Performance Analysis of Double EWMA Controller Under Dynamic Models With Drift

This paper mainly focuses on the performance of the double exponentially weighted moving average (DEWMA) controller applied to a first-order dynamic input-output model with process drift. Under a general process disturbance, based on Jury test, the necessary and sufficient conditions involving model parameters, the estimated process gain, and weighting factors are established to ensure that the closed-loop system is asymptotically stable. In addition, we explore why we should use the DEWMA controller to adjust the process rather than the single exponentially weighted moving average controller when the process drifts considerably. Later, under the stability condition, the control performance analysis of the closed-loop system is performed in terms of the asymptotical variation and the transient deviation. Then, an optimization model is established to figure out the appropriate weighting factors for reducing the overall variation of the process output during the process of operation. Finally, extensive simulations are carried out to demonstrate the validity of the theoretical analysis in the context of chemical mechanical polishing process.