Study on intelligent reasoning of product assembly sequences and simulation

This paper proposes a layer-built and degree relation assembly (LDRA) model after analyzing some existing assembly models and new DFA method, which has 15 key-part criteria. It can evaluate the product assembly quality by calculating the assembly efficiency and the ratio of the part function value to its assembly difficulty and it is integrated into CAAPP. A prototype system has been developed for reducing gears and it can generate assembly sequences by AI reasoning, evaluate the assembly quality by the proposed DFA method and check the assembly interference by 3D simulation in UGII environment. In the proposed assembly reasoning method, the goal is not to obtain an optimum assembly sequence but a better one, it is divided into two steps: (1) to determine the relational priority between the parts, (2) to obtain topology sequences; and it places emphasis on the assembly reliability.