Abstract: Based on variation propagation, a method of feasible robust design tolerance is developed for fixture system in two-dimensional multi-station assembly processes. The relationship of dimensional variation propagation for two-dimensional multi-station assembly processes is analyzed, and the state space model of variations stream is built, in which the fixture locating variations are caused by the tolerances of pin and part hole/slot. By means of number-theoretical net method, the tolerances sample of pin and part hole/slot is presented which is used to substitute into the variations model of fixture locating to obtain the sample space of fixture locating variations. Based on state space model, taking fixture locating variations for input vectors, an evaluation method of assembly yield is proposed for two-dimensional multi-station assembly processes. Furthermore, Taguchi orthogonal experiment method is adopted to analyze the main factors that influence the assembly yield and the design of regression-orthogonal combination is applied to fit the response surface model of assembly yield. With the objective to satisfy the lowest manufacturing cost of pin and part hole/slot, taking the assembly yield for constraint, the model of feasible robust tolerance design for fixture system in two-dimensional multi-station assembly processes is established. Automotive body floor assembly process with three-station is given as an example to build the model which is calculated and analyzed. The results show that the robustness of tolerance design is improved significantly with robust constraint while assembly cost increases slightly. It provides a new way of feasible robust tolerance design for fixture system in two-dimensional multi-station assembly processes.

Key words: Assembly yield, Multi-station assembly, Number-theoretical net, Robust design, Tolerance, Variation propagation