局部放电图像组合特征提取方法  

METHOD FOR EXTRACTION COMBINATION FEATURES OF PARTIAL DISCHARGE IMAGES

作　　者：李剑[1] 孙才新[1] 杜林[1] 崔雪梅[1] 李道武[1] 

机构地区：[1]重庆大学高电压与电工新技术教育部重点实验室,重庆400044

出　　处：《高电压技术》2004年第6期11-13,共3页High Voltage Engineering

基　　金：重庆大学骨干教师资助计划项目

摘　　要：研究了局部放电图像组合识别特征提取和反向传播算法神经网络分类器设计方法 ,根据变压器局部放电在线监测的要求 ,设计了 5种放电模型并进行了模拟实验。 5种放电模型数据识别结果说明 :与分别采用分形特征和统计特征的识别结果相比 。PARTIAL DISCHARGE (PD) PATTERN RECOGNITION IS AN IMPORTANT METHOD FOR INSULATION DIAGNOSIS OF ELECTRICAL EQUIPMENT. IN THIS PAPER, THE COMBINATION FEATURES AND BACK PROPAGATION NEURAL NETWORK(BPNN)ARE STUDIED FOR PD PATTERN REMOTE RECOGNITION SYSTEM. ACCORDING TO THE REQUIREMENT OF ON LINE PD MONITORING FOR TRANSFORMER, SEVERAL DISCHARGE MODELS ARE DESIGNED AND THE RELEVANT EXPERIMENT METHODS ARE PROJECTED. WITH DISCHARGE MODEL TESTES, A LOT OF DISCHARGE SAMPLE DATA IS ACQUIRED. IT CAN BE SHOWN FROM ANALYSIS OF THE RECOGNITION RESULTS OF LARGE QUANTITIES OF THE PD SAMPLES THAT THE HIGHER RECOGNITION RATIO IS ACHIEVED IN USE OF COMBINATION FEATURES THAN THAT IN USE OF FRACTAL FEATURES OR STATISTICAL FEATURES SEPARATELY.

关 键 词：局部放电 模式识别 组合特征 反向传播算法 神经网络 

分 类 号：TM835[电气工程—高电压与绝缘技术]