Fault detection and diagnosis in a closed-loop nonlinear distillation process: application of extended Kalman filters

We evaluate the effectiveness of the extended Kalman filter (EKF) for fault detection and diagnosis in closed-loop systems. A distillation column is used to illustrate the procedure. We describe the strategy for fault detection and diagnosis in the closed-loop system. We present the process models and dynamic analysis of accuracy of the model. We describe the implementation of EKF. We apply the strategy to a 30-stage binary distillation column with a partial condenser and a reboiler