摘要
提出了基于T-S模糊神经网络的电晕放电模式识别方法,设计制作了三类电晕放电实验模型,并从采集的电晕放电信号中提取最大值、最小值、均值及其分形维数作为网络的输入特征向量,根据特征向量维数、隶属度函数类型及隶属度函数个数对T-S模糊神经网络的拓扑结构进行分析,将输入神经元个数为4、隶属度函数层为3个高斯型函数的网络确定为电晕放电的模式识别网络,并对此类网络进行训练和测试,结果表明此网络用于电晕放电模式识别是有效的.
T-S fuzzy neural network(T-S FNN) is presented to recognize the types of corona discharge.Three types of corona discharge model are made,the maximum,minimum,mean and fractal dimension are extracted from sampling corona discharge signals and then used as the input vectors of neural network.T-S FNN topological structure is investigated in accordance with the number of input vectors,membership functions and the number of membership functions.The T-S FNN,with 4 input neurons and 3 Gaussian functions,is confirmed as the pattern recognition network,and the network are trained and tested with the extracted features.Results indicate that the method is effective.
出处
《哈尔滨理工大学学报》
CAS
北大核心
2011年第4期45-50,55,共7页
Journal of Harbin University of Science and Technology
基金
国家自然科学基金(51077032)
黑龙江省科学技术攻关项目(GC05A511)
关键词
T-S模糊神经网络
隶属度函数
电晕放电
模式识别
T-S fuzzy neural network
membership function
corona discharge
pattern recognition