In addition, the

impact of the uncertainty of wind power as well as the level

for demand-bid price sensitivity is also analyzed.

The next sections of the article are organized as follows.

In section 2, the authors present bi-level optimization model

to determine LMP intervals. In addition, the authors also

describe the solution to this bi-level optimization problem

including the procedure of transferring it into a Mathematical

Program with Equilibrium Constraints (MPEC) problem and

the conversion from MPEC to a Mixed-Integer Linear

Programming (MILP). Section 3 demonstrates the simulation

results and numerical analyses of PJM 5-bus system. Some

conclusions are given in section 4.

2. LMP interval under wind power uncertainty

2.1. Market clearing model

Economic Dispatch (ED) considering elastic demands

in wholesale electricity market is carried out by

Independent System Operators (ISOs) to clear market as

well as determine LMPs and output of generating units. In

this paper, the DCOPF-based approach without losses is

employed to model the electricity market and estimate

LMPs. This DCOPF is a linear programming (LP) problem

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 31

presented as follows:

1

minN

EGi Gi Wi Wi Di Di

i

c P c P c P

(1)

1 1

. :N N

F EGi Wi Di Di

i i

s t P P P P

(2)

1

min max: , , 1,

NF E

l l i Gi Wi Di Di

i

l l l

Limit GSF P P P P

Limit l M

(3)

min max min max: , , 1,Gi Gi Gi i iP P P i N (4)

max min max0 : , , 1,Wi Wi i iP P i N (5)

max min max0 : , , 1,E EDi Di i iP P i N (6)

where N is the number of buses; M is the number of lines;

cGi and cWi are energy prices offered by conventional

generation and wind power, respectively; PGi and PWi are

power outputs of the conventional generating unit and wind

power, respectively; cDi is the price bid by demand i;E F

Di DiP and P are the elastic power and fixed power of

demand i, respectively; GSF is the generation shift factor

matrix; minGiP and max

GiP are the upper and lower bounds of

the convention generation output; maxWiP is the maximum

available wind power output; maxEDiP is the maximum

price-sensitivity demand at bus i; and the variables on the

right side of the colon are the dual variables associated with

the equality and inequality constraints on the left.

The LMP at bus i can be calculated from the Lagrange

function of the above ED problem. This function and LMP

are given by

1

1

min

1 1

max

1 1

min min max max

1 1

NE

Gi Gi Wi Wi Di Di

i

NF E

Gi Wi Di Di

i

M NF E

l l i Gi Wi Di Di l

l i

M NF E

l l l i Gi Wi Di Di

l i

N N

i Gi Gi i Gi Gi

i i

c P c P c P

P P P P

GSF P P P P Limit

Limit GSF P P P P

P P P P

min max max

1 1

min max max

1 1

N N

i Wi i Wi Wi

i i

N NE E E

i Di i Di Di

i i

P P P

P P P

(7)

min max

1

M

i l i l lFlDi

LMP GSFP

(8)

2.2. LMP interval and its bi-level optimization form

Traditionally, the intervals of LMP are usually

evaluated using Monte Carlo Simulation (MCS) approach.

However, this approach requires a huge amount of

computation time in comparison with the bi-level

optimization approach in terms of the same level of

accuracy.

The problem for calculation LMP interval is formulated

as follows:

: max mini iUpper Level LMP or LMP (9)

. : mod 1 6s t Lower level ED optimization el (10)

max

, ,wf i Wi wf iP P P (11)

where ,wf iP and ,wf iP is the forecast upper and lower

bounds of the maximum wind power output. In other

words, interval constraints are used to model the maximum

wind power output in the upper level.

In the above formulation, we see that if there are N

buses in the power system, 2N optimization runs should be

carried out.

2.3. Formulation as a MPEC

Given that the lower level ED is an LP problem, the bi-

level can be transformed into a Mathematical Program with

Equilibrium Constraints (MPEC) by recasting the lower

level problem as its Karush-Kuhn-Tucker (KKT) optimality

conditions, which are added to the upper level problem as

the additional complementary constraints [14] - [16]:

OBJ: (9) (12)

s.t. Constraints in (2) and (11) (13)

min max min max

1

M

Gi l i l l i i

l

c GSF

(14)

min max min max

1

M

Wi l i l l i i

l

c GSF

(15)

min max min max

1

M

Di l i l l i i

l

c GSF

(16)

min

1

0 0N

F El l l i Gi Wi Di Di

i

Limit GSF P P P P

(17)

max

1

0 0N

F El l l i Gi Wi Di Di

i

Limit GSF P P P P

(18)

min min0 0i Gi GiP P (19)

max max0 0i Gi GiP P (20)

min0 0i WiP (21)

max max0 0i Wi WiP P (22)

min0 0i DiP (23)

min max0 0E Ei Di DiP P (24)

32 Pham Nang Van, Nguyen Duy Cung, Nguyen Duc Huy

2.4. Mixed-Integer Linear Programming (MILP)

The MPEC model depicted in (12) – (24) is nonlinear

on account of the slack complementarity constraints (17) –

(24). These slack complementary constraints are

compactly written as 0 F x x 0 , which is stated

equivalently in vector form as:

T

F x 0, x 0, F x x 0 (25)

With the method in [15], however, this MPEC problem

can be converted to a mixed-integer linear programming

(MILP), which can be solved by CPLEX [18]. The MILP

model is presented as follows:

OBJ: (9) (26)

s.t. Constraints in (13), (14), (15) and (16) (27)

min min min,0 l lM (28)

1

min min,

0

1

NF E

l l i Gi Wi Di Di

i

l

Limit GSF P P P P

M

(29)

max max max,0 l lM (30)

1

max max,

0

1

NF E

l l i Gi Wi Di Di

i

l

Limit GSF P P P P

M

(31)

min min min,i0 i M (32)

min min min,0 1Gi Gi iP P M (33)

max max max,i0 i M (34)

max max max,0 1Gi Gi iP P M (35)

min min min,i0 i M (36)

min min,0 1Wi iP M (37)

max max max,i0 i M (38)

max max max,0 1Wi Wi iP P M (39)

min min min,i0 i M (40)

min min,0 1E

Di iP M (41)

max max max,i0 i M (42)

Emax max max,0 1E

Di Di iP P M (43)

where min max min max min max min, , , , , , ,M M M M M M M

maxM are large enough constants and min max min, , ,, , ,l l i

max min max min max, , , , ,, , , ,i i i i i are the auxiliary binary

variables [14].

3. Results and discussions

In this section, the bi-level optimization approach is

performed on the modified PJM 5-bus system [17]. The

MILP problem is solved by CPLEX 12.7 under MATLAB

environment.

3.1. System data

The test system is modified from the PJM 5-bus system

[10], as shown in Figure 1. Two wind plants (WF1 and

WF2) with the same capacity are added into the system at

buses A and C while one original generator is removed

from bus A. The total fixed and maximum elastic demand

is 1200 MW equally distributed among buses B, C and D.

E D

A B C

Limit=240 MW

Brighton

Park

City Load

CenterSolitude

Sundance

WF1 WF2

100MW

$14

600MW

$10

200MW

$35

520MW

$30

Limit=400 MW

Figure 1. PJM 5-bus system with two wind farms

3.2. Impact from wind power uncertainty

This subsection shows the impact of wind power

forecast uncertainty on LMP interval. The simulation

parameters are shown in Table 1. In addition, the only fixed

demand is considered in this subsection.

Table 1. Uncertain parameters of wind power

Wind Power Model Normal Distributed

WF wind power mean (MW) 180

WF2 wind power mean (MW) 180

Wind power standard deviation range (%) 0-30%

It should be emphasized that the findings calculated in

this work are exactly the same in comparison with the MCS

method (with 5000 samples), while the simulation time for

bi-level optimization-based approach is dramatically lower

than that of MCS. Table 2 shows results achieved across

all buses from both approaches when the standard

deviation equals 15% from the mean.

Table 2. LMP result intervals from MCS method and

Bi-level optimization method

Bus Bi-vel optimization method MCS method

A [15.24, 16.98] [15.24, 16.98]

B [23.68, 28.18] [23.68, 28.18]

C [26.70, 30.00] [26.70, 30.00]

D [35.00, 39.94] [35.00, 39.94]

E [10, 10] [10, 10]

Table 3 shows LMP intervals when the indicator of

wind forecast uncertainty (σ) changes. According to the

results shown in Table 3, in general, when the standard

deviation of forecasting wind power increases, the

difference between the upper and lower bound of LMP at

every bus also rises. Moreover, these results also reveals

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 33

that despite the variation of wind power uncertainty, the

LMP intervals at bus C remain unchanged.

3.3. Impact from Demand-Bid Price Sensitivity

To investigate the influence of variation in the level of

demand price sensitivity, [11] proposed the coefficient

which is defined as Eq. (43).

max

max

EDi

i F EDi Di

PR

P P

(44)

From the formula (43), the R values range from 0.0 (100%

fixed demand) to 1.0 (100% price-sensitive demand).

Figure 2 illustrates the construction of R for the special

cases R = 0.0, R = 0.5 and R = 1.0.

In this paper, it is assumed that the ratio R of four

demands is similar. Additionally, the forecast mean value

of both wind powers is 180 MW and it follows a normal

distribution with a standard deviation of 10% from the

mean.

Table 4 presents the LMP internal results with the

various R values from the proposed approach. According

to this Table, the R ratio markedly affects the locational

marginal prices at every bus in terms of the maximum and

minimum values as well as the difference between these

two items. In particular, when R varies from 0.4 to 0.6,

there is a decline in the maximum and the minimum values

of LMP at bus D; however, these values of LMP at bus E

grows considerably. Furthermore, when the R ratio equals

0.6, 0.8 and 1.0, respectively, the LMP intervals remain

stable at all buses. Price ($/MWh)

PD (MW)

Price ($/MWh)

PD (MW)maxE

DPFDPF

DP PD (MW)maxE

DP

Price ($/MWh)

R=0.0 R=0.5 R=1.0

Figure 2. Illustration of the R ratio construction for

the experimental control of demand-bid price sensitivity

Table 3. LMP interval results with the various standard deviation

Bus LMP interval

σ = 0% σ = 5% σ = 10% σ = 15% σ = 20% σ = 25% σ = 30%

A [15.24, 15.24] [15.24, 15.83] [15.24, 16.98] [15.24, 16.98] [15.24, 23.45] [15.24, 23.45] [15.24, 23.45]

B [28.18, 28.18] [23.68, 28.18] [23.68, 28.18] [23.68, 28.18] [23.68, 28.18] [23.68, 28.18] [23.68, 28.18]

C [30, 30] [26.70, 30.00] [26.70, 30.00] [26.70, 30.00] [26.70, 30.00] [26.70, 30.00] [26.70, 30.00]

D [35, 35] [35, 35] [35.00, 39.94] [35.00, 39.94] [35.00, 39.94] [35.00, 39.94] [35.00, 39.94]

E [10, 10] [10, 10] [10, 10] [10, 10] [10.00, 19.94] [10.00, 19.94] [10.00, 19.94]

Table 4. LMP interval results with R different demand-bid price sensitivity

Bus LMP interval

R = 0 R = 0.2 R = 0.4 R = 0.6 R = 0.8 R = 1

A [15.24, 16.98] [15.65, 15.83] [14.00, 15.23] [14, 14] [14, 14] [14, 14]

B [23.68, 28.18] [23.68, 25.00] [19.39, 22.29] [18.51, 25.00] [18.51, 25.00] [18.51, 25.00]

C [26.70, 30.00] [26.70, 27.67] [21.47, 25.00] [20.24, 25.00] [20.24, 25.00] [20.24, 25.00]

D [35.00, 39.94] [35, 35] [27.17, 32.46] [25, 25] [25, 25] [25, 25]

E [10, 10] [10, 10] [10, 10] [10.66, 11.68] [10.66, 11.68] [10.66, 11.68]

4. Conclusion

This paper presents an approach to determine the

intervals of locational marginal prices (LMPs) based on

bi-level optimization model. Moreover, authors also

present the conversion of this model to a mathematical

program with equilibrium constraints (MPEC), then to a

mixed-integer linear programming (MILP), which can be

easily solved by available software tools. The results of this

bi-level optimization problem reveal that the wind

uncertainty and the demand-bid price sensitivity level have

a remarkable impact on LMP intervals. In the

computational aspect, the bi-level optimization-based

method is more efficient compared with Monte-Carlo

simulations although the calculated results using both

approaches are identical.

REFERENCES [1] Pham Nang Van, Nguyen DucHuy, Nguyen Van Duong, Nguyen

The Huu, “A tool for unit commitment schedule in day-ahead pool-based electricity markets”, Journal of Science and Technology, The

University of Danang, no. 6, pp. 21-25, June, 2016.

[2] Pham Nang Van, Nguyen Dong Hung, Nguyen Duc Huy, “The

impact of TCSC on transmission costs in wholesale power markets

considering bilateral transactions and active power reserves”, Journal of Science and Technology, The University of Danang, no.

12, pp. 24-28, December, 2016.

[3] Zou J, Lai X, Wang N, Time series model of stochastic wind power

generation, Power System Technology, pp. 2416-2421, 2014.

[4] Zhang Y, Ka WC, “The impact of wind forecasting in power system

reliability”, Third international electric utility deregulation and

restructuring and power technologies, Nanjing.

[5] Tuohy A, Meibom P, Denny E, “Unit commitment for systems with

significant wind penetration”, IEEE Trans. Power Syst, vol. 24, pp.

592-601, 2009.

34 Pham Nang Van, Nguyen Duy Cung, Nguyen Duc Huy

[6] Kroposki B, Sen PK, Malmedal K, “Selection of distribution feeders

for implementing distributed and renewable energy applications”, IEEE Rural electric power conference, Fort Collins, CO, 2009.

[7] Yan Y, Wen F, Yang S, “Generation scheduling with fluctuating

wind power”, Automation of Electric Power Systems, vol. 34, pp.

79-88, 2010.

[8] Juan M. Morales, Antonio J. Conejo and Juan Perez-Ruiz, “Simulating

the impact of wind production on locational marginal prices”, IEEE

Trans. Power Systems, vol. 26, no. 2, pp. 820-828, May 2011.

[9] Hu Y, Study on wind power storage technology and the intelligent

dispatch model, Master thesis, Lanzhou University of Technology, 2013.

[10] Xin Fang, Qinran Hu, Fangxing Li, Beibei Wang and Yang Li,

“Coupon-based demand response considering wind power uncertainty: a strategic bidding model for load serving entities”,

IEEE Trans. Power Syst., vol. 31, no. 5, pp.1025-1037, May 2016.

[11] Hongyan Li, Leigh Tesfatsion, “ISO Net surplus collection and

allocation in wholesale power markets under LMP”, IEEE Trans.

Power Systems, vol. 26, pp. 627-641, April 2011.

[12] V. Sarkar and S. A. Khaparde, “Optimal LMP Decomposition for

the ACOPF Calculation”, IEEE Trans. Power Systems, vol. 26, pp. 1714-1723, May 2011.

[13] F. Li and R. Bo, “Congestion and price prediction under load

variation”, IEEE Trans. Power Syst., vol. 24, no. 2, pp. 911-922,

May 2009.

[14] L.Baringo and A. J. Conejo, “Strategic offering for a wind power

producer”, IEEE Trans. Power Syst., vol. 28, no. 4, pp.1645-1654,

November 2013.

[15] Zhi-Quan Luo, Jong-Shi Pang and Daniel Ralph, Mathematical

Programs with Equilibrium constraints, Cambridge University Press, 2004.

[16] R.C.Bansal, Optimization Methods for Electric Power Systems: An

Overview, International Jounal of Emerging Electric Power

Systems, 2005.

[17] Fangxing Li, Rui Bo, “Small test systems for Power System

Economic Studies”, IEEE PES General Meeting, pp. 1-4, 2010.

[18] CPLEX optimization studio. [Online] Available:

https://www.ibm.com/bs-en/marketplace/ibm-ilog-cplex

(The Board of Editors received the paper on 25/4/2017, its review was completed on 15/5/2017)

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 35

AN ITERATIVE SUBPROBLEM METHOD FOR

THIN SHELL FINITE ELEMENT MAGNETIC MODELS

Dang Quoc Vuong

Hanoi University of Science and Technology; vuong.dangquoc@hust.edu.vn

Abstract - An iterative sub-problem method for thin shell finite element magnetic models is herein performed to correct the inaccuracies near edges and corners arising from thin shell models (e.g., cover of transformers, thin shells, steel laminations). Volume thin regions are replaced by surfaces but neglect border effects in the vicinity of their edges and corners. This leads to errors for solving the thin shell finite element magnetic models. A sub-problem method allows users to split a complete problem (e.g., a system composed of stranded inductors and conducting and magnetic possibly thin regions) into a series of sub-problems that define a sequence of changes, with the complete solution expressed as the sum of the sub-problem solutions. Each sub-problem is solved on its own domain and mesh, which facilitates meshing and may increase computational efficiency.

Key words - Eddy current; finite element method (FEM); sub-problem; sub-problem method (SPM); thin shell (TS)

1. Introduction

Thin shell (TS) finite element (FE) models [1-2] are

commonly used to avoid volumetrically meshing thin

regions (Figure 1, left). Indeed, the fields in the thin regions

are approximated by priori known 1-D analytical

distributions, that generally neglect end and curvature

effects (Figure 1, right). Their interior is thus not meshed

and is rather extracted from the studied domain, being

reduced to a zero-thickness double layer with interface

conditions (IC) linked to the inner analytical distributions

[1-2]. These ICs lead to inaccuracies on the computation of

local electromagnetic quantities in the vicinity of

geometrical discontinuities. Such inaccuracies increase

with the thickness, and is exacerbated for quadratic

quantities like force and Joule losses, which are often the

primary quantities of interest.

Figure 1. From volume thin region to thin shell model

In order to cope with this problem, the authors have

recently proposed a sub-problem method (SPM) for

correcting edge and corner errors and to simplify meshing

in one-way coupling sub-problems (SPs), where no

iteration between the SPs is necessary [3-6].

In this paper, the SPM is extended to correct the inherent

inaccuracies of the field distributions and Joules edge and

corner errors in two-coupling SPs, where each solution is

influenced by all the others, which thus must be included in

an iterative process. The method allows users to correct the

inherent inaccuracies of the filed distributions and Joule

losses near edges and corners appearing from the TS models.

In the proposed SP strategy [3-5], a reduced problem

with only inductors is first solved on a simplified mesh

without thin and volume regions. Its solution gives surface

sources (SSs) as ICs for added TS regions, and volume

sources (VSs) for possible added volume regions. The TS

solution is further improved by a volume correction via SSs

and VSs that overcome at the TS assumptions, respectively

suppressing the TS model and adding the volume model.

2. Iterative sequence of sub-problems

2.1. Canonical magneto-dynamic or static problem

A canonical magneto-dynamic problem i, to be solved

at step i of the SPM, is defined in a domainΩi, with

boundary 𝜕Ωi = Γi = Γh,i ∪ Γb,i. The eddy current

conducting part of Ωi is denoted Ωc,i and the non-

conducting one Ωc,iC, with Ωi = Ωc,i ∪ Ωc,i

C. Stranded

inductors belong to Ωc,iC, whereas massive inductors

belong to Ωc,i. The equations, material relations and

boundary conditions (BCs) of SP i are

curl hi = ji, div bi = 0, curl ei = – 𝜕t bi (1a-b-c)

hi = 𝜇i–1 bi + hs,i, ji = 𝜎i ei + js,I (2a-b)

n h = jf,i, n × biΓb,i = ff,i, (3a-b)

n eiΓe,iΓ𝑏, 𝑖 = kf,i (4)

where hi is the magnetic field, bi is the magnetic flux

density, ei is the electric field, ji is the electric current

density, 𝜇i is the magnetic permeability, 𝜎i is the electric

conductivity and n is the unit normal exterior to Ωi. The

fields jf,i and kf,i in (3a) and (3b) are surface sources (SSs)

and generally equal zero for classical homogeneous BCs.

Equations (1b-c) are fulfilled via the definition of a

magnetic vector potential ai and an electric scalar potential

vi, leading to the ai-formulation, with

curl ai=bi, ei= –𝜕t ai –grad vi, n ai|Γ𝑖,𝑏= af,i. (5a-b-c)

For various purposes, also for a TS representation,

some paired portions of Γ𝑖 can define double layers, with

the thin region in between exterior to Ω𝑖 [1]-[3]. They are

denoted 𝛾i+ and 𝛾i

– and are geometrically defined as a

single surface 𝛾i with ICs, fixing the discontinuities

[∙]𝛾i = ∙𝛾+ – ∙ 𝛾i–).

corner

corner

corner

edge

corner

edge

edgeedge

volume thin region surface or thin shell

from volume

to surface

36 Dang Quoc Vuong

[nhi]𝛾𝑖 , [nbi]𝛾𝑖 , [nei]𝛾𝑖and [nai]𝛾𝑖p. (6a-b-c-d)

With the definitions

(nini)= – (n+ ni+) = (n– ni

–) for the normal

n in different contexts, one has, e.g. for (6a),

[nhi]i = nihii+ – nihpi

–

= – (n+hii+ – n–hii

–). (7)

The field hs,i and js,i and in (2a) and (2b) are volume

sources (VSs) that can be used for expressing changes of a

material property in a volume region [3]. The changes of

materials in a region, from SPu(i = u) to SPp(i = p) are

defined via VSs hs,p and js,p, i.e.

hs,p = (p–1 – u

–1) bu, js,p = (p – u) eu, (8a-b)

for the total fields to be related by the updated relations

hu + hp = p–1 (bu + bp) and ju + jp = p (eu + ep).

The surface fields jf,i, ff,i and kf,i in (3a-b) and (4), and

af,i in (5c), are generally zero for classical homogeneous

BCs. The discontinuities (6a-d) are also generally zero for

common continuous field traces. If nonzero, they define

possible SSs that account for particular phenomena

occurring in the thin region between i+ and i

– [5]-[9].

This is the case when some field traces in a SPu are forced

to be discontinuous. The continuity has to be recovered

after a correction via a SPp. The SSs in SPp are thus to be

fixed as the opposite of the trace solution of SPu.

2.2. Series of coupled sub-problems

The solution x(x≡ h, b, e, j…) of a complete problem

is to be expressed as the sum of SP solutions xi supported

by different meshes [4]. An appropriate series of SPs is

worth being defined via successive model refinements of

an initially simplified model. Physical considerations

usually help to construct a series. For an ordered set Pof

SPs, the summation of their solutions gives the total

solution, i.e.

x =∑ 𝒙𝑖𝑖∈𝑃 with x≡ h, b, e, j…

Each SP is governed by static or dynamic equations and

constrained with SSs (3a-b) and (4c), and VSs (2a-b). As a

consequence, each SP i is influenced by all the other SPs

q in P to calculate each solution 𝒙𝑖 as a series of corrections

𝒙𝑖,𝑗, i.e.

lim 𝑛→∞

𝒙𝑖𝑛 = x =∑ 𝒙𝑖,𝑗 .𝑖∈𝑃

The total solution at iteration n is thus

𝒙𝑛= ∑ 𝒙𝑖𝑛

𝑖∈𝑃 = ∑ ∑ 𝒙𝑖,𝑗𝑛𝑗=1𝑖 ∈𝑃 .

The error 𝜖𝑛 of a solution 𝒙𝑛 is defined by

𝜖𝑛 = ‖𝒙𝑛 − 𝒙reference‖

‖𝒙reference‖

where 𝒙reference is a reference solution (calultated for a

classical numerical method). However, the reference

solution is usually not known. Thus, an estimated error

𝜖estimated𝑛 of a solution 𝒙𝑛 at iteration n has to be defined,

e.g. as

𝜖estimated𝑛 =

‖𝒙𝑛 − 𝒙𝑛−1‖

‖𝒙𝑛‖

The computation of the conversions 𝒙𝑖,𝑗 in a SP i,j

(SPi with particular constraints at iteration n) is kept on till

convergence up to a desired accuracy. Each correction

must account for the influence of all the previous

corrections 𝒙𝑖,𝑗 of other SPs, with j the last iteration index

for which a correction is known, i.e. j = n or n – 1. Initial

solutions 𝒙𝑖0 are set to zero.

2.3. Sub-problems: Two-way coupling

The coupled SP sequences are considered in several SPs:

Figure 2. Decomposition of a complete problem into five

SPs:SP u +SP𝑝1 +SP𝑘1 + 𝑆𝑃 𝑝2 + SP𝑘2

A problem (SP u) involving current driven stranded

inductors is first solved on a simplified mesh without any

thin regions in Figure 2. Its solution gives surface sources

(SSs) for the added TS 1 (SP p1) (Figure 2, middle left)

through TS ICs based on 1-D approximations. The solution

of SP p1 is then corrected by a correction problem (SP k1)

(Figure 2, middle right) via SSs and volume sources (VSs),

that suppress the TS representation and simultaneously add

the actual volume of the thin region. Two new added SPs

are respectively called TS 2 (SP𝑝2) (Figure 2, bottom left)

and volume correction 2 (SP𝑘2) (Figure 2, bottom right).

Here, SP𝑝2 and SP𝑘2 are independently solved in their own

domain that do not include all previous SP regions

anymore. Once obtained, the solutions of all the previous

SPs then give SSs for the new added TS SP𝑝2 through ICS

[1-2]. The TS solution of SP𝑝2 is then corrected by an

SP𝑘2, that also suppresses the TS representation and

simultaneously add the sources of SP𝑝1 and SP𝑘1. SP 𝑢

need no artificial sources and is therefore not influenced.

This leads to changes of all the previous corrections [3-6].

Therefore, each solution has to be calculated as a series of

corrections by iterating between SPs.

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 37

3. Finite Element Weak Formulation

3.1. Magnetic Vector Potential Formulation

The weak bi -formulation (in terms of ai) of SP i

(i u, p or k) is obtained from the weak form of the Ampère

equation (1a), i.e. [3], [4]

1, ,( curl ,curl ') ( ,curl ') ( , ')

i i ii i i s i i s i i

a a h a j a

, , , ,( , ') , ' [ ] , '

i h i t i t i t ii t i i i i i i a a n h a n h a

1,( , ') , ' ( )

is i i i i iF j a a (9)

where Fi1(i) is a curl-conform function space defined in

i, gauged in c,iC, and containing the basis functions for

a as well as for the test function ai' (at the discrete level,

this space is defined by edge FEs; the gauge is based on the

tree-co-tree technique); (·, ·) and < ·, · > respectively

denote a volume integral in and a surface integral on

of the product of their vector field arguments. The

surface integral term on h,i accounts for natural BCs of

type (3a), usually zero. At the discrete level, the required

meshes for each SP i in the SPM totally differ.

3.2. Inductor alone – SP u

The weak form of an SP u with the inductor alone is

first solved via the first and last volume integrals in (9)

(i u) where js,u is the fixed current density in on s.

3.3. Thin shell FE model- SPp

The TS model is defined via the term

,[ ] , 'p pp d p n h a in (9)(i p). The test function ap' is

split into continuous and discontinuous parts a'c,p and a'd,p

(with a'd,p zero on p) [2]. One thus has

,[ ] , ' [ ] , 'p p p pp p p c p n h a n h a

,| , 'p p

p d p n h a (10)

The terms of the right hand side of (10) are developed

using (4) and (7) respectively, i.e.

, ,[ ] , ' [ ] , 'p p p pp c p c p n h a n h a

, , ,(2 ), 'pp p t c p d p c p a a a (11)

, ,[ ] , ' | , 'pp p p

p d p u d p n h a n h a

, , ,1 1

(2 ) , '2 pp p t c p d p d p

p p

a a a (12)

The last surface integral term in (12) is related to a SS

that can be naturally expressed via the weak formulation of

SP u (9), i.e.

1, ,| , ' ( curl ,curl ') (13)

p p pu d p u u d p

n h a a a

At the discrete level, the volume integral in (13) is thus

limited to a single layer of FEs on the side p touching

p+, because it involves only the associated trace

n ad,p'|p. Also, the source au, initially in the mesh of SP

u, has to be projected on the mesh of SP p, using a

projection method [5], [6].

3.4. TS Correction- VSs in the Actual Volume Shell and

SSs for Suppressing the TS representation - SP k

The TS SP p solution is then corrected by SP k via the

volume integrals ,( ,curl ')ps p h a and ,( , ')

ps p j a in

(11). The VSs js,k and hs,k are given in (9).

Simultaneously to the VSs in (9), SSs have to suppress

the TS discontinuities, with ICs to be defined as

[ ] [ ]k kk p n h n h and [ ] [ ]

k kpa a n n .

The trace discontinuity [ ]kk n h occurs in (9) via

[ ] , ' [ ] , 'k p k kk k p k n h a n h a (14)

and can be weakly evaluated from a volume integral

from SP p similar to (13). However, directly using the

explicit form (4) for [ ]kp n h gives the same contribution,

which is thus preferred.

4. Applications

Let us consider a convergence test of the two-way

coupling with a simple didactic example (f = 50Hz, 𝜇𝑟 = 1, 𝜎 = 59 MS/m) (Figure 3).

Figure 3. 2-D geometry of an inductor and two plates

(d = 5mm, H1 = 120mm, H2 = H3 = 45mm, H4 = 80mm,

H5 = 67.5mm, dx = dy = 12mm)

𝐂𝐨𝐧𝐯𝐞𝐫𝐠𝐞𝐧𝐜𝐞→ 𝒂 =∑𝒂𝒊

𝒊∈𝑷

= 𝒂𝒖,𝟎 + ∑𝒂𝒑𝟏,𝒋

𝒏

𝒋=𝟏

+∑𝒂𝒌𝟏,𝒋

𝒏

𝒋=𝟏

+∑𝑎𝑝2,𝑗

𝑛

𝑗=1

+∑𝑎𝑘2,𝑗

𝑛

𝑗=1

The test at hand is considered in five SPs. It is first

solved via an SP u with the stranded inductor alone, then

adding a TS FE SP 𝑝1 that does not include the stranded

inductor anymore. An SP 𝑘1 then replaces the TS SP𝑝1with

an actual volume covering the plate 1. Next, another TS

SP 𝑝2 is added. An SP 𝑘2 eventually replaces the TS

SP 𝑝2 with another actual volume covering the plate 2. In

the correction process of SP 𝑝1, the fields generated by

SP 𝑝2and SP 𝑘2 are reaction fields that influence the source

solutions calculated from previous SP𝑝1. This means that

some iterations between the SPs are required to determine

an accurate solution considered as a series of corrections.

38 Dang Quoc Vuong

Figure 4. Flux lines of the z-component of the magnetic vector

potential corrections (real part) calculated in each SP, i.e.

SP𝑝1, SP𝑘1, 𝑆𝑃𝑝2 and SP𝑘2, with three interations. SP𝑝1is

chosen as the reference of source SP. The imaginary part

presents an analogous behavior

Figure 5. The total solutionsaof SPs after convergence

Figure 6. The norm of eddy current density ‖𝑗‖ (A/m) along

the plate 1 at the different iterations

Figure 7. Relative correction of the Joule power density along

the plate 1 (top) and the plate 2 (bottom) for the two-way

coupling, with effects of d, r, and f

Table 1. Comparation of direct FEM and one-way/two-way SPM.

The solutions of each SP of each mesh (ℳ𝑢,ℳ𝑝1 ,ℳ𝑝2 ,ℳ𝑘1and

ℳ𝑘2) are shown in Figure 5 leading to the solution of linear

system of nu, np and nk equations, respectively. The number of

iterations of two-way n. coupling is n

Variation of position

of the plate

Classical method

Subproblem method

Full problem One-way coupling One-way coupling

N times

- N times full

mesh ℳ

- N solutions

of nf nf LS

SP u:

- 1 times full

mesh ℳ

- 1 solution of

nu nu LS

SP p:

- 1 times full

mesh ℳ

- 2 solution of

np np LS

SP k:

- 1 times full

mesh ℳ

- 2 solution of

nk nk LS

SP u:

- 1 times full mesh ℳ

- 1 solution of nu nu LS

SP p1:

- 1 times full mesh ℳ𝑝1

- 2 n solution of np1 np1 LS

SP k1:

- 1 times full mesh ℳ𝑘1

- 2 n solution of nk1 nk1 LS

SP p2:

- 1 times full mesh ℳ𝑝2

- 2 n solution of np2 np2 LS

SP k2:

- 1 times full mesh ℳ𝑘2

- 2 n solution of nk2 nk2 LS

Figure 4 illustrates an iterative process (with three

iterations) of four SPs (i.e. SP𝑝1, SP𝑘1, SP𝑝2 and SP𝑘2) for

the magnetic vector potential a, where SP𝑝1 is chosen as a

source SP. Note that the source problem SP u does not need

to be corrected, because it only contains the current driven

inductor and needs no SS or VS. Figure 5 gives the

tolerance of the total solution a of the convergence (8

iterations). Figure 6 represents the convergence of the

volume correction SP𝑘1 along the plate 1, for different

iterations. The TS solution is also pointed out as a function

of the number of iterations. The error on the Joule power

0

2

4

6

8

10

12

14

-0.06 -0.04 -0.02 0 0.02 0.04 0.06

No

rm o

f ed

dy c

urr

ent

(A/m

)

Position along the plate (m)

thin shell, iter1volume , iter1thin shell, iter2volume, iter2thin shell, iter8volume, iter8reference model

0.01

0.1

1

10

100

-40 -30 -20 -10 0 10 20 30 40C

orr

ecti

on o

f Jo

ule

pow

er d

ensi

ty (

%)

Position along the plate (mm)

d=5mm, mr=1, s=5.9 107 W

-1m

-1, f=50Hz

d=5mm, mr=1, s=5.9 107 W

-1m

-1, f=300Hz

d=5mm, mr=100, s= 107 W

-1m

-1, f=50Hz

d=1.25mm, mr=100, s= 107 W

-1m

-1, f=50Hz

0.01

0.1

1

10

100

-60 -40 -20 0 20 40 60

Corr

ecti

on o

f Jo

ule

pow

er d

ensi

ty (

%)

Position along the plate (mm)

d=5mm, mr=1, s=5.9 107 W

-1m

-1, f=50Hz

d=5mm, mr=1, s=5.9 107 W

-1m

-1, f=300Hz

d=5mm, mr=100, s= 107 W

-1m

-1, f=50Hz

d=1.25mm, mr=100, s= 107 W

-1m

-1, f=50Hz

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 39

density in the plate 1 and plate 2 depends on several

parameters, as depicted in Figure 7. It can reach 60% in the

end region of plate 1 (Figure 7, top) and 40% in the end

region of plate 2 (Figure 7, bottom). Table 1 summarizes

the computational effort required by the direct FEM and

the one-way and two-way SPM.

5. Conclusion

The proposed correction scheme of TS models via a

SPM in two-way coupling leads to accurate field and

current distributions in critical regions, the edges of plates,

and so of the ensuing forces and Joules distributions. In

particular, SPs in the SPM allow users to use previous local

meshes instead of starting a new complete mesh for any

position of the plate. This can drastically reduce the overall

computation time when many variations of the problem

have to be solved e.g. optimization problems.

REFERENCES

[1] C. Geuzaine, P. Dular, and W. Legros, “Dual formulations for the

modeling of thin electromagnetic shells using edge elements”, IEEE Trans. Magn., vol. 36, no. 4, pp. 799–802, 2000.

[2] T. Le-Duc, G. Guerin, O. Chadebec, and J.-M Guichon “A new

integral formulation for eddy current computation in thin conductive

shells”, IEEE Trans. Magn., vol. 48, no. 2, pp. 427–430, 2012.

[3] Vuong Q. Dang, P. Dular, R.V. Sabariego, L. Krähenbühl,

C. Geuzaine, “Subproblem approach for Thin Shell Dual Finite

Element Formulations”, IEEE Trans. Magn., vol. 48, no. 2, pp. 407–

410, 2012.

[4] Vuong. Dang Quoc “Calculation of distributions of magnetic fields

by a subproblem methods – with application to thin shield models”,

ISSN 1859-3585 – Hanoi University of Industry, Journal of Science

and Technology, No 36 (10/2016).

[5] P. Dular, Vuong Q. Dang, R. V. Sabariego, L. Krähenbühl and

C. Geuzaine, “Correction of thin shell finite element magnetic models via a subproblem method”, IEEE Trans. Magn., Vol. 47, no.

5, pp. 158 –1161, 2011.

[6] Patick Dular, Laurent Krähenbühl, Ruth. Sabariego, Mauricio

Ferreira da Luz, and Christophe Geuzaine, “A Finite Element Subproblem Method for Position Change Conductor System”, IEEE

Trans. Magn., vol.48, no. 2, pp. 403-406, 2012.

[7] P. Dular, R. V. Sabariego, M. V. Ferreira da Luz, P. Kuo-Peng and

L. Krähenbühl, “Perturbation Finite Element Method for Magnetic

Model Refinement of Air Gaps and Leakage Fluxes”, IEEE Trans. Magn., vol.45, no. 3, pp. 1400-1403, 2009.

[8] C. Geuzaine, B. Meys, F. Henrotte, P. Dular and W. Legros, “A

Galerkin projection method for mixed finite elements”, IEEE Trans.

Magn., Vol. 35, No. 3, pp. 1438-1441, 1999.

[9] P. Dular and R. V. Sabariego, “A perturbation method for computing

field distortions due to conductive regions with h-conform

magnetodynamic finite element formulations”, IEEE Trans. Magn., vol. 43, no. 4, pp. 1293-1296, 2007.

(The Board of Editors received the paper on 19/05/2017, its review was completed on 01/08/2017)

40 Nguyen Van Hieu, Nguyen Quang Bau

THE DEPENDENCE OF AN ACOUSTOMAGNETOELECTIC FIELD ON THE

TEMPERATURE IN A CYLINDRICAL QUANTUM WIRE AlGaAs/GaAs

Nguyen Van Hieu1, Nguyen Quang Bau2 1University of Education - The University of Danang; nvhieu@ued.udn.vn

2College of Science - Hanoi National University; nguyenquangbau54@gmail.com

Abstract - The acoustomagnetoelectric (AME) field in a cylindrical quantum wire with an infinite potential (CQWIP) is investigated theoretically in the presence of an external magnetic field (EMF) by using the quantum kinetic equation method. We obtain the quantum kinetic equation for the distribution function of electrons interacting with internal and external acoustic phonons. We calculate the AME current in a CQWIP and then receive analytical expressions for the AME field in the CQWIP in the presence of the EMF. Theoretical results for the AME field is numerically evaluated, plotted and discussed for the specific CQWIP AlGaAs/GaAs. The result shows that the dependence of the AME field on the temperature of the system is nonlinear. These results are compared with those of normal bulk semiconductors and quantum well to show the difference.

Key words - cylindrical quantum wire; acoustomagnetoelectric field; electron-acoustic wave interaction; electron-acoustic phonon scattering; quantum kinetic equation.

1. Introduction

When an acoustic wave propagates through a conductor,

its momentum, and its energy, are attenuated by the electrons

which may give rise to a current usually called the

acoustoelectric current, in the case of an open circuit called

acoustoelectric field. The presence of an external magnetic

field (EMF) applied perpendicular to the direction of the

sound wave propagation in a conductor can induce another

field, the so-called AME field. Calculations of the AME

field in bulk semiconductor [1-2] in both cases the weak and

the quantized magnetic field regions have been investigated.

In recent years, the AME field in a superlattice structure and

in a Graphene Nanoribbon have been extensively studied

[3-4]. However, the work obtained by using the Boltzmann

kinetic equation method, thus, is limited to the case of the

weak magnetic field region and the high temperature. In the

case of the quantized magnetic field (strong magnetic field)

region and the low temperature, using the Boltzmann kinetic

equation is not correct. Therefore, we use quantum theory to

investigate both the weak magnetic field and the quantized

magnetic field region. The essence of the AME effect is due

to the existence of partial current generated by the different

energy groups of electrons, when the total acoustoelectric

(longitudinal) current in specimen is equal to zero. When this

happens, the energy dependence of the electron momentum

relaxation time causes average mobilities of the electrons in

the partial current, in general, to differ, if an EMF is

perpendicular to the direction of the sound flux, the Hall

currents generated by these groups will not compensate for

one another, and a non-zero AME effect will result.

In low-dimensional systems, the energy levels of electrons

become discrete and different from other dimensionalities [5].

Under certain conditions, the decrease in dimensionality of the

system for semiconductors can lead to dramatically enhanced

nonlinearities [6]. Thus the nonlinear properties, especially

electrical and optical properties of semiconductor quantum

wells (QWs), compositional superlattices (CSLs), doped

superlattices (DSLs), quantum wires, and quantum dots (QDs)

have attracted much attention in the past few years. For

example, calculations of the nonlinear absorption coefficients

of an intense electromagnetic wave by using the quantum

kinetic equation for electrons in bulk semiconductors [7], in

quantum wires [8] have also been reported. Throughout [7-8],

the quantum kinetic equation method have been seen as a

powerful tool. So, in a recent work [9-10] we have used this

method to calculate the quantum AME field in the QW and

the quantum acoustoelectric in the QW. The present work is

different from previous works [1-4] because: (1) the AME

field is a result of not only the electron-external acoustic wave

interaction but also the electron-acoustic phonon scattering in

the sample; (2) we use the quantum kinetic equation method;

(3) we show that the dependence of AME field on the

temperature T of the CQWIP is nonlinear.

This paper is organized as follows. In Section 2, starting

from the Hamiltonian of the electron-external phonon

interaction and electron-acoustic phonon scattering system

in a CQWIP, we use the quantum kinetic equation method

to obtain the quantum kinetic equation for electron in

CQWIP in the presence of an EMF. Solving the equation,

we obtain the solution of the quantum kinetic equation for

electrons in CQWIP. We calculate the AME current in a

CQWIP and then receive analytical expressions for the

AME field in the CQWIP in the presence of the EMF. In

Section 3 we discuss the results, and in Section 4 we come

to conclusions.

2. The analytic expression for the AME field in a

CQWIP

We consider a CQWIP structure of the radius R and

length L with an external magnetic field. Due to the

confined potential, the motion of electrons in the Oz

direction is free, while the motion in the (x-y) is plane

quantized into discrete energy levels called subbands. The

electron energy spectrum in the CQWIP in the presence of

an EMF is expressed as 2

, ,

, , ,

1, ( ),

2 2 2 2z

B n l n lz

N n l p N N c

p n lA A N

m

where m is the effective mass of the electron, c is the

cyclotron frequency, l = 1,2,3,... is the radial quantum

number, N = 0,1,2,... is the index of the Landau, n = 0, 1,

2,... is the azimuth quantum number, L is the length of the

CQWIP, (0,0, )zp p is the electron's momentum vector

along z-direction. In the presence of an external acoustic

wave with frequency q , the Hamiltonian of the electron-

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 41

external phonon interaction and electron-acoustic phonon

scattering system in a CQWIP in second quantization

representation can be written as

, , , , , , , , ,

, , ,

', ', '

, , , ' , , , , , ,

, , , ', ', ', ,

', ', '

, , , , , , , ,

, , , ', ', ', ,

( )

exp( )

z z z

z

z z

z

z z

z

B

N n l p N n l p N n l p k k k

N n l p k

N n l

N n l k N N N n l p k N n l p k k

N n l N n l k p

N n l

N n l q N n l p q N n l p q q

N n l N n l q p

H a a b b

I C J a a b b

U C a a b i t

(1)

where Ck is the electron - internal phonon interaction

factor, 2 3 / 2q l qC i v FS is the electron - external

phonon interaction factor, is the deformation potential

constant, , , , , , ,( )

z zN n l p N n l pa a is the creation (annihilation)

operator of the electron; ( )k kb b is the creation

(annihilation) operator of internal phonon and qb is the

annihilation operator of the external phonon,

2 2(1 ) / 2 / 2 (1 ) / 2l t l t t tF q

2 2 1/2(1 / )l s lv v , 2 2 1/2(1 / )t s tv v , vl (vt) is the

velocity of the longitudinal (transverse) bulk acoustic

wave, , ,

, ,

N n l

N n lU is the matrix element of the operator

U = exp(iqy - klz) (kl = (q2 – (ωq/vl)2)1/2).

To set up the quantum kinetic equation for electrons in

the presence of an ultrasound, we use equation of motion

of statistical average value for electrons

, , , , , ,

, , , , , , ,z z

z z

N n l p N n l pt

N n l p N n l p

a ai a a H

t

(2)

where the notation t

X mean the usual thermodynamic

average of the operator X , and

, , , , , , , , , ( )z z zN n l p N n l p N n l p

ta a f t is the particle number

operator or the electron distribution function.

Using the Hamiltonian of the electron-external

phonon interaction and electron-acoustic phonon

scattering system in a CQWIP in second quantization

replaced into the equation of motion of statistical average

value for electrons and realizing operator algebraic

calculations, the acoustic wave will be considered as a

packet of coherent phonons. We obtain the quantum

kinetic equation for electrons in the single (constant)

scattering time approximation in CQWIP in the presence

of an EMF, and then we have the equation for the partial

current density , , , , , ,

, , ,

( )z z

z

zN n l p N n l p

N n l p

pj e f

m , and we

find the density of AME current in the presence of an

EMF in CQWIP ij ij ij( )i j jj E , where

ij is the

electrical conductivity tensor, ij and

ij are the acoustic

conductivity tensors, respectively

2

1 2 3ij ij C ijk k C i jb b h b h h (3)

21 2 3ij ij C ijk k C i jc c h c h h (4)

2

20

1 2 3 .ij ij C ijk k C i j

e na a h a h h

m (5)

We consider a situation whereby the sound is

propagating along the Ox axis and the magnetic field B is

parallel to the Oz axis and we assume that the sample be

opened in all directions so that ji = 0. Therefore, we obtain

the expression of the AME field, which appears along the

Oy axis of the sample

2 2

zy zy yy zz zz yz x

AME

yy zy

E

(6)

Eq. (6) is the general expression to calculate the AME

field in a CQWIP in the case of the relaxation time of

carrier ( ) depends on carrier energy. By using the

expression of the Eqs. 3-6 and carrying out manipulations,

we derive the expression for the AME field in a CQWIP in

the presence of an EMF as follows

2 2 2 2 2

1 3 2

2

1 1 2 3 2 2 1 3

2 ( )

( )( ) ( )( )

c

AME

c c

c c

Ee m a a a

b c a a b c a a

(7)

here,

2, 0

2 2

0 0

0

12 2

0

( )( ). ,

1 ( )

( ). ,

1 ( )

gn l

g N

c

g

g

c

fma A d

n

fb A d

0

22 2

0

( ). ,

1 ( )

g

g

c

fc A d

2 32 2

', ' ' , 21 ,4

, , ', ', , '

, 3 , 3

1 2

( ) 24

( 2 ) ( 2 ) ,

B n l N n l

n l N N

n l n l N Nk s

n l n l

N N

e k TA I J u mA

v LS

mA mA

24 3 2 3

2', ' , 2

2 ,

, , ', ', , '

2 , '

, , ', '

2 , '

, , ', '

82

{ ( 2 ( ))

( 2 ( ))},

s q n l n l

n l N

n l n l N N

N N

n l n l k q

q

N N

n l n l k q

e vA U mA

FS

q q m

q m

and , ' ,

1 , , ', '2 ( ) ,N N n l

n l n l N km A

, ' ,

2 , , ', '2 ( ) ,N N n l

n l n l N km A , ' , ', '

, , ', ' ' .N N n l n l

n l n l N NA A

From Eq. (7), we can see that the dependence of the

AME field on the intensity B of the EMF, the radius R of

the CQWIP, the temperature of the system T and the

external acoustic wave of frequency are nonlinear. These

results are different compared to those obtained in bulk

semiconductor [1,2], the quantum well [9].

3. Numerical results and discussion

In order to clarify the results that have been obtained, in

this section, we consider the AME field in a CQWIP. This

quantity is considered as a function of the external magnetic

field B, the frequency of ultrasound, the temperature T of

system, and the parameters of the CQWIP AlGaAs/GaAs.

The parameters used in the numerical calculations are as

follows [8-10]: 0=10-12s, =104Wm-2, =5,3×103kg/m3,

=20,8×10-19J, vl= 2,0 × 103 m/s, vt = 1,8 × 103 m/s.

42 Nguyen Van Hieu, Nguyen Quang Bau

Figure 1. Dependence of the AME field on the temperature at

different values of the radius R=35 nm (dashed line), R=30 nm

(solid line). Here B=2.0T

Figure 2. Dependence of the AME field on the temperature at

different values of the external magnetic field B=2.0T (dashed

line), B=2.2T (solid line). Here R=30 nm

Figure 1 and Figure 2 investigate the dependence of

AME field on the temperature T for the case of the low

temperature at different values of the radius of the CQWIP

and the external magnetic field, respectively (in case of the

quantized magnetic region), which have many distinct

maxima. The result shows the different behaviour from

results in bulk semiconductor [1-2], the quantum well with a

parabolic potential (QWPP) [9]. Different from the bulk

semiconductor and QWPP, these peaks in this case are much

sharper. In addition, Figure 1 shows that the positions of the

maxima nearly do not move as the radius of CQWIP is

varied. It reaches a maximum value at T of about 15 K with

the intensity of the EMF B = 2.0 T, the AME field of

approximately 2.1 V/m at the radius R = 35 nm and 2.8 V/m

at R = 30 nm. In contrast, Figure 2 shows that the peaks

move to the higher temperature when the external magnetic

field increases because the condition for peaks to appear do

not depend on the radius but depend on the external

magnetic field. Therefore, we can use these conditions to

determine the peak position at the different values of the

external magnetic field or the parameters of the CQWIP.

From the numerical result, we have a maximum value of the

AME field of approximately 2.7 V/m at T=15 K, B=2.0 (T)

and 3.3 V/m at T=18 K, B=2.2 (T). This means that the

condition is determined mainly by the electrons energy.

4. Conclusion

In summary, we have obtained analytical expressions for

the AME field in the CQWIP. There is a strong dependence

of AME field on the acoustic wave number q, the frequency

of external acoustic wave, the radius R of the CQWIP, the

temperature T of system, the cyclotron frequency of the EMF

and the intensity of the EMF. The result show that there are

many distinct maxima in the quantized magnetic field region

and the sound waves traveling in parallel to the EMF and it

reaches the maximum value then. The result of the numerical

calculation is done for the CQWIP AlGaAs/GaAs. This result

has shown that the dependence of the AME field on the

temperature has the peaks move to the smaller temperature

when the magnetic field increases. We want to emphasize that

the condition for the positions of the peaks to appear in the

dependence of AME field on the intensity of the EMF in

CQWIP is not dependent on the temperature T. This result

shows the difference from the results obtained in normal bulk

semiconductors [1-2] and in quantum well [9]. These are

important results of the present work.

Acknowledgments

This work is funded by Ministry of Education and

Training of Viet Nam under Grant number B2016.DNA13.

REFERENCES

[1] G. M. Shmelev, G. I. Tsurkan, N. Q. Anh, Phys. Stat. Sol. 121 (1984) 97.

[2] M. Kogami,Sh. Tanaka, J. Phys. Soc. Japan 3 (1970) 775.

[3] S. Y. Mensah, F. K. A. Allotey, J. Phys. Condens. Matter. 8 (1996) 1235.

[4] K. A. Dompreh, S. Y. Mensah, S. S.Abukari, R. Edziah, N. G.

Mensha and H. A. Quaye, Nanoscale Syst.:Math.Model.Theory

Appl. 4 (2015) 50.

[5] Y. Zhang, K. Suenaga, C. Colliex, S. Iijima, Science. 281 (1998) 973.

[6] S. S. Rink, D. S. Chemla, D. B. Miller, Adv. Phys. 38 (1989) 89.

[7] V. V. Pavlovich, E. M. Epshtein, Sov. Phys. Semicond. 11 (1977) 809.

[8] N. Q. Bau, H. D. Trien, J. Korean Phys. Soc. 56 (2010) 120.

[9] N. Q. Bau, N. V. Hieu, N. V. Nhan, Superlattices and

Microstructures. 52 (2012) 921.

[10] N. V. Nhan, N. V. Nghia, N. V. Hieu, Material Transactions, 56

(2015) 1408.

(The Board of Editors received the paper on 06/11/2017, its review was completed on 20/11/2017)

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 43

A STUDY ON SYNTACTIC, SEMANTIC AND PRAGMATIC FEATURES OF

HYPERBOLIC EXPRESSIONS IN CEO NEWS ITEMS IN ENGLISH VERSUS

VIETNAMESE

Dinh Thi Minh Hien, Chu Thi Mui

University of Foreign Language Studies - The University of Danang; minhhien02@yahoo.com, chuthimui@gmail.com

Abstract - The article examines the syntactic, semantic and pragmatic features of hyperbolic expressions (HEs) in Chief Executive Officials (CEO) news items in English and Vietnamese. Its design is a combination of both qualitative and quantitative research methods using the theory of hyperbole by Galperin (1981) and Đinh Trọng Lạc (1994) based on which HEs are extracted from various sources of daily up-dated CEO news on internet. The findings of the study show that HEs of English and Vietnamese CEO news have some differences and similarities in terms of semantics, syntax and pragmatics. The findings are expected to contribute a beter insight into the explanation of the characteristics of HEs in English and Vietnamese language teaching, communication and Translation.

Key words - hyperbolic expressions (HEs); CEO news; syntactic; semantic; pragmatic features

1. Overview

Hyperbole is seen as one of stylistic devices having a

great power of expressive meaning. People use it as a

useful tool to make their words more impressive and attract

the reader’attention. Hyperbole can be found popularly in

literature and in oral communication, in poetry as well as

in other discourse types; however, in the area of CEO news,

hyperbolic expressions (HEs) remain as a mystery that

needs to be further discovered. Let us look at the examples

below of hyperbole employed in English and Vietnamese

CEO news items:

(1) Universal's theme park business is hotter than a fire-

breathing dragon.

Obviously, the bolded expression is an overstatement.

The speaker used the hyperbolic expression (HE)

“Universal's theme park business is hotter than a

fire-breathing dragon” to emphasize that universal's theme

on park business grows rapidly.

(2) ADSL xuất hiện với các quán cafe Internet mọc lên

như nấm, giúp phổ biến Internet đến từng khu phố; giai

đoạn Internet văn hóa xã len lỏi về vùng nông thôn; và 3G

xuất hiện.

To express the idea that many coffee shops have been

built since ADSL appeared in the consuming market, the

hyperbole expression (HE) “quán café mọc lên như nấm”

is used to exaggerate the large number of coffee shops that

have internet.

For the above-mentioned reasons, we have decided to

conduct a study with the topic “A study on syntactic,

semantic and pragmatic features of hyperbolic expressions

in CEO news items in English versus Vietnamese”.

Hopefully, this study can contribute a meaningful part to

the learning and teaching of English and help Vietnamese

learners comprehend and use hyperbole effectively in the

areas of economics, communication and translation.

2. Focus of the study

In order to achieve the aim of the article I intend to

fulfil the objectives (i) to examine the hyperbolic

expressions in English and Vietnamese CEO news items in

terms of syntax, semantics and pragmatics; (ii) to find out

the similarities and differences in syntactic, semantic and

pragmatic features of hyperbolic expressions in English

and Vietnamese CEO news items; (iii) to propose effective

ways to comprehend and translate hyperbolic expressions

effectively in English and Vietnamese CEO news.

3. Theoretical Background

3.1. Stylistics

According to Galperine, (1981) stylistics is a branch of

general linguistics which deals mainly with the functional

styles of language and the special media of language to

secure the desirable effect of the utterance, and is called the

stylistic devices and expressive means.

3.2. Stylistic Devices

3.2.1. Definition of Stylistic Device

Galperin defines stylistic device as “a conscious and

intentional literary use of some of the facts of the language

(including expressive means) in which the most essential

features (both structural and semantic) of the language

forms are raised to a generalized level and thereby present

a generative”.

Đinh Trọng Lạc (1995) in “99 Phương tiện và Biện

pháp Tu từ tiếng Việt” stated that stylistic device is one of

the most decisive factors influencing the typical style of a

writer or a lecturer. Besides, readers are able to discover

particular styles of each writer or convey messages thanks

to different types of stylistic devices used in his work.

3.2.2. Function of Stylistic Devices

Stylistic devices are used in writing to make it more

effective and persuasive. Without stylistic devices, writing

would be plain and shallow. The more stylistic devices you

know, the more unique your writing can be.

3.2.3. Classification of Stylistic Devices

Galperin (1971) offered a classification of expressive

means and stylistic devices based on the level-oriented

approach: Phonetic expressive means and stylistic devices,

Lexical expressive means and stylistic devices and

Syntactical expressive means and stylistic devices.

3.3. Hyperbole

3.3.1. Definitions and Functions of Hyperbole

According to Galperin (1977) hyperbole can be defined

as “deliberate overstatement or exaggeration of a feature

44 Dinh Thi Minh Hien, Chu Thi Mui

essential (unlike periphrasis) to the object or phenomenon.

Furthermore, Galperin (1977) stated that “hyperbole is a

device which sharpens readers’’ ability to make a logical

assessment of the utterance. Hyperboles are perfect for

fictional works, especially to add color and depth to a

character or humor to the story.

3.4. Hyperbole and Metaphor

Hyperbole is often confused with metaphor because a

metaphor does not use “like” or “as” to make a comparison,

which is similar to hyperbole. However, the main

difference is that hyperbole is an exaggeration to

emphasize or make readers pay more attention to and

understand more deeply the meaning of utterances.

3.5. Hyperbole and Simile

Simile and hyperbole are the terms for two different figures

of speech. Simile is the use of words such as “like” to compare

an object, concept, or person to something else. Simile and

hyperbole can often be confused, because hyperbole sometimes

fancifully compares its subject to something else.

3.6. Semantic Features

Semantics can be defined as the study of language

meaning; that is the meaning of words, phrases, sentences

and larger units of discourse (termed texts,or narratives).

According to Richards et all (1987), “semantic features are

“the smallest units of meaning in a word.”

Take a look at the following examples:

(3) … GE’s exit from its vast finance arm …

(4) … những tên tuổi khổng lồ của thế giới dường như

là một kỳ tích.

In examples (3) and (4), the HEs “vast finance arm”,

“những tên tuổi khổng lồ của thế giới ” are hyperboles of

immensity and they are used to make strong impressions

on readers.

3.7. Syntactic Features

Chomsky (1965) stated that syntax is the study of the

principles and rules for constructing phrases and sentences

in a language. Syntax focuses on the word order of a

language and the relationships between words.

Let us look at the following syntactic feature in

hyperbolic expressions in CEO news items in English

versus in Vietnamese:

(5) Single –handly wiped out around 90,000 jobs. [N-4]

It is obvious that the syntactic features of hyperbolic

expressions (3) are recognized in clause: S+ V+O

(6)“CEO Vietjet là người phụ nữ có bàn tay sắt bọc nhung”.

In example (6), the syntactic features of HE (3) are

recognized in clause: S+ V

3.8. Pragmatic Features

Crystal (1997) defines pragmatics as the study of

communicative action in its sociocultural context. In

addition to using speech acts such as apologizing,

complaining, complimenting, and requesting,

communicative action includes also engaging in different

kinds of discourse and participating in speech events of

varying length and complexity.

The following examples are hyperboles expressing

statement in terms of pragmatics in English and

Vietnamese CEO news items.

(7) Amazon founder Jeff Bezos, seen here in this Jan. 28,

2016, file photo, says technology is entering a golden era.”

(8) Vinh nói” “Tình làng xóm của người miền Tây Nam

bộ đã ăn sâu trong máu, nhà nào có việc thì cả làng cùng

đến”, Võ Văn Châu chia sẻ.”

4. Research Design and Methodology

The research design of this project is a combination of

both qualitative and quantitative research methods which

are used to collect and analyze HEs in Vietnamese and

English CEO news items.

5. Findings and Discussion

5.1. Syntactic Features Of Hyperbole Expressions In

English And Vietnamese CEO News Items

Table 1. Examples of Syntactic Feature of HEs in English and

Vietnamese CEO News Items

HEs Examples

NP = pre-modifier + N “ tremendous growth"

NP = pre-modifier + N

+ post- modifier “nguồn nhân lực vàng”

NP = N+ post-modifier “mùi hương mê hoặc của cà phê Shin”

VP = V+NP “has blown us away”

VP = V intransitive “gây nghiện cho khách hàng”

VP = V + post-modifier “chia sẻ chóng mặt,nổi như cồn”

VP = pre-modifier + V

+ post-modifier “bị dội bom tin nhắn”

AdjP = pre-modifier +

Adj “crystal clear”

PP = Pre-modifier + Prep

+ NP + Post-modifier

“In a very short amount

of time…”

Adv = Adverb “…extremely well”

S1 + V + Comp Adj +

than + S

“Universal's theme park business is

hotter than a fire-breathing dragon.”

One of the + Superlative

Adj + N + PrepP

“One of the most powerful media

titans in U.S. history.”

S + Adv + V+ Superlative

Adj + PrepP + that + clause

“He really is best in class on all

the criteria that we looked at.”

NP + as/like + NP

“Crazy like a fox.”

“Con đường đầy chông gai như kinh

doanh smartphone ở Việt Nam.”

NP + that + V + look

like + NP + VP

“His energy level is so high that

looks like his eyes might pop out of

their rocket.”

NP + Pre + as + NP “…nó như "tình yêu sét đánh"

Conditionals

(IF + present simple +

simple future)

“Nếu người sáng lập không phát

triển nhanh được chắc chắn

startup sẽ chết hoặc thoi thóp, bị

đối thủ mạnh hơn đè bẹp.”

S + V “Với chiến lược “gây nghiện thị trường”

S + V + O “It’s killing people’s minds.”

S + V + C “Watson has become the silver thread

tied into much of IBM’ success.”

Existence Clause “There are enormous global

pressures.”

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 45

5.2. Semantic Features of Hyperbolic Structures (HEs)

in English and Vietnamese CEO News Items

Table 2. Examples of Semantic Feature of Hyperbolic

Expressions (HEs) in English and Vietnamese CEO News Items

HEs Examples

Immensity “những tên tuổi khổng lồ của thế giới”

Strength

“Whitman has single-handedly wiped out

around 90,000 jobs at the firm”

Người lèo lái con thuyền DHG vượt qua

bão táp đến thành công

Enthusiasm

"his tireless work promoting low-cost investing”

“đam mê nghiên cứu và sáng tạo đã dần

ngấm vào máu của ông”

Growth "Digitization is coming at us like a rocket.”

“Thị trường bất động sản đang lên cơn sốt”

Competition “Amazon.com was going to "kill" Best Buy…”

“Săn lùng chất xám …”

CEO’s Talent

“He handles complexity extremely well.”

“Người lèo lái con thuyền DHG vượt qua

bão táp đến thành công.”

5.3. Pragmatic Feature of HEs in English and

Vietnamese CEO News Items

Table 3. Examples of Pragmatic Feature of HEs in English and

Vietnamese CEO News Items

HEs Examples

Promise

"We're bringing all of our energies to bear

there," he said.

“Con gà Vietjet sẽ đẻ thật nhiều những quả trứng

vàng hạnh phúc và thịnh vượng, cho hành khách,

cho cộng đồng và cho chính mỗi chúng ta.”

Warning

“…waiting for consensus can kill you, because

of speed matters, too.”

“Khi khởi nghiệp, đa số các nhà sáng lập đều tin

rằng, chỉ cần có ý tưởng điên rồ, khác biệt hoặc

độc đáo là sẽ thành công nhưng đó lại là một

trong những sai lầm chết người…”

Assertion

“General Motors has been a giant in the auto

industry for more than one hundred years.”

“CEO Vietjet là người phụ nữ có bàn tay sắt bọc nhung”

Statement

“Amazon founder Jeff Bezos says when it comes to

technology, "We're on the edge of the golden era”

"Tôi cho rằng xe điện là ngành hàng có ‘gen

khủng long’, anh Long cho biết.”

Prediction

“Tech startups seem to be in boom times with

the huge valuations we're seeing.”

“10-15 năm nữa là thời điểm rất kim cương.”

Asking

“What does it feel like to be the CEO of a super-

hot company ….?”

“M&A có thực sự là “chiếc đũa thần” cho sự

phát triển của doanh nghiệp?”

5.4. Quantitative results of hes in English and

Vietnamese ceo news items.

5.4.1. Frequency of Syntactic Features of HEs in English

and Vietnamese CEO News

Both languages include HSs of two types: phrases and

clauses.In terms of phrases, hyperbolic expressions are

recognized in NPs, VPs, AdjP, AdvPs, and PPs.

Table 1. Frequency of Syntactic Features of HEs in

English and Vietnamese CEO news

Types of Hes English Vietnamese

Oc % Oc %

HP

NP 11 25 18 21.4

VP 1 2.3 20 23.8

AdjP 2 4.55 0 0

PP 1 2.3 0 0

AP 2 4.55 0 0

ComS 13 29.5 10 11.9

HC Cond 0 0 1 1.2

Others 14 31.8 35 41.7

Total 44 100% 84 100%

Noun phrases and verb phrases which have

pre-modifiers and post- modifiers in English and

Vietnamese syntactic features are found to be similar.

Vietnamese syntax has a number of distinctive features

which are different from English syntax. First of all, there

is a distinct difference in the order of nouns within noun

phrase between two languages. There are also no cases of

hyperbolic expressions in form of AdjPs, PrepPs and

AdvPs compared to their English counterparts. Regarding

conditionals, among the samples of HEs in English CEO

news items, there are no cases of HEs in the form of

conditionals. In terms of clauses, there are significant

differences between the two languages. In English, clauses

commonly used are structures S+V+O and S+V+C, while

in Vietnamese, clauses commonly used are structures S+V.

Figure 1. Proportion of HEs in English and Vietnamese CEO

News Items in Terms of Syntax

5.4.2. Frequency of Semantic Features of HSs in English

and Vietnamese CEO News

The hyperbolic structures such as hyperboles

expressing immensity, hyperboles expressing strength,

hyperboles expressing enthusiasm, hyperboles expressing

growth, hyperboles expressing competition, hyperboles

expressing CEO’s talent are denoted in CEO news in

English and in Vietnamese.

There are some differences between English and

Vietnamese used in CEO news items. Vietnamese writers

Proportion of HEs in English CEO News Items

Hyperbolic Phrases

Comparison Structures

Hyperbolic Clauses

Proportion of HEs in Vietnamese CEO News Items

Hyperbolic Phrases

Comparison Structures

Hyperbolic Clauses

46 Dinh Thi Minh Hien, Chu Thi Mui

consider “Enthusiasm of CEOs” and “strength” as

interesting topics to hyperbolize; these aspects, on the other

hand, are rarely paid attention to by the English ones.

Another distinction worth mentioning is that English

writers employ hyperboles expressing immensity and

hyperboles expressing growth more commonly than

Vietnamese ones do.

Table 2. Frequency of Semantic Features of HEs in English and

Vietnamese CEO News

No. Hyperbolic

expressions

English Vietnamese

Oc % Oc %

1 Immensity 6 25 3 8.8

2 Strength 3 12.5 4 11.8

3 Enthusiasm 2 8.35 20 58.8

4 Growth 6 25 2 5.9

5 Competition 2 8.35 2 5.9

6 CEO’s Talent 5 20.8 3 8.8

TOTAL 24 100% 34 100%

Figure 2. Proportion of HEs in English and Vietnamese CEO

News Items in Terms of Semantics

5.4.3. Frequency of Pragmatic Features of HSs in English

and Vietnamese CEO News

Hyperbolic structures in both languages are used with

the same purposes of showing promise, warning,

assertion,, statement, prediction, and asking. It can be said

that among all of hyperboles investigated, a large number

of hyperboles expressing conclusion are used in

Vietnamese while there are fewer hyperboles expressing

conclusion found in English expressing this function.

Table 3. Frequency of Pragmatic Features of HEs in English

and Vietnamese CEO News

No. HEs English Vietnamese

Oc % Oc %

1 Promise 4 9.6 3 4.9

2 Warning 2 4.8 2 3.3

3 Assertion 21 50 35 56.4

4 Statement 11 26 15 24.1

5 Prediction 3 7.2 4 6.5

6 Asking 1 2.4 3 4.8

TOTAL 42 100% 62 100%

Figure 3. Proportion of HEs in English and Vietnamese CEO

News Items in Terms of Pragmatics

6. Conclusions

There are 6 groups of hyperbolic structures in both

languages that emphasize and highlight a certain feature of the

mentioned object such as hyperboles expressing immensity,

hyperboles expressing strength, hyperboles expressing

enthusiasm, hyperboles expressing growth, hyperboles

expressing competition, and hyperboles expressing CEO’s

talent. Also hyperbolic structures in both languages are used

with the same purposes of showing promise, warning,

assertion,, statement, prediction, and asking.

7. Implications

7.1. For CEOs

This study will provide CEOs with experience of how

to resolve problems in business. Moreover, information in

CEOs news items will help CEOs predict economic

situation in the future and work out suitable strategies for

their enterprises.

7.2. For language teachers and language learners

7.2.1. For language teachers

Language teachers should ask students to do different

kinds of hyperbolic exercises from different sources to help

learners raise their awareness about the differences of HEs

between English and Vietnamese in terms of syntax,

semantics and pragmatics. By doing this, they can help

students apply what they understand about this language

device into reality as well as to make their lessons more

lively,which can contribute a meaningful part to the

language learning process.

Proportion of HEs in English CEO News Items

Immensity Power

Enthusiasm Growth

Competition CEO’s Talent

Proportion of HEs in Vietnamese CEO News Items

Immensity Power

Enthusiasm Growth

Competition CEO’s Talent

Proportion of HEs in Vietnamese CEO News Items

Hyperboles expressing Promise

Hyperboles expressing Warning

Hyperboles expressing Assertion

Hyperboles expressing Statement

Hyperboles expressing Prediction

Hyperboles expressing Asking

Proportion of HEs in English CEO News Items

Hyperboles expressing Promise

Hyperboles expressing Warning

Hyperboles expressing Assertion

Hyperboles expressing Statement

Hyperboles expressing Prediction

Hyperboles expressing Asking

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 47

7.2.2. For language learners

Knowledge on HEs might help language learners pay

special attention to hyperbolic syntactic structures to avoid

errors made due to lack of knowledge about sentence

elements in both languages.

7.2.3. For translators

In many cases we cannot find the complete equivalents

in the two languages; therefore, knowledge on HEs might

help translators and interpreters be aware of the

grammatical construction as a single unit, compare the

meanings of the sentences in the two languages, determine

where the clauses begin and base on context to get the most

suitable equivalents.

7.2.4. For learners of Economics

Thanks to the similarities and differences found out

between CEO news items in English and Vietnamese,

leaners of Economics might get a lot of information about

English and Vietnamese CEO news from which they can

appreciate a general view of CEOs in general and a global

view of economic situations of other countries in

particular.

REFERENCES

[1] Chomsky, N. (1957). Syntactic Structures. The Hague: Mouton.

[2] Claridge, C. (2011). Hyperbole in English: A corpus-based study of

exaggeration. Cambridge University Press.

[3] Galperin, I.R. (1977). Stylistics. Moscow Higher School.

[4] McCarthy, M., & Carter, R. (2003). “There’s millions of them”:

hyperbole in everyday conversation. Journal of Pragmatics,

36(2004),149-184.

[5] Đinh Trọng Lạc (1995). 99 phương tiện và biện pháp tu từ Tiếng

Việt. Nhà xuất bản Giáo Dục.

(The Board of Editors received the paper on 29/11/2017, its review was completed on 21/12/2017)

48 Phan Tran Minh Hung, Phan Nguyen Bao Quynh

THE IMPACT OF COMPETITION ON CREDIT RISK:

THE CASE OF VIETNAM COMMERCIAL BANKS

Phan Tran Minh Hung1, Phan Nguyen Bao Quynh2 1Ba Ria Vung Tau; phantranminhhung@gmail.com

2Binh Dinh; quynhpnb2655@gmail.com

Abstract - This main purpose of this research is to investigate the influence of competition on credit risk in Vietnam commercial banks over the period 2006 – 2016. Both Lerner indicator and Herfindahl Hirschman Index (HHI) are employed to measure competition degree while non-performing loan (hereafter, NPL) ratio is a proxy for credit risk. The main results indicate that competition positively impacts on the probability of loan non-payment. However, more specifically, expanding lending products also denotes a positive effect on the capability of non-repayment, supported by the “competition – instability” prevalent view. Otherwise, We further find strong evidence that the relationship between competition and credit risk is non-linear with U-shape.

Key words - competition; risk; credit risk; Lerner index; commercial banks.

1. Introduction

One of the extremely essential roles of competition is

to enhance operational quality for the ultimate purpose of

value maximization. However, we should not conclude that

the competitive strategies do not lead to negative aspects.

For example, banks that intend to compete excessively may

lead to face NPLs, even results in going bankrupt. For that

reason, the relation between competition and credit risk has

received scholars’ attentions. This is reflected in a series of

studies published recently. However, these researches have

not had a high consensus because the effect of competition

on loan recovery is mixed.

On the one hand, the prevalent view point “competitive

– instability” supposes that there is a positive correlation

between competition and credit risk. This can be explained

that profit margins are narrowed and banks might take

excessive risks to maximize returns when a large number

of banks expand competitiveness extent. Hence, expanding

activities to compete contributes to eroding brand value,

consequently leading to collapse (Keeley, 1990) or

competition is one of the main sources of bank instability

(Boyd et al, 2005) and therefore higher competitiveness

will lead banks to more volatility (Soedarmono et al, 2011).

On the other hand, the recent empirical results

supporting the "competition-stability" opinion document

the less intensive the competition, the greater the credit

risk. Allen and Gale (2004) further argue that the more

competition will lead to a reduction in bad debt or "banks

become more powerful in expanding profitability and

mitigates NPLs" (Koetter and Poghosyan, 2009).

Moreover, the reduction of interest rate or even lowering

appraisal criteria encourages banks to be approachable and

control each client segment easily. Sometimes, this can

eliminate adverse selection and moral hazard encountered

by customers, thereby contributing to the lower probability

of loan payment.

In Vietnam, the integrated progress has helped Vietnam

banking system to develop in line with international

standards and become stronger. However, Vietnam

commercial banks also confront certain obstacles. One of

them is competitive forces among Vietnam commercial and

foreign banks. In the context of global integration, banking

system plays an essential role in economy and therefore

banks are forced to enhance their competitiveness and have

their appropriate strategies to ensure their roles. It is certain

that the impact of competition on credit risk is theoretically

and practically significant. However, it remains silent on

whether credit risk is a function of competition in Vietnam.

Hence, in order to contribute theoretical and practical

evidence to bank managers and policymakers, this impact

needs to be examined in Vietnam.

2. Literature review and hypothesis development

2.1. Literature review

The nexus of competition and the probability of loan

payment discussed in empirical studies in countries around

the world indicate mixed results.

On the one hand, the overwhelming opinion of

"competition - risk" suggests that diversification is one of

the main sources of credit risk. The interests in the

relationship between competition and stability in banking

sector were triggered by Keeley (1990), who initiated an

academic debate that product diversification to compete

contributes to eroding brand value, consequently leading to

collapse (Keeley, 1990). As the quality of the loan portfolio

is likely to deteriorate due to debt holders’ more marginal

benefits requirements and thereby increase bank fragility.

In addition to this, recent studies have illustrated that

enhancing competitiveness makes banks reduce

borrowers’ loan-related information, their motivation to

manage loans, resulting in a worse effect on bank stability

(Allen and Gale, 2004). Furthermore, banks with high

market power in lending sector are under pressure of

increasing risk because high interest costs generate

difficulties for customers to repay, leading to exacerbate

adverse selection and moral hazard. Hence, greater

competition encourages banks to accept more diversified

risks, making banking system more vulnerable to shocks

(Anginer et al, 2014).

We hypothesize that competition impacts positively

credit risk of commercial banks (H1).

On the other hand, the "competition - stability"

perspective favors the existence of a positive relationship

between competition and credit risk. Enhancing

competitiveness is encouraged to minimize the probability

of increasing risk because the lack of competitive

operations can exacerbate the instability of banks. Mishkin

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 49

(1999) paid attention to the notion of "too big to fail",

documents that large banks exist moral hazard established

by managers who usually accept risky deals under the

patronage of the central bank. Furthermore, these banks are

generally supported by governmental policies that

encourage them to take more risks that destabilize the

banking system (Acharya et al, 2012). Additionally, the

nexus of bank concentration and NPL ratio indicates that

more market power associates with riskier loan portfolios

(Berger et al, 2009). Higher interest rate leads to the poorer

loan portfolio’s risk due to adverse selection and moral

hazard (Stiglitz and Weiss, 1981).

We hypothesize that competition effect negatively

on credit risk of commercial banks (H2).

Moreover, Martinez-Miera and Repullo (2010)

document a non-linear relationship between competition and

credit risk. This is because the ultimate purpose of enhancing

competiveness is to divergence bad effects with the

immediate step of product quality improvement. Therefore,

in the first period, improving competiveness delivers banks

to a better situation. However, a negative aspect of this issue

is that banks tend to focus on operational diversification but

they neglect intrinsic resources leading easily to unexpected

risks. In detail, they find the evidence of a U-shaped

relationship between competition and bank risk. The

probability of default goes up following an increase in bank

competition but it has a downward trend after reaching a

threshold. The idea was supported by Berger et al (2009),

Kasman and Kasman (2015).

We hypothesize that the nexus of competition and

credit risk is nonlinear (H3).

2.2. Methodology

2.2.1. Methodology

The two-step System GMM method is utilized to

examine whether credit risk is a function of competition.

Using benchmark estimators, such as Pooled Ordinary

Least Square (OLS), Fixed-effects (FE) or random effect

(RE) results in biasedness, leading to potentially

misleading inferences. This is because OLS considers

banks to be homogeneous. However, in reality, each bank

has different characteristics, such as attention level to risk,

competitiveness and corporate governance. Thus, OLS can

lead to biased estimates if these bank fixed effects are not

controlled. Otherwise, the other methods FE and RE cannot

cover potential endogenous concerns. There are two main

factors leading to endogeneity. Firstly, simultaneous

effects indicate that the casual nexus in the specification

can occur in two dimensions, so regression of these

explanatory variables may be correlated with error term,

leading to endogenous concern. Secondly, omitted-

variable bias explains that FE and RE estimations do not

take into account the external factors that are assumed in

error terms and are not correlated with explanatory

variables. However, these factors, namely, inflation,

economic crisis could explain changes in banks’ operation.

In addition, these traditional econometric techniques above

could not address all endogenous concerns with the

visibility of the lagged dependent variable.

The System Generalized Method of Moments (S-GMM)

initiated by Blundell and Bond (1998) uses the lagged

explanatory variables to establish instruments. The

conditions for the S-GMM estimation include: (1) the

visibility of over-identifying restrictions in order to ensure

the suitability of instrumental variables and no correlation

between instrumental variables and error term; (2) no

second-order autocorrelation in first-order differences.

Therefore, Hansen and Arellano-Bond tests are employed

with the aim of checking the suitability of two conditions

above. Besides, the two-step GMM method is better than the

one-step GMM because of using covariance-matrix in case

of existing serially correlated errors in the second-order or

heteroscedasticity. For these reasons, the two-step SGMM is

the most appropriate method to regress this relationship.

2.2.2. Empirical model

The model to assess the impact of competition on credit

risk in Vietnam commercial banks is as follows:

NPLi,t = β0 + β1NPLi,t-1 + β2COMi,t + β3CONi,t +ui,t (1)

Where NPLi, t-1 is the one period-lagged NPL rates,

COM and CON denote vectors of competition and control

variables, respectively.

The study also adds one period-lagged value of NPLs

as an independent variable in the model for the purpose of

indicating that the rate at which bank risk converges toward

a long-run level (Kasman and Kasman, 2015).

Moreover, to investigate the nonlinear relationship

between competition and credit risk, the squared

competition indices are added to the equation as follows:

NPLi,t = β0 + β1NPLi,t-1 + β2COMi,t + β3COMi,t2 + β4CONi,t + ui

(2)

2.2.3. Variable construction

Credit risk

Credit risk is as a ratio of loans in groups 3, 4 and 5 to

total bank loans or NPL ratio. If NPL is high and cannot be

controlled it will lead to failures. Hence, NPL is an

important factor that should be strictly followed because

NPLs are mainly employed to describe credit quality. In

the meanwhile, credit risk is one of the major risks. Hence,

credit risk is a concern of interest in terms of bank stability

(Kasman and Kasman, 2015). If the more the bad debt ratio

to total outstanding loans is, the riskier the lending

portfolios (Berger et al, 2009). Furthermore, the higher in

NPL ratio, the more probable in bank insolvency (Kabir

and Worthington, 2017).

Competition variables

The Lerner index initiated by Lerner (1934) is

employed to measure bank competitive extent because the

unstructured approach can evaluate market power of banks

with the concentration on the difference of price and

marginal costs (Tusha and Hashorva, 2015). Specifically,

the Lerner index defined as the difference between output

price and marginal cost exhibits that whether banks

evaluate their products higher than marginal cost (Berger

et al, 2009), If Lerner = 0, the market is perfectly

competitive and vice versa. if Lerner = 1, the market is

completely monopoly. The Lerner index is calculated as

50 Phan Tran Minh Hung, Phan Nguyen Bao Quynh

follows: Lerneri,t =Pi,t – MCi,t

Pi,t

Where Pi,t is the output price of bank i at time t which

is the ratio of total revenue to total assets and MCi, t is the

marginal cost of bank i at the end of period t.

Since the marginal cost of banks cannot be directly

observed, the MC is calculated based on total cost. The

bank’s total cost (TC) is calculated by the logarithm of cost

with one output factor (total assets (Qi, t)) and three inputs

(Wj) including: labor cost (W1 - the ratio of employee cost

to total asset); material cost (W2 - the ratio of non-interest

expense to fixed asset); capital cost (W3 - the ratio of

interest cost to total bank deposits (Berger et al, 2009).

Specifically, the specification of total cost is as follows:

lnTC = β0 + β1lnQi,t + β2

1

2lnQi,t

2 + ∑(γktlnWk,it)

3

k=1

+ ∑(ϕklnQitlnWk,it)

3

k=1

+ ∑ ∑(lnWk,i,tlnWj,i,t)

3

k=1

3

j=1

Following this, the marginal cost equation is computed

by taking the first derivative of the total cost function, by:

MC =TCit

Qit[β1 + β2lnQit + ∑ (ϕklnWk,it)

3

j=1]

Where (β) and (ϕ) coefficients are determined from the

regression outcomes of the total cost specification

constructed above.

Additionally, we also approach the traditional measure

of HHI (Herfindahl-Hirschman Index) in order to consider

as a proxy of competitive degree because this index is

employed to assess the contribution extent of each

individual in a population (competitive degree). According

to HHI approach, the competitive extent will be classified

as: HHI <0.01 (perfectly competitive); 0.01 <HHI <0.1

(highly competitive); 0.1 <HHI <0.18 (medium

competitive) and HHI> 0.18 (highly concentrated and

tending to be monopolistic). In this study, the HHI index is

to reflect concentration extent of loans (HHI_L) and

calculated as follows: HHI_L = ∑ Siki=1 ; where, Si is

calculated as market share of bank i on total loans of banks

in the banking system.

Control variables

Control variables include bank-specific and

macroeconomic conditions to control the net impact of

competition variables on bank risk. Bank variables include:

SIZE-the natural logarithm of total assets; TA_GRO-the

growth rate of total assets value of the current year

compared to the previous year and L_TA-the ratio of total

loans to total assets. Macroeconomic characteristics

contain: LN(GDP) - the natural logarithm of gross

domestic product and INF- inflation rate.

2.2.4. Data

A set of secondary data is collected from audited

financial statements, annual reports, prospectuses of

Vietnam commercial banks in the period of time from 2006

to 2016 through banks’ websites or stock exchanges and

some other websites. Database employed for this research

is unbalanced dynamic panel data because of the lack of

data of a few merged and acquired banks. In order to avoid

the adverse effects of insufficient data, banks with

consecutive five-years or more are chosen. Finally, there

are 27 banks selected with 207 observations. Moreover, we

reduce the effect of outliers by winsorizing all ratios at the

fifth and ninety-fifth percentiles.

3. Results and discussions

3.1. Descriptive statistics and correlations

Table 1 presents the summary statistics for the entire

sample. On average, a bank in the entire sample has NPLs

ratio of 0.022, being in the range of bad debt ratio. With

respect to competition variables, the means of LERNER

index and HHI are 34.3% and 0.132, respectively,

indicating that the competition extent is extremely serious.

An average bank in the sample has total asset logarithm of

18.025 million VND, a total asset growth speed of 39%, a

loan to total asset ratio of 51.2%.

In terms of macroeconomic characteristics, the means

of natural logarithm of GDP and inflation rate are

9.434 million VND and 8.5%, respectively.

Table 1. Descriptive statistics of variables

STT Variable Mean Std.

Error Min Max

1 NPL 0.022 0.016 0.000 0.096

2 LERNER 0.343 0.132 -0.049 0.624

3 HHI_L 0.132 0.032 0.104 0.223

4 SIZE 18.025 1.353 13.135 20.730

5 TA_GRO 0.390 0.709 -0.392 8.355

6 L_TA 0.512 0.138 0.041 0.808

7 LN(GDP) 9.434 0.186 9.026 9.654

8 INF 0.085 0.060 0.009 0.231

Source: Author’s calculation

An important hypothesis is that there is no

multicollinearity among the explanatory variables. All of

the correlation coefficients in Table 2 are less than 0.8.

Following Klein's rule of thumb, it can be concluded that

the independent variables in the equation are not multi-

collinear. Additionally, we also test multi-collinearity via

Variance Inflation Factor (VIF). However, these indices

fluctuate from 1.09 to 3.79 (less than 5), proving that it is

unlikely to have multicollinearity (to conserve space, these

VIF indexes are unreported in the paper)

Table 2. Correlation matrix

NPL LER

NER

HHI

_L SIZE

TA

_GRO L_TA

LN

(GDP) INF

NPL 1

LERNER -0.12 1

HHI_L -0.20 -0.08 1

SIZE 0.00 0.38 -0.34 1

TA_GRO -0.2 -0.00 0.32 -0.31 1

L_TA 0.00 0.22 0.05 0.17 -0.17 1

LN(GDP) 0.14 0.14 -0.78 0.47 -0.41 0.04 1

INF 0.06 -0.38 0.25 -0.21 0.03 -0.16 -0.44 1

Source: author’s calculation

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 51

3.2. Results and discussion

The results in Table 3 indicate that the models are

satisfactory in terms of serial correlation with the P-value

of AR (1) less than 0.05 and the P-value of AR (2) not

statistically significant. Therefore, there is no second order

autocorrelation. Moreover, Hansen test’s results record a

high P-value value which is over 0.1, hence it is impossible

to disprove the hypothesis that the instruments are

appropriate. This demonstrates that the instruments solve

the endogeneity. Therefore, the beta coefficients of the

regression model can be used for analysis.

Table 3. Regression results

Dep. Variable NPL

Model (1) (2) (3) (4)

L.NPL 0.845*** 0.555*** 0.369** 0.390**

LERNER 0.041*** 0.431***

LERNER_2 -0.637***

HHI_L -0.058* -2.747***

HHI_L_2 9.767***

SIZE -0.018*** -0.026*** -0.019*** -0.017***

TA_GRO 0.001 0.001* 0.002 0.003***

L_TA 0.027*** 0.041*** 0.017* 0.033**

LN(GDP) 0.087*** 0.109*** 0.137*** 0.094**

INF 0.144*** 0.131*** 0.109*** 0.085***

_cons -0.529*** -0.581*** -1.102*** -0.395***

No. instrument 23 23 23 23

Hansen test 0.238 0.307 0.310 0.472

AR1 0.015 0.007 0.024 0.011

AR2 0.970 0.691 0.356 0.259

Note: ***, **, * denote significance at 1%, 5%, 10%.

Source: author’s calculation

Table 3 shows that the coefficient of the Lerner index is

positive and highly significant at the 1% level with

magnitude of 0.041. This result suggests that the competition

is positively related to credit risk, implying that when banks

diversify their products to compete, the bad debt ratio is

larger, leading banks to become more unstable. The

conclusion is to support the "competition - financial

stability" opinion and, is consistent with our both predictions

(H1) and earlier findings in the literature (Berger et al, 2009;

Moch, 2013; Fiordelisi and Mare, 2014).

Turning to the impact of competition in lending

operation on the possibility of loan repayment, the

coefficient of HHI_L is negative and significant at the 10%

level. This indicates that there is a negative effect of the

expansion of lending operations on the proxies of credit risk.

However, the competition is positively associated to credit

risk because the proxy and the competitive extent have the

same magnitudes but are opposite in sign. This finding

suggests that the higher in outstanding loans, the more

serious in bad debt ratio. The reason for this trend is that

banks tend to lower evaluation standards, leading to take

more risk in order to maximize their profits. The conclusion

is consensus with the perspective of "competition - risk" and

in line with both what we anticipate (H1) and Kasman and

Kasman (2015). In fact, Vietnam commercial banks have

competed mainly based on traditional interest rate-related

activities such as loans, deposits. Basically, lending

operations account for a large proportion of total assets.

Therefore, increased competitive degree is commonly

attributed to strengthening loans. For this reason, riskier

loans is followed by expanding lending activities in

Vietnamese context because our banking system cannot

control the effects of overheated loans development and the

consequences of the 2008 financial crisis.

In particular, following Berger et al (2009), Martinez-

Miera and Repullo (2010), Kasman and Kasman (2015), the

nonlinear correlation (the U-sharp curve) between

competition and credit risk is next investigated. The results in

model 3 exhibit that the negative coefficient on the squared

Lerner index is statistically significant at the 1% level. The

new finding proves that the presence of the U-Shaped curve

between competition and borrowers’ affordability is

practical, is in line with what we anticipate above (H3).

To arrive at a more complete picture, we continue to find

consistent evidence of nonlinear influence of expanding in

lending sectors on credit risk. Model 4 documents that the

positive coefficient on the squared HHI is statistically

significant at the 1% level. However, the relationship

between the competitive extent and the proxy of this variable

is opposite. Therefore, we further point out the nonlinear

effect of loans competition on credit risk. In other words, the

U-shaped is the most suitable curve to describe the nonlinear

nexus, consistent with our expectation (H3)

In order to explain the non-linear relationship above in

the context of Vietnam, we document that 2012 is the

bottom of the U-sharped graph. The period 2010 - 2012 not

only witnessed the most competitive period but also

denoted that bad debt rose significantly. The reason could

be explained is that the banks aggressively competed not

based on internal forces in the worse macroeconomic

conditions, leading to more NPL rates. Furthermore, the

barriers in mobilization for foreign banks were removed

and they started to participate in the more equal

competition environment with domestic banks.

On the other hand, Vietnam commercial banks’

inefficiency also led to an increase of competition and

credit risk. In the period 2006-2010, many established

banks generated the more competitive environment among

banks along with the proliferation of Vietnam economy.

However, too many small-scaled banks existed because

they started from rural commercial banks and were

converted into urban commercial banks which had a rapid

growth of assets and loan portfolios. As a result, the bad

debt increased in the next phase and destabilized banking

system. In addition, the application of the maximum

lending rate makes competitiveness among banks more

stressful. The reason is that Vietnam commercial banks

have developed by focusing on two main traditional

products namely deposit and loans activities. Hence, banks

mainly compete with one another for interest.

Realizing the repercussions of extremely quick

development in lending sector, a series of policies was

established to be contributive to address NPLs. In the

meanwhile, competition extent remained even fiercer,

generating the greater credit risk in the period of time from

2013-2016. (In order to conserve space the effects of

52 Phan Tran Minh Hung, Phan Nguyen Bao Quynh

control variables are not reported in this paper)

4. Conclusions and implications

This study provides empirical evidence that the less fierce

the competition is, the greater the credit risk is in Vietnam

commercial bank sector. To arrive at a more complete picture,

we also further find that the more intense the banks is in

lending sector, the poorer the credit risk is. Overall, this

indicates that expanding comprehensive or lending operations

to compete is one of the main sources of increasing credit risk.

In other words, enhancing competitiveness in whole or

lending operations will boost credit risk. In the meanwhile,

this research also points out that the correlation between

competition and credit risk is non-linear with U-Shaped curve,

implying that the positive and negative impact only happen at

the right or left of the bottom, respectively.

Based on the research results, some implications are

proposed to alleviate credit worthiness when Vietnam

commercial banks tend to be more competitive as follows:

Firstly, although the expansion of products is considered

as one of the main reasons of increasing loan-related risks,

this does not mean that banks have to stop competitive

strategies. In sharp contrast, banks need to be encouraged to

compete to other both domestic and foreign banks more

aggressively because competition is the dispensable trend to

obtain the ultimate goal of value maximization. This requires

each bank to have appropriate strategies, including: not

lowering lending evaluation standards, promoting quality and

applying cutting-edge technologies. In addition, Vietnam

commercial banks need controls lending operation-related

risks in order to partially alleviate NPLs, take measures to

detect and address the threat of lending activities.

Secondly, the aftermath of the financial crisis of 2008

in Vietnam banking system is that NPL rate surges,

exhibiting that debt loans is influenced by macroeconomic

conditions. Therefore, in order to ensure the safe range of

bad debts, macroeconomic factors such as inflation,

unemployment, and economic growth must be maintained.

To obtain this, the state bank of Vietnam plays a role in

framing the most appropriate policies for the government.

More specifically, the monetary policies must be suitable

with the context of Vietnam in each period to control

inflation but ensure high economic growth for the ultimate

purposes of increased competition and decreased NPLs.

Thirdly, the determination of the U-curve bottom is an

extremely essential intermediate step because this is a

background to consider the possible implications for

competition in each period of time. Specifically, the

competitive strategies should be enhanced in short term in

order to improve credit risk. However, this trend will not

be encouraged if credit risk reaches the bottom because of

the serious repercussion of competition.

REFERENCES

[1] Acharya V., Gromb D., Yorulmazer T. (2012), Imperfect competition in the interbank market for liquidity as a rational for

central banking, American Economic Journal: Macroeconomic,

Volume 4, Issue 3, Pages 184–217.

[2] Allen F., Gale D. (2004), Competition and financial stability, Journal

of Money, Credit and Banking, Volume 36, Issue 3, Pages 433-480.

[3] Anginer D., Demirguc-Kunt A., Zhu M. (2014), How does

competition affect bank systemic risk?, Journal of Financial Intermediation, Volume 23, Issue 1, Pages 1-26.

[4] Arellano M., Bond S. (1991), Some Tests of Specification for Panel

Data: Monte Carlo Evidence and an Application to Employment

Equations, The Review of Economic Studies, Volume 58, Issue 2,

Pages 277-297.

[5] Berger A N., Klapper L F., Turk-Ariss R. (2009), Bank competition

and financial stability, Journal of Financial Service Research, Volume 35, Issue 2, Pages 99-118.

[6] Blundell R., Bond S. (1998), Initial conditions and moment

restriction in dynamic panel data models, Journal of Econometrics,

Volume 87, Issue 1, Pages 115-143.

[7] Boyd J H., De Nicolo G. (2005), The theory of bank risk taking and

competition revisited, Journal of Finance, Volume 60, Issue 3,

Pages 1329-1343.

[8] Fiordelisi F., Mare D. S. (2014), Competition and financial stability

in European cooperative banks, Journal of International Money and Finance, Volume 45, Pages 1–16.

[9] Kabir M N., Worthington A C. (2017), The ‘competition-

stability/fragility’ nexus: A comparative analysis of Islamic and

conventional banks, International Review of Financial Analysis,

Volume 50, Pages 111-128.

[10] Kasman S., Kasman A. (2015), Bank competition, concentration and

financial stability in the Turkish banking industry, Economic Systems, Volume 39, Issue 3, Pages 502-517.

[11] Keeley M. (1990), Deposit Insurance, Risk and Market Power in

Banking, American Economic Review, Volume 80, Issue 5, Pages

1183-1200.

[12] Koetter M., Poghosyan T. (2009), The identification of technology

regimes in banking: Implication for the market power-fragility nexus, Journal of Banking and Finance, Volume 33, Issue 8, Pages 1413-1422.

[13] Lerner A P. (1934), The concept of monopoly and the measurement

of monopoly power, The Review of Economic Studies, Volume 1,

Issue 3, Pages 157–175.

[14] Mishkin F S. (1999), Global Financial Instability: Framework,

Events, Issues, Journal of Economic Perspectives, Volume 13, Issue

4, Pages 3-20.

[15] Moch N. (2013), Competition in fragmented markets: New evidence

from German banking industry in the light of the subprime crisis, Journal

of Banking and Finance, Volume 37, Issue 8, Pages 2908-2919.

[16] Martinez-Miera D., Repullo R., (2010), Does competition reduce the

risk of bank failure? Review of Financial Studies, Volume 23, Issue

10, Pages 3638-3664.

[17] Soedarmono W., Machrouh F., Tazari A. (2011), Bank competition,

crisis and risk taking: Evidence from emerging markets in Asia,

Journal of International Financial Markets, Institution and Money, Volume 23, Pages 196-221.

[18] Stiglitz J E., Weiss A. (1981), Credit Rationing in Markets with

Imperfect Information, The American Economic Review, Volume

71, Issue 3, Pages 393-410.

[19] Tusha A., Hashorva A. (2015), Lerner Index and Boone indicator to

Albanian banking sector, International Journal of Scientific and

Engineering Research, Volume 6, Issue 10, Pages 467-471.

(The Board of Editors received the paper on 07/11/2017, its review was completed on 11/12/2017)

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 53

INSIGHTS INTO THE IMPLEMENTATION OF ENGLISH MEDIUM

INSTRUCTION AT VIETNAMESE UNIVERSITIES: EXPLORING THE

LEARNING EXPERIENCES OF BUSINESS STUDENTS

Le Thi Thuy Nhung

Banking University of Ho Chi Minh City; nhungltt@buh.edu.vn

Abstract - This study critically examines the effectiveness of English-medium instruction (EMI) at Vietnamese universities through students’ perspectives. Data is collected via 10 focus group interviews with undergraduate students enrolling in Business courses at four universities in Ho Chi Minh City, Vietnam. The findings show that most of the students favor EMI. However, they face difficulties in dealing with demands of their programs. Students take the challenges for granted, having a strong belief in the social and economic returns of EMI. Still, the extent to which students obtain prestigious, well-paid jobs as a result of their EMI experiences has yet to be demonstrated. The move to EMI programs at universities in Vietnam and in Asian countries is gathering momentum. The imperative now is to make them as effective as possible.

Key words - English Medium Instruction for non-language subjects; students; perceptions; challenges; globalization; internationalization; Vietnam

1. Introduction

The status of English as an international language exerts

a considerable impact on educational policies and practices

in educational systems all over the world (Nunan, 2003).

Adopting English as a Medium Instruction (EMI) is one of

the major educational trends in non-English speaking

countries in the past decades. This trend is prominent in

higher education institutions not only in European but also

in Asian countries where English is traditionally taught as a

stand-alone foreign language subject.

In Vietnam, EMI has recently been introduced into

university programs. Despite the initial appeal of EMI, the

effectiveness of EMI programs appears to vary because

many Vietnamese universities are under-resourced and are

not adequately prepared for EMI. Anecdotal evidence

shows that implementation of EMI in Vietnamese

universities is problematic because of shortage of qualified

teaching staff, study materials, and resources (Ha Anh,

2013; Pham, 2010). There is a lack of research into the

current practices of EMI in Vietnamese universities to add

weight to the anecdotal evidence.

The present study investigates the learning experiences

of students undertaking EMI courses in the field of

Business and Management studies at four universities in

Ho Chi Minh. It examines their attitudes towards EMI,

their perceptions of the benefits of EMI; and their

perceptions of the challenges facing them in doing EMI

courses. The study contributes to the limited research on

EMI in Vietnamese universities. It provides insights into

current implementation of EMI in Vietnamese higher

education contexts.

2. EMI policy and development of EMI in Vietnamese

higher education

The Vietnamese government in the past decades has

encouraged higher education institutions to introduce EMI

into university programs in a number of proposals. Some

of them are: Foreign Language Education 2008-2020’;

Vietnam’s Educational Strategies 2009-2020; and

Fundamental and Comprehensive Reform of Higher

Education 2006-2020. These documents provide reasons

for the introduction of EMI in Vietnamese higher

education: to enhance graduates’ employability in the

international workplace and thus enhance the quality of

human resources; to enhance the quality of Vietnamese

higher education through a process of internationalization;

and to improve the English proficiency of Vietnamese

lecturers through using English for teaching, research, and

professional exchange.

In response to this EMI policy, many EMI courses and

programs have been offered in both public and private

universities in Vietnam since the early 2000s. There are 35

Advanced Program initiatives supervised by Ministry of

Education and Training (MOET) in 23 Vietnamese

universities between 2008 and 2015. These Advanced

Programs allow selected Vietnamese universities to import

curricula from prestigious universities abroad. The

International Standard Programs (ISP) were introduced at

the Vietnam National University which offered 16 training

courses in English in 2008 (Vu & Burns, 2014). The High

Quality Programs promulgated by MOET were launched

in 2014 in several Vietnamese universities. These

programs share the following features: they model partly

or entirely undergraduate degrees’ curricula from

prestigious Western universities; they use English as the

medium of instruction to deliver course content; and they

are instructed by highly qualified staff. The intention of

these programs is to enhance the quality of university

teaching and to improve students’ English proficiency,

thereby enhancing graduates’ employability.

EMI policy also triggers the establishment of English

medium universities. In 2003, the International University,

the first public English-medium university, was established

in Ho Chi Minh City to offer all its programs in English. In

2009, the Government launched the New Model University

Project to build four ‘world-class’ universities with US$400

million in loans from the Asian Development Bank and the

World Bank. These universities were to be public, research-

oriented, and foreign-managed institutions. They would be

modelled after prestigious universities in the United

Kingdom and the United States and maintain high quality

standards, academic excellence, and assessment and

governance procedures that match international standards.

These four universities were to be English medium

universities. Two of them are currently in operation.

54 Le Thi Thuy Nhung

3. Review of relevant literature

The medium of instruction refers to the language used

to deliver the content of non-language subjects other than

the language studied as an object in itself. English Medium

Instruction (EMI) is widely adopted at the tertiary level in

non-dominant English speaking countries to teach subjects

in disciplinary areas such as Economics, Science,

Engineering, Medicine and Information Technology. EMI

focusses less on pedagogical strategies and more on

content, whereas other bilingual approaches focus on both

content and language (Hamid, Jahan, & Islam, 2013). The

language-learning goals in EMI class often are not made

explicit. In other words, EMI is a content-driven approach

which uses a foreign language, and EMI courses are

targeted at post-secondary students (Park, 2007). The

growth of EMI in higher education in non-English

speaking countries is driven by two major forces:

globalization and internationalization (Tollefson & Tsui,

2004). EMI is being introduced into universities in Europe

and East Asia because of national strategies for human

capital development and internationalization of education.

There is a large volume of research into EMI. Many

studies on EMI in Europe and North America used surveys

to identify attitudes of students and lecturers towards EMI.

EMI appears to generate favorable attitudes among

students. EMI prepares graduates for the international

workplace and further education, enhances students’

English proficiency and increases their self-confidence.

Similar results were found in EMI studies in universities in

East Asian countries where students’ preference for EMI

was mainly driven by career and academic purposes (Byun

et al., 2011; Manakul, 2007). Researchers also looked into

the impact of EMI on students’ performance and learning

outcomes. EMI appears to increase students’ English

language proficiency, especially their specialized

vocabulary knowledge (Costa & Coleman, 2012).

The effects of EMI on student comprehension of

subject content have been examined. Studies show mixed

experience of EMI lectures for students. While some

students report they donot have problems in understanding

EMI lectures, others experience problems understanding

lectures delivered in English (Airey & Linder, 2006).

Problems with EMI lectures include complicated

specialized vocabulary and concepts, lecturers’ unclear

pronunciation and unfamiliar accents, fast delivery rate,

and students’ poor concentration during EMI lectures.

Apart from having difficulties understanding lectures,

students report difficulty with academic reading, especially

understanding new terminology and concepts in textbooks.

For writing, mastering relevant academic writing

conventions are challenging for many students. Spoken

interaction is demanding for students in English-medium

courses. Interaction remains limited because of students’

lack of competence and confidence in speaking English.

There is evidence that EMI increases students’

workload (Tatzl, 2011; Vinke, Snippe, & Jochems, 1998).

Time is spent on reading course materials, looking up

vocabulary in preparation for lectures and revising for

examinations. Despite students’ positive attitudes towards

EMI, there remains dissatisfaction with EMI (Evans &

Morrison, 2011). Many students call for improvement in

the quality of teaching and better English proficiency

among EMI lecturers.

4. The study

With an aim to provide insights into the current practice

of EMI in Vietnamese universities, the study seeks to

answer the following questions:

• What are students’ general perceptions of EMI?

• What do students perceive as the most challenging

and satisfying aspects of EMI?

• What are the implications of students’ reported

experiences of EMI for the practice of EMI in Vietnamese

universities?

Ten focus group interviews are conducted with 46

students enrolling in Business courses at four universities in

Ho Chi Minh City, Vietnam. There are four to six students per

group. At the time of the interview, the students are in the third

or fourth year of their Business programs so they have

extensive experiences with EMI. The interviews are recorded,

transcribed verbatim, and translated from Vietnamese into

English. The data is analyzed through a process of pattern-

coding to identify themes (Miles & Huberman, 1994)

corresponding to each research question. The themes and

subthemes are grouped to establish an overall picture of

students’ experiences of English-medium instruction.

5. Findings and discussion

5.1. Students’ general perception of English Medium

Instruction

When asked about their perceptions of EMI, many

students share the view that EMI is important for job

seeking, job performance, and career advancement. The

students believe that EMI will provide them with a

satisfactory English proficiency level to meet job

requirements, given that most professional positions in the

labor demand English competence. One of the students

express her opinion in this respect:

Many job interviews nowadays are conducted in

English. I think students who do EMI tend to do better in

interviews than those who don’t.

The students believe that they will gain a competitive

edge over graduates from Vietnamese-medium programs.

Many expect to possess a high level of English proficiency

and professional knowledge by the end of their study

programs. These ‘dual assets’ will enable them to compete

successfully with other Vietnamese university graduates

whose English proficiency tends to be lower. EMI will

enable them to compete against foreign job-seekers in the

future labor market. EMI is seen to open better job

opportunities because graduates can not only work for local

companies but also for ‘foreign’ enterprises where English

is the working language. Many students share the same

view with the following student:

… Vietnam has opened its employment market to foreign

workers as part of the trade agreement with World Trade

Organization. It means, when we graduate, we have to

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 55

compete not only with Vietnamese graduates but also with

foreigners for jobs right here in our country. Therefore, EMI

would provide a head start for us to survive such competition.

Most students are generally in favor of English-medium

instruction. EMI is associated with an advantage in

employment and pursuit of higher education. Students’

favorable attitudes towards EMI are also found in previous

studies conducted internationally. In Vietnam, English is

recognized as an important tool for “job seeking, job

promotion and overseas studies” (Do, 2006, p.8). Since

most professional positions require English proficiency,

people are increasingly aware of the role that English plays

in their future career and individual success. In this study,

most students assume that undertaking EMI will increase

their English proficiency and thus contribute to their

academic and career advancement. EMI will bring students

better career prospect and greater economic returns.

Students also assume that EMI is a guarantee of a high-

quality education. However, to date in Vietnam there is

little research data or report to show that EMI students do

move into high paying, prestigious careers.

5.2. Challenges of English-medium instruction

The focus group discussions reveal the following

challenges: comprehending textbooks; understanding EMI

lectures; expressing ideas in English; and dealing with

written assignments.

5.2.1. Comprehending English textbooks

The students report that the textbooks used for Business

studies are mainly American and British publications. They

acknowledge several benefits of using imported textbooks

such as frequent updates of knowledge, originality and

accuracy of content in comparison with its translation

versions in Vietnamese. However, the imported English

textbooks without any modification and simplification pose

several challenges such as heavy load of reading, difficult

terminology, concepts and unfamiliar content knowledge.

Several students share the following ideas:

…I find it difficult to understand courses books on

Micro or Macroeconomics written by Mankiw (Greg

Mankiw is an American macroeconomist – information

added by the researcher). The books are about the US

economy from the perspective of an American economist.

Things are different.

Similarly, some students claim that since they study

American Accounting instead of Vietnamese Accounting,

therefore, they are worried if they will work for a

Vietnamese company and do not have much knowledge

about how the Vietnamese accounting system works. In

addition, some students are concerned that they will

perform poorly in Vietnamese companies because they

study overseas-based content at university.

Use of imported English textbooks written for English-

native students is problematic for EMI students. A heavy

reading load, foreign structure and presentation of

information, and unfamiliar academic writing styles in

English course books present problems for many students.

In addition, a majority of students do not have sufficient

exposure to academic reading before university. An

examination of EFL textbooks at upper-secondary schools

in Vietnam show that the reading passages are quite short.

This indicates that Vietnamese students and their overseas

counterparts in EFL contexts have limited experience and

insufficient proficiency of reading before university

(Hellekjær, 2005). Students enter university with poor

reading skills. They rely heavily on dictionaries and

struggle with the reading requirements.

5.2.2. Understanding EMI lectures

In general, the students indicate that their understanding

of the content of EMI lectures is hindered because of

problems understanding Business terminology, lecturers’

English, and ineffective lecturing skills. Some students find

it difficult to understand accents of lecturers, especially those

who have lived overseas for an extended period of time.

Students complain about lecturers’ fast delivery rate and

unclear pronunciation. They are critical of lecturers’

pronunciation, especially English spoken by senior lecturers.

The students describe some lectures as ‘very brief’ and

‘dry’ because of some lecturers’ unwillingness to elaborate

or clarify the content. Some lecturers are reluctant to provide

examples or anecdotes to enhance understanding. As a

result, some lectures are boring because the lecturers cannot

use English sufficiently well. The students claim that they

miss out some main points of the lectures because the

lecturers do not use any English language signposts to

indicate important content. One of the students is frustrated:

… He speaks with the same tone from the start till the

end. I can’t take notes at all because I don’t know which

point is the main point.

In general, students indicate that listening to EMI

lectures is more demanding cognitively because they have

to pay close attention. In the first few courses of EMI, many

students do not understand the content of the lectures

completely because of the complexity of the Business

specialization.

Students’ problems with understanding lectures may be

attributed in part to lecturers’ lack of command of English

and experience with EMI teaching. Most lecturers in the

current study hold at least a Master’s degree mainly from

English-speaking countries. Most of them are young. They

may not have had an opportunity to take part in an English

language training courses designed for teaching EMI

because such courses are new in Vietnam. It seems that these

lecturers are assumed by program administrators to have

sufficient English proficiency. Nevertheless, given students’

frequent complaints about lecturers’ oral English

proficiency, the assumption that most EMI lecturers are well

qualified linguistically is questionable. In addition, many

students lack the English proficiency to do EMI courses.

5.2.3. Expressing ideas in English

The students experience difficulties expressing their

ideas in English because they lack specialized vocabulary,

and a deep understanding of the content. The inability to

elaborate or explain in English stops them getting involved

in classroom discussions.

...When we have a discussion in Vietnamese, we had

lots of things to say to each other. However, when a group

56 Le Thi Thuy Nhung

member reported to the class what we discussed, he often

fails to translate the content from Vietnamese into English.

Most students say that they only speak English when

exchange students are present. In other circumstances, they

speak entirely in Vietnamese. A few students want to speak

English but ‘no one wants to’ and some others are ‘afraid that

group members will not understand’ what they say in

English. Students expect that EMI will create an environment

for them to practise oral English. In fact, there is limited

English interaction outside the EMI class because of

students’ lack of desire to speak English on campus. Previous

research shows that EMI students do not take opportunities

to speak English outside class. They resort to local languages

for ease of communication (Bolton & Kuteeva, 2012).

5.2.4. Dealing with written assignments

Dealing with written assignments in English is difficult

for the majority of students in the study. Three areas of

writing are problematic: academic writing conventions;

choosing appropriate academic writing style; and using

accurate terminology. The students are not prepared to

write assignments in their EMI courses. Most students in

the study come straight to university from high schools

where they have very limited experience of writing in

English. English writing at high schools is much more

‘simpler’ and different from English writing at university.

Meanwhile, the English courses offered at their universities

provide only ‘general knowledge of academic writing’

such as writing paragraphs and essays. They do not cover

academic writing conventions in their EMI courses such as

writing literature reviews, business reports and essays, and

referencing their work. As a result, most students have to

consult lecturers and sources on the Internet to help them

deal with writing assignments. Problems with academic

writing has been reported as a major issue for students in

EMI courses in Scandinavian countries (Hellekjaer, 2010)

and in non-dominant English speaking contexts elsewhere

(Campbell & Li, 2008) and Vietnam is not an exception.

5.3. Students’ satisfaction with EMI

There is a relatively high level of satisfaction with EMI

courses. Sources of satisfaction are improvement in

English language proficiency, increase confidence in

learning, opportunities for cultural exchange with

international students, and highly qualified academic staff.

The students from the English-medium university

appreciate the presence of teaching assistants (TAs) in their

courses because the TAs are very supportive and helpful to

assist students in dealing with course exercises.

The students consider the EMI curriculum as ‘more

advanced’ than the Vietnamese curriculum. The EMI

curriculum includes an international dimension by

providing new courses such as International Marketing

and fieldtrips which are not available to mainstream

students. In addition, EMI lecturers are described as highly

trained, understanding, supportive, and friendly. Young

lecturers particularly are considered as ‘highly proficient in

English and technologically sound’ because they use their

knowledge of information technology more actively and

effectively. For example, some junior lecturers design

interactive webpages to post lecture slides, study materials,

and discussion boards with their students. Students know

they can understand course content better when delivered

in Vietnamese. However, EMI courses are ‘more

challenging but more interesting and worthwhile’. The

perceived benefits of EMI seem to outweigh its challenges.

6. Conclusion

The students in the study see the potential benefits of

EMI for their future careers and show positive attitudes

towards EMI. Still, EMI pose numerous challenges to

these students. They seem to be inadequately prepared for

EMI academically and linguistically. The students face

English language-related problems during their study.

These problems are alleviated as students progress

through the program. The majority of students indicate

that they adapt to the EMI environment over time by

developing coping learning strategies to deal with the

demands of EMI courses.

The use of English as a medium of instruction in higher

education in non-dominant English speaking countries

continues to increase. As Byun et al. (2011) indicate, the

issue is not whether or not to adopt EMI but how to

implement EMI successfully. The threshold level of

English proficiency for EMI students should be increased.

English language training courses should be geared

towards disciplinary and academic needs to help students

cope with the requirements of EMI courses. Imported

textbooks should be modified or simplified to suit

Vietnamese students. Vietnamese universities should do

more to encourage students to speak English on campus

and outside EMI classes. More English language support

and training courses designed to enhance EMI lecturers’

performance are recommended. University administrators

should take lecturers’ oral English proficiency into

consideration when selecting academic staff. If these

preliminary conditions are established, the implementation

of EMI in Vietnamese universities will yield more

promising outcomes. Universities with insufficient

resources can consider setting up bilingual programs in two

languages, Vietnamese and English, before offering

English medium-only programs.

REFERENCES

[1] Airey, J., & Linder, C. (2006). Language and the experience of learning university Physics in Sweden. European Journal of

Physics, 27(3), 553-560.

[2] Bolton, K., & Kuteeva, M. (2012). English as an academic language

at a Swedish university: Parallel language use and the ‘threat’ of

English. Journal of Multilingual and Multicultural Development, 33(5), 429-447.

[3] Byun, K., Chu, H., Kim, M., Park, I., Kim, S., & Jung, J. (2011).

English medium teaching in Korean higher education: Policy,

debates and reality. Higher Education, 62(4), 431-449.

[4] Campbell, J., & Li, M. (2008). Asian students’ voices: An empirical study

of Asian students’ learning experiences at a New Zealand university. Journal of Studies in International Education, 12(4), 375-396.

[5] Costa, F., & Coleman, J. A. (2012). A survey of English-medium

instruction in Italian higher education. International Journal of

Bilingual Education and Bilingualism.

[6] Do, H. T. (2006). The role of English in Vietnam's foreign language

policy: A brief history. Paper presented at the 19th Annual English

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 57

Australia Conference, Perth, Western Australia.

[7] Evans, S., & Morrison, B. (2011). Meeting the challenges of English

medium higher education: The first-year experience in Hong Kong.

English for Specific Purposes, 30(3), 198-208.

[8] Ha Anh (Producer). (2013). Chương trình tiên tiến đang..thụt lùi

[Advanced programs are moving backwards]. Retrieved from http://thanhnien.vn/giao-duc/chuong-trinh-tien-tien-dang-thut-lui-

8583.html

[9] Hamid, M. O., Jahan, I., & Islam, M. (2013). Medium of instruction

policies and language practices, ideologies and institutional divides:

Voices of teachers and students in a private university in Bangladesh. Current Issues in Language Planning, 14(1), 144-163.

[10] Hellekjær, G. (2005). The acid test: Does upper secondary EFL

instruction effectively prepare Norwegian students for the reading of

English textbooks at colleges and universities? (Unpublished doctoral dissertation), University of Oslo, Oslo, Norway.

[11] Hellekjaer, G. (2010). Language matters: Assessing lecture

comprehension in Norwegian English medium higher education. In

C. Dalton-Puffer, T. Nikula & U. Smit (Eds.), Language use and

language learning in CLIL classroom (pp. 233-258). Amsterdam: John Benjamins.

[12] Manakul, W. (2007). English in engineering education for Japanese

graduate students. Australasian Association of Engineering

Education, 13(2), 53-63.

[13] Miles, M., & Huberman, A. (1994). Qualitative data analysis: An

expanded sourcebook (2nd ed.). Thousands Oaks, CA: Sage.

[14] Nunan, D. (2003). The impact of English as a global language on

educational policies and practice in the Asia-Pacific region. TESOL Quarterly, 37(4), 589-613.

[15] Pham, T. L. (2010). Should we use English as the medium of

instruction in Vietnamese universities? Paper presented at the Phat

trien va giu gin su trong sang cua tieng Viet trong thoi ky hoi nhap

quoc te hien nay [Promoting and preserving the purity of Vietnamese language in the current era of global integration], Hochiminh

University of Foreign Languages and Information Technology

(HUFLIT), Hochiminh City, Vietnam. http://lypham.net/joomla/index.php?option=com_content&task=vie

w&id=190

[16] Tatzl, D. (2011). English medium masters’ programmes at an Austrian

university of applied sciences: Attitudes, experiences and challenges. Journal of English for Academic Purposes, 10(4), 252–270.

[17] Tollefson, J. W., & Tsui, A. B. M. (2004). Contexts of medium of

instruction policy. In J. W. Tollefson & A. B. M. Tsui (Eds.),

Medium of instruction policies: Which agenda? Whose agenda? (pp.

283-294). Mahwah, NJ: Lawrence Erlbaum.

[18] Vinke, A. A., Snippe, J., & Jochems, W. (1998). English medium

content courses in non-English higher education: A study of lecturer experiences and teaching behaviours. Teaching in Higher

Education, 3(3), 383-394.

[19] Vu, T. T. N., & Burns, A. (2014). English as a medium of

instruction: Challenges for Vietnamese tertiary lecturers. The

journal of Asia TEFL, 11(3), 1-31.

(The Board of Editors received the paper on 14/11/2017, its review was completed on 07/12/2017)

58 Quy H. Nguyen

WORK-INTEGRATED LEARNING CURRICULUM: AN EFFECTIVE DESIGN TO

ENHANCE GRADUATE EMPLOYMENT RATE IN VIETNAM

Quy H. Nguyen

University of Foreign Language Studies, The University of Danang; nhquy@ufl.udn.vn

Abstract - Universities are improving quality of education to attract more students. Students will choose to study at a university that not only equips them with good knowledge of disciplines but also with skills to work in the workplace. Ministry of Education and Training of Vietnam requires universities to conduct a survey about employment rate of graduates within 12 months of graduation. The graduate employment rate of a university cognitively reflects the teaching quality of the institution as well as its reputation. This puts pressures on Vietnamese universities to apply strategies to improve student learning outcomes and prepare them to become work-ready employees. However, there are gaps between what students learn at universities and what they are expected to perform at work. Currently, many graduates do not meet the demands of enterprises and become unemployed after graduation. Therefore, a university-enterprise partnership in the implementation of work-integrated learning (WIL) curriculum is necessary. WIL is a course focusing on the integration of disciplinary knowledge and workplace practice. The paper suggests the need to design WIL into university curriculum in Vietnam although it has some challenges. The paper also informs university leaders, policy makers, and employers in Vietnam about the importance of WIL.

Key words - work-integrated learning; graduate employment; employability skills; graduate attributes; curriculum

1. Introduction

1.1. Social and Cultural Context of University-Enterprise

Partnerships in Vietnam

In a knowledge-based economy and society of the 21st

century, education and training plays an important role in the

development of all countries. The missions of Higher

Education system and universities are greatly emphasized.

Obviously, university stakeholders such as employers

around the world express their growing concern about

teaching and learning performance of universities in terms

of learning outcomes of students. This gives pressures to

universities as they are expected to develop not only

theoretical knowledge, technical skills but also generic skills

of students. Generic skills are used to refer to graduate,

professional, transferable, work ready employability. There

is no exception in the context of Vietnam.

Many employers across the country have expressed

their concerns regarding gaps in graduate skills. They

require graduates to have a certain level of language

proficiency (especially English) and other generic skills

such as good communication skills, teamwork and

personal skills, critical thinking, presentation or decision-

making. Most of them perceive that Vietnamese students

have a good knowledge base but lack the ability to

intelligently apply theoretical knowledge to the work

setting. In practice, Vietnamese enterprises not only invest

little time and effort in shaping the curriculum or processes

for learning of students but also have little care about the

learning processes and products of universities. However,

they often place great emphasis on the outcomes of student

learning and even complain about the quality of teaching

and training of the universities. Graduates are often

complained to lack disciplinary knowledge, foreign

languages, and generic employability skills (VL, 2016;

Vương, 2017).

Not only employers, Vietnamese graduates admit that they

are not sufficiently provided generic skills at the

universities. Students often complain that they have few

chances or activities to develop skills and practical

knowledge. As employability skills are under-developed,

the professional preparation for future career of students

attracts much attention from students, their parents,

employers, and society. The under -preparation for

graduates in terms of professional knowledge and the soft

skills required at the workplace brings about disadvantages

for Vietnamese graduates in the international market.

While many foreign companies rate teamwork, initiative

and communication skills as important skills at work, these

are usually the weak capabilities of Vietnamese graduates.

This problem is a challenge for Vietnamese graduates in

the international labour market where they are equally

expected to perform the same work skills as graduates of

other developed countries to be successfully recruited.

There are reflections from students that they must take

some compulsory courses that do not directly relate to their

disciplines at the expense of critical thinking and other

skills expected by employers. Facing skill shortage, it is

very difficult for Vietnamese graduates to integrate into the

international labour market. Therefore, an increased rate of

graduate unemployment is a result of this problem.

1.2. Graduate Unemployment Rate in Vietnam

According to statistics of the Ministry of Labour,

Invalids and Social Affairs of Vietnam, in the fourth

quarter of 2016, some 471,000 graduates failed to find a

job in their chosen field (Vietnamnet, 2017). One of the

popular reasons for this high unemployment rate is that

graduates do not meet the demands of the workplaces.

Nguyen (2017) reported the opinion of a 25-year-old

unemployed graduate that students only received heavy

theoretical training. The statement of the above-mentioned

student is like the previous observation of Professor

Nguyễn Minh Thuyết who is a legislator and Head of the

Committee of Culture, Education, and Youth of Vietnam

National Assembly. According to Professor Thuyet, the

training of impractical subjects in universities should be

reduced. However, the progress is still very slow and not

much has been done. Consequently, “unemployment

among college graduates is the highest in the nation”

(Nguyen, 2017, p.1). According to that report, universities

in Vietnam have been failing to prepare graduates for

higher-level employment. While Vietnam’s economy

needs a more highly qualified workforce, such a high rate

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 59

of unemployment among university graduates slows down

the development of the country.

Accordingly, T. Nguyen (2016) reported the opinion of

Emanuela di Gropello, an economics expert in the human

development department of the World Bank’s East Asia and

Pacific region: “Higher education in Vietnam has made

significant progress. However, it still needs improvement in

training and providing students with soft skills and in

cooperating with research institutes and especially

workplaces.” Also, “90% of Vietnamese graduates do not

possess the soft skills necessary for work and that this is one

of the most important reasons why every year more than

400,000 university graduates fail to get a job” (opinion of Mr.

Tran Trong Thanh, chairman of one of the biggest companies

in Vietnam, Vinapo Inc cited in T. Nguyen, 2016).

1.3. An Emergence of Changes in the Changing World of

Higher Education

In the context of Vietnam where the number of

unemployed graduates is increasing, the pressures that

students, their parents, and society put on universities are

overwhelming. Although universities are becoming more

conscious of the above-mentioned needs of enterprises,

they are not able to meet the demands of enterprises

because enterprises are traditionally passive partners in the

process of education and training of universities. In the new

context of the competitive markets, enterprises must

become active partners of universities. Their collaboration

and contribution to universities are considered as an

investment in the universities in which they will get a

return in terms of skilled workforces.

According to the Law on Higher Education 2012 of

Vietnam, all universities must be accredited by 2020. This

strategic plan is not only to ensure the institutional quality

but also to help universities approach the international

standards of higher education. Being accredited

institutions, universities have more responsibility to

society in terms of authority, power, and accountability

over the contents of teaching and learning. More

importantly, the outcomes of student learning in terms of

graduate attributes should meet the national academic

standards and even international ones to improve the

employment rate of graduates.

2. Work-Integrated Learning Curriculum

2.1. Work-Integrated Learning

Work-integrated learning (WIL) is a curriculum design

in which students combine theory at universities and practice

at the workplaces that are relevant to their training programs

and future career. More specifically, “WIL is a curriculum

strategy that incorporates high-authenticity real-world

experiences into the curriculum” (Smith & Worsfold, 2014,

p. 1070). Similarly, Jackson (2015) defines WIL as the

practice of combining traditional academic study, or formal

learning, with student exposure to the world-of-work in their

chosen profession, has a core aim of better preparing

undergraduates for entry into the workforce.

As a strategy of placing students in real workplaces or

highly authentic simulated environments and training

workplaces, WIL is widely used to help students. In this

practice-based approach of teaching programs, course

content is fused with work tasks.

Although WIL is not a new term in the world of higher

education, this jargon does not seem to be familiar with

many universities in Vietnam. In Vietnam, final-year

students often have a period of practice at workplaces to

experience work and to learn professional practices. But

that is not enough to develop their professional skills.

Therefore, the development of WIL programs is necessary

to show the responsibility of universities to society in terms

of responding to the needs of society. This is a very

important mission of universities.

There are several forms of WIL which encourage

students to experience authentic work practices, learn and

practice applying skills and knowledge in a real-world

context such as work placements, internships, field work,

sandwich course degrees, or even service learning (Jackson,

2015). For these collaborative programs, regular, intensive

contact sessions are envisaged, typically held onsite at the

company premises on a regular basis of schedules.

2.2. Benefits of Work-Integrated Learning

It cannot be denied that the implementation of WIL in

the university curriculum brings many benefits not only to

students but also to enterprises. Actually, when students

focus their attention on contextual and tacit areas of

workplace practice, their experiences contribute to the

development of employability skills (Francis, 2008). By

studying at the workplaces, students can explore their

understanding of what they have been taught in class in

different ways. Furthermore, involving in a variety of

employability skills such as team working, problem-

solving, communication information literacy and

professionalism accelerates maturity and enhances

motivation and accountability in students (Mandilaras,

2004). Previous studies have found that students who have

participated in WIL programs become critical and rational

graduates. They have good ability to function effectively in

the modern workplace as well as their career development.

These advantages of WIL improve employment prospects

of graduates (Freudenberg, Brimble, & Cameron, 2011;

Jackson, 2015).

Through WIL programs, students’ confidence in their

workplace capabilities is built as WIL provides students

with a better understanding of the nature and standard of

industry required skills (Billet, 2011). Having work

experiences, WIL results in student innovation which then

drives work productivity of graduates in the future.

For the benefits of enterprises, students participate in

WIL programs are required to demonstrate an

understanding of new knowledge and apply that

knowledge in ways that offer benefits to the enterprises in

which students are practicing. Smith (2012) mentioned that

WIL offers a better return on investment to students; meets

employer demands for work-ready graduates, and provides

employers with skilled employees at low cost.

2.3. Work-Integrated Learning in Australian Universities

In Australia, since 2000s, many universities have

60 Quy H. Nguyen

designed WIL courses and successfully implemented them

to better support student learning. The country has

established an organization named Business, Industry, and

Higher Education Collaboration Council (BIHECC)

oversees the connections between universities and

enterprises. According to McLennan and Keating (2008),

such establishment of BIHECC reflects an appreciation

and recognition of the role of enterprises as a unique and

valuable learning environment for students.

In its summary report published in 2015, BIHEC

emphasized the important role of higher education in

creating a strong and sustainable future of Australia.

Particularly, an increased collaboration and partnerships

between education, research, business and government is

crucial to generate and act on ideas which aim to ensure

prosperity and economic growth of Australia.

Recognizing the importance of this relationship,

businesses in Australia have been trying their best effort to

establish relationships with universities to make a

knowledge transfer and research development. The

Australian Industry Group reports that it has good

relationships with universities and the relationship is getting

better (Howard, 2016). The results of interviews with

Australian businesses reveal that most enterprises that are

working with universities do not want to dictate curriculum

or research but they want universities to take into account

the challenges that they are facing. In other words,

Australian enterprises do not see universities as contractors

but as organizations interested in their business. Successful

collaborations are not dictated by one side or the other but

are based on discussion, engagement and, above all, trust.

In Australia, the following five organizations have

closely worked together to promote WIL programs:

Universities of Australia, Australian Chamber of Commerce

and Industry, Business Council of Australia, The Australian

Industry Group, and Australian Collaborative Education

Network Limited. These organizations believe that strong

partnerships between universities, industry, academics and

the community improve the quality and capacity of

Australian education systems and the innovation, breadth

and competitiveness of the Australian economy as well as

shape future prospects of Australia (ACEN, 2015). Upon

this fundamental awareness, all universities in Australia

have a wide range of WIL courses. WIL courses’

components enable students to put their knowledge into

practice in real settings. Students have more opportunities to

contact and work with a wide range of employers. The

synergistic collaborations are crucial to help students meet

the challenges and opportunities required by rapidly

changing global realities at the workplace.

3. Lessons for Vietnam

3.1. University-Enterprise Training Partnership

In Vietnam, most students spend much time to study

theory at universities and do not have many chances to

practice at the workplaces. The awareness of educational

leaders of Vietnam about the importance of practice at the

workplace has been increased in recent years. Some

universities have established relationships and partnerships

with enterprises and organizations to get support for

students. Such approach is good but not enough and

effective. The most important thing is to develop or design

curriculum which are integrated in student learning.

To effectively implement WIL in university curricula,

the real partnership between universities and enterprises is

crucial. Being considered an innovative model of teaching

and learning at universities, WIL requires much effort and

contribution of enterprises. Universities and enterprises

need to work closely and appropriately to design and

facilitate learning in the workplace.

While university academics and students have

theoretical knowledge of the disciplines, people at the

workplaces have practical expertise that significantly

contributes to the productivity of WIL programs. In this

partnership with enterprises, academics are no longer the

dominant artists at universities and business workers are no

longer the dominant ones at the workplaces. Both

academics and enterprise staffs share responsibilities and

roles in the development and teaching of WIL programs.

In WIL courses, students are normally required to

complete an industry placement or an internship (paid or

unpaid) as part of the course. This is particularly common

in hands-on fields such as hospitality, engineering, IT and

communications. For example, in Australia, industry

placements may be offered in both degrees and Vocational

Education and Training (VET) qualifications. The courses’

length varies from a few weeks up to a year. Even if some

university degrees do not offer a compulsory or optional

industry placement option, students may be able to

organize one independently with industry.

The partnership between universities and enterprises

confronts with challenges facing our world. How university

graduates can apply their knowledge to real challenges of the

world motivates universities in this partnership. According

to Smith (2012), when universities develop WIL courses in

the curriculum, universities show their responsibilities to

society and enterprises in producing work-ready graduates.

Universities value the contribution and cooperation of

stakeholders in creating a sustainable growth for the

universities. Without such cooperation, it is difficult for

universities to fulfill their missions in this highly competitive

market of Higher Education.

3.2. Challenges of Work-Integrated Learning

Implementation

Although the university and enterprise partnership has

many benefits for both universities and enterprises, it has

some difficulties in the implementation. For example,

Choy and Delahaye (2011) indicate that the development

of a WIL curriculum relies on genuine partnerships

between the universities and organizations. These types of

partnerships require lengthy processes of negotiating the

curriculum and pedagogies to support learning based on the

workplace. Such requirement presents challenges both in

its formation and implementation for organizations due to

limited resources of universities and enterprises, and large

and diverse student cohorts. Furthermore, within the

universities, the pressures of ‘publish or perish’ draws

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 61

academics’ attention away from teaching and learning

activities (Freudenberg et al., 2011).

Also, in this partnership, there is a transition in the role

of academics, from traditionally pure teaching roles to the

different ones. For example, academics are expected to

understand the social-cultural environment of the

workplace and to develop knowledge about the enterprise’s

business (Choy & Delahaye, 2011). The design of learning

experiences for students requires a substantial involvement

of both academics and the workplace’s managers. It is

critical that enterprises consider them as important

stakeholders of universities. Their partnership with

universities not only brings benefits for universities and

students but also has good outcomes for the enterprises.

From the view of pedagogy, Billet (2011) raises the

need for strong pedagogic practices of WIL design, rather

than resources and infrastructures. It is necessary that

enterprises as the host organizations provide adequate

access to academics who work as supervisors of students

at the workplaces, offer learning support and induction

processes for students. Also, enterprises should clearly

establish, design, and articulate their expectations of

graduates in WIL programs to create authentic learning

activities that are well aligned to learning objectives,

expected learning outcomes, and course assessment. These

factors should be well integrated together, so that students

can make links between their learning in the different

contexts (classroom, on-campus, off-campus). Such

experiences offer students better understanding of what is

expected for effective practice of targeted skills and

knowledge at the workplace.

4. Implications for a Successful Implementation of

Work-Integrated Learning in Vietnam

4.1. For Universities in Vietnam

The development of WIL is a strategic response of

universities to the demands of enterprises and society. It is the

responsibility of universities to find measures that improve

employability skills of graduates. If the alumni’s career is

good, the universities will be able to attract more students as

well as funding, sponsorship from different industries.

During the implementation, WIL content should be

systematically integrated into curriculum of university

degrees. It might be a formal course or a professional

program assisting theoretical studies of students.

Researchers have found that students who involve in WIL

demonstrate significant gains in both disciplinary

knowledge and generic skills compared to those who do

not participate in WIL (Choy & Delahaye, 2011; Smith,

2012). These results highlight the potential gain for

universities from investing the necessary resources to

develop WIL opportunities for their students to assist

learning their outcomes. Therefore, universities should be

more active to establish partnership with different

industries to design WIL programs for their students.

Besides, the universities should discuss with

enterprises to build more makerspaces within the

universities or at the workplace. These places will function

as good learning environment for students. The recent

opening of the Maker Innovation Space at the University

of Danang, Vietnam is an evidence for this necessity of the

partnership between universities and enterprises. Through

such space, students of the University of Danang have good

opportunities to discuss with their friends, apply

knowledge to the real context, identify problems, learn and

apply skills, and share experiences with others.

4.2. For Enterprises in Vietnam

Choy and Delahaye (2011) emphasize the importance

of creating a long-term partnership between universities

and businesses. Senior management of organizations

should be highly and proactively involving in it. Each

organization should involve its partners in its strategies for

development, and not merely in technical tasks or short-

term activities. Facilitating this partnership, enterprises

need to create the authenticity of WIL from which students

can engage in a real environment, observe, interact and

respond to the complexities of real practice. Such

requirement indicates a fact that the role of enterprises is

very important in making WIL programs effective.

4.3. For Government of Vietnam

The government of Vietnam and ministries such as

Ministry of Education and Training and the others should

support the development of WIL programs to connect

undergraduate students with businesses to strengthen the

development of graduate employability. Without the above

suggested support of the government and relevant

ministries, the development of students’ employability

skills and ties with business becomes very difficult.

5. Conclusions

WIL is an umbrella term for a range of approaches and

strategies that integrate theories with practices within a

purposefully designed curriculum. The implementation of

WIL provides students with valuable practical experiences

related to their courses, demonstrates required characteristics

as taking initiative and being proactive at the workplaces.

WIL also assists a smooth transition from the university to

the workplace as well as productivity outcomes for the

enterprises and the whole economy. The implementation of

WIL in university curriculum is a response to labor and skill

challenges of industry. It is considered not only a pathway for

work-readiness in professional education but also one of the

key opportunities for improving the work-readiness of all

graduates. For those purposes to be achieved, the following

fundamental dimensions need to be well considered:

authenticity of the program, integrated learning supports both

at university and the workplace, and the alignment of

teaching and learning activities and assessments with

integrative learning outcomes. Cates and Jones (1999)

emphasize the consideration of skill transfer in the design of

assessments and activities as well as the use of formative

assessment in WIL courses.

Potentially, the implementation of WIL would serve the

needs of universities, graduates, and enterprises. This

collaborative design is only successful when all three

parties in this triangle highly and actively demonstrate their

commitment to the activities such as teaching of university

academics, learning of students, and instruction of the

62 Quy H. Nguyen

workplace’s staff. In this partnership, academic knowledge

should be indeed integrated into the workplaces, so it takes

much time and effort for every involved parties. Through

the partnership, academics gradually learn about the

complexities of the enterprises, the students as worker-

learners in the WIL courses understand the theories and

ascertained ways to apply such knowledge to the specific

context of the workplaces.

We believe that WIL is an effective approach/design that

can improve employability skills of graduates. In a broader

sense, the partnership between universities and enterprises

will help to increase workplace participation and

productivity and to reduce the unemployment record. In the

rising global competition nowadays, WIL enhances national

productivity of Vietnam. As human capital underpins the

future capability and competitiveness of all countries, the

extent to which Vietnam can adapt, develop, and deliver a

quality of life compared with that of other developed

countries in the world, will depend in large part on how well

Vietnamese universities and enterprises collaborate.

REFERENCES

[1] ACEN. (2015). National Strategy On Work Integrated Learning In

University Education. Retrieved from http://cdn1.acen.edu.au/wp-

content/uploads/2015/03/National-WIL-Strategy-in-university-education-032015.pdf

[2] BIHECC. (2007). Graduate employability skills. Retrieved from Canberra:

https://aces.shu.ac.uk/employability/resources/GraduateEmployabilitySkil

lsFINALREPORT1.pdf

[3] Billet, S. (2011). Curriculum and pedagogical bases for effectively

integrating practice-based experiences. Strawberry Hills, NSW:

Australian Learning and Teaching Council (ALTC).

[4] Cates, C., & Jones, P. (1999). Learning outcomes: The educational

value of cooperative education. Columbia, MD: Cooperative Education Association.

[5] Choy, S., & Delahaye, B. (2011). Partnerships between universities

and workplaces: some challenges for work-integrated learning.

Studies in Continuing Education, 33(2), 157-172. doi:10.1080/0158037X.2010.546079

[6] Francis, H., and J. Cowan. (2008). Fostering an action-reflection

dynamic amongst student practitioners. Journal of European Industrial Training, 32(5), 336-346.

[7] Freudenberg, B., Brimble, M., & Cameron, C. (2011). WIL and

Generic Skill Development: The Development of Business Students'

Generic Skills Through Work-Integrated Learning. Asia-Pacific

Journal of Cooperative Education, 12(2), 79-93.

[8] Howard, J. (2016). Business insight into more successful collaboration.

Retrieved from http://www.theaustralian.com.au/higher-education/opinion/business-insight-into-more-successful-

collaboration/news-story/495cf2267d60eb0249db99cb52c7458e

[9] Jackson, D. (2015). Employability skill development in work-

integrated learning: Barriers and best practice. Studies in Higher Education, 40(2), 350-367. doi:10.1080/03075079.2013.842221

[10] Mandilaras, A. (2004). Industrial placement and degree

performance: Evidence from a British higher institution.

International Review of Economics Education, 3(1), 39-51.

[11] McLennan, B., & Keating, S. (2008). Work-Integrated Learning

(WIL) In Australian Universities: The Challenges Of Mainstreaming

WIL. Paper presented at the ALTC NAGCAS National Symposium, Melbourne.

[12] Nguyen, D. T. U. (2017). In Vietnam, the best education can lead to worse job

prospects. Retrieved from https://www.bloomberg.com/news/articles/2017-

08-20/in-vietnam-the-best-education-can-lead-to-worse-job-prospects

[13] Nguyen, T. (2016). Employers lament lack of soft skills in graduates.

University World News, (397). Retrieved from

[14] Smith, C. (2012). Evaluating the quality of work-integrated learning

curricula: a comprehensive framework. Higher Education Research &

Development, 31(2), 247-262. doi:10.1080/07294360.2011.558072

[15] Smith, C., & Worsfold, K. (2014). WIL curriculum design and

student learning: a structural model of their effects on student satisfaction. Studies in Higher Education, 39(6), 1070-1084.

doi:10.1080/03075079.2013.777407

[16] Vietnamnet. (2017). Graduate unemployment rises. Retrieved from

http://english.vietnamnet.vn/fms/society/175579/graduate-

unemployment-rises.html

[17] VL. (2016, 12/10). Sinh viên mới ra trường thất nghiệp do đâu? Lao

động. Retrieved from https://laodong.vn/ky-nang-mem/sinh-vien-moi-ra-truong-that-nghiep-do-dau-600529.bld

[18] Vương, T. (2017, 28/4). Thiếu ngoại ngữ và kỹ năng là lý do sinh

viên ra trường thất nghiệp? Giáo dục Thời đại. Retrieved from

http://giaoduc.net.vn/Giao-duc-24h/Thieu-ngoai-ngu-va-ky-nang-

la-ly-do-sinh-vien-ra-truong-that-nghiep-post176178.gd

(The Board of Editors received the paper on 12/10/2017, its review was completed on 14/12/2017)

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 63

IDEATIONAL MEANING IN IELTS SAMPLE ESSAY INTRODUCTIONS IN

LIGHT OF SYSTEMIC FUNCTIONAL GRAMMAR

Nguyen Thi Trung1, Nguyen Thi Quynh Hoa2 1PhD Student of the English Language Course 31; dtrungtk@gmail.com

2University of Foreign Language Studies - The University of Danang; hoaufl@gmail.com

Abstract - It is shown that IELTS, which is recognized and accepted by more than 10,000 organizations in the world, has proved its significant role in education and occupation nowadays. In this article, M.A.K Halliday (2004)’s Systemic Functional Linguistics theory of ‘Clause as Representation’ is applied to investigate how ideational meanings are expressed in IELTS sample essays. By focusing on the transitivity system of the structure of the clauses, particularly the study of process types and the logico-semantic relation, this paper sheds light on the ideational meaning in the introductions of IELTS sample essays. The findings of this article are expected to enable IELTS candidates to take into account the relationship between the transitivity choice of processes, logico-semantic relation and the salient features of high-quality IELTS academic writing introductions to produce their own appropriate introductions, which, in turn, increase their scores in the test.

Key words - ideational meaning; transitivity; process; logico-semantic relation; clause complex

1. Rationale

Nowadays, achieving an international English certificate

indicating one’s English proficiency level is considered to

be one of the most essential requirements in applying for a

good job or tertiary education. As a result, the number of

candidates all over the world taking part in the international

English language tests, such as TOEFL, IELTS and TOEIC,

has been increasing more and more. Among these tests, the

International English Language Testing System (IELTS),

involving the assessment of four language skills, namely

listening, reading, writing and speaking, has been selected as

the preferred test to assess the ability to use language

correctly and appropriately in different countries. Getting a

high score in this international exam is the key to enable

candidates to open the door to their future career.

Among the four language skills of English assessed by

IELTS, academic writing occupies a crucial part whether

in school or workplace. How to enhance IELTS writing

outcomes, therefore, has been a matter of great concern for

language researchers and educators in different countries

where English is taught as a foreign language or second

language. There has been a great amount of research

carried out to investigate the IELTS writing tasks For

instance, Bagheri [1] investigates EFL graduate students’

IELTS writing problems and students’ and teachers’

beliefs and suggestions regarding writing skill

improvement, while Panahi [6] focuses on the strengths

and weaknesses of Iranian IELTS candidates in Academic

Writing Task 2. However, the ideational meaning of IELTS

sample essays has not been investigated from the

perspective of clause as a message. Such research would

provide the test-takers with an in-depth description of the

linguistic features of IELTS essays in terms of process as

well as logico-semantic relation and their connection with

the band score. In this paper, the main focus is on the

IELTS sample essay introductions with a view to giving a

brief analysis of ideational meaning in the light of

functional grammar.

The theoretical framework used to analyse and explain

how ideational meanings are expressed in clause structures

in IELTS essays’ introductions is the Systemic Functional

Theory developed by Halliday [4], Halliday & Matthiessen

[5]. Halliday [4, p.41] states that “The aim of Systemic

Functional Grammar has been to construct a grammar for the

purpose of text analysis: one that would make it possible to

say sensible and useful things about any text, spoken or

written, in modern English”. In fact, Halliday’s grammar has

been acknowledged as “the most widely employed

methodology for analyzing texts” Finch [3, p.191]. The

ideational metafunction, which is related to ‘clause as

representation’, construes the external and internal

experiences of the world. The ideational meanings are

further classified into the experiential and logical meanings.

The experiential subfunction encodes the experiences while

the logical one focuses on the relationships between ideas.

The experiential function allows users of language to present

the ‘content’ meaning of ‘who did what to whom’ through

the system of transitivity or process types associated with

different participant roles in different configurations. The

study investigates the introductions of IELTS sample essays

with a focus on experiential meaning in terms of six

processes: material processes, mental processes, relational

processes, verbal processes, behavioural processes,

existential processes, together with logico-semantic relation,

thereby making significant methodological and practical

contributions to achieving success in IELTS writing essays

from the beginning.

2. Theoretical Background

2.1. The System of Transitivity

The concept of Transitivity, in Halliday’s terms, is a

central component in interpreting the experiential meaning

of the clause in a linguistic text through the study of process

types; that is, what kind of event or state is being described

(realized by types of verbs) and the roles of participants; the

entities involved in the process, such as Actor and Goal;

Senser and Phenomenon; Carrier and Attribute (which are

part of the nominal group); and the circumstances in which

participants and processes are involved (realized by the

adverbial and prepositional phrases). As Halliday and

Matthiessen [5, p.170] mention, “the transitivity system

construes the world of experience into a manageable set of

process types”. Each process type provides its own model or

schema for interpreting a particular domain of experience as

a figure of a particular kind.

64 Nguyen Thi Trung, Nguyen Thi Quynh Hoa

Types of process

Within Systemic Functional Theory, a number of

models of process types have also been postulated.

However, it can be said that the most comprehensive and

perceptive theory for uncovering experience, which most

of the descriptive works of English and other languages

have drawn upon, is proposed by Halliday and

Matthiessen. Halliday [4] and Halliday & Matthiessen [5]

states three major process types: that is, MATERIAL,

MENTAL and RELATIONAL in the English transitivity

system. Apart from the three mentioned, Halliday

recognises three other intermediate process types which are

located at the boundaries between these process types.

These categories, according to him, are not clearly set

apart, but, nevertheless, are recognisable in the grammar as

intermediate between the different pairs - sharing some

features of each, and thus acquiring a character of their

own: BEHAVIOURAL process, VERBAL process, and

EXISTENTIAL process.

Material process

Material process typically construes ‘doings’ or

‘happening’ in the physical universe; that is, concrete and

tangible actions in the real world.

Material Pr: material; + Actor; (+Goal) (+Range)

(+Beneficiary)

Mental process

Different from describing actions as the material

process does, the mental process represents the meanings

of thinking or feeling.

Halliday [4, p.118] divides mental process verbs into

three classes: cognition (verbs of thinking, knowing,

understanding), affection (verbs of liking, fearing) and

perception (verbs of seeing, hearing). A mental process

should have two participants involved: the Senser (the one

who feels, thinks, perceives), and the Phenomenon (what

is felt, thought, perceived).

Mental Pr: mental; + Senser; + Phenomenon

Relational process

Relational process is a process of being: that is,

something is being said to ‘be’ something else rather than

as ‘doing’ or ‘sensing’. There are two different types of

relational process: the attributive relational process with

the two participants, namely the Carrier and the Attribute,

and the identifying relational one consisting of two

nominal phrases as participants, namely Token and Value.

The identifying one is reversible while the attributive one

is not reversible.

a. Identifying: Pr: identifying: +Token; +Value

b. Attributive: Pr: attributive: +Carrier; +Attribute

Behavioural process

Halliday [4] describes a behavioural process as a “half-way

house” between material and mental processes or a process of

‘physiological and psychological behaviour’ Halliday &

Matthiessen, [5, p.248]. Being partially like the material

process and partially like the mental process, the behavioural

process has to be experienced by a conscious being.

Behavioural Pr: behavioural;+ Behaver;

(+Behaviour) (Phenonmenon)

Verbal process

Verbal process accounts for the process of ‘saying’.

There are four participant roles of verbal processes;

namely, a Sayer (the one who is responsible for the

verbalization), a Receiver (the one to whom the

verbalization is addressed), a Verbiage which corresponds

to what is said; content or name of saying, and a Target, at

which the Verbal process may be directed.

Verbal Pr:verbal; +Sayer; (+Receiver) (+Verbiage)

Existential Process

Existential process represents something that exists or

happens. Eggins [2, p.214] states that “there when used in

existential processes, has no representational meaning: it

does not refer to a location”. Existential processes are

typically recognized by the verb ‘be’ or other verbs to

express the existence or happening such as ‘exist, remain,

arise or occur’. The function of existential processes is to

construe being as simple existence, therefore, there is only

one obligatory participant known as Existent. An Existent

can be an entity, event or action.

Existential Pr: existential; +Existent

2.2. Unit of Analysis

In systemic functional linguistics, the clause, the

highest unit of grammatical analysis has received a special

status. English clause can be conceptualized along the

dimensions of stratification, rank, and metafunction.

According to Thompson [8], a clause is any stretch of

language centred on a verbal group.

In Halliday’s theory of functional grammar, a sentence

can be interpreted as a clause complex: a Head clause

together with other clauses that modify it. In the article, a

clause complex is a combination of two or more clauses

into a larger unit, with their interdependence normally

shown by explicit signals such as conjunctions while clause

simplex consists of single clause.

2.3. Logico-Semantic Relation

In analyzing the logical structure of the clause, there are

two systems that need to be taken into consideration: one

of taxis (how two or more adjacent clauses are liked to each

other through dependency and interdependency relations)

and logico-semantics (the types of meanings that allow

adjacent clauses to project or expand on each other).

2.3.1. Taxis System

Taxis system is concerned with interdependency relation

between clauses linked into a clause complex. There are two

options within the system of taxis: parataxis (where clauses

are equally and independently related) and hypotaxis (where

one clause is dependent on or dominates another).

Parataxis is the relation between elements of equal

status. In other words, the elements are independent. For

paratactic structures, a numerical notation (1, 2, 3…) will

be used for illustration.

Hypotaxis is the relation between a dependent element

and its dominant, the element on which it is dependent.

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 65

Hypotaxis will be represented by the Greek letter notation

and a single slash in which an alpha (α) is used for

dominant clauses and a beta (β) or (γ…) is used for

dependent clauses.

2.3.2. Logico-Semantic System

Logico-semantic system refers to the specific type of

meaning relationships between clauses. There are two

basic types: projection (where one clause is reported or

quoted by another clause) and expansion (where one clause

develops or extends on the meanings of another).

As far as the projection is concerned, there are two

options: locution (where what is projected is speech) and

idea (where what is projected is thoughts)

In terms of expansion, there are three main choices:

elaboration (by restating in other words, specifying in

greater detail, commenting, or giving illustration),

extension (by adding new information, giving an exception

to it, or providing alternative), enhancement (by

embellishing it with some circumstantial features of time,

place, cause, or condition).

Expansion:

- elaboration = (equal)

- extension + (is added to)

- enhancement x ( is multiplied by)

These symbols combine with those for parataxis and

hypotaxis:

=2 = ß +2 + ß x 2 x ß “2 “ß ‘2 ‘ß

Halliday [4, p. 218]

2.4. IELTS Academic Writing Task 2

The Writing Task 2 of the IELTS Academic Test

requires candidates to write an essay to express a view and

produce an argument clearly and persuasively in English.

The topics in task 2 are of common interest relating to

environment, technology, education, media, health, public

transport, globalization or traditional culture suitable for

and easily understood by candidates. The way to deal with

IELTS Writing Task 2 is similar to an academic-style essay

on a common topic.

An academic essay has three parts: the introduction, the

body and the conclusion. The most common diagram of an

introductory paragraph is the ‘funnel’. Spencer [7, p.78]

illustrates this ‘funnel’ in the following diagram.

The introduction introduces the topic of the essay, explains

the academic problem, and outlines how the writer intends to

handle it. It is a general rule that an introductory paragraph

should contain the thesis statement, which is generally located

at the end of the introduction. In the thesis statement, the writer

tells the reader what to expect in the essay.

3. Research Methods

The aim of the article is to investigate the transitivity

and logico-semantic relation in order to identify the types

of ideational meaning in the introductions of IELTS

samples. In order to achieve this aim, different research

methods are employed as follows:

The descriptive method is used to give a detailed

description of the main areas of ideational meaning,

particularly processes and clause relation in the light of

functional grammar.

Qualitative and quantitative approaches are applied as

the purpose of this research is to examine how and to what

extent are ideational meaning conducted in the introduction

of IELTS sample essays.

The population of the study is 30 essays of which the

introductions are picked out to build up the data as as this

is a pilot study on a small scale where the primary focus is

on the introductions of IELTS sample essays.

Two criteria for selecting samples have been

established: 1-IELTS sample essays from books published

by Cambridge University Press 2-IELTS sample essays

with the band score from 7.5 to 9 or ones prepared by an

examiner as an example of a very good model, the level

which most candidates would like to achieve.

4. Findings and Discussion

4.1. Logico-Semantic Relation Analysis

The analysis of IELTS sample essay introductions into

clauses and clause complexes from SFL perspective can be

done as follows:

Sample 1 extracted from Cambridge IELTS 9

(A) Traditionally, children have begun studying foreign

languages at secondary school, || but introducing them

earlier is recommended by some educationalists. ||| (B)

This policy has been adopted by some educational

authorities or individual schools, with both positive and

negative outcomes. |||

Clause division of the text

CC A cl.1 |||Traditionally, children have begun

studying foreign languages at secondary school, ||

cl.2 || but [[introducing them earlier]] is

recommended by some educationalists. |||

cl. 3 [[introducing them earlier]]

CC B cl.4 |||This policy has been adopted by some

educational authorities or individual schools, with both

positive and negative outcomes. |||

Logico-semantics

Clause simplexes: 01 (B); Clause complexes: 01 (A)

A |||Traditionally, children have begun studying foreign 1

languages at secondary school, || but [[introducing +2

them earlier]] is recommended by some

educationalists. |||

Sample 2 extracted from Cambridge IELTS 9

(A) It has been suggested || that high school students

should be involved in unpaid community services as a

compulsory part of high school programs. ||| (B) Most of

the colleges are already providing opportunities to gain

Hook

General statement

Thesis statement

66 Nguyen Thi Trung, Nguyen Thi Quynh Hoa

work experience, ||however these are not compulsory. |||

(C) In my opinion, sending students to work in community

services is a good idea || as it can provide them with many

lots of valuable skills.

Clause division of the text

CC A cl.1. |||It has been suggested ||

cl.2 || that high school students should be

involved in unpaid community services as a compulsory

part of high school programs. |||

CC B cl.3 ||| Most of the colleges are already

providing opportunities ||

cl.4 [[to gain work experience,]]

cl.5 ||however these are not compulsory. |||

CC C cl.6 |||In my opinion, sending students to

work in community services is a good idea ||

cl.7 [[to work in community services]]

cl. 8 || as it can provide them with many

lots of valuable skills. |||

Logico-semantics

Clause simplexes: 0; Clause complexes: 03 (A, B, C)

||| It has been suggested ||that high school students α ’β

should be involved in unpaid community services as a

compulsory part of high school programs. |||

||| Most of the colleges are already providing

opportunities α 1

|| to gain work experience, ||however these are not

compulsory x β + α 2

||| In my opinion, [[sending students to work in

community services]] is a good idea ||as it can provide them

with α xβ

many lots of valuable skills. |||

The analysis shows that in 30 introductions, there are

74 sentences, of which 30 sentences consist of clause

simplexes, the remainder (44) are clause complexes,

including 32 with two-clause complexes and 12 with more-

than-two complexes. The writers chose clause complexes

over clause simplexes as clause complexes enable

language users to construe logical connections between

experiential events, thus expressing ideational meaning. It

is clear that the choice of clause complexes creating a

closer logico-semantic tie between clauses gives fluency to

writing. Table 1 represents the clause simplexes and clause

complexes of the text.

Table 1. Distribution of clause simplexes and clause complexes

Type Clause

simplex

Clause complex Total

No. of

CC Two-clause

complex More-than-two-clause complex

Total 30 32 12 74

Percentage 40,5% 59.5% 100%

As far as logico-semantic dimension is concerned,

projection and expansion are used to show a very wide

range of possible logico-semantic relationships between

clauses. In terms of taxis, the clauses in the clause

complexes are a mixture of both paratactic and hypotactic

sequences, but there are more clauses in hypotactic relation

than those in paratactic relation, which shows clauses’

dependency. Their semantic relations are mainly of

extension and enhancement. The distribution of logico-

semantic taxis is given in Table 2.

Table 2. Logico-semantic taxis of clause complexes

Logico-

semantic

taxis

Expansion Projection Total

Elabo

ration

Exten

sion

Enhance

ment

Idea Locu

tion

Parataxis 0 16 0 0 0 16

Hypotaxis 3 7 13 11 6 40

Total 3 23 13 11 6 56

Although embedded clauses do not participate in the

clause-complex structure of the texts, it is found that the

number of embedded clauses is frequently used in the

introductions with 67 embedded clauses out of the total 204

clauses, accounting for 32.8%. It can be indicated that

embedded clauses are used to function as either part of a

group constituent structure of a clause (qualifier in a

nominal group or postmodifier in an adverbial group) or as

a constituent of the clause in their own right (embedded

clause as Participant). For instance,

As a result, urban citizens have access to jobs and

facilities [[that improve their living standards

considerably.]]

Post-modification (qualifier)

[[Indulging children]] has been a common parenting

problem

Participant

in many modern families.

As Eggins [2] states the principle behind embedded is

compression; that means embedding is more static. In

IELTS writing, the writers tend to use quite a lot of

embedded clauses to narrow down the meaning

constituents, indicating their forethought in the

construction of the clause as Halliday [4] suggests that

embedding associates with formal and careful written text.

In terms of the length of the introduction, it is shown

that there are 1667 words used in the total of 30

introductions, of which the longest introduction has 98

words and the shortest one has 27 words. The average

words used per introduction is about 55 words. In writing

an IELTS essay introduction, it is suggested that an

introduction should not be too long as the purpose of an

introduction is to leave a good first impression of the

candidate’s essay writing skills on the examiner. It may be

a better idea to write a short introduction with a clause

complex introducing the topic and another telling the

reader what the writer is going to writing about called the

thesis statement. A good thesis statement will show the

reader how the essay will be structured.

4.2. Transitivity

From the view of transitivity, of the 204 clauses 79 are

of relational processes which have been illustrated in Table

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 67

3. They are used to present the state of beings with frequently

used verbs such as be, become, mean, have, feature,

represent, seem, as illustrated in the following extracts.

(1) I believe that child-rearing should be the

responsibility of both parents and that, whilst the roles

within that partnership may be different, they are

nevertheless equal in importance.

(2) … they have Internet access to information in a way

that was unimaginable for earlier generations.

73 clauses are material process describing the actions

with such verbs as sell, provide, influence, do, teach, adopt,

send, work, stop, travel, give, educate, improve, boost, buy,

advertise… 34 of mental processes projecting the writer’s

opinion, beliefs or feeling (believe, choose, decide, reflect,

feel, need, view, understand, think, appreciate, mean, find,

content…); 10 clauses which are verbal process,

accounting for 4.9 % of the total processes used in the

introductions show sayings of the writers. The verbal

processes are mainly used to quote sayings from other

people, which makes the saying more persuasive,

indicating the writer’s stance at the same time with verbs

like say, support, be recommended, suggest, agree,

argue… . 6 of existential processes positing that things or

people are in existence with the easily identified structure

involving the use of the word ‘there’. Only 2 clauses are

behavioural processes describing the physiologic and

psychological behaviours of the characters. Below are

some examples taken from the IELTS samples’

introductions.

(3) They also believe that the influence on individuals

is wholly beneficial as it is a long established way of …

(mental process, relational process, relational process)

(4) Others believe that highways, busses and regular

trains should be improved before new, high-speed lines are

added. (mental process, material process, material process)

(5) There is no doubt that some students in schools behave

badly and… (existential process, behavioural process)

Table 3. Transitivity in IELTS sample introductions

Process

types (PT) M

ater

ial

Rel

atio

nal

Men

tal

Ex

iste

nti

al

Ver

bal

Beh

avio

ura

l

To

tal

of

PT

Total 73 79 34 6 10 2 204

Percentage 35.8% 38.7% 16.7% 2.9% 4.9% 1% 100%

5. Conclusion

From the analysis, we may come to conclusion that

functional grammar is proved to be a very useful analytic

tool, enabling us to look at the whole text and its

characteristic features. By examining the transitivity

patterns in the introductions, we can explain how

experiential meaning is being constructed and analysing

the logical structure of the clause complex which

complement transitivity choices, we are able to be aware of

logical connections between experiential events.

As shown in the findings, it is suggested that IELTS

academic writers should begin the introduction with a

general statements, then narrows to the thesis. It is obvious

that the writers should choose clause complexes over

clause simplexes as clause complex systems provide them

with structural resources to construe logical connections

between experiential events. It is also recommended that

IELTS academic candidates should use different types of

interdependency between linked clauses to build on the

ideational meanings of the clauses. The test-takers are

advised to use paratactic conjunctions to express the logical

relationship between two clauses of equal structural status,

and hypotaxis to relate one clause to another in a modifying

or dependency relationship. Although embedded clauses

do not take their place as full constituent partners in the

clause complex, it is proposed that an IELTS introduction

should involve an intricate mix of both taxis and

embedding as embedding associates more with formal,

careful written text. In terms of transitivity, the writers

have a strong tendency to use two main types of process;

namely relational and material processes as an instrument

to describe action, events or states of being related to their

own experience regarding the topic discussed in the essays.

To sum up, the transitivity system and logico-semantic

relation occupy a crucial position in the realization of

linguistic features of IELTS sample introductions. It is

clear that writing with better awareness of the relationship

between the transitivity choice of processes and logico-

semantic relationships between clauses may enable non-

native English speaking IELTS writers to produce

stylistically appropriate IELTS introductions, thus

achieving higher score for the test.

REFERENCES

[1] Bagheri, M. S. (2016), “EFL Graduate Students’ IELTS Writing

Problems and Students’ and Teachers’ Beliefs and Suggestions Regarding Writing Skill Improvement”, Journal of Language

Teaching and Research, Vol. 7. No. 1, 198-209.

[2] Eggin, S. (2004), An Introduction to Systemic Functional

Linguistics, London: Printers.

[3] Finch, G. (2000), Linguistics Terms and Concepts, Houndmills:

Macmillan Press Ltd.

[4] Halliday, M. A. K, (1994), An introduction to Functional Grammar,

London: Edward Arnold.

[5] Halliday, M. A. K.& Matthiessen, C. (2004), An Introduction to

Functional Grammar, London: Edward Arnold.

[6] Panahi, R. (2015), “The Strengths and Weaknesses of Iranian IELTS

Candidates in Academic Writing Task 2”, Theory and Practice in

Language Studies, Vol. 5, No. 5, 957-967

[7] Spencer, C.M. & Arbon, B. (1997), Foundations of Writing:

Developing Research and Academic skills, Lincolnwood. IL: National Textbook Company.

[8] Thompson. G. (2004), Introducing Functional Grammar, Oxford:

Oxford University Press.

[9] IELTS, Cambridge IELTS 1,2,3,4,5,6,7,8,9,12, Cambridge

University Press.

[10] Jakeman, V. & Mc Dowell, C (2001), IELTS Practice Test Plus1 &

2, Pearson Education Limited.

[11] Jakeman, V. & Mc Dowell, C (2003). Insights into IELTS,

Cambridge University Press.

(The Board of Editors received the paper on 05/9/2017, its review was completed on 24/11/2017)

68 Ho Si Thang Kiet

CORRELATION BETWEEN SELF, PEER AND TEACHER ASSESSMENT:

A CASE STUDY IN A TRANSLATION COURSE

Ho Si Thang Kiet

University of Foreign Language Studies - The University of Danang; kiet.ho@ufl.udn.vn

Abstract - Assessment plays an essential role in teaching and learning process. This paper investigates the correlation between three different forms of assessment including self, peer and teacher assessment in a case study of a Translation course. Thirty Vietnamese students of English participated in the study. Data collection comes from the assessment scores of the three forms of assessment. The results show that the students tend to value their own work higher than their peers and the teacher. Although self-assessment has a positive and significant correlation with peer and teacher assessment, peer assessment seems to be a better alternative assessment as it has a stronger correlation with teacher assessment. In addition, the three forms of assessment can support one another to enhance the quality of the teaching and learning process and develop learner autonomy in EFL classrooms.

Key words - self assessment; peer assessment; teacher assessment; correlation; translation.

1. Introduction

Assessment plays an essential role in measuring

learners’ knowledge and achievements in educational

settings. It is a process that is supposed to be carried out as

an inseparable part of the teaching and learning process.

However, in a traditional EFL classroom, teacher

assessment tends to be the only form of assessment, while

other forms of assessment such as student self-assessment

and peer assessment are of minor importance.

Consequently, learner autonomy becomes underdeveloped

in the learning process. This does not reflect the learner-

centred approach in language teaching and learning in

which learners should take the initiative not only in their

learning, but also in the assessment of their work.

In order to develop learner autonomy that can improve

the learning process, forms of formative assessment such

as peer assessment and self-assessment are crucial.

Especially, in the translation course, there has been only

teacher assessment at the end of the course and the scores

are usually drawn from the marker’s subjective judgments

due to unclear marking criteria. This study is conducted on

the basis of some specific assessment criteria on

translation, which can result in more accurate scores. In

addition, the study also tries to make students familiarize

with peer assessment and self-assessment as forms of

formative assessment so that they can achieve better results

of study. This study aims to investigate the correlation

between three different forms of assessment including self-

assessment, peer assessment and teacher assessment in

orderto find out which assessment validates one

assessment better than the other.

2. Theoretical background

2.1. Definition of assessment

Assessment is a broad term that refers to “a process of

collecting information about something that we are

interested in, according to procedures that are systematic and

substantially grounded” (Bachman, 2004, pp. 6-7). In other

words, assessment involves a set of procedures to gather

information about students’ performances or achievements

of knowledge and skills in a particular course or area.

Among different forms of assessment, peer and self-

assessment are important aspects of formative assessment

for learning practice. Gattullo (2000) categorizes formative

assessment as “an ongoing multi-phase process that is

carried out on a daily basis through teacher–pupil

interaction”, “providing feedback for immediate action” and

“modifying teaching activities in order to improve learning

processes and results” (p. 279, cited in Ketabi and Ketabi,

2014, p.437). Assessing their own work or that of others can

help students develop their understanding of learning

objectives so that they can achieve success in learning.

Research has shown that students can make more progress

when they are actively engaged in peer and self-assessment.

2.2. Self-assessment

Self-assessment is an important aspect of formative

assessment. According to Hedge (2000), self-assessment is

“the ability [of students] to assess their own performance

and the ability to be self-critical” (p.94). Andrade and Du

(2007) defines self-assessment as “a process of formative

assessment during which students reflect on and evaluate

the quality of their work and their learning, judge the

degree to which they reflect explicitly stated goals or

criteria, identify strengths and weaknesses in their work,

and revise accordingly” (p.160). Both definitions indicate

that self-assessment requires students to reflect on their

own work and make judgments on their performance in

order that they can improve themselves. Such an emphasis

on student responsibility and judgments is “a necessary

skill for lifelong learning” (Boud, 1995, p.11).

2.3. Peer assessment

Apart from self-assessment, another important aspect

of formative assessment is peer assessment. There are a

variety of definitions of peer assessment. According to

Strijbos and Sluijsmans (2010), peer assessment is referred

to as “an educational arrangement where students judge a

peers’ performance quantitatively and/or qualitatively and

which stimulates students to reflect, discuss and

collaborate” (p. 265). Topping (2010) also defines peer

assessment as “an arrangement for learners to consider and

specify the level, value, or quality of a product or

performance of other equal-status learners” (p.62).

The benefits of peer assessment have been underscored

in education. With peer assessment, student learning can

improve through “a sense of ownership and responsibility,

motivation, and reflection of the students’ own learning”

(Saito and Fujita, 2009, p. 151). Cheng and Warren (2000)

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 69

argue that peer assessment can facilitate student learning and

provide teachers with a clear picture of learners’

performance. Especially, peer assessment has shifted

teacher-centered assessment in language teaching to learner-

centered assessment (Abolfazli Khonbi and Sadeghi, 2012),

which results in learner autonomy in their learning process.

2.4. Correlation between teacher assessment and student

assessment

Previous studies show that there is a correlation

between teacher assessment and student assessment.

Zakian, Moradan and Naghibi (2012) show that there is a

high correlation between teacher and student assessment

and that the teacher assessment can be supplemented with

student assessment in the context of productive language

skills. Kilic (2016) show that peer assessment of pre-

service teachers’ performances is found to be significantly

higher compared with teacher assessment and self-

assessment. These studies show that the three forms of

assessment can support one another to enhance the quality

of the teaching and learning process.

3. Research method

The study applies quantitative approach to reveal the

correlation between self, peer and teacher assessment in a

translation course. A Translation course at the Department

of English for Specific Purposes, University of Foreign

Language Studies, the University of Danang, is selected as

a case study for this research. The Translation course is

chosen as the study is done in the researcher’s own class

and the third-year students have enough ability for self-

assessment and peer assessment. Thirty students in the

Translation course participated in this research. The study

consists of 26 female students and 4 male students. This is

because female students tend to outnumber male students

in a foreign language major in Vietnam. To deal with the

issues above, the study aims to answer the following

research questions:

(1) What type of assessment is valued most in the

Translation course?

(2) What are the correlations between self, peer and

teacher assessment in the Translation course?

The case study is conducted in four phases: (1)

translating process, (2) peer assessment, (3) self-

assessment, and (4) teacher assessment. The first phase

engages the students in a translating process with a

translation test. The translation test consists of 10 sentences

in Vietnamese that the students have to translate into

English. The translation texts involve different topics

including tourism, education, health, society, environment

and technology that the students have been familiar with in

the course material. The students are given one hour to do

the translation thoroughly. Before the students participate

in the peer and self-assessment, they are provided with

translation assessment criteria with all clear explanations

from the researcher. The translation assessment criteria

used for this study are taken from Lee (2013) covering

translation accuracy (6 points) and translation expression

(4 points) (Appendix). The criteria have been translated

into Vietnamese to facilitate students’ understanding. The

assessment criteria are applied for each translated sentence

with 10 points per sentence for a total score of 100.

Peer assessment is introduced first in order that the

students are fully comfortable with it before moving on to

self-assessment. The second phase involves the peer

assessment process in which the students’ translated texts

are randomly distributed to their peers for peer assessment,

using Lee’s (2013) assessment criteria and the translation

key provided by the researcher. The researcher then

collects the students’ peer assessment scores.

Now that the students have been familiar with peer

assessment, they move on to the third phase of self-

assessment of their own translated texts which is also based

on the same assessment criteria and translation key as in peer

assessment. Having assessed the work of others, the students

will find it easier to identify weaknesses in their own work

and to see how they can make improvements. The researcher

then collects the students’ self-assessment scores.

The last phase involves the teacher assessment of the

student’s translated texts that is completed by the

researcher himself, using the given assessment criteria and

translation key.

For data analysis, the means, maximum scores and

minimum scores of self, peer and teacher assessment are

first calculated. Then statistical computation using Pearson

Product Moment Correlation is utilized to find out the

correlation between the three forms of assessment.

4. Findings and discussions

Table 1 provides descriptive statistics on the overall

result of the self, peer and teacher assessment. It includes

the number of students, minimum and maximum scores,

means and standard deviation. The table shows that self-

assessment has the highest means (M=72.50, SD=5.25)

compared to that of peer assessment (M=69.43, SD=7.19)

and teacher assessment (M = 66.80, SD=6.03). This proves

that most of the students value themselves higher than the

assessment criteria. The self-assessment also has the

highest scores (Max=90), followed by peer assessment

(Max=88) and teacher assessment (Max=80) in the last

position. This order is the same for the minimum scores.

This leads to another interpretation that the students tended

to assign higher value for their translation work than their

peers and the teacher. However, the peer assessment shows

a wider range of standard deviation than the self and

teacher assessment.

Table 1. Descriptive statistics for self, peer and teacher assessment

Variable N

Minimum

scores

(Min)

Maximum

scores

(Max)

Means

(M)

Standard

Deviation

(SD)

Self-

assessment 30 65 90 72.50 5.25

Peer

assessment 30 55 88 69.43 7.19

Teacher

assessment 30 50 80 66.80 6.03

In order to interpret more the data in Table 1, inferential

statistics are carried out using Pearson Product Moment

70 Ho Si Thang Kiet

Correlation to find out the correlation between self, peer and

teacher assessment. The Pearson correlation coefficient, also

referred to as the Pearson's r, is a measure of the linear

correlation between two variables. It has a value between +1

and −1, where 1 is total positive linear correlation, 0 is no

linear correlation, and −1 is total negative linear correlation.

The results are given in Table 2.

Table 2. Correlation between self, peer and teacher assessment

Self-

assessment

Peer

assessment

Teacher

assessment

Self-

assessment

Pearson

correlation

Sig. (2-tailed)

N

1

30

.561**

.001

30

.678**

.000

30

Peer

assessment

Pearson

correlation

Sig. (2-tailed)

N

.561**

.001

30

1

30

.712**

.000

30

Teacher

assessment

Pearson

correlation

Sig. (2-tailed)

N

.678**

.000

30

.712**

.000

30

1

30

Note: **Correlation is significant at the 0.01 level (2-tailed).

The statistical analysis in Table 2 shows that self-

assessment has a positive correlation with peer

assessment and teacher assessment with r = 0.561 and r

=.678 respectively and the correlation is significant at the

0.01 level. Similarly, peer and teacher assessment also

have a positive and stronger correlation with r =.712,

which is also significant at the 0.01 level. This confirms

the assumption that the students are able to assess their

peer work properly. Peer assessment seems to provide a

better alternative assessment in addition to teacher

assessment. This result echoes Kilic (2016) who shows

that peer assessment of pre-service teachers’

performances is significantly higher than teacher

assessment and self-assessment. The result also implies

that peer assessment tends to validate teacher assessment

better than self-assessment to determine the quality of a

good translation. This means that peer assessment can be

used as an alternative assessment that is valid and reliable

about the translation work as long as it is in conformity

with the assessment criteria. In addition, as there is a

positive and significant correlation between the three

forms of assessment, they can support one another to

enhance the assessment of students’ work. The

assessment criteria seem to match both the students’ and

the teacher’s perspectives in assessing the translation

work. This shows that a clear set of assessment criteria is

extremely important to assess students’ work properly.

5. Conclusions

The study reveals some interesting insights into

different forms of assessment, especially self and peer

assessment as forms of formative assessment which have

not got much attention in EFL classrooms. The study

shows that students tend to value their own work higher

than their peers and the teacher. Although self-assessment

has a positive and significant correlation with peer and

teacher assessment, peer assessment seems to be a better

alternative assessment as it has a stronger correlation with

teacher assessment. In addition, the three forms of

assessment can support one another to enhance the quality

of the teaching and learning process. It is the use of forms

of formative assessment that will shift teacher-centered

assessment in language teaching to learner-centered

assessment, which results in learner autonomy in their

learning process.

The development of effective self and peer assessment

takes considerable time and effort and both need to be

managed carefully. However, once fully embedded into the

teaching and learning process and combined with teacher

assessment, these forms of formative assessment can

facilitate the teaching and learning process and motivate

students to move forward to achieve their learning goals. It

is expected that the results of this case study may be used

for other courses as well so that students can engage in self-

assessment and peer assessment to achieve better results

for their studies and develop their learner autonomy in the

learning process.

REFERENCES

[1] Abolfazli Khonbi, Z. and Sadeghi, K. (2012). The Effect of

Assessment Type (self vs. peer vs.teacher) on Iranian University EFL Students’ Course Achievement. Language Testing in Asia,

2(4), 47-74.

[2] Andrade, H. and Du, Y. (2007). Student responses to criteria-

referenced self-Assessment. Assessment and Evaluation in Higher Education, 32 (2), 159-181

[3] Bachman, L. (2004). Statistical analysis for language assessment.

Cambridge: Cambridge University Press.

[4] Boud, D. (1995). Enhancing learning through self-assessment. London: Kogan Page

[5] Cheng, W., and Warren, M. (2000). Making a difference: Using

peers to assess individual students' contributions to a group project.

Teaching in Higher Education, 5(2), 243-255.

[6] Gattullo, F. (2000). Formative assessment in ELT primary (elementary) classrooms: an Italian case study. Language Testing,

17 (2), 278–288.

[7] Hedge, T. (2000). Teaching and learning in language classroom.

Oxford: Oxford University Press.

[8] Ketabi, S and Ketabi, S. (2014). Classroom and formative assessment in second/foreign language learning. Theory and

Practice in Language Studies, 4 (2), 435-440.

[9] Kilic, D. (2016). An Examination of Using Self-, Peer-, and Teacher

assessment in Higher Education: A Case Study in Teacher Education. Higher Education Studies, 6(1), 136-144.

[10] Lee, T. (2013). Incorporating translation into the language

classroom and its potential impacts upon L2 learners. In Tsagari, D.

And Floros, G. (Eds), Translation in Language Teaching and Assessment (pp3-22). UK: Cambridge Scholars Publishing.

[11] Saito, H., and Fujita, T. (2009). Peer-assessing peers’ contribution

to EFL group presentations. RELC Journal, 40(2), 149–171.

[12] Strijbos, J. W., & Sluijsmans, D. (2010). Unravelling peer

assessment: Methodological, functional, and conceptual developments. Learning and Instruction, 20(4), 265-269.

[13] Topping, K. J. (2010). Peers as a source of formative assessment. In:

H. L. Andrade and G. J. Cizek (Eds.), Handbook of formative

assessment (pp. 69–75). New York, NY: Routledge.

[14] Zakian, M., Moradan, A. and Naghibi, S.E. (2012). The relationship between self-, peer-, and teacher assessments of EFL learners’

speaking. World J Arts, Languages, and Social Sciences, 1(1):1-4.

ISSN 1859-1531 - THE UNIVERSITY OF DANANG, JOURNAL OF SCIENCE AND TECHNOLOGY, NO. 12(121).2017 71

Appendix:

Assessment criteria for the Translation test

Students’ translation work is assessed by three

independent channels as follows:

(1) Students assess their peer work based on the given

assessment criteria.

(2) Students assess their own work themselves based on

the given assessment criteria.

(3) The teacher assesses students’ work based on the

given assessment criteria.

The assessment criteria are applied on a sentence-by-

sentence basis. There are 10 sentences in the test. Each

sentence has a maximum score of 10. The translation test

has a maximum score of 100.

Assessment criteria (Lee, 2013)

1. Translation accuracy

Score Criteria

6 Messages in translation almost match the original with no

errors.

5 Messages in translation almost match the original, but

with one minor error.

4 Messages in translation almost match the original, but

with two minor errors.

3 Messages in translation almost match the original, but

with one major error and three minor errors.

2 Messages in translation almost match the original, but

with two major errors or only pile-up descriptions.

1 Messages in translation almost match the original, or with

a lot missing in translation.

2. Translation expression

Score Criteria

4 Translation is clear and understandable with appropriate

usages, register, collation and punctuation.

3 Translation is generally clear and understandable, but

with one or two errors in usage and expression, typos, or

redundant word.

2 Translation is barely understandable, but with syntactic

errors and inappropriate usage and expression.

1 Translation is ungrammatical and difficult to

understand, with many omissions.

(The Board of Editors received the paper on 20/11/2017, its review was completed on 06/12/2017)