An analytical method for fault detection and isolation

The fault detection and isolation (FDI) problem is an inherently complex one. The process control requires data management for a strategically behaviour and the uninterrupted maintenance. So, it is necessary to assure the systempsilas safety and availability. In this paper, the author proposes a practical design method useful in control and supervising of control system in presence of process faults. This paper proposes to solve the FDI problem by developing a model-based approach using a parity equation. The procedures was applied to detect the failed sensor or failed actuator and suppose the use of mathematical model of process and the correlation (covariance) function calculated in real time between the signals, which represent the commands, and the measurements. The fault detection and isolation represents the next level up necessary for the realisation of robust and fault tolerant systems.