Development and Test of Methods for Fault Detection and Isolation

Almost all industrial systemns are automated to ensure optimal production both in relation to energy consumtion and safety to equipment and humans. All working parts are individually subject to faults. This can lead to unacceptable economic loss or injury to people. This thesis deals with a monitoring scheme which detects a fault as it occurs, which identifies tha faulty component, and accommodates the system to the faulty situation, so a satisfactory operation continues, or the system fails to a safe state. The focus in this work is on faults in the instruments of control loop systems, because they are especiallu crucial for the entire operaiton of a closed loop system. The purpose of the thesis is to investigate, deveop, and verify methods for fault detection and isolation on control loop systems. An Industrial Position Controller, (IPC), laboratory setup is used as an application example throughout the thesis. The IPC offers prospects of repeated fault scenarios, and support studies in robustness issues. The thesis contributes with a numerical fault analysis representation, practical applications of existing methods for FDI, and a method for robust FDI for practical applications.