Fault Diagnosis in a Power Generation Plant Using a Neural Fuzzy System with Rule Extraction
In this chapter, FMM is endowed with a rule extraction algorithm. With the rule extraction algorithm, FMM is able to explain its predictions using fuzzy if-then rules, thus overcoming the “black-box” phenomenon as suffered by most NN models. Applicability of FMM to fault diagnosis tasks in a power generation plant has been examined. The potential of FMM in learning and predicting faults in complex processes as well as in providing a comprehensible explanation for its predictions has been demonstrated in two experiments. The proposed rule extraction algorithm is able to yield a comprehensible rule set. The extracted rules have been verified as meaningful and are in line with the domain knowledge as well as experts’ opinions. Further research work will concentrate on the aspects of implementation, validation, and verification of FMM as a useful, robust, and intelligent fault diagnosis tool in a variety of application domains.
KeywordsFault Diagnosis Rule Extraction Heat Transfer Condition Condenser Tube Power Generation Plant
Unable to display preview. Download preview PDF.
- 3.Carpenter GA and Tan AH (1995) Rule extraction: From neural architecture to symbolic representation. Connection Science 7:3–27Google Scholar
- 4.Craven MW and Shavlik JW (1994) Using sampling and queries to extract rules from trained neural networks. In: Machine Learning: Proceedings of the Eleventh International Conference, San Francisco, CA, USAGoogle Scholar
- 5.Dellomo MR (1999) Helicopter gearbox fault detection: a neural network based approach. Journal of Vibration and Acoustics 121:265–272Google Scholar
- 6.Denny G (1993) F16 jet engine trending and diagnostics with neural networks. In: Proceedings of SPIE, vol. 1965, pp. 419–412Google Scholar
- 7.Dietz WE, Kiech EL and Ali M (1989) Jet and rocket engine fault diagnosis in real time. Journal of Neural Network Computing 1: 5–18Google Scholar
- 9.Efron B and Tibshirani RJ (1993) An introduction to the bootstrap. Chapman & HallGoogle Scholar
- 10.Freitas JFG, MacLeod IM and Maltz JS (1999) Neural networks for pneumatic actuator fault detection. Transactions of the SAIEE 90:28–34Google Scholar
- 11.Fu LM (1994) Rule generation from neural networks. IEEE Transactions on Systems, Man, and Cybernetics 28: 1114–1124Google Scholar
- 12.Jack LB, Nandi AK and McCormick AC (1999) Diagnosis of rolling element bearing faults using radial basis function networks. EURASIP Journal on Applied Signal Processing 6:25–32Google Scholar
- 17.System description and operating procedures, Prai Power Station Stage 3, vol. 14, 1999Google Scholar
- 18.Tickle B, Orlowski M and Diederich J (1996) DEDEC: A methodology for extracting rule from trained artificial neural networks. In: Proceedings of the Rule Extraction from Trained Artificial Neural Network Workshop, Society for the Study of Artificial Intelligence and Simulation of Behaviour Workshop Series (AISB’96), University of Sussex, Brighton, UK, pp. 90–102, 1996Google Scholar