Fault Isolation and Fault-Tolerant Control Design for Non-Gaussian Stochastic Distribution Control Systems With Multiple Sensor Faults

A fault isolation, estimation and fault-tolerant control algorithm is proposed for non-Gaussian stochastic distribution control systems with disturbance and multiple sensor faults. Sensor faults are represented as actuator faults virtually, and an observer is devised to detect the sensor fault occurrence time. Then two subsystems are separated by the expanded system through introducing the coordinate transformation matrices. One subsystem contains only sensor faults and does not contain disturbance and the other contains sensor faults and disturbance, which provides convenience for fault isolation. The faults are estimated respectively by the multiple fault isolation observers with the same number of sensors. A fault-tolerant control scheme is proposed after getting the fault information to compensated sensor faults and track the desired probability density function. Finally, a MATLAB simulation example is used to verify the feasibility of the algorithm. The graphical abstract is divided into seven parts. The first is Variable entry. A new state variable is introduced to represent the form of the sensor faults as the form of the actuator faults virtually. The second is fault detection augmentation system. The third is residual judgment. Checking whether the system residual is greater than the threshold to determine whether the fault occurs. The fourth is coordinate transformation matrices. The system disturbance is decoupled from the faults to be separated by the coordinate transformation matrices. The fifth is fault isolation. Designing a set of fault isolation observers for two subsystems. The sixth is fault estimation. The fault time and amplitude are estimated by using the fault isolation residuals. The final is fault-tolerant control. The fault tolerant controller is constructed according to the fault information after the sensor fault is estimated. image