Computer Simulation of Batch Grinding Process Based on Simulink 5.0 Computer Simulation of Batch Grinding Process Based on Simulink 5.0

Computer Simulation of Batch Grinding Process Based on Simulink 5.0

  • 期刊名字:中国矿业大学学报(英文版)
  • 文件大小:521kb
  • 论文作者:LI Xia,YANG Ying-jie,DENG Hui-
  • 作者单位:School of Mineral Processing & Bioengineering
  • 更新时间:2020-11-22
  • 下载次数:
论文简介

Jun. 2005J. China Univ. of Mining & Tech.(English Edition)Vol.15 No.2Computer Simulation of Batch GrindingProcess Based on Simulink 5.0LI Xia, YANG Yingjie, DENG Hui yong ,HUANG Guang-yaoSchool of Mineral Processing & Bioengineering,Central South niversity,Changsha ,Hunan 410083, ChinaAbstract: How to use Simulink software in grinding system was studied. The method of designing batch grinding sub-system and the steps of building batch grinding blockset were introduced. Based on batch grinding population balancemodel, batch grinding was simulated with Simulink. The results show that the simulation system designed with Simulinkexplain reasonably the impersonal rule of batch grinding. On the basis of batch grinding simulation, the computer simula-tion of mineral processing system with Simulink of grinding and classification, comminution, etc, can be properly ex-plored.Key words: simulation; Simulink; mathematic model; batch grindingCLC number: TD 941 Introductionhigh performance equipments and advanced tech-Since MATLAB was issued by Mathworks com-nologies in mineral processing. With the advantagespany in America in 1984, MATLAB has become theof nondestructive measurement, iterance, security,most excellent software applied in science and tech-economic, controllability, no limit of the climate andnology all over the world in the past few years.space, computer simulation will bring a profoundSimulink is a branch product of MATLAB, whichtransformation to the exploiting process of traditionalprovides the system simulation with a visual andmineral resource. The three key components ofpowerful operation platform. For modeling in simu-computer simulation are system studied, mathematicJink, the models as block diagrams are built by justmodel and computer. The equation that Von Rittingerusing click-and-drag mouse operations with graphicalproposed to describe the energy change duringuser interface (GUI). With GUI, the models just ascrushing was the pioneer mathematic model inyou would with pencil and paper or as most text-crushing 2, but until the advent of digital computerbooks depict them are built. So the people who en-in 1960, the simulation of mineral processing processgage in scientific research are disengaged from thebecame true. Computer simulation has provided anheavy programming work with simulink. Now Simu-effective and economical means to resolve the prob-link has been used in simulation in the fields oflems in the optimization of flowsheet [3-5), the designpower system, signal control, communication design,of control scheme 16-8), the forecast of ore dressingfinance, account, biology and medicine, etc“dataetc. Grinding is an important unit opera-Mineral resources are important strategical re-tion-3sing. The quality ofsources for the national economy and defence. With中国煤化工the development of science and technology, the needgrindYHCNMH Grctly concentrationindex! research at all times.of various mineral products is increasing ceaselessly,however the grade of ore is falling gradually, so theResearch on batch grinding is the foundation of re-amount of ore needed to be disposed is rising in-search on grinding. In this paper, the software wascreasinglyunder the situation, it is imperative to useapplied to study the batch grinding process.Received 7 October 2004; accepted 7 November 2004LI Xia etalComputer Simulation of Batch Grinding Process Based On Simulink 5.01492 Principle of Simulation(Fig.2) and the output part. The batch grinding sub-system consisted of 5 constant blocks, 2 productThe principle of simulation in this paper is batchblocks, 1 output block. The output part consisted of 1grinding population balance model. The model is asscope block and 1To Workspace block.followsdf,(t)Batch=-Sf(1)+ ZbysjfS,(), (i= .2...n),.(1)ftdt4L mil」Size distributionwhere f; is the mass fraction of a particular size classBatch grindingof productsubsystemi; S; is the selection function or fraction breakage rateof size class i; bj is the breakage distribution functionScopeof the size class.Fig.1 Simulation model of batch grinding systemharstyConstantf f()][s、0...0BH Sum14 Matrix「MatixMatrixf2(t)0 SzBreakageMultiply1 Multply+Ddistribution matrix| Product 1f(t)=S=Product 2Intcgratordistrnbution[S]--f0 P of product[f,()]0 0...S.Selctionof feedFig.2 Batch grinding subsystem「00....00]b2; 03.2The application of blocks in Simulink to the8=|batch grinding simulation systemlbm bn2- ba.n 0]3.2.1IntegratorIntegrator block outputs the integral of its inputThe expression (1) can be transformed into the fol-at the current time step. Its input may be a scalar or alowing matrix differential equationvector. In the batch grinding simulation system, thedf(t)2=-(1- B)Sf(t),(2)input of Integrator block is the product's size distri-bution in the mill at different time. So, it is a vector.Where f([t) is the size distribution matrix of product; IIntegrator block has 10 parameters in all. It is neces-is the identity matrix of n rank; B is the breakage dis-sary to emphasize the usage of initial conditiontribution matrix; S is the selection matrix.source. Integrator gets the state’ initial condition3 Computer Simulation of Batch Grindingfrom Initial condition parameter (if set to interal) orfrom an extermal block (if set to external). When ini-Processtial condition source is set to extermal, it makes initial3.1 Simulation methodcondition parameter invalid and adds an input port toThe key to use the computer simulation is ana-the integrator block. In this paper, initial conditionlyzing the characters of the object studied and thensource was set to external, so we got the integratorestablishing the simulation model. On the base ofblock in Fig.2.batch grinding population balance model, the only3.2.2To workspacething to do is to denote rightly and concisely theTo Workspace block writes its input to themathematic model with Simulink. As the model in-workspace in MATLAB. In the simulation system,creases in size and complexity, the model can behe parameters required to set were V ariable Namesimplified by grouping blocks into subsystems andthen masking subsystems. In this paper, some blocksand Sa中国煤化工;ameter determinedwere grouped into batch grinding subsystem, andthe stdHCNMHGlariableNamewasthen batch grinding subsystem was masked. There-set to be fi and Save format was set to be array. So thefore, the simulation model of batch grinding systemsize distribution of product was saved as an aray(Fig.1) was made up of the batch grinding subsystemnamed f in workspace. The simulation results can be150J. China Univ. of Mining & Tech.(English Edition)Vol.15 No.2put in the MATLAB workspace for post- processingmodel established previously. Secondly, the start timeand visualization.of simulation was set to 0 min and the stop time of3.3 The simulation trialssimulation was set to 40 min using simulation pa-Firstly, the values of selection matrix S, break-rameters menu. Other parameters such as solver op-age distribution matrix B, identity matrix I and thetion, Refine factor, Save to workspace, Load fromsize distribution matrix of feed fo (Table 1) were en-workspace kept their default values. Finally, thetered to simulate batch grinding with the simulationsimulation was run.Table 1 Parameters used in the batch grinding simulationclass_fB0.579 30.5127 0.00000.208 10.087 30.292 60.1796 0.5875 0.000 00.044 60.221 1 .0.0832 0.1796 0.5875 0.00000.026 70.049 70.083 20.01740.12620.0324 0.0497 0.0832 0.1796 0.5875 0.0000.011 70.095 30.0218 0.0324 0.0497 0.0832 0.1796 0.5875 0.000 00.008 00.014 80.0218 0.0324 0.0497 0.0832 0.1796 0.5875 0.00090.00540.05440.0100 0.0148 0.0218 0.0324 0.0497 0.0832 0.1796 0.5875 0.00001(0.011 50.00000.0210 0.0320 0.0458 0.0676 0.1060 0.1497 0.2329 0.4125 1.000 0.00004 Results and Discussion5Establishment of Batch GrindingBlocksetFig.3 was the result using the variables t andf(t)in MATLAB workspace to plot with Origin 7.0.Blocksets supplied by Mathworks are the pro-From Fig.3, we can leam how the percentage weightfessional blocksets oriented to the special fields. It isin each size fraction changes with the time. Threethe expansion of the basic functions of Simulink. It istrends of change were derived: 1) the percentagevery convenient for user to build models in al kindsweight in coarse size fraction descended monoto-of fields. Though Simulink has included the block-nously; 2) the percentage weight in middle size frac-sets of Communications Blockset, DSP Blockset,tion firstly increased monotonously, and then de-Fuzzy Logic Toolbox, Neural Network Blockset,scended monotonously; 3) the percentage weight inStateflow, Control System Toolbox, Virtual Realityfine size fraction increased by degrees monotonously.According to the simulation results, it is very con-Toolbox, etc, it doesn't provide blocksets related withvenient to determine the optimal batch grinding time.mineral processing system. So the development ofOn the assumption that the grinding fineness is thatblocksets related with mineral processing system will20% passing through -37 μ m sieve, the batchfacilitate user to simulate mineral processing processgrinding time is around 12 min according to thein Simulink. In the following, the process of exploit-curve labeled 10 in Fig.3.ing batch grinding blockset in simulink will be in-Class umtroduced simply.0.74010Setp l: Use the menu command Simulink Li-0.brary Browser/FileNew Library in Simulink to cre-e 0.4ate a new model named batchgrinding, and then saveit int中国煤化工is workmyb-lkset-demo:IH.CNMHGStep 2: Drag tne Datcn grnding subsystem block,204(Scope block, Sum block, Constant block, Product/minblock, To Workspace block and Integrator block toFig.3 Results of simulationthe batchgrinding model,LI Xia et alComputer Simulation of Batch Grinding Process Based On Simulink 5.0Step 3: Select all the blocks in the batchgrindingDuring the course of startup of Simulink, Simu-model, and then use the command Create subsystemlink will search every file named slblocks in MAT-in the shortcut menu to unite all the blocks into aLAB search path, and then bring blocksets defined insubsystem.files named slblocks into Simulink.Step 4: Double click the subsystem, and thenStep 6: Use the menu command MATLAB/delete all the input and output ports (Fig.4).ile/Set Path/Add Folder in MATLAB to add the pathd:matlab6p5\worklmyblksetdemo to the MATLABsearch path.Step 7: Close Simulink Library Brower, andthen reopen it. Batch grinding blockset has appearedin Simulink Library Brower. Now, the user can usethe blocks in batch grinding blockset as other blocksBalchgrindingin normal blockset.Fig.4 Batch grinding blockset masked6 ConclusionsStep 5: Close the batchgrinding model, and thenwrite an M file named slblocks. The content it1) The simulation results of batch grindingslblocks is as follows:process explain reasonably the impersonal rule offunction blkStruct = slblocksbatch grinding.Browser( 1).Library = batchgrinding' ;2) On the basis of batch grinding simulation, it%define the filename of blockset ( without theis feasible to investigate deeply the computer simula-extended name mdl)tion of mineral processing system of grinding andBrowser(1).Name = 'batch gringing Blockset' ;classification, comminution, etc. and the application% define the name of blockset displayed inof MATLAB/Simulink to mineral processing simula-Simulink Library Browsertion is of a very lightful future.References[1] Wang M R. Modeling and Simulation Dynamically with Simulink4.Bejing:Publishing House of Electronics Industry,2002. l II. (In Chinese)[2] Lynch A J. The past, present and future of mineral processing simulation. Metallic Ore Dressing Abroad, 2000.9(18): 2-6.[3] Munoz C, Cipriano A. An integrated system for supervision and economic optimal control of mineral processing plants.Minerals Engineering,199,12(6):627- 643.[4] Espig D, Reinsch V. Computer aided grinding circuit optimization utilizing a new mill efficiency curve. Mineral Processing,1996 (44 45): 249 -259.[5] Morrell S, Man Y T. Using modelling and simulation for the design of full scale ball mill circuits. Minerals Engineering,1997, 10(12):1311-1327.[6] Fermandes C, Peres A E C. Optimization of the grinding circuit at arafertil, Brazll. Minerals Engineering,1999,12(8):969- -984.[7] YuJQ, Xi A M, FuJ H. The application of fuzzy adaptive learning control(FALCON) in milling casification operationsystem. Jouranl of Xi an University of Architecture & Technology, 2000, 32(2):175-178. (In Chinese)[8] Wang H Q, Zhang S Y. Application of predictive fuzzy control in gr中国煤化工: and Meallurgical En-gineering, 2002, 22(3):60- 62. (In Chinese)YHCNMHG[9] Evertsson C M. Modelling of flow in cone crushers. Minerals Engineering, 1999, 12(12):1479- 1499.[10] Ferreira J P, Loveday B K. An improved model for simulation of flotation circuits. Minerals Engineering, 2000,13(14 -15):1441-1453..

论文截图
版权:如无特殊注明,文章转载自网络,侵权请联系cnmhg168#163.com删除!文件均为网友上传,仅供研究和学习使用,务必24小时内删除。