Comparative Assessment of a Chemical Reactor Using Extended Kalman Filter and Unscented Kalman Filter

Abstract

Abstract Continuous Stirred Tank Reactor is a typical chemical reactor system with complex nonlinear dynamic characteristics. The variables which characterize the quality of the final product in Continuous Stirred Tank Reactor are often difficult to measure in real-time and cannot be directly controlled using feedback configuration. In this work, a comparative study of the performance analysis of Extended Kalman Filter with respect to Unscented Kalman Filter for Continuous Stirred Tank Reactor that rely solely on concentration estimation of Continuous Stirred Tank Reactor via measured reactor temperature is done. The performance of these two filters is analyzed in simulation with Gaussian noise source under various operating conditions and model uncertainties. It is shown that in the presence of large model parameter mismatch and initial state mismatch, Unscented Kalman Filter is able to provide more reliable estimate, when compared to Extended Kalman Filter. However, under ideal conditions the performance of Extended Kalman Filter is found to be better.