Multiple-model adaptive estimation of a hydraulic wind power system

Nonlinear model of hydraulic wind power system operates on a wide spectrum of operating points such as random wind speed disturbances and applied control commands. Thus, one way to linearize this model is to use multiple linear models representing the whole range of operating points. This paper introduces a minimal number of fixed linear models in a multiple model adaptive estimation (MMAE) framework to reduce the state estimation error. System parameters such as pressures of the pump and motors can be estimated while the overall error in entire operating points is reduced. The algorithm is composed of a bank of Kalman filters, each of which is modeled to match particular real world operating condition. Simulation results demonstrate that the adaptive approach can optimally estimate the state variables in a wide range of operating points.