A Survey on Fault Detection and Diagnosis Methods

Abstract

Fault detection and diagnosis in modern control systems have been of constant interest in recent publications. Its progress is a consequence of the requirements imposed by the development of other technologies through demands on security operations, guarantee of the required functions execution, reduction of costs, and optimization of maintenance tasks. In order to provide the survey in this area, the article discriminates the main fault detection and diagnosis techniques, allowing the reader to acquire, in different practice scenarios, an ability to discern the possibilities of applying the methods in focus. The text is divided in signal analysis-based methods, model-based methods, data-driven methods, and hybrids methods. The conclusion exposes the main global limitations in the area as possible subjects for future works.