Observer-Based Robust Fault Detection for Fuzzy Multi-model Jumping System

Abstract

To enhance the reliability and security of dynamics, the subject of model-based FD has been actively studied in the past twenty years, such as [257, 258]. The main method used in this chapter is to select an appropriate threshold and a suitable evaluation function. When the value of the evaluation function exceeds the selected threshold, a fault alarm is generated. Using this method, the main problem to be solved is how to distinguish faults and disturbances when the system has model uncertainties. Usually we will adopt the observer-based robust FD method, which is considered to be the robustness scheme of the detected system. In the robust FD system, the residual generator not only includes the influence of disturbance and modeling errors, but also maintains the sensitivity to faults and robustness to unknown inputs. In [259], a scheme to measure the performance between robustness and sensitivity is proposed. Inspired by this, more attention has been paid to the robust FD observer (FDO) design problem for nonlinear dynamics, such as $H_{\infty } / H_{-}$ approach [118] and $H_{2} / H_{\infty }$ scheme [260].