Key Technology Research On Assembly Planning With Units Partition

With the increasingly fierce competition of market, the functions and structures of products become more and more multiple and complex. Design of mechanical products has a great effect on the development of national economy and national defense industries. In this dissertation, method by combing assembly planning with assembly unit for mechanical product is proposed, which centers on the assembly modeling, takes assimilability analysis of product as realized purpose and investigates key technologies of assembly planning. The correlative key technologies in this research are as follows:Firstly, assembly modeling of mechanical products for assembly planning is established. Based on the analysis of traditional assembly model, assembly semantic relation model (ASRM) is proposed, which is based on engineer semantic information and has expansionable performance. By combining drive information, connector and geometric constraints into ASRM and quantifying difficulty of assimilability, it's aim to provide information to assembly unit partitioning and assembly sequence planning. And the compositive form combining component information, assembly constraints and engineer semantic information into ASRM is introduced.Secondly, the method of assembly units partitioning of mechanical product is studied. By decomposing mechanical product into smaller units, it is aim to decrease the difficulty and avoid combination explosion in assembly planning. And it will support assembly design and planning in parallel way. By analyzing the degree of assembly relation among components, the assembly units partitioning based on universal gravitation is proposed, which establishes assembly liaison intensity between functional components with gravity. Internal clustered degree and external detached degree are treated as selected criteria.Thirdly, a method of matching of assembly units based on subgraph isomorphism and case based reasoning is proposed. Assembly models are constructed by ASRM, which simplifies storage structure. Similarity calculation formularies between assemblies are provided. This method lays the foundation to assembly planning problems by case based reasoning, and simplifies the assembly planning among assemblies with similar structure.Fourthly, a hybrid algorithm for assembly sequence planning combining bacterial chemotaxis with genetic algorithm (BGHA) is proposed and tested. An optimization model of assembly sequences is modeled, and the mapping rules between the optimization model and the BGHA model are presented. This algorithm can keep diversity of the populations during evolution process, upgrade the quality in solution searching and decrease the probability of trapping into local optimal solutions. BGHA is less dependent upon the influence of initial population, parameters of crossover, and mutation probabilities, which is a novel and high efficiency hybrid optimization algorithm for assembly sequences planning.Finally, the assembly sequence merging by using nested partitions method (NPM) is implemented after optimal subsequences have been achieved and the interfaces among assembly units have been analyzed, which is computer aided reasoning method. Considering complexity of interfaces among assembly units, minimum cost spanning tree and topological sorting give assistances to realize the main steps in NPM of assembly sequence merging, which are partitioning, sampling, calculate the promising index and backtracking.