Knowledge Reuse Technologies And Its Application For Laege Scaled Hydraulic Turbine Scheme Design

Knowledge reuse, a key technology of complex mechanism scheme design, has aroused more and more abroad attention by theory circle. Under the support of the key technologies R&D program of Heilongjiang province (Grant No. WH05A01) and the application background of large scaled hydraulic turbine scheme design of Harbin Electric Machinery LTD., several theories and technologies of knowledge reuse for complex mechanism scheme design, such as design requirement analysis and function translation, the methods of knowledge mining and design process reuse, multi-hierarchy case decomposition, multi-hierarchy case-based reasoning and design scheme evaluation and optimization, are discussed in this dissertation.As for the design requirement of large scaled hydraulic turbine scheme design, the inapplicability of conventional Quality Function Deploy (QFD) to complex mechanism scheme design is analyzed. Then, the diversified design requirement information is formally expressed by building a basic element model, the decomposition process and the hierarchies of the design requirement are expressed based on multilevel basic element implication tree. After that, the conventional House of Quality (HOQ) is improved from several sides, such as basic element model of design requirement, translation between multilevel function characteristics and multilevel structure characteristics, weight of basic element of design requirement, weight of function characteristic and structure characteristic and evaluation method of design scheme. Eventually, the basic element of the design requirement is translated into the diversified design information of scheme design based on improved House of Quality.Due to various complicated design problems in the process of scheme design, the methods of knowledge mining and design process reuse based on basic element are researched. The deep-seated knowledge of complex mechanism scheme design is processed based on the basic element of design knowledge, several methods of knowledge mining, such as extension data mining method, association rule mining method and association rule extension method, are presented. The design process implication reuse is achieved based on those implication information, including association rules and extension rules, which are attained by the methods of knowledge mining. And then, by analyzing the characteristic and characteristic quantity of the target basic element and the extension basic element synthetically, the computation model of extension degree of knowledge reuse and the searching and matching method of basic element extension reuse based on the extension of basic element are put forward. Then, the hierarchical knowledge and the construction of knowledge base of basic element are accomplished. Based on these, the design process extension reuse model is presented and the design process extension reuse is achieved.After that, the methods of multi-hierarchy case reuse for large scaled hydraulic turbine scheme design are discussed, including multi-hierarchy case decomposition and multi-hierarchy case-based reasoning. The decomposition granularity, the decomposition sequence and the decomposition mapping relation among coterminous domains of multi-hierarchy case are studied based on axiomatic design, the judgment method of multi-hierarchy case decomposition is proposed, and then the process model and the reuse model of multi-hierarchy case are developed. Based on multi-hierarchy case decomposition, the expression model of multi-hierarchy case class is presented, and the corresponding indexing model of multi-hierarchy case class and matching model of design attribute are proposed based on the subspace method. Through analyzing the process of multi-hierarchy case reasoning (MCBR), the model of MCBR is developed; the optimum multi-hierarchy case and the corresponding optimum combined multi-hierarchy case are achieved.In order to deal with and use the incomplete and the uncertain design information in the procedure of the complex mechanism scheme design, the evaluation and optimization problem of the complex mechanism scheme design is studied. By analyzing the multi-attribute and uncertain interval attribute evaluation methods, an improved multi-attribute gray interval relevant evaluation model is put forward. On this condition, an improved multi-attribute grey fuzzy evaluation model is presented based on the grey system theory and the fuzzy theory. Based on multi-attribute gray interval relevant evaluation model, with the completeness and integrity of the design information, a multi-attribute fuzzy optimization model based on optimal membership degree and the information axiom of axiomatic design is presented from fuzzy information value side. Aiming at the different design schemes and evaluation indexes, the rational evaluation and optimization model is selected to process the uncertain interval attribute information and the optimum design scheme is achieved.Finally, based on the former theories and technologies, HEC scheme design support system (HEC-CDSS) is developed, which includes modules: type scheme design, structure scheme design and design scheme optimization. Though the application case of large scaled hydraulic turbinescheme desig of Harbin Electric Machinery LTD., the efficiency and feasibility of the technology, theory and system is validated.