A nonlinear model predictive control approach for robust end‐point property control of a thin‐film deposition process

Abstract

AbstractAlthough an increasing number of simulation and experimental evaluation studies show the advantages of nonlinear model predictive control (NMPC), generally in practice processes are operated using open‐loop input and reference trajectories. NMPC strategies can address explicitly constraints and nonlinearities during the feedback control of batch processes. However, the control performance is strongly dependent on the quality of the model used in the controller. Although generally in the nominal case good performance can be achieved, disturbances and model uncertainties can have drastic influence on the control performance, and might even lead to instability. The paper presents an NMPC algorithm that explicitly takes parameter uncertainty into account in the controller designs. The shrinking horizon NMPC algorithm minimizes a weighted sum of the nominal performance objective, an estimate of the variance of the performance objective, and an integral of the deviation of the control trajectory from the nominal optimal control trajectory. The robust performance is quantified by estimates of the distribution of the performance index along the batch run obtained by a power series expansion about the control trajectory. The potential advantages of the proposed robust NMPC in the microelectronics industry are demonstrated using an example process of thin‐film deposition. Copyright © 2007 John Wiley & Sons, Ltd.