| Semiconductor manufacturing industry is the core and basis of the information technology industry. Because it is an advanced manufacturing and high value-added industry, it is very significant for promoting the development of national economy. Because of the development of semiconductor manufacturing technology, it makes a great development in the progress of global information, the networking, and knowledge economy such that semiconductor manufacturing technology becomes an important symbol of national economic development and scientific and technological progress for almost every country. Thus, many countries take it as the backbone of the national industry. With the wafer size becoming more and more larger, the process of the wafer fabrication becomes more complex. Therefore, cluster tools are widely used to process wafers in semiconductor manufacturing. In a cluster tool, there is no buffer between the process modules. This makes that it is very complicated to schedule the cluster tools. Advanced scheduling methods can improve system efficiency and the quality of the completed wafers, reduce costs. Thus, it is very important and significant to developed a efficient scheduling strategy for the cluster tools. In order to do further research for scheduling time-constrained cluster tools to obtain better production strategy, this paper will study the scheduling problem of the cluster tools deeply.The most studies on the scheduling analysis of the cluster tools are based on earliest strategy. For these studies, the activity time variation is not considered. However, in fact, the process and robot tasks are subject to random variations. Therefore, it is necessary to consider the activity time variation in scheduling the cluster tools. In order to reduce the effect caused by the activity time variation on the wafer sojourn time, the robot waiting time can be used to offset the random disturbance. Thus, it is crucial to know where the robot waits when the system is scheduled based on earliest strategy. Hence, in this paper, the transient process is analyzed. With the analysis for the transient process, we can obtain where the robot waits when the system reaches the steady. Thus, we can check if the schedule obtained by earliest strategy is applicable for the cluster tools with residency time constraints and activity time variation. This motivates us to analyze the transient process of the single-arm cluster tools with residency time constraints and activity time variation.In this paper, the Petri net model for the single-arm cluster tools with residency time constraints is introduced. Based on the model, the dynamical behaviors of the system under the transient and steady state are described. Then, this paper analyze the time properties of the system. With the properties, two scheduling strategy are proposed for the system. They are earliest scheduling strategy and robot waiting time control (RWTC) scheduling strategy, respectively. Based on earliest strategy, under the steady state, the waiting time at every step are found by using some expressions. Then, based on RWTC strategy, the methods for assigning robot waiting times are presented. It is very flexible to control the system by using RWTC strategy. Finally, with the activity time variations considered, by comparing the two scheduling strategy, it is found that, it is better by using RWTC strategy to schedule the system because it can offset the effects caused by activity time variations as much more as possible.To test the results in the paper, based on the Petri net model, the simulation model are developed by using eM-plant software. Then, examples are given to the application of the results in the paper. |
PDF Full Text Request