Contribution to state and sensor fault estimation applied to PEM Fuel Cell Systems

This paper aims firstly at contributing to the problematic of nonlinear Takagi-Sugeno (TS) approach for state and fault estimation with unmeasurable premise variables. The second objective is the application of this method on Fuel Cell Systems (FCSs). In order to achieve the estimated value of the fault, an observer based method is essential. The development and optimization of Fuel Cell parameter estimation in parallel with fault estimation is not treated in an exhaustive manner in literature. The estimation is based on a PI observer where the estimation error is bounded, having an additional corrective sliding terms. This eliminates the inconvenient generated by the Lipschitz constant.