Failure Response Policy For Dual-arm Cluster Tools Based On Petri Net

As the basis of networking and information technology, development of semiconductor manufacturing is an important symbol of the embodiment of a country in modern economy and high-tech power. With the semiconductor wafer diameter increase, the processing technology becomes more and more complicated, the traditional batch processing is unable to meet the needs of wafer manufacturing. Thus, semiconductor manufacturers tend to adopt cluster tool as an integrated equipment to produce wafers. Considering the production efficiency, the working space and pollution factors, cluster tool is composed of a several processing modules and a transfer module, and is placed in a vacuum environment. Since there is no intermediate buffer between the process modules, the schedule of a cluster tool is very complicated. Furthermore, a processing module in a cluster tool is prone to error and when a processing module is broken down, it may last for several hours. Hence, it is very important for system to respond correctly to such events such that cluster tool can continue its operation if scheduling is still possible. Hence, this thesis studies the failure response strategy of dual-arms cluster tool based on Petri Net.In the whole wafer fabrication process, there usually exsiste a bottleneck process, multiple processing modules are arranged to match the whole process to increase the wafer processing flexibility and meet process requirements. Petri Net can describe the parallel and conflict in discrete event dynamic system, is an effective tool model for event driven, and it also as a tool to make model of mathematical and graphical. It is widely used in discrete event simulation. This thesis presents the Petri Net model of the dual-arms cluster tool that can describe the stay time of wafer processing and the wait time for mechanical hand unload wafer from the vacuum lock or processing module. Using this Petri Net model, the schedulability of the system can be achieved when the time is determined, to verify the feasibility of scheduling when time has fluctuation.Since there exsit parallel module in the dual-arms cluster tool, when a module of the parallel module has failure, system maybe continue to finish the wafer processing. However, how to control the dual-arms cluster tool when failure happens is a significant problem to be study. Through develope the Petri Net model of the dual-arms cluster tool, we analyze the feasibility of scheduling when there has one failure module in parallel module. When the parallel module has failure, if control method of dual-arms cluster tool is not proper, it maybe violate the sojourn time constraint, and deteriarate the wafer quality seriously, and even make the wafer becomes scrap. In fact, it is the key for how the fault response transit from the steady state before failing to the steady state after failing. Thus, we propose corresponding control strategies for different failure scenarios, such that makes the system continue to complete the wafer processing, and ensures wafers meet the quality requirements. Finally, we use several examples to illustrate the validate response strategies.