An intelligent chiller fault detection and diagnosis methodology using Bayesian belief network

A generic intelligent fault detection and diagnosis (FDD) strategy is proposed in this study to simulate the actual diagnostic thinking of chiller experts. A three-layer Diagnostic Bayesian Network (DBN) is developed to diagnose chiller faults based on the Bayesian Belief Network (BBN) theory. The structure of the DBN is a graphical and qualitative illustration of the intrinsic causal relationships among causal factors in Layer 1, faults in Layer 2 and fault symptoms in Layer 3. The parameters of the DBN represent the quantitative probabilistic relationships among the three layers. To diagnose chiller faults, posterior probabilities of the faults under observed evidences are calculated based on the probability analysis and the graph theory. Compared with other FDD strategies, the proposed strategy can make use of more useful information of the chiller concerned and expert knowledge. It is effective and efficient in diagnosing faults based on uncertain, incomplete and conflicting information. Evaluation of the strategy was made on a 90-ton water-cooled centrifugal chiller reported in ASHRAE RP-1043.

► The generic intelligent has the ability to simulate the actual diagnostic thinking of chiller experts.
► A generic framework for the intelligent chiller FDD strategy is represented.
► A three-layer diagnostic Bayesian network (DBN) is proposed based on the Bayesian Belief Network (BBN) theory.
► This strategy is effective and efficient in diagnosing faults based on uncertain, incomplete and conflicting information.