Identification and Control of a Fuel Cell System in the Presence of Time-Varying Disturbances

Abstract

For most fuel cell power systems, the control implementation is essential because it can cope with electrical loads and prevent the negative effects due to over temperature. In this work, a time-varying Hammerstein model is developed to identify a fuel cell system in the presence of time-varying disturbances since the load variations are usually time-varying and unknown. The static nonlinearity part is approximated by a series of piecewise linear functions. A moving-horizon least-squares estimator for parameter estimation is added to update the linear dynamic elements as well as identify the dynamic behavior of time-varying disturbances. The composite feedback control, which consists of the inverse function and the internal model control (IMC) structure with updated parameters, is implemented to control the stack temperature. Simulations show that the accuracy of system identification is guaranteed and the control performance is superior to the conventional model-based control.