Kalman Filter-Smoothed Random Walk Based Centralized Controller for Multi-Input Multi-Output Processes

In this paper, a novel centralized controller is presented to control multi-input multi-output industrial processes with heavy interactions and significant time-delays. The system model equations are represented in a non-minimal stochastic state-space form. Also, the state and measurement equations respectively use smoothed random walk model and finite impulse response model of the plant. To design the controller, a quadratic cost function is considered. A standard Kalman filter algorithm is used to estimate the state vector of the controller and solve to the discrete algebraic Riccati equation simultaneously. By using the smoothing parameter, the controller behavior can be changed between the Kalman filter random walk controller and the Kalman filter integrated random walk controller. To evaluate the effect of the smoothing parameter the proposed controller is first applied to a single input single output system. Then an industrial-scale polymerization reactor which has the two-input and two-output system is used to investigate the performance of the designed controller. The simulation results indicate that the controller has a good performance in tracking the set point and robust due to changing the system parameters.