Machining process model based on virtual reality environment

Vol.12 Suppl: 2J. CENT. SOUTH UNIV. TECHNOL.Oct.2005Article ID: 1005 - 9784(2005)S2 -0034 - 07Machining process model basedon virtual reality environment"WANG Tairyong(王太勇)'，WANG Wen-jin(汪文津) , WANG Wen-ying(汪文颖)2，FAN Sheng bo(范胜波)', LUO Jun(罗琊)'(1. School of Mechanical Engineering, Tianjin University, Tianjin 300072, China;2. College of Mechanical Engineering, Tianjin Universityof Science and Technology, Tianjin 300222, China)Abstract: Virtual manufacturing is fast becoming an affordable technology with wide ranging applications in mod-ern manufacturing. Its advantages over existing technology are primarily that users can visualize,el involvementand interact with virtual representations of real world activities in real time. In this paper, a virtual cutting system isbuilt which can simulate turning process, estimate tool wear and cutting force using artificial neural network etc. U-sing the simulated machining environment in virtual reality (VR), the user can practise and preview the operationsfor possible problems that might occur during implementation. This approach enables designers to evaluate and de-sign feasible machining processes in a consistent manner as early as possible during the development process.Key words: virtual manufacturing environment; NC graphics verification; artificial neural network; tool wear esti-mation model; cutting force prediction modelCLC number: TG519. 1; TP391.73Document code: A1 INTRODUCTIONmanufacturing planning phase is finished simulta-neously with the design-planning phase, whichIn recent years, computers have increasinglyleads to reduce development time[1].been integrated into various production facilities inTraditional manual process planning involvesindustry. Computer-aided design (CAD)，comput-several steps. The first step is the interpretation ofer-aided engineering (CAE), computer- automatedthe design data which are usually displayed blue-process planning ( CAPP )，and computer- aidedprints, or CAD system. The second step is the se-manufacturing (CAM) are now commonly used.lection of manufacturing operations and suitableAdopting advanced manufacturing technology hasmachines. The third step is the determination ofbecome necessity if firms have to maintain theiroperation sequences. The forth step is the selectioncompetitive advantage.Virtual manufacturingof clamping devices and the orientation of the cut-method have been recently considered as essentialting tools. Finally, the overall machining time andways to further improve product quality and shrinknonmachining time are calculated and processthe time to bring a product to market. Virtualsheets, operation sheets and route sheets are pre-manufacturing method is the term used to stresspared.that design phase is developed concurrently withVirtual manufacturing method change tradi-planning for other phase such as planning of thetional sequence design process. V irtual manufac-manufacturing phase. Geometry simulation andturing technology is a rapidly developing computerphysical simulation has been integrated into virtualinterface that strives to immerse the user complete-manufacturing model so as to estimate the result ofly within an experimental simulation， therebymachining. The benefits of virtual developmentgreatly enhancing the overall impact and providinga much more intuitive link between computer andmethod include the relative ease of altering designuser2. The philosophy of the virtual manufactur-decisions, since it is the least costly time to makeing is to consider all aspects of product during thedesign changes, with the least impact on manufac-early stage of design, in order to avoid the cost andturing investment. Another advantage is that thetime consuming activities downstream associated中国煤化工D Foundation item: Projects (50475117, 50175081) supported by the NationYHC N M H Gna; projet (33181611)supported by the Science and Technology Comission of Tianjin MunicipalReceived date; 2005 -07-15; Acepte date; 2005 -08 - 16Correspondence: WANG Wen-jin, PhD; E mail: wangeiin71@ yahoo. com. cnWANG Tair-yong, et al; Machining process model based on virtual reality environmentwith traditional design and manufacturing process.to utilize human sense and interaction to engineer-It requires quick information exchange so that de-ing applications in a virtual environment. A num-sign engineers can be aware of the manufacturingber of researchers have employed VM technologyapproaches and be able to select the design withto deal with various aspects and issues in designthe lowest manufacturing cost. Engineers may es-and manufacturing. Schulz et al410] presented theirtimate quality of virtual product using all kinds ofanalysis method in virtual manufacturing environ-research in developing a virtual environment foment. Therefore, virtual manufacturing model iscar-body engineering applications by VRML scenesaiming to remove the wall and enhance the real-uising the Internet as communication platform fortime communication between design and manufac-analyzing engineering simulation. The use of VMturing engineers. Over the years, there is a lot ofto support industrial applications is rapidly increas-active research in both conceptual constructive vir-ing[i5,1l0]. BMW uses VM to explore the visualiza-tual manufacturing systems. WANG et a[43] pres-tion of new car layouts to enhance communicationented an immerse virtual turning system based onand the simulation of processes, while Rolls Royceglobal illumination model. LI et a1C4] developed anAeroengines and associates use VM technology inobject oriented prototype system based on virtualheir early verication of maintenance procedures,reality for maintaining training. At Dalian RailwayInstitute, a NC lathe machining simulation systemtraining maintenance engineers in these proce-has been developed[5.6].dures, enhanced visualization of complex environ-Physical simulation of machining processments and rapid prototyping of plant control andwhich aimed at disclosing physical characteristic ofinstrumentation[172machining process is focused by researchers. Ow-An engineer can design plan, verify and rede-ing to machining parameters and cutting dynamicsign the object and the machining process usingcharacteristics are considered, it can accurately im-virtual reality technology. When the design is sat-itate real machining process so as to avoid expen-isfactory, the NC program of part may then besive cutting experiments. Recently, Cutting forcetransferred into the CNC machines to generate aprediction model and tool wear prediction modelreal prototype.have been studiedC-9. There are some researchIn short, VM technology is seen as an effec-work reported in either geometric modeling of thetive tool which can provide flexible visualizationprocessC1o] or physical modeling11-3]，however, litand simulation, and provide not only improved de-tle work has been performed on the integration ofsign and machining process but also training of de-the two in such a way that can be applied to a widesigners and workforce-18].range of machining process.The goal of this research is the development of3 GENERAL STRUCTURE OF VIRTUAL MA-.an approach that integrates the decisions of the de-sign and the manufacturing process phase in a vir-CHINING PROCESS MODELtual manufacturing environment to create virtualmachining modelof product. Therefore, cuttingA virtual machining model was construcforce prediction model and tool wear estimationted based on the virtual reality technique (shown inmodel were integrated into virtual machiningFig. 1).The first step to run the system is to load NCprocess. On the basis of physical simulation mod-program of part. The second step is to select theel, the integration modeling system of virtual ma-virtual machine tool, virtual clamp, virtual cutterchining simulation is developed and the integrationand virtual workpiece. An interactive process al-of real machining and virtual machining based onlows the user to either select machine tools from aagile manufacturing is realized.menu with the mouse button, or type or to selectan input from data files. The position of the ma-2 APPLICATION OF VMchine and workpiece can be moved or rotated sothat the user' s viewpoint can be positioned close tothe-; to simulate machi-VM technology has shown its advantages overni中国煤化Iufacturing eniron-conventional simulation processes where it can aidmC N M H Gistics of machiningthe understanding of sophisticated engineeringprocess and the dimensions ot the part can be inspectedproblems. A VM system can provide not only basicusing the system display menu during operations. Thecomputerized functions, but also an efficient wayfinally step is to evaluate the machining result so as toJournal CSUT Vol. 12 Suppl .2 2005(VMS). The construction of machining system| Product modelmodel is shown in Fig.2. In the VMS, machiningsystem including machine tool, cutter, and fixturePart programmust be modeled. Machining system models aredesigned according to the structure and specifica-tions of real machines. In short, their geometricalMachineMaterialmodels can also be modeled by measuring the realtoolclampCuttingmachining system.simulaticSimulation configuration| Modifica-database,一tion.GeometryPhysicalsimulationL simulation( A ) Horizontal bedVR machining envirormertTool post system) Moving axisNor EvaluationCarriageChuck↓PassCross slideMachining) WorkpiceTool pooFig. 1 Structure of virtual machining modelmodify the existing NC program.The system allows the user to create a 3Doriginalworkblank, simulate the machining processbased on NC program, and pass the code to CNCFig.2 Geometrical model of CNCmachines in the real world to create the real parts.machining systemThe machining process of a part can be viewedfrom different directions in a 3D simulation4.2 Virtual cutter modelprocess, so that the user can easily find failures orDuring the machinig process, the cutter isweakness in NC code planning. Therefore, it is adriven by the machinig tool to remove the materialuseful tool for either training or trial run purposesfrom the raw stock. Its geometric characteristicsin order to detect any errors that might exist in NChave direct influence on the machining accuracyprogram.and surface quality. In order to model a cutter,some aspects may be considered such as geometric4 VIRTUAL MACHINING MODELdimensions , microtopography of cutting edge, etc.In virtual machining process, virtual interme-The geometrical model of cutter is shown in Fig. 3.diate product takes the form of virtual raw stockbefore virtual machining is machined on virtual ma-chine driven by NC program. The behavior of vir-tual machine and virtual intermediate product arefeedback to user visually by animation, and usercontrols the machining process via the simulationinterface. During the virtual machining process,the geometric model of virtual intermediate work-piece is modified and refreshed in real- time underthe assumption of ideal machining conditions.Fig.3 Cutter modelMeanwhile, the collision and interference betweencutter, workpiece, fixture, or other parts of ma-chine tool are checked. At last, the evaluation andl.3 Virtual material removal algorithm modelsuggestion reports of process and machining condi-Aimed at feature of turning operation, a real-tions are fedback to user.time中国煤化工I that is based onfeatuimulating material4.1 Geometrical model of virtual CNC machiningremoYHCNMHGhispaper.Allin-systemterpolation types are based on line interpolationCNC machining system modeling is an impor-and circular interpolation in turning process.tant step to establish the virtual machining systemTherefore, complex product module can be ob-WANG Tai-yong, et al; Machining process model based on virtual reality environment●37.tained combining cylinder unit in the form of allF路+臣where F。 is cutting force, F, is thrustkinds of size. Virtual material removal algorithmFmodel is shown in Fig. 4.force, Fi is feed force).Experiment conditions are listed in Table 1.Fig. 6 shows flank wear prediction capability of thetrained neural network. The results of computersimulations are in good agreement with that of toolwear experiments.(2)Input layer，Cutting speed via↑Output layer(b)Cutting depth apofTool wearTimeT- dForce ratio F.d(CMiddle layerFig. 4 Workpiece obtained through featureInformnation streammodeling during turning process(a)-Workblank; (b)- -Feature modeling in tuming process;Fig.5 Neural network for tool wear(c) - -Finished workpieceTable 1 Cutting condition5 PHYSICAL SIMULATION MODELMachineImproved CA616 CNC Lathe5.1 Tool wear prediction model based on artificialWorkpiece material45 Carbon steelneural networkHardnessHB187Tool wear measurement can be obtainedTool materialCarbide to(YT15)through indirect measurement method. Tool wearBack-rake angle/<*)12affects the cutting force is well known. Tool wearSide relief angle/<*)7sensing using cutting force signals is obtained fromPlan approach or entering angle/()9dynamometers during turning process. In this pa-per, a method to predict flank wear from cuttingAuxiliary angle/<)gforce signals using a neural network technique isSlope angle/<*)0presented. A neural network approach is employedNose radius/ mm0.02in order to cope with the stochastic characteristicCutting fluidnot usedof the cutting forces as well as the varied cuttingconditions. The strategy is to develop a neural net-5.2 Cutting force model based on artificial neuralwork model to predict flank wear from force rationetworkinformation obtained from a tool dynamometer.Cutting force is a basic parameter and it di-The neural network used in this work is based on arectly influences relative displacement between toolmultilayer feed- forward model which consists of in-and workpiece, tool wear and surface quality input, hidden, and outputs layers, and error-back-turning operation. Therefore, simulation model ofpropagation algorithm for learning is used1920.cutting force is an important part of machiningphysi中国煤化工is paper, changeThe structure of BPNN is shown in Fig. 5. Thevalueartificial neuralneural networks were trained with following pa-netwqYHc N M H Gof turning exper-rameters: cutting speed U, feed rate F, cuttingiments. Cutting process can be completely de-depth ap, time t, cutting force ratio FR ( FR = scribed by using the model, which is a research●38.Journal CSUT Vol. 12 Suppl 2 20050.20300z2500.12-c 200●一Test value兰0.08一Simulation value150- - Simulation valuev=191.5 m/min0.04-F=75 mm/min一Test valuev=130 m/minv=75 mm/min10a=2 mmCutting time/min102030405060Fig.6 Comparison of measured andTime/minpredicted flank wearFig. 9 Comparison of measured andplatform for constructing other models which therepredicted thrust force F,are close relationship with cutting force.The structure of BPNN is shown in Fig. 7.T he neural networks were trained with following1150parameters: cutting speed U, feed speed U, cuttingdepth ap, cutting force F,，thrust force F, feed1100-force F.Experimental conditions are listed in Table 1.1050Figs. 8 - 10 show cutting force prediction capabilityof the trained neural network. The results of com-望1000>士<一Simulation value :◆- Test valueputer simulations are in good agreement with thatof dynamic cutting force experiments.ap=2 mm900)2030405060Error infomation,Fig. 10 Comparison of measured andCutting speedpredicted cutting force F,0- -Feed force FxFeed speed-Thrust force Fy5.3 Production time modelCutting deptht=t+te●tm/T+t.(1)0- Cutting force FzTool wear___where a is the machining time; ta is the changingOutput layerInput layercutter time; T is the tool life; t。 is the subsidiarytime; t is production time.Middle layerInformation stream5.4 Production cost modelFig. 7 Neural network for cutting forceC=tm●C.+te●Ce●tm/T+Ce●tm/T+t。●C。(2)where C。is the production cost per second, Ce is700the cutter cost, c is production cost.子5006 APPLICATION8 400-●一Simulation valueCutting force model and tool wear model are◆一Test value豆200integrated into geometry machining simulation sys-100-a。=2 mmtem so as to realize the integration of geometry0~10 203040一5060simu中国煤化工ion. The virtualmacher to interact withtherMYHcNMHGmetime,usercenFig. 8 Comparison of measured and predictedobserve the dynamic characteristics of turningfeed force F:process and evaluate cutting result. Figs. 11 - 13WANG Tar-yong, et al; Machining process model based on virtual reality environment●39●show the turning simulation process.eled and verified before they can be actually imple-mented. Virtual manufacturing environment offersthe user's new ways to not only visualize theirproblems but also to interact with the environmenteffectively and efficiently. These visualizations,combined with interaction can improve the deci-sion- making capabilities of engineers thereby im-proving quality and reducing the development timefor new products. The following conclusions canbe drawn.1) Virtual machining system can verify NCprograms and ensure the safety of machines. Dur-Fig.11 Turning geometry simulation processing virtual manufacturing environment, machiningprocess can be predicted and interference and colli-sion are detected to ensure the reliability of NCprograms and safety of machines.↑5:802) Virtual machining system can predict the三tool wear and cutting force. The user can observedynamic characteristics of turning process in virtu-al turning environment. Cutting conditions can al-so be optimized according to simulation results.9E0e4.61913) The users can practise all operations, pro-co8cedures and skills in the virtual machining environ-ment.Fig.12 Result of turning simulationREFERENCES[1] PAN Jun, MA Deng-hua. JIANG Zu-hua.Researchon virtual product development and its simulation mod-el [J]. Computer Integrated Manufacturing System,2002, 8(9): 684 - 689.[2] Mujber T s, Secsi T, Hashmi M s J. Virtual realityapplications in manufacturing process simulation J].Journal of Materials Processing Technology 2004,日0155: 1834 - 1838.[3] WangT Y, WangGF, Li H w, et al. Constructionof a realstic scene in virtual turning based on a globalillumination model and chip simulation [J]. Journal ofFig. 13 Geometry simulation and physicalMaterials Processing Technology, 2002, 129(1): 524simulation in virtual turning environment- 528.(v= 155 m/min; V1 = 180 mm/ min;[4] LiJ R, Khoo LP, Tor s B. Desktop virtual reality forap =1.35 mm)maintenance training: An object oriented prototypesystem ( V-REALISM) [J]. Computers in Industry,7 CONCLUSIONS2003, 52(2): 109 - 125.[5] XU Li, YUAN Bin, SHI Zhi-bui. Study and realia-Virtual machining system can be a powerfultion of NC lathe machining simulation system UJ].Journal of Dalian Railway Institute, 2002, 23(2): 37 -tool for testing and evaluating new parts, decrea-39. (in Chinese)sing the time to market and reducing the productcost. Virtual machining environment has been ap-中国煤化工，svirnment technologyplied to solve client' s real-world problems, andCNMH Ghanical Enincrinehas increased profitability, decreased the lead timeCHuwL，1J0/t uIL- uiu. IL uuunese)to market, increase the worker safety. Manufac-[7]Chungchoo C, Saini D. A computer algorithm forturing processes and design can be defined, mod-flank and crater wear estimation in CNC turning opera-Journal CSUT Vol. 12 Suppl 2 2005tions [J]. International Journal of Machine Tools &[14] Schulz M, Reuding T, Ertl T. Analysing engineeringManufacture, 2002, 42(13); 1465- 1477.simulations in a virtual environment UJ]. IEEE Com-[8] Yen Y C, SohnerJ, Lilly B, et al. Estimation of toolputer Graphics Appl, 1998, 18(6): 46-51.wear in orthogonal cutting using the finite element[15] WilsonJ R, Cobb s, D'Cruz M, et al. Virtual realityanalysis[J]. Journal of Materials Processing Technolo-for industrial application; opportunities and limita-gy, 2004, 146(1): 82 -91.tions [M]. Nottingham: Nottingham University[9] Orel T, Nadir A. Prediction of flank wear by usingPress, 1996.back propagation neural network modeling when cut[16]Rosenblum L J, Macedonia M R. Projects in VRting hardened H-13 steel with chamfered and honed[J]. IEEE Computer Graphics Appl, 1998, 18(6): 6CBN tools [J]. International Journal of Machine Tools9.& Manufacture, 2002, 42(2): 287 - 297.[17] Peng Q, HallF R, Lister P M. Application and eval-[10] Tost D. Puig A, Pe' rez-Vidal L. Boolean operationsuation of VR-based CAPP system [J]. Journal offor 3D simulation of CNC machining of driling toolsMaterials Processing Technology, 2000, 107(1): 153[J]. Computer-aided Design, 2004, 36(4): 315 -- 159.323.[18] YAO Ying-xue, LI Jianr-guang, Lee W B. VMMC: atest-bed for machining [J]. Computers in Industry,tion of chatter in milling using a predictive force mod-2002, 47(3); 255 - 268.el [J]. International Journal of Machine Tools &[19] Sikdar S K, CHEN Ming yuan. Relationship betweenManufacture, 2000, 40(14) :2047 - 2071.tool flank wear area and component forces in single[12] YangJ A, Jaganathan v, DU Ru-xu. A new dynamicpoint turning [J]. Journal of Materials Processingmodel for driling and reaming processes [J]. Interna-Technology, 2002, 128(1): 210 - 215.tional Journal of Machine Tools & Manufacture,[20] JIAO Li-cheng. Implementation and application of2002, 42(2); 299-311.neural network [M]. Xi' an: XIDIAN University[13] LeeK Y, Kang MC, Jecong Y H, et al. Simulationof surface roughness and profile in high-speed end(Edited by CHEN Can-hua)milling [UJ]. Journal of Materials Processing Technol-ogy, 2001, 113(1); 410-415. .中国煤化工MYHCNMHG