Kalman filtering algorithm for on-line identification of input-output curves for thermal power plant

This paper presents the applications of Kalman filtering algorithm for the on-line identification of the input/output (I/O) curves for the thermal power plants. The advantage of this approach is the capability to keep tracking the state of the thermal unit by making the coefficients of the I/O curve continuously adaptive to the real characteristic of the unit. Different models for the I/O curve of the thermal units are used, including polynomial models having different orders. The proposed algorithm is tested for several examples from the literature