Intermittent Fault Detection and Fault-Tolerant Control Algorithm Based on Neural Network in Nonlinear Systems

This is the authors' abstract. We don't add key points for this paper.

In the field of system engineering, nonlinear system fault detection and fault-tolerant control are important research areas, and a number of efficient algorithms have been proposed. Therefore, this study proposes a neural network-based nonlinear system inter-instance fault detection and fault-tolerant control algorithm. First, from the system design level, under the offset property of the system as a random time-delay linear system, this study completed the model definition of inter-instance faults. At the same time, a fully visible quantitative analysis was carried out. Compared with permanent faults, intermittent faults have the characteristics of random appearance time and disappearance time. Based on the proposed preliminary model, this study proposed a state observer based on a constrained neural network. This observation period can complete the fault signal estimation and system state error analysis to the greatest extent. Then, this study conducted a comprehensive analysis using the mean-sufficient model and a graph-based nonlinear system modeling method. At the same time, a dynamic system model was constructed, which allows more convenient addition of components. Through quantitative and estimation analysis, it was demonstrated that this method can effectively detect the location and extent of intermittent faults. Finally, this study uses the ticket high-throughput model to predict control and minimize the cost function algorithm for final optimization. The experimental results show that the proposed algorithm can effectively improve the real-time nature of fault detection and queue uniformity of the fault-tolerant system.