Moving window RtR fault/disturbance tolerant control approach in high-mix semiconductor manufacturing process

In this work, a new state estimation method based on linear model for high-mix semiconductor manufacturing processes is proposed. IMA(1,1) is a general disturbance model adopted in semiconductor manufacturing process, which can reflect the practical process disturbance. Taking IMA(1,1) disturbance into account, our algorithm exerts moving window approach on the linear model of the process and employs minimum norm estimation method to provide better estimates for the states without the IMA(1,1) model parameters. Furthermore, a step fault is considered and our approach can achieve fault-tolerant control. Comparing to the t-EWMA algorithm, this method provide better control performance, which is proved by simulation results.