Fault diagnosis in a heat exchanger using process history based-methods
Chapter in Scopus
- Additional Document Info
- View All
A comparison of fault diagnosis systems based on Dynamic Principal Component Analysis (DPCA) method and Artificial Neural Networks (ANN) under the same experimental data is presented. Both approaches are process history based methods which do not assume any form of model structure, and rely only on process historical data. The comparative analysis shows the online performance of both approaches when sensors and/or actuators fail. Robustness, quick detection, isolability capacity, false alarm rates and multiple faults identifiability are considered for this experimental comparison. An industrial heat exchanger was the experimental system. ANN showed instantaneous detection for actuator faults; however, with greater (22%) false alarm rate. ANN can isolate multiple faults; whereas, DPCA did not show this property, but required a minor training effort. © 2010 Elsevier B.V.