Design of Fault Detection Observer Based on New-Type Neural Networks

In this paper, the convex combination algorithm (CCA) is proposed to optimize new-type neural networks. This method updates the weights by iterating to massage the information in the hidden layer. And a new error function is set up to measure the performance of the neural networks. The optimized parameters can be obtained by decoupling the weights, which improves the calculating speed of the parameters. On the basis, a fault detection and diagnosis method is proposed based on the observer for the nonlinear modeling ability of neural network. Finally, this method is applied to the nonlinear systems, and the sensitivity of the neural networks fault detection observer to nonlinear systems failure is proved by simulation.