Research On Fusion Optimization Model For Barrel Finishing Process Parameters

Process intelligent decision-making is the key condition of manufacturing transformation and upgrading.Therefore,this paper takes the barrel finishing process as the research object,and studies the intelligent process decision-making system on the database platform established by the research group.Currently in the barrel finishing process,the specific features and machining requirements of parts,are based on the expert experience in a large number of trials to choose suitable technological scheme,because there are many different kinds of parts,and processing requirements of the same parts also variety,process parameter optimization problem of the new complex parts lack of unified standard and effective theoretical guidance,can not meet the requirements of rapid development.In order to make use of case knowledge and expert experience of barrel finishing process database platform,it is necessary to study the intelligent optimization model of process parameters.In the barrel finishing process,the selection of technological parameters such as abrasive media,finishing equipment and grinding fluid has an important influence on the processing effect.Therefore,this paper studies the process parameter optimization model of the barrel finishing process based on case-based reasoning,expert reasoning and transfer learning,and utilizes the case data of database platform from different perspectives.For the same part to be machined,there may be conflicts among the above three optimization results,which increases the difficulty of process selection.Therefore,a process parameter fusion decision model is needed to study.In order to make full use of three kinds of optimization model of the optimization results,at the same time solve the problem of three optimization result of conflict,this paper proposes a fusion decision model based on improved D-S evidence theory of the process parameters of the barrel finishing process,it can make full use of the case information,expert experience and difference information of different types of knowledge,which is obtained by fusion decision-making process more believable.The main contents of this paper are as follows:(1)The principle and process of barrel finishing process and the key factors affecting the effect of barrel finishing process were studied.On this basis,the case characteristics of process parameter decision-making were determined,and the types of commonly used process parameters and the corresponding description methods were summarized and sorted out.Then,an Entity Relationship(E-R)diagram was designed for process parameter decision of barrel finishing process.Finally,a case base for process parameter decision of the barrel finishing process was built.(2)A process parameter optimization model for barrel finishing process was established based on case-based reasoning,expert reasoning and transfer learning.The process parameters include: abrasive media,finishing equipment and grinding fluid;For case-based reasoning,by calculating the similarity between cases,take the process plan of the similar case as the alternative for the new problem.For expert reasoning,firstly,the optimization rule construction of finishing equipment and grinding fluid was studied emphatically,and then the process parameter optimization reasoning machine was designed.For transfer learning,the finishing equipment and the grinding fluid were numerically described by vector and binary values respectively,and then the process parameters were optimized by joint distributed adapt-support vector machine.Finally,the case data in the database platform is used to simulate the three optimization models.(3)Aiming at the possible conflicts among the optimal results of the three models,a process parameter fusion decision model based on improved D-S evidence theory was established.Firstly,the process parameter decision identification framework was dynamically constructed according to the optimal selection results and the basic probability assignment of the three kinds of evidence was determined.Then,aiming at the disadvantages of the original D-S evidence theory in the fusion of high conflict evidence,a D-S evidence theory synthesis method was proposed to allocate conflict information according to the proportion of basic probability assignment.Finally,the optimized results of the three models were fused with the improved D-S evidence theory,and the fused process scheme was obtained.The simulation results show that the improved fusion decision model can better solve the problem of conflicts between the optimization results of different models,and through the fusion of different types of knowledge,the decision results have a higher accuracy than the three methods used alone.(4)Based on the database platform built by the project team,the human-machine interface of process parameter fusion decision system based on improved D-S evidence theory is designed and developed by C# language,which provides decision-making guidance for the process personnel in the barrel finishing process.