摘要
紫外成像技术可以检测到污秽绝缘子的局部放电影像,是一种非接触的检测方法,而等值附盐密度(ESDD)可用来划分绝缘子的污秽等级。为更准确地进行绝缘子污秽度检测,提出1种基于紫外成像技术的绝缘子附盐密度检测方法,据此判别绝缘子的污秽状态。在人工气候室进行了绝缘子人工污秽实验,基于视频与图像处理技术研究了ESDD分别为0.05、0.08、0.15、0.25、0.35、0.40 mg/cm2,相对湿度分别为60%、65%、70%、75%、80%、85%、90%、95%条件下绝缘子的放电情况,提出了描述污秽绝缘子紫外图像中放电特性的稳定性光斑平均面积、间歇性光斑平均面积、光斑平均面积3个特征参量。结果表明:随着湿度和ESDD的增加,绝缘子污秽放电光斑面积增大,放电更加剧烈,当相对湿度为95%且ESDD为0.40 mg/cm2时,光斑平均面积占总面积高达18.4%;利用最小二乘支持向量回归机建立绝缘子ESDD检测模型,并用粒子群算法优化参数,当迭代进行到第55次时收敛,得到较好的适应度为0.002 7;最后通过实验验证,该模型准确率达到95%。
The UV imaging technology is a non-contact charging detection method that can easily and quickly detect contamination insulator discharge situation, while equivalent salt deposit density（ESDD） can divide pollution levels of insulators. To improve the accuracy of insulator contamination detection, we put forward a prediction method for insula- tors ESDD based on the UV imaging technology to identify insulator pollution status. First, we performed insulator artificial contamination experiments in the artificial climate chamber. The video and image processing technology was adopted to study the insulator discharge under ESDD of 0.05, 0.08, 0.15, 0.25, 0.35, 0.40 mg/cm2 and humidity of 60%, 65%, 70%, 75%, 80%, 85%, 90%, and 95%. Then we put forward three characteristics parameters of UV imaging insula- tor discharge, they are the stability spot area, the intermittent spot area, and the average spot area. The results show that, with the increase of humidity and ESDD, the spot area of insulator contamination discharge will increases, and the dis- charge is more severe. Humidity is 95% and ESDD is 0.40 mg/cm2, the average area of the spot is as high as 18.4%. Moreover, the ESDD prediction model is established by using least square support vector regression model and the par- ticle swarm optimization（PSO） algorithm to optimize the parameters. When the iteration is carried out to fifty-fifth times, the degree of adaptation is 0.0027. Finally, the experiments verify the model with a high accuracy rate of 95%.
出处
《高电压技术》
EI
CAS
CSCD
北大核心
2015年第11期3561-3568,共8页
High Voltage Engineering
基金
国家自然科学基金(51177109)
电力设备电气绝缘国家重点实验室项目(EIPE14211)~~
