An approach to robust fault detection for nonlinear system based on RBF neural network observer. (English) Zbl 1007.93032

The paper presents a robust fault detection and isolation (FDI) method for affine nonlinear systems, which is based on the use of radial basis function (RBF) neural networks (NNs). The RBF NN is employed to approximate the nonlinear item of the monitored system in order to improve the accuracy of state estimation. It is shown that the resulting state estimation error tends to zero asymptotically, i.e., the method eliminates the effect of modeling errors on the residuals. In addition, an NN classifier is employed to complete the fault isolation. A new training (weight adjustment) algorithm is derived to enhance the robustness of FDI. A simulation example is included.


93B51 Design techniques (robust design, computer-aided design, etc.)
92B20 Neural networks for/in biological studies, artificial life and related topics
93B07 Observability
93C10 Nonlinear systems in control theory
90B25 Reliability, availability, maintenance, inspection in operations research