Sensor fault detection and isolation in a thermal power plant steam separator

A fault detection and isolation approach, specially designed for steam separators in thermal power plants, is presented in the paper. The first step of the proposed algorithm is to identify the process. Because of the presence of sporadic high-intensity measurement noise (outliers), the paper proposes a robust version of recursive identification. Starting from the parameter vector of the identified model, the second step of the proposed procedure has the form of a data-driven fault detector. This particular fault detection and isolation approach was implemented at TEKOB1 Kostolac Thermal Power Plant in Serbia, whose nominal power output is 330 MW.