Simplification techniques for EKF computations in fault diagnosis: Model decomposition

Abstract

The extended Kalman filter ( E m )

A Heuristic Extended Kalman Filter Based Estimator for Fault Identification in a Fluid Catalytic Cracking Unit

“…Among the existing quantitative model-based methods, the EKF has found widespread use and is one of the most popular methods because of its simplicity in implementation and its ability to handle a reasonable degree of nonlinearities. − There are, however, a number of practical challenges in designing such systems as a result of several factors such as the complexity of process dynamics, lack of adequate models, incomplete and uncertain data, diverse sources of knowledge, and amount of effort and expertise required to develop and maintain them. Hence, in the chemical engineering community, application of the EKF has only been reported for small (in terms of complexity and scale) case studies, most often for the continuously stirred tank reactor and the bioreactor. ,− For large deviations from the reference state trajectory, the EKF performs poorly or becomes unstable.…”

Sensor Fault Detection Using Noise Analysis

“…Analytical Redundancy. This approach exploits the implicit redundancy in the static and dynamic relationships between measurements and process inputs using a mathematical process model. − Different models of residual generation have been proposed: 5 parity space approach, − dedicated observer approach, , fault detection filter, parameter identification method, and artificial neural networks. − Implementation of the AR fault detection algorithms is usually expensive because of the time and effort needed to develop, test, and identify a good process model. The method is plant-dependent, that is, for each process plant, a unique model must be constructed.…”