Nonlinear Methods for Fault Diagnosis

This chapter provides an overview on different fault diagnosis strategies, with particular attention to the fault detection and isolation methods related to the dynamic processes and application examples. Model-based fault diagnosis methods usually use residuals that indicate changes between the process and the model. The chapter discusses some commonly used residual generation and evaluation techniques and presents their mathematical formulation. It considers the model-based residual generation schemes relying on parity space (relation) methods and observer-based approaches for their application to dynamic processes and technical systems, and summarizes the classical nonlinear geometric approach technique. The data-driven approach has started from the algebraic results, with the purpose of investigating the possibility of extending estimation schemes to dynamical systems, determining the whole family of models compatible with noisy sequences. The chapter provides a brief introduction on the general structure of a fault diagnosis system relying on fuzzy systems and neural networks.