The Study Of Sliding Mode Observer-based Fault Reconstruction Method Of Nonlinear System

The automation degree of modern industrial production is more and more high; This requires that the control system has higher reliability. To prevent the loss and hazard caused by fault, the real-time monitoring for the state of control system must be carried out. Based on that residual error method cannot carry out estimation for the fault, so this method is gradually being replaced by fault reconstruction. Because the nonlinear system is closer to the control system in the actual production, the study of nonlinear system is more practical.Based on Lyapunov stability theory and sliding mode variable structure control system, this paper mainly studies a kind of nonlinear system including multiple faults and multiple unknown input disturbances.From improving the robustness of modeling errors and interference,improving the diagnosis sensitivity of the fault and improving the fault reconstruction precision and other aspects, this paper carries out in-depth study for robustness fault diagnosis problem of uncertain nonlinear system based on the observer. The main research work of this paper is as follows:Specific to nonlinear system including multidimensional actuator faults and multidimensional unknown input disturbance, this paper proposes a fault diagnosis scheme. Firstly, the original system is carried out linear coordinate transformation to obtain two subsystems. For the above two subsystems, the sliding mode observer and sliding mode input signal are designed, thus to realize the precise reconstruction of multidimensional actuator faults and the estimation of multidimensional unknown input disturbance. The simulation result using single joint manipulator as an example proves the reliability of this method.Specific to nonlinear system including multidimensional actuator faults, multidimensional sensor faults and multidimensional unknown input disturbance, this paper proposes a fault diagnosis scheme. Through carrying out linear transformation for the output equation and design of low pass filter, the sensor fault is transformed to equivalent actuator fault.Then through a linear coordinate transformation, the original system istransformed into three subsystems. Using sliding mode control theory,based on the realizability of actuator fault reconstruction, the sensor fault reconstruction can be carried out, and the unknown input disturbance can be estimated. The simulation results of two simulation examples illustrate that this method is feasible.