基于改进不变矩与概率神经网络的水电机组轴心轨迹特征提取研究  

Research on the Feature Extraction of Hydropower Units Shaft Orbit Based on Improved Moment Invariants and PNN

作　　者：袁喜来 刘东 胡晓 刘冬[2] YUAN Xi lai;LIU Dong;HU Xiao;LIU Dong(Production Technology Department, Hubei Energy Group Co., Ltd., Wuhan 430072, China;School of Power and Mechanical Engineering, Wuhan University, Wuhan 430072, China)

机构地区：[1]湖北能源生产技术部,武汉430072 [2]武汉大学动力与机械学院,武汉430072

出　　处：《中国农村水利水电》2019年第6期149-152,158,共5页China Rural Water and Hydropower

基　　金：国家自然科学基金项目(51379160)

摘　　要：在水电机组状态监测与故障诊断中,轴心轨迹是反映机组运行状态的重要特征。提出了将图形改进不变矩算法与概率神经网络相结合的方法,运用改进不变矩算法对水电机组几种不同运行状态下的转子轴心轨迹进行特征提取,得到相应的特征矩向量,构建概率神经网络进行训练分类,并结合电站实测数据进行了验证。结果表明,该特征提取与分类方法简单稳定,对不同形状的轴心轨迹具有较高的区分度和较好的识别率,可以为水电机组故障诊断提供有效依据。In the state monitoring and fault diagnosis of hydropower units,the shaft orbit is an important feature that reflects the operating state of the unit. In this paper,the method of combining the improved invariant moment algorithm with the PNN is proposed. The improved invariant moment algorithm is used to extract the shaft orbits of hydropower units under different operating conditions,and the corresponding characteristic moment vectors are obtained. A PNN is constructed for classification. The method is verified with the measured data of the power station. The results show that the feature extraction and classification method is simple and stable,and has a high degree of discrimination and good recognition rate for different shapes of the shaft orbit,which can provide an effective basis for fault diagnosis of hydropower units.

关 键 词：水电机组 不变矩 概率神经网络 轴心轨迹 特征提取 

分 类 号：TV734.21[水利工程—水利水电工程]