Research On Process-Oriented Intelligent Collaborative Product Design

To shorten the time-to-market of the product, an intelligent computer aided product design tool should be built to provide an integrated product design perspective for designers, and to improve the efficiency of design decision–making and collaboration. In this thesis, the research on the intelligent collaborative design is detailed introduced.Firstly, based on product design modularization and parameterization, an engineering product design process model is presented with the foundation of the available design theories and design paradigms. In the model, several design entities and their relations are formally described, while two kinds of design tasks (activities) e.g. design-decomposition tasks and design-reasoning tasks are analyzed in detail. In view of the serious influence of the design repetition on the product design process, two methods are introduced to reduce these repetitions according to their specific reasons, and then the process planning procedure is described. On the basis of the process model, a quantitative method for computing the dependency degree between two design tasks is given, and the formula for computing design time of different task connecting patterns are derived, so the total time for finishing the whole product design process can be estimated.Next, with the help of the above product design process model, a process-oriented intelligent collaborative product design framework is proposed. In this collaborative design framework, the product design process is divided into several module tasks and part tasks. The module task is solved by the algorithm-based design paradigm, while the part task is solved by the ASE (Analysis-Synthesis-Evaluation) design paradigm. To solve the module design task, basic decomposition actions such as parameter distribution and inheritance, surrogate decomposition model and multi-level decomposition technique are utilized; while to solve the part design task, a method based on the widely used ASE design paradigm and design parameterization is built, the method can adapt well to many different part design tasks.Based on the above design framework, a computer–aided design system is constructed and implemented. The system is composed of some nodes. Each node consists of a Web Server, a Module Server, a Mobile Agent Engine, Tool Agent Server(s), Module Task Agent(s), and Part Task Agent(s). In a Module Task Agent, the genetic algorithm is integrated to decompose the design task in a multi-objective way, different solutions can be provided for designers to select, while in a Part Task Agent, the working mechanism of expert systems is incorporated to strengthen its reasoning ability and adaptability. To check the support of the Agent system for the product design, a coaxial gear reducer and a transimission are taken as examples to demonstrate its functions. It can be illustrated that the system can reduce designers' workload and increase the efficiency of design decision-making and collaboration. The last part of the thesis is the conclusion of research and anticipation of some future work needed.