A novel substation fault diagnosis and maintenance method based on ontology and Bayesian network

Based on ontology and Bayesian network, a novel substation fault diagnosis and maintenance method is proposed in the paper. The method firstly constructs the ontology model of substation fault diagnosis, then, generalizes more local ontologies. In general, to mine high-level ontologies are more significant than to identify low-level concepts in process of fault diagnosis. Hence, hierarchical substation fault diagnosis is constructed based on ontology and local ontologies as well as more low-level concepts. Then according to hierarchical ontology structure and fault patterns, Bayesian network is constructed. After the key risk factors are identified and minor risk factors are ignored, a simplified Bayesian network is generated. Based on such simplified Bayesian network, the key risk factors are identified. Hence, the maintenance works become easy and effective. The study indicates the proposed method can improve diagnosis levels and serviceability, and is an effective substation fault diagnosis and servicing method. In the end, a practical application in substation fault diagnosis shows the effectiveness of the method.