The lack of onboard gas path measurements combined with the measurement errors leads the gas path analysis to an underdetermined problem with uncertainty. Incorporating additional information such as the domain knowledge and heuristics, as well as information derived from other diagnostic assessment methods, has become a promising consideration. In this paper, a Bayesian network-based multiple diagnostic information fusion mechanism is proposed to improve the performance of the gas path analysis. The domain knowledge and constraints regarding the component degradation pattern are incorporated into the network by setting an informative prior for the health parameters; furthermore, a fault mode prior probability table is developed to incorporate additional diagnostic information to narrow down the candidate faulty components to a possible set. The effectiveness of the proposed method is demonstrated on a simulation case study of a typical turbofan engine. As more information is incorporated into the network, the diagnostic result is improved. The faulty components can be successful isolated, and the fault magnitude can be identified with less uncertainty.
Bayesian Network-Based Multiple Sources Information Fusion Mechanism for Gas Path Analysis
Journal of Propulsion and Power ; 32 , 3 ; 611-619
2016-01-08
9 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Bayesian Network-Based Multiple Sources Information Fusion Mechanism for Gas Path Analysis
Online Contents | 2016
|Bayesian Network-Based Multiple Sources Information Fusion Mechanism for Gas Path Analysis
Online Contents | 2016
|British Library Conference Proceedings | 2018
|