摘要
传统的电气设备放电强度评估基本是通过相应的放电检测技术结合模糊推理算法实现的,但模糊推理算法中的模糊规则和隶属度函数往往根据专家经验确定,缺乏自适应能力,影响评估结果的准确性。为此,提出一种基于自适应神经模糊系统(ANFIS)算法的电气设备放电强度评估方法。该方法将神经网络的自学习机制引入模糊系统,通过对大量样本数据的训练确定最佳的隶属度函数参数,并自动产生模糊规则,使模型具有较高的准确性和适应性。
Traditional discharge intensity evaluation of electrical equipment is basically achieved by the corresponding discharge detection technology combined with the fuzzy reasoning algorithm.However,the fuzzy rules and membership functions in the fuzzy reasoning algorithm are often deter mined based on expert experience,lacking adaptive ability,which affects the accuracy of evaluation results.This paper presents an evaluation method for the discharge intensity of electrical equipment based on adaptive neuro-fuzzy inference system(ANFIS)algorithm.This method introduces the self-learning mechanism of neural network into the fuzzy system,deter mines the best membership function parameters by training a large number of sample data,and automatically generates fuzzy rules,which makes the model have higher accuracy and adaptability.
作者
罗宇鹰
LUO Yuying(Jiangjin Power Supply Branch,State Grid Chongqing Electrical Power Company,Chongqing 402260,China)
出处
《电工技术》
2019年第19期32-35,40,共5页
Electric Engineering
关键词
局部放电
紫外放电检测
脉冲电流
ANFIS
partial discharge
ultraviolet discharge detection
pulse current
ANFIS
