基于人工神经网络的变压器绝缘模型放电模式识别的研究.PDF

21 1 Vol. 21 No. 1 Jan. 2001 2001 1 Proceedings of t he CSEE 2001 Chin. Soc. for Elec. Eng . (2001) 0100210 1 1 1 1 1 2 2 2 姜 磊 , 朱德恒, 李福祺 , 谈克雄 , 覃刚力 , 金显贺, 王昌长, T . C. Cheng ( 1. 清华大学电机工程系, 北京 10008 ; 2. 南加州大学, 洛杉矶 CA900089) ANN BASED DISCHARGE PATTERN RECOGNITION OF INSULATION MODELS OF ELECTRICAL TRANSFORMERS 1 1 1 1 1 JIANG Lei , ZHU Deheng , LI Fuqi , TAN Kexiong , QIN Gangli , 2 2 2 JIN Xianhe , WANG Changchang , T . C. Cheng ( 1. Tsinghua U niversity, Beijing 10008 , China; 2. University of Southern California, Los Angeles CA900089, U. S. A. ) ABSTRACT:The main discharge types in insulation of electrical , t ransformers were analysed, 7 kinds of experimental models , simulating discharges in electrical transformers and 3 kinds of , models simulating interfering discharges in air were designed , and model experiments under some circumstances were per , ( qn formed. Using digital measuring device, the quantityphase in , , q formation of discharge pulse current of models were obtained. T he feature of discharge w as extracted using the 3D pattern , n ) , chart and the artificial neural net works was used to recognize , ANN t he discharge models. The investigation shows that ANN has e [ 1, 2] qn nough ability to recognize different types of discharge of oilpa , per insulation in transformers. , KEY WORDS: transformer insulation; p