Study On Extension Case-Based Reasoning Method For Mechanical Product Redesign Process

Under the trend of current industrial development,the product design stage as a prerequisite for manufacturing is becoming more and more important in order to cope with the challenges of quality assurance,cost control and rapid response in China's manufacturing industry.Among the existing design activities,the redesign process based on the original products has become the mainstream.As an emerging technology in the field of artificial intelligence,case-based reasoning has a broad application prospect,which uses the past solutions and experience of similar problems to obtain the current solutions.However,due to many factors such as the large number of products,high retrieval dimension and the complexity of its internal coupling relationship,there are many problems in the application of case-based reasoning to product redesign,which are embodied in the two core elements of case retrieval and case revise.In this context,the extension case-based reasoning method for the redesign process of mechanical products is proposed,which combines the methods of correlation function,basic element theory,modeling and solving contradiction problems in Extenics with case-based reasoning,aiming at achieving high-quality,low-cost and fast-paced design.The research contents are as follows:(1)Aiming at the contradiction between index accuracy and speed in case retrieval,as well as the problem of retrieval utility,a case retrieval method based on correlation function and D-HS index is proposed.Combining fuzzy number with correlation function,the formula of the fuzzy correlation function is derived to measure the similarity between the fuzzy parameters.Based on the analytic hierarchy process,the method named unidimensional correlation function combination similarity measurement model is established for high-dimensional retrieval to reconcile the contradiction between index accuracy and speed.The product attribute values are transformed non-linearly by using Sigmoid function,which improve the accuracy of traditional D-HS index.On this basis,combines the D-HS index with the correlation function model to avoid the retrieval utility problem.(2)Aiming at the problem of change propagation in case revise,a method of product change risk prediction based on MDP is proposed.With the help of DSM and genetic clustering algorithm,the network diagram of product parts is constructed,then it is modeled as MDP by analyzing the Markov property of change propagation in the network diagram,and the key elements(mainly transfer probability and reward)are quantified based on the methods or indicators such as Boolean DSM numeralization,change propagation index,modification difficulty coefficient,etc.The MDP model is solved by Dynamic Programming or Temporal-Difference,and the risk of change is predicted according to the result.The prediction granularity is refined to the specific characteristics of parts based on composite element,in order to realize the intelligent recognition of Hub feature pairs in the process of product change propagation.(3)Aiming at the problem of design contradiction in case revise,the extension integration model is constructed and the corresponding solution method is given.The extension analysis of design contradiction problem is carried out in order to complete the division of user demand by extension set,obtain the structure to be modified,and point out the shortcomings of incompatibility problems and opposite problem model in the design contradiction problem modeling.On this basis,an extension integration model is established.The extension transformation operator is introduced into the case revise,and its modification difficulty and selection priority are given.The extension integration model is solved by combining the extension strategy generation method and the transformation bridge method.At the same time,the risk of change is introduced for the initial screening and evaluation,and the extension knowledge which involved in this process is extracted to realize the reuse of design knowledge.Based on the above research work,with the help of programming software,database technology,3D modeling software and related hardware equipment,the extension case-based reasoning system of product redesign process is developed,and the vacuum pump redesign process is taken as an example to verify the feasibility and effectiveness of the study.