DIAGNÓSTICO DE FALLOS EN EL CONTROL Y COMPENSACIÓN DE LA MEDICIÓN POR SENSORES VIRTUALES EN UNA CENTRAL TERMOELÉCTRICA

In this work the system of fault diagnosis in control and measurement compensation by virtual sensors in a thermal power plant is presented. The diagnostic system is applied to control systems of the power plant, to compensate for faulty sensor measurements. The design is based on Luenberger-like observers for Linear Parameter-Varying (LPV). An LPV model of the steam generator and turbine is parameterized with real data from the operation of a power plant and represents the dynamic behavior of the pressure in the first stage of the turbine, drum pressure, superheated steam pressure and electrical power. The main contribution of this work is to show that the proposed diagnostic system is suitable for real plant applications. In this sense, the performance of the diagnostic system is evaluated adequately in a scenario by using a simulator that emulates the real operation of a power thermal plant. Key Words: Observer, Fault diagnosis, Nonlinear model, LPV model, Virtual sensor, Power thermal plant.