Fault detection based on SIR state estimation and smoothed residual

It is often needed for many fault detection approaches based on model for stochastic systems that the model should be linear and Gaussian. But the performance of these approaches usually be awful when the model is nonlinear and non-Gaussian. Now one of the better approaches for nonlinear and non-Gaussian stochastic systems is the fault detection approach based on SIR likelihood. In this paper a novel approach named fault detection based on SIR state estimation and smoothed residual is proposed. In this approach the estimate value of the state of the nonlinear stochastic system is obtained firstly using SIR. Then the difference between these ideal observation of the estimate value and those observation of the state of the system is smoothed. The fault detection is done according to these smoothed difference. The simulation results indicate that the false alarm rate of the two approaches is like while the missing alarm rate of the novel approach is better the one of the fault detection approach based on SIR likelihood.