The Research Of Sub-assembly Recognition Algorithm In Assembly Sequence Planning

With the rapid development of the country’s industrialization, assembly problem has become one of the reasons for high product costs. Sub-assembly recognition is one of the key problems of assembly sequence planning. With the increasing complexity of assemblies, assembly process often requires people from different departments to work together. The recognition of sub-assembly assemblies in an assembly can effectively achieve parallel development and planning of the assembly. In the paper, the focus is concentrated on a sub-assembly recognition algorithm. It is used to recognize the sub-assemblies in various assemblies in order to simplify assembly sequence and decrease assembly difficulty and costs.(1) The sub-assemblies in assembly sequence planning are defined and analyzed. Combined with fuzzy sets and clustering methods, the principles of sub-assembly recognition algorithms are explored. Weighted undirected connection graph is used to describe the information model of an assembly. It is then converted into an adjacency incidence matrix to judge the membership relations between parts in the assembly. Then, the sub-assemblies are found.(2) According to the sub-assembly recognition process and principles, its mathematical model is established.In Matlab, it is programmed and verified through three examples. The first example verifies whether the algorithm can recognize autonomously. The second example verifies whether the algorithm can accurately judge the incidence relations between the parts in the assembly. The third example verifies the accuracy of the algorithm when the part number is large.(3) The main factor of the algorithm, the number of sub-assembly c and fuzzy exponent m, are selected to analyze its influence on recognition results. When the factors change, validity and convergence rate of the algorithm vary. By analyses and comparisons, the rational selection of the factor is suggested.