Spot-welding sequence planning and optimization using a hybrid rule-based approach and genetic algorithm

Performing assembly planning to find a valid hierarchical assembling structure of a product (i.e. Manufacturing Bill of Materials or MBOM) based on the constraints and necessities inferred from or declared by different sources is potentially complicated. On the other hand, Engineering Changes (EC) may drastically affect the constraints and necessities which the planning of an MBOM was based on. Managing ECs to evaluate and propagate their effects on the upstream data used in assembly planning and downstream activities and information is crucial but problematic. Often it is possible to define a set of rules for the constraints and necessities of assembly planning and find solutions or check validity of solutions based on the rule-set. This paper proposes a rule-based assembly planning method and introduces the concepts and standard notations on how structured rule-sets can be derived from descriptive rules and then used in an algorithm for generating or validating MBOMs. The method was partially automated and successfully employed along with a commercial Virtual Manufacturing package integrated with an in-house developed GA-based sequence optimizer and applied to the sequence optimization in minimizing the cycle time of the robotic spot welding operations for a sheet-metal assembly found in automotive industry.