Wear, 114 (1987) 29 - 39

MODELLING DURABILITY OF COAL GRINDING SYSTEMS

29

S. F. SCIESZKA

Technical University of Silesia, 44-101 Gliwice (Poland)

(Received December 13,1985; accepted April 4,1986)

Summary

An attempt has been made to demonstrate the wear test procedure, which simulates the tribological conditions in the ball-race-type coal mill and which enables the laboratory results to be applied with a reasonable degree of certainty, in evaluating the service life of grinding elements. The tribological conditions in the ball-race system, namely the pressure distribution, sliding velocity gradient and temperature inside the grinding zone, were identified. The design of the laboratory rig and the outline of the operating procedure were based on principles of similarity. A com- parison between industrial data and simulative tribotesting results shows a very good correlation. This proved that the results from a relatively simple rig may be used in predicting the service life of the machine ele- ments.

1. Introduction

In all tribological investigations model tests are frequently used in the development of new materials for certain applications since original indus- trial elements involve high expenditure as regards both duration and energy [ 1, 21. However, with model tests of tribological processes the problem of- ten arises as to what extent the results obtained under laboratory conditions are applicable to the industrial conditions. This problem arises since tribo- logical processes are usually very complicated and are influenced by many factors [ 11. As the operating costs of a coal pulverizing plant comprise main- ly grinding element replacement due to wear, the importance lies firstly with the correct choice of wear-resistant materials which will result in minimum wear of the elements and secondly with the accurate estimation of their service life. In this paper an attempt has been made to show the wear test procedure which simulates tribological conditions in a ball-race-type coal pulverizer and which allows the results of the test experiments to be applied with a reasonable degree of certainty.

0043-1648/87/$3.50 0 Elsevier Sequoia/Printed in The Netherlands

30

2. Identification of tribological conditions in ball-race coal pulverizers

In ball-race mills, which at present comprise a significant number of pulverizers used for coal grinding, the predominant mechanisms of commi- nution are crushing and attrition. In the ball-race system shown in Fig. 1 the tribological conditions are described by the pressure distribution, gra- dient of sliding velocity and temperature inside the grinding zone of the coal as well as on the surface between the coal and the grinding elements. The geometry of the ball-race system is constantly changing since a pro- gressive wear process is taking place. To identify the pressure condition at the interface between the ball and the bed of fine coal, the model shown in Fig. 1 was applied. Loading and geometrical data were taken from the Babcock 12.9 E-type mill as follows: load per ball L = 120 kN; 2r = d = 0.98 m; b = 0.32 m; p = 16”, di = 0.742 m. Since s = r sin 0 and s1 = rl sin /3 then s = 0.135 m and s1 = 0.102 m (Fig. 1).

The area of contact A was calculated using eqn. (1)

A=bsi+rsarcsin = 0.08 m2 (1)

Then the average pressure on the interface between the ball and the coal layer jj = L/A = 1.5 MPa was obtained. Assuming a certain pressure distribution inside the contact zone, the maximum nominal pressure was calculated. The value of maximum pressure pmax was taken as representative for the given mill. Similar pressure conditions were created in the labora- tory test rig [3, 41 by a normal force N = 2000 N which results in pm,, = 2.83 MPa.

A A

Fig, 1. Simplified diagram of the contact between the ball and the upper race and the ball and the coal layer where A-A is the instantaneous axis of rotation, 1 is the area of contact and 2 is the pressure distribution at the interface between the ball and the bed of coal.

31

The value of the maximum slip velocity u,,,, between the ball and the race (or coal layer) was calculated based on the pattern shown in Fig. 1 [ 51

(2)

where u = (n/30)7r, R = 4.67 m s-l, h = 0.12 m, 1 = 0.06 m, R = 1.65 m and n = 27 min-‘.

Controlled thermal conditions inside the coal mills must be maintained since all the moisture contained in the coal must be evaporated before igni- tion can take place. The drying capability of a given pulverizer design de- pends on the extent of the circulating load within the mill (the ability to mix the dry classifier is restored rapidly with incoming raw wet coal feed) and on the air weight and air temperature within the design toleration limits. Some pulverizers are designed to operate satisfactorily with an inlet air temperature of up to 700 K. Within the shear zone of crushed coal the temperature is usually 100 - 200 K lower than the inlet air temperature. The calculated value u,,,, = 1.55 m s-l applied in the recommended labo- ratory rig [3, 41 generated an equilibrium temperature which was too high and which exceeded the temperature inside industrial mills. The tempera- ture inside the grinding zone of the coal was taken as a more important similarity criterion than the velocity. By means of preliminary experiments [3] the velocity was fixed at a level which guaranteed the same temperature in the laboratory rig as in industrial mills. This was important since for

simulation criterions

Fig. 2. Simulative tribotesting of phenomena inside the coal pulverizer where 1 is the simulation of the conditions on the ball-upper race interface and’ 2 is the simulation of the tribocondition in the ball-coal layer system.

32

temperatures above the critical point coals start to burn and fundamentally change their properties. Finally, u,,,, = 0.157 m s-l (n = 100 min-‘) was chosen for the proposed tester [3,4]. The procedure for adjusting the operating variables in simulative tribotesting is shown in Fig. 2.

3. Laboratory tribotesting

The simulative tribotesting of the phenomena inside the coal pulverizer was divided into two parts, dealing first with the interactions between the ball and the upper race and secondly concerning the ball-coal layer inter- face. A schematic diagram of the procedure for simulative tribotesting is shown in Fig. 2. Only that part of the investigation which simulates the ball-coal layer system is presented in this paper.

3.1. Basic mechanical properties of selected coals In order to attain the greatest efficiency in processing coal, it is essen-

tial to obtain a thorough knowledge of the fundamental behaviour of coal when it is subjected to various kinds of forces such aa those occurring during compressive strength tests, impact strength tests [6] or hardness tests. Once these properties of the coals have been determined, an attempt to establish a correlation with other properties such as grindability can be made and a way of predicting coal behaviour inside a pulverizer can be found. Coal has to be broken down to sizes appropriate for combustion. Owing to the ever-increasing mechanization of coal processing, it has become important to have more knowledge about the strength of coal and an insight into its mode of breakage under stress. Coals vary widely both in chemical constitution and physical structure. The breakage of coal occurs as a result of applied stresses on weaknesses due to cracks, ranging in size from sub- microscopic to macroscopic, that run through the coal. Five coals, namely, Camden, Duvha, Grootvlei, Hendrina and Kriel, were selected for the tests. All the results are summarized in Table 1. The relation between the various properties of the selected coals in the form of regression equations may be presented as follows:

IS1 = 60.4 + 0.236 ucl r = 0.870

0 ,.. = 3 VHSI - 729 r = 0.463 (3)

ISI = 46 + 1.32 VH5, r = 0.770

where r is the coefficient of correlation.

3.2. Simulation of conditions at the ball-coal layer system The analysis of the comminution processes in ball-race mills showed

that the relative redistribution of fine coal layers in front of the rotating balls is the main agent in the grinding process. Pulverization takes place on the shear planes between loose particles of coal under stress and in the sector

33

Fig. 3. Schematic diagram of the tribotester: 1, drill chuck; 2, drive shaft; 3, cylindrical chamber; 4, thermocouple; 5, torque indicator; 6, thrust bearing; 7, normal force indica- tor ; 8, displacement indicator; 9, recorder.

of semiherztian contact (if the stress in this sector is higher than the com- pressive strength of coal). Wear on both grinding elements (i.e. the balls and the race) originates mainly from the sliding interaction between the hard coal components and the grinding elements. The relative displacement of quartz and pyrite particles produces abrasion-like grooves by cutting and ploughing. Assuming that the predominant wear phenomena result from the relative rubbing between the layers of coal and the grinding elements and that the grinding processes take place mainly in the shear zone of fine coal, the pulverizer’s ball-race configuration could be simplified and replaced by a simulated test such as that shown in Fig. 3. In this apparatus, various wear and grinding conditions may be simulated (Fig. 2) by adjusting the operating variables (N and n) and the constant quantities such as di, d, and h [3, 41. The proposed test method described below was chosen to simulate as closely as possible the grinding-wearing action in a full-scale ball-race pulverizer. The apparatus (Fig. 3) consists of a drive shaft and disc rotating in a closed cylindrical chamber. Normal pressure can be applied to the disc through the drive shaft. The wear element consists of a rectangular blade fixed to the underside of the disc. The space in the bottom of the cylinder is filled with a given mass of the coal sample. The disc is then rotated for a given number of revolutions and the mass loss of the blade is determined. During the grinding-wear process the power input is measured. The use of the proposed method enables a number of parameters of interest in tribology and comminution to be determined. These are expressed mathematically in Appendix B.

TA

BL

E

1 w

fi

Bas

ic

mec

han

ical

pr

oper

ties

of

fi

ve

sele

cted

co

als

Num

ber

Coa

l pr

oper

ties

C

amde

n D

uvha

G

root

vlei

H

endr

ina

Kri

el

1 V

Hs

(loa

d 5

kgf

) 1

21.4

23

.8

18.5

21

.5

18.2

2

VH

s (l

oad

5 k

gf)

11

17.2

19

.9

17.8

17

.2

16.2

3

G

(MP

a)

80.3

54

.0

46.0

51

.6

41.5

4

oc

(MP

a)

59.0

50

.3

44.0

25

.3

37.6

5

IS1

78.4

75

.6

71.3

73

.7

67.9

6

App

roxi

mat

e pe

rcen

tage

of

qu

artz

(%

) 14

.2

2.5

2.0

21.5

15

.5

7 A

ppro

xim

ate

perc

enta

ge

of

pyri

te

(%)

4.7

2.7

2.0

2.0

2.5

8 IC

(mgJ

-‘)

0.73

4 0.

755

0.57

5 0.

673

0.64

5 9

Wi

(J P

-l)

1533

15

32

2057

18

88

1974

10

A

F,

(bla

de

of

stan

dard

st

eel)

(m

g k

g-‘)

1547

34

0 28

3 58

6 39

5 11

A

F,

(bla

de

of

cast

ir

on)

(mg

kg-

‘) 13

5 38

.6

43.2

79

.7

56.5

12

IA

, (b

lade

of

st

anda

rd

stee

l)

(mg

rnp2

ss

’) 13

00

284

163

391

251

13

IA,

(bla

de

of

cast

ir

on)

(mg

me2

s-

l)

99

32.3

22

.6

49.1

34

.4

14

7,

(MP

a)

2.30

4 2.

176

1.68

9 2.

096

1.97

8 15

C

(M

Pa)

0.

958

1.11

5 1.

287

1.07

3 1.

224

16

G

39.1

37

.5

30.8

36

.4

34.8

35

A series of tests have been performed on five selected coals to deter- mine their grindability and abrasiveness. The results of the tests are sum- marized in Table 1. As only those results from tests which completely simulate the operational and material conditions in the industrial pulverizers can be directly applied to the design calculations, the same grade of cast iron was used for the blades as is currently used for the races.

4. Correlation between the mill grinding component life and the abrasive- ness of coals determined under laboratory conditions

A comparison has been made between the industrial data collected by the Electricity Supply Commission (ESCOM) Power Station Performance Monitoring Services and the laboratory results for the following five coals: Camden; Duvha; Grootvlei; Hendrina; Kriel (Table 2). A very good correla- tion has been found between simulative tribotesting results and full-scale industrial mill data (Fig. 4). The correlations may be presented in the form of the regression equations as follows:

TABLE 2

Comparison between mill grinding component life and abrasiveness of selected coals

Number Coal MSL

(h)

ASL

(h) AFc (mg kg-‘)

IA, (mg rnsz s-l)

1 Camden 13260 5700 135.0 99.0

2 Duvha 20000 15300 38.6 32.3

3 Grootvlei 25435 16009 43.2 22.6

4 Hendrina 22000 14000 79.7 49.1

5 Kriel 22000 15500 56.5 34.4

t 25xlGOOh

ASL 8 MSL

Fig. 4. Correlation between IA, and MSL and ASL, where 1 and 2 are the regression lines of the equations MSL = 27064 - 137 IA, and ASL = 19976 - 140 IA,.

36

MSL = 27 287 - 95.6 AF,

MSL = 27 064 - 137 IA,

ASL = 20 725 - 105 AF,

ASL = 19 976 - 140 IA,

r = -0.834

r = -0.923

r = -0.960

r = -0.988

(41

5. Conclusion

The comparison between the industrial data and the laboratory results indicates that there is a very good correlation between factors describing the abrasiveness of coal according to the new proposed method of testing and the maximum and average service life for the bottom ring in Babcock vertical mills. The abrasive factor and intensity of abrasion can be applied in the calculation of the maximum and average service life of the rings. In this case, the results from a relatively simple laboratory apparatus, designed and operated according to the principles of similarity, may be used to pre- dict the service life of machine elements in industry. The simulative tribo- testing procedure and rig can be used to evaluate the abrasiveness of any granular material and for testing the wear resistance of any material in any two-body abrasive action.

Acknowledgments

This Paper is based on research supported by the Mechanical Engi- neering Department of the University of Cape Town and by ESCOM, Johan- nesburg. The author wishes to thank Professor R. K. Dutkiewicz for his valuable comments during the whole research programme and the ESCOM Power Station Performance Monitoring Services for providing power plant data.

References

1 H. Krause and T. Senuma, A contribution towards improving the applicability of laboratory wear tests in practice. In S. K. Rhee, A. W. Ruff and K. C. Ludema (eds.), Proc. Int. Conf. on Wear of Materials, San Francisco, CA, March 30 -April 1, 1981, American Society of Mechanical Engineers, New York, 1981, pp. 753 - 763.

2 H. Czichos, Tribology, Elsevier, Amsterdam, 1978, pp. 264, 277. 3 S. F. Scieszka, New concept for determining pulverizing properties of coal, Fuel, 64

(1985) 1132 - 1142. 4 S. F. Scieszka, A technique to investigate pulverizing properties of coal, Powder

Technol., 43 (1985) 89 - 102. 5 F. J. Halling, Principles of Tribology, Macmillan, London, 1975, p. 117. 6 I. Evans and G. Pomeroy, The Strength, Fracture and Workability of Coal, Pergamon,

London, pp. 55,65.

37

Appendix A: Nomenclature

a AFc ASL b C

d, 4 EI F h

;A ICC IS1

ml

2,

;

P P PC R S t

td

T

TP G

UVH, W

wi AW WR

diameter of the hertzian contact area (m) abrasion factor (mg kg-‘) average service life (h) diameter of the hertzian contact area (m) apparent cohesion (MPa) diameter of disc (m) diameter of cylinder (m) energy input (J) feed size modulus (pm) height of blade (m) number of revolutions intensity of abrasion (mg rnp2 s-l) index of comminution (mg J-‘) impact strength index initial mass of blade (g) final mass of blade (g) maximum service life (h) rotational speed normal force (N) pressure (MPa) product size modulus (pm) pulverized fraction of coal below 75 I.trn (g) radius (m) area of surface (m2) time (s) duration of test (s) average integral value of torque (Nm) peak value of torque (Nm) residual value of torque (Nm) velocity (m s-i) Vickers’ hardness (load 5 kgf) work input (J g-‘) work index (J g-‘) wear of blade (g) wear resistance (MJ g-i)

relative wear resistance compressive strength (MPa) normal stress (MPa) residual value of shear strength (MPa) ultimate friction angle (deg)

Appendix B

B.l. Presentation of results The following formulae are used:

38

wear of blade

AW=m,-mmz

energy input

EI = 2nTi

where

Tc’j td o

T(t) d2

abrasion factor of coal

AF = $J6

intensity of abrasion

IA= ; d

where S is the area of blade surface exposed to wear (m2)

work index

index of comminution of coal

IC= ;103

wear resistance of material

WR = ;10-6

relative wear resistance of material

E’ WRspecimen

WRstandard

peak value of shear strength of crushed coal

6

residual value of shear strength of crushed coal

6

031)

W)

033)

034)

U35)

036)

U37)

w3)

UW

@lo)

39

apparent cohesion of interlocked coal particles

C= &VP -Cl lo+

ultimate frictional angle of internal resistance