内蒙古电力技术2025，Vol.43Issue(3)：54-61,8.DOI:10.19929/j.cnki.nmgdljs.2025.0035
可逆式抽水蓄能机组故障诊断方法及应用
Fault Diagnosis Method of Reversible Pumped Storage Units and Its Application
摘要
Abstract
This paper proposes a fault diagnosis method for reversible pumped storage units.Firstly,a causal diagram network is constructed using the attribute relationship between unit faults and symptoms.By introducing hidden variable nodes and probability assignments,a fault diagnosis model based on structural causal model(SCM)is established that follows the Leaky Noisy-Or assumption.Then,a counterfactual reasoning diagnosis strategy based on twin network is proposed.According to the topology structure of the fault model,combined with the construction and reduction of twin network,counterfactual reasoning is carried out to calculate the sufficient and necessary factors of the fault candidate set.All candidate faults are sorted to obtain a clear fault causes.Finally,taking the rotor thermal bending fault of a reversible pumped storage unit as an example,the process of establishing a fault diagnosis model based on causal graph network and the counterfactual reasoning diagnosis method based on twin network are introduced in detail.The diagnosis results are consistent with the actual fault cause of the unit,therefore proving the effectiveness of the fault diagnosis method proposed in this paper.关键词
可逆式抽水蓄能机组/故障诊断/结构性因果模型/反事实推理/孪生网络Key words
reversible pumped storage unit/fault diagnosis/structural causal model(SCM)/counterfactual reasoning/twin network分类
信息技术与安全科学引用本文复制引用
李晓波,张利慧,云杰,俎海东,何成兵..可逆式抽水蓄能机组故障诊断方法及应用[J].内蒙古电力技术,2025,43(3):54-61,8.基金项目
内蒙古电力(集团)有限责任公司内蒙古电力科学研究院分公司科技项目"抽水蓄能机组全工况水力激振分析与对策研究"(2023-ZC-10) （集团）
