Neural network approach for sensor fault detection and accommodation

Abstract

This thesis seeks to investigate if a neural network based fault detection and accommodation scheme is able to limit the influence of sensor faults on the performance of a nonlinear dynamic process. After a failure examination of MLP network based approach, an Elman network based fault detection and accommodation approach was proposed. Dynamic systems can be modelled using Elman network directly by employing the simplified Dynamic Backpropagation (DBP) algorithm proposed. The simulation result obtained from a SISO plant suggests that the Elman network based scheme is successful. Then, a modified Elman network with delay blocks and the corresponding algorithm to learn the delay is proposed to handle the system with dead-time. Finally, experiments on a nonlinear coupled tank system were implemented to demonstrate the effectiveness of the Elman network based fault accommodation scheme. Sensor fault tolerant experimental results demonstrate the feasibility and effectiveness of the proposed scheme.