Research On A Methodology Of Knowledge Acquisition For Computer-aided Process Innovation

In the increasingly intensive competition of global market, manufacturing companies pay more attention to innovation and the knowledge that enables innovation. Computer-Aided Process Innovation(CAPI), which can stimulate creative thinking of process designers and assist them to implement systematic process innovation, is of great significance for reducing manufacturing costs, improving product quality and enhancing competitive edge. Process innovation design is a structured innovative implementation process based on knowledge, thus rational, efficient and formalized knowledge acquisition is the prerequisite and basis for CAPI. However, the existing knowledge of innovative knowledge sources is not organized according to CAPI, thus it can not be applied directly to process problems solving. Aiming at these problems of knowledge acquisition for knowledge-driven CAPI, based on the analysis of knowledge needs in the innovation design, a knowledge acquisition technology architecture for CAPI is established. Then several key technologies, knowledge accumulation based on bilayer social wiki network, process contradiction matrix construction based on process patent mining and knowledge evaluation using fuzzy linguistic computing, are systematically explored. The main research work and contributions are as follows:(1) A technology architecture of process innovation knowledge acquisition for CAPI is constructed. The concept, connotation and framework of knowledge-driven CAPI are proposed. Then a detailed study is performed for the sources, classifications and contents of process innovation knowledge(PIK). On this basis, by analyzing the technology needs for knowledge acquisition in process innovation design, a PIK acquisition technology architecture is built. This chapter provides a general framework for the subsequent study of the theories and methods.(2) A PIK accumulation approach based on bilayer social wiki network is proposed. By combining the technical characteristics of social network with wiki,a novel PIK accumulation schema based on bilayer social wiki network is introduced. Aiming at the multi-source knowledge fusion, the process innovation knowledge fusion algorithm based on semantic elements reconfiguration and corresponding conflict resolution rules are raised. To ensure the credibility and orderliness of knowledge refinement, credible groups based on knowledge social trust degree are built, and collaborative refinement control processes and corresponding editing rules are given. In addition, a collaborative control process for PIK holistic refinement and collaborative evolution is explored. The study of this chapter realizes formalized PIK accumulation based on collective intelligence.(3) A construction approach to process contradiction matrix based on process patent mining is studied. Firstly, a presentation model of process contradiction matrix is built, and the detailed construction process of process contradiction matrix is presented. On this basis, an automatic classification method of process patents is studied by taking process method, manufacturing object and manufacturing feature as references. A mining method of process contradiction parameters based on typical semantic patterns is proposed, and a mining method of process contradiction solving principles based on SAO(Subject-Action-Object) is proposed. The study of this chapter realizes the reasonable mapping between process patents and process contradiction matrix.(4) A knowledge evaluation method of process innovation using fuzzy linguistic computing is introduced. Through analyzing evaluation requirements and process of PIK in an open environment, a comprehensive evaluation index system including objective level, criteria level and subcriteria level is designed. Then a PIK evaluation algorithm using 2-tuple fuzzy linguistic computing is proposed. To obtain more reliable criteria weights of evaluation system, a multilevel fuzzy comprehensive weights determination method is introduced by combining fuzzy linguistic computing weights and AHP(Analytic Hierarchy Process) weights. The study of this chapter realizes that the experts’ evaluation information for criteria weights and candidate knowledge can be effectively aggregated.(5) Based on the above research, a prototype system of PIK acquisition for CAPI, named pi Pioneer-PIKAS, is designed and developed. According to applications on PIK acquisition of micro-cutting and process problem solving of a micro-turbine, the proposed methods and technologies of this dissertation are validated.