基于信息融合技术的大型水轮发电机故障诊断-中南大学学报.PDFVIP

第 38 卷第 2 期 中南大学学报( 自然科学版) Vol.38 No.2 2007 年 4 月 J. Cent. South Univ. (Science and Technology) Apr. 2007 基于信息融合技术的大型水轮发电机故障诊断 贺建军，赵 蕊 ( 中南大学 信息科学与工程学院，湖南 长沙，410083) 摘 要：为了能够从多方面反映水轮发电机组系统状态，实现对水轮发电机组故障模式的自动识别与准确诊断， 将信息融合技术应用于水轮发电机组故障诊断系统。根据故障特征量将故障进行分类处理，采用多个并联的BP 子神经网络进行水轮发电机组故障的局部诊断，获得彼此独立的证据，再运用D−S证据理论融合算法对各证据进 行融合，最终实现对水轮发电机组故障的准确诊断。诊断测试实验证明：采用该诊断系统可有效地提高诊断可信 度，减少诊断的不确定性。 关键词：水轮发电机；故障诊断；信息融合；证据理论；神经网络 中图分类号：V263.6 文献标识码：A 文章编号：1672-7207(2007)02−0333−06 Hydroelectric generating sets fault diagnosis based on information fusion technology HE Jian-jun, ZHAO Rui (School of Information Science and Engineering, Central South University, Changsha 410083, China) Abstract ：Hydroelectric generating sets(HGS) information fusion diagnosis system was built for reflecting the HGS system state in multi-aspects, realizing automatical identification of HGS fault patterns and accurately diagnosing the faults. After fault feature data were classified and processed, several shunt-wound BP networks were used to carry on local HGS fault diagnosis and acquire independent evidences each other. Then D-S evidence theory fusion algorithms were used to fuse evidences. Accurate HGS fault diagnosis was fulfilled finally ．The diagnostic tests prove that the system is good to improve the reliability of the diagnosis and decrease the uncertainty markedly. Key words ：hydroelectric generating sets; fault diagnosis; information fusion; evidence theory; neural network 信息融合的过程包括多传感器监测目标(信息采集) 、