摘要
文中分析35 kV SF_(6)断路器一相触头的动态电阻曲线的特点,提取出主触头接触行程、主触头电阻平均值、弧触头接触行程、弧触头电阻平均值和主触头分离时刻动态电阻极大值5个特征参数,分析得到随着烧蚀程度加深,主触头接触行程和弧触头接触行程减小,主触头电阻平均值、弧触头电阻平均值和主触头分离时刻动态电阻极大值增大。提出电寿命指示参数描述触头状态,采用数值表示法,1表示电寿命完好,0表示电寿命完结。以5个特征参数为输入,电寿命指示参数为输出,基于BP神经网络建立电寿命评估模型。使用思维进化算法优化BP神经网络的初始权值和阈值,使用L-M算法优化BP网络学习过程中的权值调整量,优化后测试集电寿命指示参数的预测误差为0.0183,相比原始BP网络下降61.31%。最后基于电寿命指示参数提出了触头电寿命的评估分级方法,并给出了对应的检修策略。
In this paper,the characteristic of dynamic resistance curve in one phase of contact of 35 k V SF_(6)circuit breaker is analyzed.Such five characteristic parameters as contacting travel and average resistance of the main contact and of the arcing contact as well as the maximum dynamic resistance at separation of the main contact are extracted.It is concluded through anaysis that with the deepening of the ablation degree,the contacting travel of the main contact and arcing contact reduces,while the average resistance of the main contact and arcing contact as well as the extreme maximum dynamic resistance at the main contact at instant of separation increase.It is proposed to describe contact condition with electrical endurance indication parameters and the numerical representation method is used:1and 0 represent good and end of electrical endurance respectively.The electrical endurance assessment model is set up based on BP neural network with five characteristic parameters as inputs and the electrical endurance indication parameters as output.The mind evolutionary algorithm is used to optimize initial weight and threshold of BP neutral network.The L-M algorithm is used to optimize the weight adjustment during the learning period of BP network.After optimization,prediction error of the collector endurance indicator parameter is 0.0183,which is decreased by 61.31%compared with the original BP network.Finally,the assessment and and grading method of electrical endurance of contact is proposed based on the electrical endurance indicator parameters and corresponding maintenance strategy is given.
作者
马飞越
黎炜
王尧平
王羽
刘博
马文长
刘北阳
黄河
MA Feiyue;LI Wei;WANG Yaoping;WANG Yu;LIU Bo;MA Wenchang;LIU Beiyang;HUANG He(Power Research Institute of State Grid Ningxia Electrical Power Co.,Ltd.,Yinchuan 750001,China;Maintenance Company of State Grid Ningxia Electrical Power Co.,Ltd.,Yinchuan 750001,China;School of Electrical Engineering and Automation,Wuhan University,Wuhan 430072,China;China Electric Power Research Institute,Beijing 100192,China)
出处
《高压电器》
CAS
CSCD
北大核心
2023年第3期44-52,60,共10页
High Voltage Apparatus
基金
国家电网有限公司总部科技项目(5500-202135125A-0-0-00)。
