Research On Methodology Of Extension Knowledge Evolution For Product Low-Carbon Design

Low-carbon design,driven by the low-carbon demand,coordinates the relations among carbon footprint(CFP),cost(C),and performance(P)in each phase of product life cycle from the beginning of design activity.It belongs to sustainable design domain,which is a complicated backtracking design process involving large amount of tacit knowledge to be discovered.Extension knowledge is used to solve contradictions in complicated design process.The purpose of researching on the methodology of extension knowledge evolution for product low-carbon design is to takes advantages of the Extenics theory to explore the techniques of extension knowledge modeling,knowledge reuse modeling and evolution algorithm,which provide a reasoning and evolutionary tool for reuse and derivation of extension knowledge.The main research contents of the thesis are as follows:(1)For the knowledge modeling in low-carbon design,an extension knowledge modeling method is proposed for product life cycle.Based on the framework of PAS2050,estimation modeling for CFP and C in product life cycle is built.The amount of CFP and C of a product is allocated to components and parts with the activity-based method,which fill up the low-carbon data for case base.The product level granularity is expanded,knowledge representation modeling for product cases is built with the basic element method,and the hierarchical low-carbon case base is established.The model of extension knowledge is built,which includes extension reasoning,transformation,and extension set knowledge.Representation modeling of extension knowledge is constructed by the element of Know_What,Know_How,Know_Why,which are packed with basic element model.With the proposed method,extension knowledge can be represented formally,shared freely.(2)For the utility problem in retrieving and adaptation difficulty for similar cases,a retrieval technique is proposed based on improved index method and adaptation information.Index method is introduced for utility problem,as the weaknesses of discretized-highest similarity method,the equal probability mesh model is built to make the uniform distribution for product cases in each interval of attributes;the rule of matching status is proposed to overcome the neglect of similar cases when the required cases locate in the corner of the cross intervals.For different data style of attributes,local similarity algorithms are built respectively;information content model is constructed based on the fuzzy AD theory,which is taken as the criterion for adaptation.Thus,it forms two criteria for similar cases retrieval.(3)For extension knowledge acquisition and reuse in low-carbon design,reuse modeling for extension knowledge is constructed.The reuse modeling for extension reasoning knowledge based on R-FBS is built;the process reasoning knowledge of function analysis,behavior model and evaluation,structure synthesis and analysis is studied.Transformation algorithm for structure reconstruction is established;transformations are performed on the value of attributes and attributes of module element,and module element types to discover evolutionary laws,and the adaptation case base is constructed.(4)Integration innovative method combining TRIZ and Extenics is proposed to solve low-carbon design problems and produce derivation extension knowledge.Integration framework is set up,contradiction problems can be described qualitatively and quantitatively with basic element model and dependent function;invention principles are extracted using analysis tools and knowledge base in TRIZ;invention principles are converted into detailed technical schemes by transformation operations in Extenics;strategies of extension knowledge derivation for technical contradiction and antithetical problem,physical contradiction and incompatible problem are constructed.(5)The evolutionary algorithm modeling based on multi-dimensional dependent function is proposed to solve the combined explosion problem of design schemes.Correlations among transformation parts are analyzed,the estimation models of CFP and C are simplified,performance constraint model is set,and fitness function is constructed.The mathematic model of multi-dimensional dependent function is established and its reduction rule is deduced.A second-choice mechanism for genetic operation based on the multi-dimensional dependent function is constructed innovatively.A detailed optimization model with GA for low-carbon design is proposed to produce the optimal design scheme.Based on the above theoretical research,an extension knowledge evolution system for low-carbon design is developed.The vacuum pump is taken as an example to testify the low-carbon case base construction technique based on extension model,similar cases retrieval technique,knowledge reuse technique of extension reasoning and transformation,knowledge reuse technique of adaptation cases,extension knowledge derivation technique by integrating TRIZ and Extenics,evolutionary algorithm technique based on multi-dimensional dependent function,and thus it demonstrates the feasibility of the above theoretical research.