| This research presents a systematic approach to expedite the generation of new assembly variant designs from mature component designs in agile manufacturing environment. It consists of fundamental research in two parts: assembly modeling and assembly variant design methodology.; The complementary assembly modeling concept is presented and two kinds of assembly models, a hierarchical assembly model (HAM) and a relational assembly model (RAM), are established. The hierarchical relationships between assembly constituent components are captured explicitly by the HAM, and the mating relationships between components are explicitly modeled by the RAM at the manufacturing feature level. These models are further specialized to cater for the needs for assembly variant design. As a result, the assembly variants model (AVM) and assembly mating graphs (AMG) are generated.; The assembly variant design methodology is tightly coupled with the proposed complementary assembly modeling concept and rests on the concept of constraint group (CG). Based on the AVM, matching components are assumed to be searched and retrieved. Then for each assembly variant, the constraint groups are identified by manipulating corresponding assembly mating graph. The variant design process is finally formulated as a mixed integer linear programming problem (MILP) which is solved using a standard solver. The optimal solution provides insights about how to modify existent component designs to form a new assembly design with optimal redesign time and cost.; A prototypical software system has been developed to identify the constraint groups and then to formulate the variant design process as an MILP which is solved using CPLEX. |
