on the design and analysis of transfer chute systems · a set of measurable material properties....

XXVI Encontro Nacional de Tratamento de Minérios e Metalurgia Extrativa

Poços de Caldas-MG, 18 a 22 de Outubro 2015

ON THE DESIGN AND ANALYSIS OF TRANSFER CHUTE SYSTEMS

ILIC, D.1, DONOHUE, T.J.2 1,2TUNRA Bulk Solids, University of Newcastle, Australia.

e-mails: [email protected], [email protected]

ABSTRACT Chute transfers are crucial components of the materials handling chain, yet too often, also the weakest link. They are typically employed in belt conveying systems to accelerate bulk materials from feeders, screens, transfer from one conveyor to another, or alternatively discharge into storage vessels including silos, bins or ship holds. Inadequate design in combination with variation of material characteristics or operational requirements often results in problems, which translate to loss in productivity, unscheduled maintenance and stalled operation. Understanding of material properties is of utmost importance and the key parameters are discussed. The analysis of transfer systems is performed using continuum mechanics based on a set of measurable material properties. With such application, overview of transfer functionality can be quickly developed for single particles. However, as material characteristics vary from idealistic free flowing and the nature of the transfer deviates from a rapidly moving stream, the interaction between individual particles becomes of increased importance. Application of Discrete Element Modelling (DEM) to analyse such interactions, although restrained by computing power, is at a stage where evaluation of large scale systems is thorough and advanced. The practicalities associated with application of both of these modelling techniques to the evaluation and optimisation of transfer chute systems is presented, as well as key design criteria developed through flow analysis which would be of extreme benefit if incorporated into the tendering process. KEYWORDS: handling; transfer; chute; modelling; optimisation; design. RESUMO Chutes de transferência são componentes cruciais na cadeia de manuseio de materiais. No entanto, frequentemente são também o elo mais fraco. Chutes são tipicamente aplicados em sistemas de transportadores de correia para acelerar sólidos granulados provenientes de alimentadores e peneiras, ou para transferir de um transportador para outro ou descarregar em silos ou porões de navios. O projeto inadequado, em conjunto com várias características do material ou exigências operacionais, frequentemente resultam em problemas, que têm como consequência perda de produtividade, manutenções não programadas e atrasos operacionais. A compreensão de propriedades do granel manuseado é de suma importância e os principais parâmetros são discutidos. A análise de sistemas de transferência é realizada através da mecânica do contínuo com base em uma série de propriedades mensuráveis do material. Com tal aplicação, uma ideia geral da funcionalidade da transferência pode ser realizada de forma rápida para partículas unitárias. Entretanto, à medida que as características do material se diferenciam da situação ideal de fluxo livre e a característica do fluxo desvia de um fluxo rápido, a interação entre as partículas individuais se torna cada vez mais importante. A aplicação do

Ilic, D.; Donohue, T.J.

Método dos Elementos Discretos (DEM) para analisar essas interações, apesar de limitada pela capacidade computacional, atingiu um estágio em que é possível avaliar sistemas de larga escala de forma avançada e detalhada. Os aspectos práticos associados à aplicação tanto dessas técnicas de modelamento quanto à avaliação e otimização de chutes de transferência são apresentados, bem como critérios de projeto desenvolvidos através da análise de fluxo, que pode ser de grande benefício se incorporada às propostas. PALAVRAS-CHAVE: manuseio; transferência; chute; modelagem; otimização; projeto.



1. INTRODUCTION The guidelines associated with transfer chute design are based on maintaining a rapidly moving stream characterized by accelerated flow conditions for which the continuum analysis approach is applied. This method directly implements material properties, obtained through standardised tests, and a general assessment of transfer functionality is obtained within hours. Without adequate assumptions, flow characteristics exhibited at other conditions cannot be evaluated. Phenomena associated with dispersed, decelerated streams, multiple co-ordinate planes of flow and build-up is extremely difficult to evaluate and relies on many assumptions. Supporting structure, prevents clear observation of flow within transfer configurations. Also, it is dangerous to freely open maintenance access hatches and adequate feedback from site, other than observations prior to and immediately following the transfer is sparse. Other types of verification may include measurements from belt weighing systems indicating throughput and forces on idler rolls in the loading area. Opportunely, DEM allows such visualization of the general mode of flow and enables qualitative analyses to be obtained. Accurate specification of DEM input parameters leads to quantitative results from which challenging systems and difficult to handle materials may be thoroughly analysed. Common materials handled vary profoundly in behavioural characteristics and include coal, iron ore, bauxite, gold ore, alumina, potash, wheat and biomass. Typical particle fractions associated with these materials range from micron size up to primary/secondary type crushed large lumps. The major drawback of DEM is that the computational time can be in days or weeks and to achieve realistic turnarounds the inclusion of fine particles is omitted. Among the wide field of its application, DEM of transfer flows belong to those that have enjoyed an increasingly large interest. Transfer bottlenecks have been recognized by industry as warranting considerable attention. Emphasis on technical specification and numerical modelling, based on the characteristics of the material handled, is being applied in the preliminary stages of design. 2. CONTINUUM MODELLING Material characterisation and flow parameter determination forms the fundamental basis for the implementation of the continuum approach and indeed transfer design. Standardised, well established macroscopic property tests such as compressibility, angle of repose, particle size distribution, wall (boundary) friction and internal strength of bulk materials are performed and then directly applied. Belt carrying capacity may be assessed using the procedure outlined in CEMA [3] for example. Head height, critical velocity and discharge angle is then evaluated, following which the trajectory path is plotted. Based on projectile motion, a number of different models for the calculation of discharge trajectories have been developed based and include, Dunlop Conveyor Manual [8] and Booth [4].


The discharging material stream is then re-directed typically through an impact plate (curved hood) and/or a subsequent transfer chute (loading spoon). Pioneering work of Roberts [23 - 26], lead to the development of the lumped parameter analysis model for flow through transfer chutes. In these accelerated flows, with shallow bed depth to chute width ratios (H/B<0.5), Roberts [23] verified experimentally that the greatest total frictional work is due to the bulk solid material sliding on the chute bottom. From the work of Roberts [23 - 25] the equations of motion are summarized

below, where R is the radius of curvature of chute and is the angular position.

Figure 1. Accelerated Chute Flow Models from Roberts [25] – Hood and Spoon.

For the hood:

)sin(cosV

RgV

d

dVEE

for sin

Rg

V2

(1)

For the spoon:

)sin(cosV

RgV

d

dVEE

(2)

Solutions of Eqn. (1) and (2) take into account the variation of equivalent friction, μE, along the chute as a function of stream velocity. For a straight inclined chute, the velocity of the material stream a distance, s, down the incline can be expressed as:

𝑉 = √𝑉𝑜2 + 2 ∙ 𝑎 ∙ 𝑠 (3)

Where acceleration 𝑎 = 𝑔 ∙ (𝑐𝑜𝑠𝜃 − 𝜇𝐸 ∙ 𝑠𝑖𝑛𝜃) and 𝜃 is the inclination of the chute from the vertical. The model used to calculate velocity immediately following impact,

Vo, based on velocity, Vi, and incidence angle 1 prior to impact, is given by:

11i

o sine1cosV

V (4)

Where the restitution factor 1e0 . Nominally, impact angles are maintained low, completely inelastic impact and no rebound is assumed. By maintaining continuity, and applying the above equations flow is analysed at specific areas by “following” a



streamline from the incoming conveyor, through the transfer to loading onto the outgoing conveyor. Loading to an inclined belt is illustrated in Figure 2.

Figure 2. Loading Velocity Components at Chute Exit and Acceleration Length.

Calculation of velocities through the transfer allows impact pressure, Pvi and abrasive wear, Wa, on the belt to be calculated:

Pvi = Vn2 (5)

Wa = µ1 Vn2 (Vb – Vp) (6)

Where µ1 is friction between the bulk solid material and conveyor belt, Vb is belt speed, Vn the normal velocity component and Vp is the velocity component parallel to belt travel direction. Roberts [26] assumes acceleration is non-uniform and velocity as a function of distance along the belt is non-linear. This is due to the material burden height, hs, decreasing with an increase in velocity. Assuming block like motion, analysis of the forces due to the belt driving the material forward between the skirt plates with width of, ws, allows acceleration to be calculated:

𝑎 = 𝜇1 ∙ 𝑔 − 𝜇2 ∙ 𝑔 ∙ 𝐾𝑉 ∙ℎ𝑠

𝑊𝑠 (7)

Where µ1 is the friction with the belt, µ2 is the friction with the skirts and typically 0.5 < µ1 and µ2 < 0.7. This then allows for the acceleration length, Lb to be determined. The internal walls of transfer chutes are generally lined with materials such as stainless steel and ceramic tile where low friction is required. Abrasion resistant metal or urethane linings have also been implemented for high abrasive wear and/or heat resistant installations. For impact resistant installations, hard faced liners are used and in situations where hard rocks are handled, ledges or micro-ledges (rock-box, dead-box) are installed. In these types of transfers, the mode of flow is dependent on the internal friction characteristics of the bulk solid material rather than the friction between the material and the wall boundary. In a previous study, Donohue et al [7] show that this type of flow can also be modelled using the continuum approach by making appropriate assumptions.


3. DISCRETE ELEMENT MODELLING (DEM)

In recent years, DEM has become an increasingly popular tool in the analysis and visualization of interactions in belt conveying systems, with transfer chutes by far the most documented. Studies by Hustrulid and Mustoe [10] and Nordell [22]) were some of the first while verification of simulated transfer flow and impact forces on the belt were measured by Katterfeld et al [16] while Kessler and Prenner [18] looked at project specific scale modelling calibration. Research into and verification of the continuum method by DEM have been studied by Ilic et al [11] and Donohue et al [7]. In contrast to the lumped parameter approach, DEM considers individual particles. The interaction of an assembly of particles and their environment is treated dynamically and, during one calculation cycle, contact forces and displacements are evaluated. Integration over a small time step, particle velocities and positions are updated. Simulated properties are defined by the contact model used are microscopic. They include shear (Young’s) modulus, bulk elastic stiffness, co-efficient of restitution, damping, particle/particle, particle/boundary friction, Poisson’s ratio, rolling friction and particle density. Due to the impracticalities of directly measuring such parameters, calibration tests which provide the link to physical bulk solid properties are performed. These include angle of repose, internal friction and boundary friction (sliding/dynamic friction) tests where DEM parameters are iterated to “fit” real macroscopic behaviour. At present, transfer chute flow simulation generally involves modelling particles in the order of 20-50mm diameter. The exact size and loading stiffness is governed by the throughput, overall application footprint, end goal of analyses and project deadlines. Depending on the nature of the material handled, the application and outcomes sought, specific calibration tests may also be required. Modelled particles are assumed rigid with deformation accounted for in the contacts between them. Particle overlap is used to calculate the normal and tangential contact forces. The total force acting on a particle is the summation of the total normal force and the total tangential force acting at that contact. Specifics regarding different contact models have been the subject of research for many years including Cundall [5], Walton [28], Wassgren [29-31], Di Renzo and Di Maio [6], Schäfer et al [27], Luding [21] and Kruggel-Emden et al [19]. From that research, a generalised overview of the model implemented in RockyTM (used by TUNRA Bulk Solids) is presented. The normal force contact is a partially latching or hysteretic linear spring, originally introduced by Walton and Braun [28]. The model is based on a specified stiffness during loading and a calculated stiffness during unloading. A diagram illustrating contact between two spheres with radius Ri and Rj, of mass mi and mass, mj is shown in Figure 3, as found in Wassgren [30]. The normal force, Fn, is given by:

res

maxres

max

resnNU

nNL

n

00

k

k

F

(8)



Where NLk and NUk are the loading/unloading spring stiffness NLNU kk and n, res

and max are the normal, residual and maximum overlap during contact. Loading stiffness is calculated by multiplying group stiffness to particle size. In this manner, group stiffness is roughly equivalent to bulk elastic modulus of the simulated material (not individual particles) based on work done by Bathurst and Rothenburg [2]:

LEkLN

and 2res

UN

LEk

Where E is bulk elastic stiffness, L the particle size and res

restitution coefficient. On the other hand, the tangential contact force model is based on a linear spring in series with a sliding friction element (without damping), is also illustrated in Figure 3. The tangential force, Fs, acting on particle i, is given by:

𝐹𝑆,𝑜𝑛 𝑖 = min(𝑘𝑆 ∙ 𝛿𝑠, 𝜇|𝐹𝑁|) (9)

Where 𝛿𝑠 is tangential (shear) overlap, 𝑘𝑆 shear stiffness, 𝜇 is either static, µs, or dynamic, µd, friction depending if sliding is or is not taking place for the contact. To reduce the number of variables, the dynamic and sliding friction, is assumed equal. For more information regarding the normal and tangential force contact model, refer to Walton et al [28] and Wassgren [29, 30]. A simple cohesion model used in RockyTM is based on an on/off switch and illustrated in Figure 3. For two particles in contact, the adhesive force is calculated by an adhesive fraction coefficient, µadh, multiplied by the minimum weight of the two corresponding particles. It is initiated when the distance between particles, zn, is lower than a defined distance, sadh. When the distance specified is exceeded, adhesive force becomes zero.

Normal Contact Force Tangential Contact Force Adhesion Contact Force

Figure 3. Contact Force Models (RockyTM).

The rolling friction model implemented in RockyTM is based on the Model Type C, elastic plastic spring-dashpot model as described in Ai et al [1] and refined by Wensrich and Katterfeld [32]. In this model, the total rolling resistance torque consists of two components: a mechanical spring torque and a viscous damping torque, incrementally calculated and related to the contact stiffness, coefficient of rolling friction, contact normal force, angular velocity, radius of the particles in contact, damping rate and inertia. Through research widely published, well established guidelines regarding selection of parameters across a number of applications including loading stiffness, damping, Poisson’s ratio, coefficient of restitution can be


found: Zhou et al [33], Li [20], Katterfeld [15Erro! Fonte de referência não encontrada.-15], Johnstone [14], Grima [9], Katterhagen et al [17] and Ji et al [13]. 4. PRACTICAL APPLICATION The design of a central transfer on a stacker machine was developed to handle coal at 8,000t/h and belt speeds in the order of 6.0m/s for an Australian high capacity coal export terminal. The transfer and velocity profile comparison obtained using both analysis tools are presented in Figure 4. Transfer components were designed using the traditional continuum mechanics approach only. Following commissioning, verification using DEM was also performed (Ilic et al [11]). A subsequent in depth study (Ilic [12]) was undertaken to further investigate each method. The burden remaining in the spoon at cessation of flow is shown in Figure 6.

Figure 4. Stacker central transfer chute.

This transfer has been in operation for over 6 years, handling tonnages above the original design capacity. Subsequently, the same design was adopted as the basis for two additional stacking machines transferring in excess of 10,000t/h, which were installed at the same terminal in 2012. The transfers were used as a platform for the development of technical design requirements later incorporated into tender documents for a terminal expansion in 2013 and replacement ship loaders this year. Identical methodology was recently employed in the design of a biomass handling facility located at a major power plant in the UK. The project was completed in collaboration with a local conveyor manufacturing company for which in excess of twenty transfers were designed within structural constraints and operational criteria. An example of one of the transfer concepts is shown in Figure5 below. The facility proved to be an aspiring one, in which fall heights in excess of 25m and throughputs of 2800t/h were assigned. Following the success of these transfers, TUNRA Bulk Solids was approached regarding development of technical design criteria which was to be incorporated into contract documents for future expansion work.



Figure 5. Modelling Biomass: continuum, DEM and site observations during installation.

5. GUIDELINES According to Roberts [26] the primary objectives of transfer chute design are to:

Match the component of exit velocity Vp as close as possible to the belt speed;

Reduce the component of the velocity normal to the belt, Vn, (i.e. abrasive wear);

Load conveyor belts centrally and evenly so as to avoid mis-tracking issues;

Ensure reliable flow without spillage or blockages, and self-cleaning of the chute (cut-off angle, selected as compromise between loading velocity and hang-up).

General guidelines which are commonly found in technical specification design criteria documents related are also summarised below:

Transfer must handle design throughput in an efficient and favourable manner;

Self-cleaning chutes, with minimization of hang-up and no spillage;

The flow is to be controlled with no blockages;

Minimization of equipment wear, areas of free fall and angle of impact;

Loading to be as gently as possible, symmetrical and central;

Modularized construction for easy replacement in case of failure;

Optimise airflow, turbulence, impact, temperature and dust generation. 6. FLOW DESIGN CRITERIA AS PART OF TECHNICAL SPECIFICATION Application of the numerical analyses approaches presented in parallel, based on properties reflective of the material handled, allows for the most accurate flow assessment up to date. Introduction of more specific design criteria as part of the technical specification used in transfer design is proposed:

Material burden shall not exceed (some maximum i.e. 65%) of the chute cross sectional area;

Stream velocity at loading point on outgoing conveyor shall match, within ±(tolerated i.e. 10)%, the outgoing belt velocity;

Head height of the transfer is to be no lower than (minimum i.e. 4.0m) vertical distance from head pulley centreline to underside of belt;


Impact angles within transfer shall be kept to within (tolerated i.e. 15-30 degrees range) or lower;

Chute inclination in flow areas is to be no lower than (maximum i.e. angle of internal friction) and greater than (minimum i.e. wall friction angle) at the exit.

7. REFERENCES 1. Ai, J. Chen, J. Rotter, M. Ooi, J.Y. Assessment of rolling resistance models in

discrete element simulations, Powder Technology, Vol 206, pp269-282, 2011.

2. Bathurst, R.J., Rothenburg, L., Micromechanical Aspects of Isotropic Granular Assemblies with Linear Contact Interactions, Journal of Applied Mechanics, Vol.55, pp17-23, ASME, 1988.

3. Belt Conveyors for Bulk Materials 6th Edition, Conveyor Equipment Manufacturers Association (CEMA), USA, 2005.

4. Booth, E.P.O., Trajectories from Conveyors – Method of Calculating them Corrected, Engineering and Mining Journal, Vol.135, No.12, pp552-554, 1934.

5. Cundall P.A., Strack O.D.L, A Discrete Numerical model for Granular Assemblies, Geotechnique, Vol. 29, No. 1, pp47-65, 1979.

6. Di Renzo, A., Di Maio, F.P., Comparison of Contact-Force Models for the Simulation of Collisions in DEM-Based Granular Flow Codes, Chemical Engineering Science, Vol. 59, pp525-541, 2004.

7. Donohue, T.J., Ilic, D., Bell, R. Newman, L., The use of DEM in the Design and Analysis of WEARBACK Transfer Chutes, Australian Bulk Handling Review, pp62-67, 11/12 2010.

8. Dunlop Conveyor Manual, Dunlop Industrial and Aviation Division, Dunlop Automotive and Industrial Group, Australia, 1977.

9. Grima, A.P., Wypych, P.W., Development and Validation of Calibration Methods for Discrete Element Modelling, Granular Matter, Vol. 13, pp127-132, 2011.

10. Hustrulid, A.I., Mustoe, G.W. Engineering Analysis of Transfer Points using Discrete Element Analysis, Proceedings of the Annual Meeting of the Society of Mining Engineers (SME), Phoenix, AZ, 1996, Chapter 2, pp9-13, 1996.

11. Ilic, D. McBride, W. Katterfeld, A. 2007. Validation of Continuum Methods Utilising Discrete Element Simulations as applied to a Slewing Stacker Transfer Chute, 9th ICBMH 2007., Conference Proceedings (USB), University of Newcastle, Australia.

12. Ilic, D., Bulk solid interactions in belt conveying systems, PhD Thesis, The University of Newcastle, 2013.



13. Ji, S., Hanes, D.M, Shen, H.H, Comparison of Physical Experiment and Discrete Element Simulations of Sheared Granular Materials in an Annular Shear Cell, Mechanics of Materials, Vol. 41, No. 6, pp764-776, 2009.

14. Johnstone, M.W., Calibration of DEM models for Granular Materials Using Bulk Physical Tests, PhD Thesis, University of Edinburgh, 2010.

15. Katterfeld, A., Donohue, T. J., Ilic D., Application of the Discrete Element Method in Mechanical Conveying of Bulk Materials, Plenary paper presented at 7th

International Conference for Conveying and Handling of Particulate Solids (CHOPS), Germany, 2012.

16. Katterfeld, A. Gröger, T. & Minkin, A. Discrete Element Simulation of Transfer Stations and their Verification. Bulk Solids & Powder Science Technology, Vol. 2, 2007.

17. Katterhagen, W.R., Curtis, J.S., Wassgren, C.R., Stress Results from Two-Dimensional Granular Shear Flow Simulations Using Various Collision Models, Physical Review E, Vol. 71, 061307, American Physical Society, pp1-11, 2005.

18. Kessler, F., Prenner, M., DEM – Simulation of Conveyor Transfer Chutes, FME Transactions, Faculty of Mechanical Engineering, Belgrade, Vol. 37, pp185-192, 2009.

19. Kruggel-Emden, H., Simsek, E., Rickelt, S., Wirtz, S., Scherer, V., Review and Extension of Normal Force Models for the Discrete Element Method, Powder Technology, Vol. 171, pp157-173, 2007.

20. Li, Y. Xu, Y. Thornton, C. A comparison of discrete element simulations and experiments for ‘sandpiles’ composed of spherical particles, Powder Technology, Vol. 160, pp 219-228, 2005.

21. Luding, S., About Contact Force-Laws for Cohesive Frictional Materials in 2D and 3D, Proceedings Issue, Behaviour of Granular Media, 2006.

22. Nordell, L.K., Palabora Installs Curved Transfer Chute in Hard Rock to Minimize Belt Cover Wear, Bulk Solids Handling, Vol. 14, No.4, pp739-743.

23. Roberts, A.W., An Investigation of the Gravity Flow of Non-cohesive Granular Materials Through Discharge Chutes, Transactions of the ASME, American Society of Mechanical Engineers, Journal of Engineering for Industry, Paper No. 68-MH-5, pp373-381, 1969.

24. Roberts, A.W., Basic Principles of Bulk Solids, Storage, Flow and Handling, TUNRA Bulk Solids, Centre for Bulk Solids and Particulate Technologies, 1998.

25. Roberts, A.W. Wiche, S.J. Ilic, D.D. & Plint, S.R. 2004. Flow Dynamics and Wear Considerations in Transfer Chute Design, Proceedings of the 8th ICBMH 2004., Institution of Engineers, Australia: 330-334.


26. Roberts, A.W., Chute Design Considerations for Feeding and Transfer, Centre for Bulk Solids and Particulate Technologies, Interim Report, University of Newcastle, 2004.

27. Schäfer, J., Dippel, S., Wolf, D.E., Force Schemes in Simulations of Granular Materials, Journal de Physique I, Vol. 6, pp5-20, France, 1996.

28. Walton, O.R., Braun, R.L., Stress Calculations for Assemblies of Inelastic Spheres in Uniform Shear, Acta Mechanica, Vol. 63, pp73-86, Springer-Verlag, 1986.

29. Wassgren, C, Sarkar A. ("PSL DEM Lecture 06: Introduction to Soft-Particle DEM and Normal Contact Force Models - Part I," http://pharmahub.org/resources/122, 2008.

30. Wassgren, C., Sarkar, A. "PSL DEM Lecture 07: Normal Contact Force Models - Part II," http://pharmahub.org/resources/124, 2008.

31. Wassgren, C., Sarkar, A. "PSL DEM Lecture 09: Tangential Contact Force Models," http://pharmahub.org/resources/253, 2008.

32. Wensrich, C.M. Katterfeld, A. Rolling Friction as a Technique for Modelling Particle Shape in DEM, Poweder Technology, Vol. 217, pp409-417, 2012.

33. Zhou, C. Ooi, J.Y., Numerical Investigation of progressive development of granular pile with spherical and non-spherical particles, Mechanics of Materials, Vol. 41, pp 707-714, 2009.

on the design and analysis of transfer chute systems · a set of measurable material properties....

Documents