| Nowadays, to keep the advantage in global intensely competitive market, all the electronics manufacturers try their best to improve production efficiency to save the cost. Although a SMT production line consists of much equipment, the speed of the line was actually determined by mounter. So, automatic process optimization technique is one important technique of multi-head gantry mounter R&D.The process optimization problem of multi-head gantry mounter is usually decomposed into two related sub-problems: feeder assignment optimization and pick and place sequence optimization, and then they are solved independently. Because their complexity and coupling, a new hierarchical solution is put forward. Main contents of this thesis are as follows:Firstly, the whole optimization problem is modeled as an extended non- symmetry TSP problem. Thus, the problem is converted into sub-problem: to construct pick-and-place cycles and determine their sequence.Secondly, based on the above model, a kind of hierarchical method is presented. That is: after the acquisition of the mounting data, needing nozzles are chosen to let the number of changing nozzle as less as possible; mounting reel-groups are constructed according to the load balance of the mounter heads; feeder assignment is solved by dynamic programming; component-groups are constructed based on the above job.Thirdly, according to the above algorithm, the programming of the simulation software is completed. By the comparison with other algorithms, this algorithm is validated very effectively and stably. The process optimization module and data conversion module are integrated into the whole mounter software system to accomplish automatic mounting after analyzing the PCB mounting process.At last, the load assignment and balance problem of SMT production line is analyzed to find its solution. |