Nonlinear State Estimation of Power Plant System using State-Dependent Kalman Filter

Abstract

In this paper a nonlinear Kalman filter estimator is presented to estimate the states of a nonlinear large scale power plant system. Power plant process is a highly nonlinear process, therefore state estimation techniques that uses a linearization-based approach has a limited performances due to linearization errors. This paper proposed an alternative nonlinear estimation approach that based on the State-Dependent-Differential-Riccati-Equation (SDDRE). This technique avoid linearization and can fully take into account the system nonlinearities and the measurement noise. An 11th order nonlinear model of power plant system was first developed and simulated using decentralized PID controllers and then used for the design of the nonlinear estimator. The performance of this estimator is demonstrated via nonlinear simulations. The results show good performance of the proposed SDDRE filter for large scale systems with high nonlinearities.