A Knowledge-based Synthesis Approach For Conceptual Design Of Multi-disciplinary Signal Systems

As the rapid development of technology and multi-disciplinarydesign requirements of products, designers are facing problems such aslimited knowledge range, knowledge can not be updated, design thinkingset, which bring challenges to design multi-disciplinary productscreatively. Especially signal transformation systems, as they involveknowledge of electrical, mechanical, magnetic, optical, thermal and otherdisciplines, the design problem solve process need to explore the broadmulti-disciplinary solution space, thus high demands are put forward fordesigners. In this regard, this dissertation proposes a conceptual designapproach for multi-disciplinary signal systems based on principle solutionknowledge synthesis. The approach includes establishing the knowledgerepresentation model and knowledge base of multi-disciplinary signaltransformation principle solutions, and developing a correspondingcomputer-aided conceptual design system. With the approach, designerscan break the knowledge barrier between diffenent disciplines, andachieve creative conceptual design of multi-disciplinary signaltransformation products objectively and efficiently.Firstly, according to the characteristics of multi-disciplinary signaltransformation knowledge, the dissertation proposes a unifiedrepresentation approach for multi-disciplinary signal flow by anattribute-value deployment way. Based on the approach, a knowledgerepresentation model for multi-disciplinary signal transformationprinciple solutions is established. The model focuses on the functionalknowledge of principal solution, and uses function, structure and application knowledge represent multi-disciplinary signal transformationknowledge formally. Based on the model, a certain size of principalsolution knowledge base is established by widely collecting signaltransformation principal knowledge from diffenent disciplines, which isused as the knowledge resource for conceptual design.Based on the model, a synthesis approach for multi-disciplinarysignal systems is proposed. The approach uses a solving method based onSense-Selection-Action, which can synthesize principle solutions’functional knowledge according to the desired function, and generateprinciple solution chain as the conceptual solution. A correspondingalgorithm is also developed to achieve automated synthesizing process.With the multi-disciplinary signal systems synthesis approach, acomputer-aided signal transformation conceputual design system(abbreviated as CASTCDS) is developed. CASTCDS is a Web systembased on ASP.NET platform, with SQL Server2008database to representmulti-disciplinary signal transformation principal solution knowledgebase formally. CASTCDS supports remote knowledge sharing andreusing, and can efficiently use existing multi-disciplinary knowledge forconceptual design synthesis according to the desired function, whicheffectively solve the problems designers facing when carrying outmulti-disciplinary signal system conceptual design.At last, the conceptual design of engine rotate speed sensor is used asan example to show the systhesis process and operation method ofCASTCDS. The result analysis demonstrates the systhesis approach andCASTCDS can synthsize multi-disciplinary signal transformationprinciple knowledge，and the conceptual solutions generated have certaininnovation and high reference value.