基于改进模糊神经网络的铣削加工参数选择

Jun.2013机 床 与 液 压Hydromechatronics Engineering Ve1.41 No.12DOI：10.3969/j.issn.1001-3881.2013.12.010Selection M illing Parameters Based on ImprovedFuzzy Petri NetsQIAO Yinhu ' ，HAN Jiang ，ZHANG Chunyan ，CHEN Jieping2J.School ofMechanical and Automotive Engineering，Hefei University of Technology，Hefei 230009，China；2.Engineering ofMechatronics and Vehicle，Anhui Science and Technology University，Fengyang 233100，China1.IntroductionAbstract：This paper focuses on the selection of machining parameters of miling operation.Acompact selection method based on expeR rules.which wore obtained frOm experimental resultsand extracted from the knowledge of skiled operators.is presented.A machining model is con-structed based on a fuzzy Petri nets.The optimal cutting parameters are subjected to an objec-tive function of maximum production rate or minimum production cost with the constraints of apermissible Iimit of surface roughness.curing force and a feasible range of curing parameters。

The fuzzy logic model has been used in the driling operation to select driling speeds for one kindof material in this paper.The relationship between a given material hardness and driling speedcan be described and evaluated by the fuzzy relation for diferent cuting tool materials and difer-ent hole diameters and feed rates.An application example is presented and solved to ilustratethe efectiveness of the presented algorithms。

Key werds：fuzzy petri nets，eficiency and reliability，curing parameter selectionDue to high capital and machining costs，thereis an economic need to operate these machines as ef-ficiently as possible in order to obtain the requiredpay back. The success of the machining operationdepends on the selection of machining process param-eters.Proper selection of process parameters play asignificant role to ensure quality of product，to reducethe machining cost，to increase productivity in com-puter controlled machiningUSA，only 38% of the toolstool-life capability，it willprocesses[1].In theare used up to their fulbe even worse in otherReceived：2013-O1-18This werk i8 partially supported by National Natural ScienceFoundation of China(61 164012)，The key construction disci-plines of Anhui Science and Technology UniversityfAKXK20102-5)$QIAO Yinhu.E.mail：qiaoyinh### 126.corncountry.This situation urges the need tofor develope-ing more scientific approaches to select cutting toolsand cutting conditions in order to obtain the for opti-mum economic and technological machining perform-ance2].A human process planner selects machi。

ning process parameters according tousing his own ex-perience or from the handbooks.But these parame-ters do not give optimal result.Various optimizationstrategies and algorithms which covers ranging fromelementary numerical search methods to more system-atic approaches by using employing nontraditionaltechniques tofor optimizeation the of process parame-ter had been reported in the literature[3].Selectionof machinability data，which includes to chooseingthe appropriate machining parameters of speed，feedrate and，depth of cut，plays an important role in theeficient utilization of machine tools and thus signifi-cantly in influences the overall manufacturing costs。

In addition to the proper selection of speeds andfeeds，the optimum performance of any machiningoperation depends on variables such as part configu-QIAO Yinhu，et al：Selection Miling Parameters Based on Improved Fuzzy Petri Netsration，condition of the machine，type of fixturing，dimensional tolerance and surface roughness. Be-cause the effects ofthese variables on tool life are notalways precisely well·known，it becomes difficult torecommend optimum conditions for a machining oper-ation.This is a multiple.variable and multiple.con。

straint optimization problem [4].It is also popularfor its ability to develop rule based expert systems。

The skilled human operators employ experimentalrules that can be cast into the fuzzy logic framework。

These observations inspired many investigators towork in this area。leading to the development of theso called fuzzy logic and fuzzy rule based contro1.Sofar，there is no standard method of choosing the prop-er shape of the fuzzy sets of the control variables。

cised.In this work.a fuzzy logle model has been de。

veloped tofor selecting cuting parameters in the drill-ing operation5-7].Based on the simulation soft。

ware of Matlab，this paper developed deals with amethod to extractif then ”rules from a smallset of machining data.This e presented method utili-zes both probabilistic reasoning and fuzzy logical rea-soning to benefit from the machining data and fromthe judgment and preference of a machinist.Machi-ning of materials(miling，i。 。

driling，etc.)is an em-pirical science. Forab out surface roughnexample，a piece of knowledgeess that if、l，t ratio is high thenRa is high”could be extracted from a given set of ex。

perimental data。 and some operational decisionscould be made during the machining process(e.g.，the proper NC codes of S and F could be determinedto ensure the desired perform ance and the controllimits of a control chart for a process could be deter。

2.1.M old of parameters selecting systemThis paper introduces a system for the selectionof milling machining parameters.Expert rules are e-valuated by the fuzzy set theory.As shown in Fig.1，the fuzzy model uses fuzzy-expert rules，triangularmembership functions for fuzzifcation and centroidarea method during the defuzzifcation proceses[8]。

Fuzy Petri nets(FPNs)are special Petri netsthat provide an eficient way to process fuzzy and an-certain knowledge[9].Due to knowledge represen-tation by a visual graphic way and inference by a con-current manner，FPNs are extensively adopted to rea-son in practical applications[10]。

L - -- . -- -- - -- -- -- - - -. -- -- - -- -- -. -- - -- -- -- - -- -- - - -- - - -- -- -- - - --Fig.1 Block diagram of fuzzifcation metal-cutting pametem selecting system2.2.The model of cutting speed selecting systemThe selection model of cutting speed is estab-lished as Fig.2. Model of fuzzy inference forcutting speedWith the support of expert rule base，the tool di-ameter d0，cutting depth of cut aw and material harddegree HB are the input parameters and V is the out-put parameter as shown in Fig.2.as input in the upfigure，Vis the output，Since the values of aw and d0have a vague relationship，after intelligenwith fuzzy logic inference model[12]，sot reasoningthe milingHydromechatronics Engineeringerational modal analysis[J].International Journal of Ma-chine Tools& Manufacture，2009，49：947-957。

[2] Jawahir I S，Wang X.Development of hybrid predictivemodels and optimization techniques for machining opera-tions[J].Journal of Materials Processing Technology，2007。185：46-59。

[3] Venkata R R，Pawar P J.Parameter optimization of amulti·-pass milling process using non--traditional optimiza-tion algorithms[J].Applied Soft Computing，2010(10)：445-456。

[4] Ganping Sun，Paul Wright.Simulation-Based CuttingParameter Selection for Ball End Miling[J].Journal ofManufacturing Systems，2005，4(24)：352-365。

[5] Hashmi K，Graham I D，Mills B.Fuzzy logic based dataselection for the driling process[J].Journal of MaterialsProcessing Technolog，2000，108：55-61。

[6] LIU Xiaodong.Novel artifcial inteligent techniques viaAFS theory：Feature selection， concept categorizationand characteristic description[J].Applied Soft Compu-ting，2010，10：793-805。

[7] Hashmia K，E1 Baradieb M A，Ryan M.Fuzzy-logicbased intelligent selection of machining parameters.Jour-nal of Materials Processing Technology ，1999，94：94- 1l1。

[8] UllahA M M S，Khalifa H H.A human.assisted knowl-edge extraction method for machining operations[J].Ad-vanced Engineering Informatics.A1 Ain.United Arab E。

[9] Sun J，Qin s，Song Y.Fault diagnosis of electric powersystems based on fuzzy Petri nets[J].IEEE Transactionson Power Systems，2004，l1(19)：53-59。

[10]YUAN H W，YUAN H B，LI X.Fuzzy Petri nets reason-ing for application of electric control system fault diagno-sis[J].IEEE Conference on Robotics.Automation andMechatronics，2001.12：1-6。

[11]z咖 Jing.Principles and Applications of Fuzzy ControL[J].cHINA MAcHINE PRESS，doi：7111046714。

[12]YUAN J.Improved basic inference models of fuzzy Petrinets[C]//Proc.7th w0rld Con.Inteligent Control andAutomation，200I7：1488-1493。

基于改进模糊神经网络的铣削加工参数选择乔英 ，韩 江 ，张春燕 ，陈杰平1.合肥工业大学 机械与汽车工程学院，合肥 230009；2.安徽科技学院 机电与车辆工程学院，风阳 233100摘要：研究了铣削加工的参数选择。采用基于专家规则和模糊神经网络的推理方法获得优化的切削参数。通过使用优化后的切削参数对切削效率、加工成本和表面光洁度进行优化组合。以-种切削材料为例，考虑其硬度因素，对其切削速度等加工参数进行优化，验证了所提出的参数优化方法的有效性。

关键词：模糊神经网络；效率与可靠性；加工参数选择中图分类号：TH16” - - - - - - - - - - - - - - - - - - - - - - 》 Correction Announcement 》June，2013《 We have misprinted the personal information of the author of the paperNumerical Investigation of the Coling《System Designing for HEV Batery Module Based on STAR-CCM”，which was published in June，2012.We，hereby，correct the personal information of the author as WANG Yaxiong.E-mail：yaxiongwang###ejnu.net”.We，here，express our deep apology for the author。

《 Editorial Department0f r0 ch口 r0n Enginering 《 。 n啪m 月 m 。 阳m∞