Litcius/Paper detail

Research on Knowledge Graph and Bayesian Network in Fault Diagnosis of Steam Turbine

Nan Yang, Guigang Zhang, Jian Wang

202020 citationsDOI

Abstract

Because the fault diagnosis of steam turbine and other important power generation equipment mostly depends on the diagnosis knowledge, this paper proposes a fault diagnosis method based on Knowledge Graph and Bayesian Network, which integrates the empirical knowledge into the intelligent diagnosis process. The diagnosis method is divided into two steps. First, based on deterministic reasoning, the possible fault devices and fault modes are separated from the Knowledge Graph. Then, based on the above inference results, the Bayesian network model is automatically generated from the Knowledge Graph. Finally, in the process of troubleshooting, the maintenance personnel interact with the inference process of Bayesian Network and finally determine the cause of the fault. The effectiveness of this method is verified by the application in two turbine fault diagnosis cases.

Topics & Concepts

TroubleshootingBayesian networkComputer scienceFault (geology)InferenceGraphArtificial intelligenceData miningProcess (computing)Steam turbineReliability engineeringMachine learningEngineeringTheoretical computer scienceOperating systemMechanical engineeringSeismologyGeologyFault Detection and Control SystemsEngineering Diagnostics and ReliabilityRisk and Safety Analysis