Diagnosis of a solar power plant using TS fuzzy-based multimodel approach

In the context of the electricity production, the solar energy is appropriate and endless. At present the technologies of solar concentration are the one which present most possibilities for commercial use. Thus, it is necessary not only to design processes of conversion of energy, but it is also important to assure an availability of these equipment by the conception of fault detection and isolation (FDI) systems. To improve the behavior of the solar power plant, we use a model based on differential algebraic equations to describe variations of the solar radiation, ambient temperature, flow rate and temperature of fluid. These phenomena are highly nonlinear. Moreover, a large class of nonlinear systems can be well approximated by T-S fuzzy models. The diagnosis scheme is based on a fuzzy observer to estimate faults and faulty system states; a proportional (P) observer to estimate constant faults in then adopted. Using descriptor redundancy property, a solution is proposed in terms of linear matrix inequalities (LMI). The performance of the proposed approach is pointed out by focusing on a model of solar power plant through numerical results.