基于紫外成像图像信息的绝缘子表面放电量化方法

Study of Insulator Surface Discharge Quantification Method Based on Ultraviolet Imaging Image Information

紫外成像法是一种有效的非接触式放电检测方法,为采用紫外成像法量化绝缘子的表面放电电信号强度,论文以紫外图像中放电光斑面积的大小作为紫外成像量化参量,试验对比研究了绝缘子放电的紫外成像图像和电脉冲信号特征,分析得到了观测距离为9m和紫外成像仪在几种典型增益下的光斑面积与脉冲电流的峰值、脉冲电流平均值的关系。研究了可见光通道与紫外通道的图像参数随距离增加时的变化规律,研究发现两者具有近似相同的衰减特性,在此基础上提出了采用可见光通道的信息用于估计成像区域面积的方法,避免了不同距离下检测到的结果没有对比性的问题,论文最后采用了最小二乘回归支持向量机实现了对放电强度的估计。

Ultraviolet （UV） imaging is an effective non-contact discharge detection method. In order to quantify the insulator surface discharge intensity with UV imaging method, in this paper the discharge facular area in UV image was selected as a UV image quantity parameter. Then the discharge UV imaging and electrical pulse signal characteristics were compared through experiment, and the relation curves of pulse amplitude, average pulse current with the facular area were obtained under the conditions of observe distance was 9m and several typical device gains. The vary characters of UV and visible images parameters with the increased detection distance is researched, and it shows that the parameters of UV and visible images had the approximate attenuation characteristics. Based on the above research work, a discharge imaging region area estimation method by means of visible image information was proposed. With this method, it avoides the problem that under different observe distance the detection result was lack of comparison. Finally, the support vector regression machine model is established and realized the estimation of discharge intensity.