Robust fault detection and isolation in semi-actively controlled building structures using a set of unknown input observers

Building structures are subject to earthquakes and unwanted vibrations which can be effectively managed via controllers. Semi-actively controlled building structures are prone to sensor and actuator faults very similar to other dynamical systems. When a fault occurs in sensors or actuators of a controlled systems, the system faces a performances degradation or even failure. Consequently, it is vitally important to detect and isolate a fault at the right time in these systems. To do so, here, unknown input observers (UIO) are proposed for robust fault detection and isolation of actuators and sensors in buildings. For the proof of concept, a three-story structure with magnetorheological (MR) dampers is taken into account. Via designing these observers, each faulty actuator and/or sensor is detected and isolated. Here, the LQR controller is also used to facilitate an optimal control strategy over the system. The obtained simulation results utter the acceptable accuracy of the proposed method in the detection of fault (time and location) and the robustness of the fault detection method against external disturbances.