Abstract
The reliability and cost-effectiveness of energy conversion in gas turbine systems are strongly dependent on an accurate diagnosis of possible process and sensor anomalies. Because data collected from a gas turbine system for diagnosis are inherently uncertain due to measurement noise and errors, probabilistic methods offer a promising tool for this problem. In particular, dynamic Bayesian networks present numerous advantages. In this work, two Bayesian networks were developed for compressor fouling and turbine erosion diagnostics. Different prior probability distributions were compared to determine the benefits of a dynamic, first-order hierarchical Markov model over a static prior probability and one dependent only on time. The influence of data uncertainty and scatter was analyzed by testing the diagnostics models on simulated fleet data. It was shown that the condition-based hierarchical model resulted in the best accuracy, and the benefit was more significant for data with higher overlap between states (i.e., for compressor fouling). The improvement with the proposed dynamic Bayesian network was 8 percentage points (in classification accuracy) for compressor fouling and 5 points for turbine erosion compared with the static network.
Highlights
Reducing maintenance costs is one of the most pressing concerns for gas turbine owners
In a dynamic Bayesian network (DBN), the prior probability P(Y) is not constant, but it depends on some hyperparameter φ as generalized in Equation (9) for a continuous distribution [25]
Three cases are presented and compared: a static BN where the prior probability distribution in the parent nodes is constant over time, a DBN where the prior probability distribution evolves over time following a Poisson distribution, and a DBN where the parent nodes are constructed as temporal nodes, i.e., conditionally dependent on the condition in the previous time step
Summary
Reducing maintenance costs is one of the most pressing concerns for gas turbine owners. Great advancements in computational power and machine learning techniques have led to a wide set of applications for gas turbine diagnostics [1,2,3] Advantages of such data-driven methods include robustness against measurement noise, fewer sensor requirements, and no knowledge requirement of proprietary information [4]. A better understanding of the advantages and disadvantages of the compared methods can be extracted from Table 1 Ageing models are another application that can benefit from DBNs. In [25], the authors present the use of hierarchical BNs for failure rate estimation, which utilize a multi-stage prior probability as a function of certain hyperparameters. In Part 2, the BN models will be applied to synthetic data and real field data to diagnose multiple fault scenarios and discriminate between gradual, time-dependent degradation and abrupt faults [27]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
