融合多源信息处理的发电设备故障智能诊断方法  

Intelligent fault diagnosis method for power generation equipment based on multi-source information processing

作　　者：丁照洋 邱敏昭 郑旭彬 朱海锋 钟伟津 DING Zhaoyang;QIU Minzhao;ZHENG Xubin;ZHU Haifeng;ZHONG Weijin(China Southern Power Grid Energy Storage Co.,Ltd.,Guangzhou 511400,China)

机构地区：[1]南方电网储能股份有限公司,广东广州511400

出　　处：《电子设计工程》2025年第7期108-112,共5页Electronic Design Engineering

摘　　要：为了智能、准确地检测出抽水蓄能电站中存在的设备故障,提出了一种融合多源信息处理技术的发电设备故障智能诊断方法。该方法采用发动机转子各部件的寿命和多种故障的表现形式等不同信息作为数据源,为故障检测算法提供数据支撑。检测算法基于MIC-LGBM和VMD-GRU设计,利用MIC-LGBM生成高精度健康模型,使用MIC选择最相关的工作参数,以LGBM构建健康模型,根据LGBM健康模型和实验数据生成电动机转子部件的寿命退化指数,通过VMD-GRU预测模型实现对发电设备故障的精确预测。在实验对比中,所提算法的故障检测准确率远优于其他主流检测算法,与表现最佳的GRU相比,准确率提高了15.65%,具有良好的工程应用价值。In order to intelligently and accurately detect equipment faults in pumped storage power plants,a power generation equipment fault intelligent diagnosis method integrating multi-source information processing technology is proposed.This method uses different information such as the lifespan of various components of the engine rotor and the manifestations of various faults as data sources to provide data support for fault detection algorithms.The detection algorithm is designed based on MIC-LGBM and VMD-GRU,using MIC-LGBM to generate high-precision health models.MIC is used to select the most relevant working parameters,and LGBM is used to construct a health model.Based on the LGBM health model and experimental data,the life degradation index of the motor rotor components is generated.Through the VMD-GRU prediction model,accurate prediction of power generation equipment faults is achieved.In the experimental comparison,the fault detection accuracy of the proposed algorithm is much better than other mainstream detection algorithms.Compared with the best performing GRU,the accuracy is improved by 15.65%,which has good engineering application value.

关 键 词：发电故障检测 多源信息处理 MIC LGBM VMD GRU 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术] TN92[自动化与计算机技术—计算机科学与技术]