The simulation of MISO MPC for gaseous pilot plant control with presence of measurement noise

This study focuses on pressure control in gaseous pilot plant using model predictive control (MPC) algorithm which consists of multi-input and measurement noise. The applicated MPC uses linear approximation. Process model is obtained from subspace system identification method with amplitude modulation pseudo random binary signal (APRBS) inputs, and its control law is derived from constrained optimization technique with quadratic programming algorithm. Simulation results demonstrated the necessity to pay attention to controller aggressivity by tuning the controller weights and apply a filter due to the noisy measurement data to prevent the oscillation of actuator in steady state condition.