摘要
对于变压器油中局部放电超高频测量系统所得到的局部放电的特征量,首先,选择优先权较高的6个特征量作为自适应神经模糊推理系统(ANFIS)的输入量,其次,构建6输入单输出的ANFIS,它采用了Takagi-Sugeno模糊系统的if-then规则,利用梯度下降和最优平方估计相结合的混合学习算法进行训练。最后,对于模型的有效性进行了检验,检验结果表明利用ANFIS系统进行局部放电的模式识别是可行的。
Features of partial discharge （PD） in transformers is extracted by Ultra-High-Frequency PD monitoring system. Firstly, according to the priorities of features, six of them are selected as the input variables of the adaptive neuro-fuzzy inference system（ANFIS）. Next, the ANFIS model of six inputs and one output is presented. Takagi and Sugeno＇s fuzzy if-then rules are used. Hybrid learning algorithm combining the gradient method and the least squares estimate （LSE） is adopted to train the ANFIS. Finally, the availability of ANFIS is tested. The results showed that the method based on ANFIS is feasible in PD pattern recognition.
出处
《电工电能新技术》
CSCD
北大核心
2005年第4期30-33,共4页
Advanced Technology of Electrical Engineering and Energy
基金
国家自然科学基金资助项目(50377034)
