基于拉曼光谱评估变压器油–屏障式绝缘老化状态
Evaluation of Transformer Oil-Barrier Based on Raman Spectroscopy Insulation Aging State

变压器油–屏障式绝缘老化状态的准确评估对电力系统的安全可靠运行至关重要。本工作基于拉曼光谱检测平台及自行搭建的老化设备，对45#绝缘油与牛皮纸组成的绝缘系统进行加速热老化实验，依据对热老化样本的取样时间，热老化样本分为6类，采用支持向量机对原始数据进行分类，并用网格法、粒子群算法、遗传算法优化支持向量机参数。由于样本数量多以及光谱维度高，优化支持向量机参数时间较长，为进一步提升分类准确率以及分类速度，对原始光谱数据使用非对称重新加权惩罚最小二乘法、Savitzky-Golay平滑、主成分分析等预处理方法，消除基线漂移与环境噪声的影响，并且降低分类时间，提高分类准确率，该工作为变压器油–屏障式绝缘老化状态的有效评估提供技术支持和理论依据。
It is critical to accurate assessment of the aging state of transformer oil-barrier insulation for the safe and reliable operation of power systems. In this study, the accelerated thermal aging experiment was carried out on the insulation system composed of 45# insulating oil and kraft paper based on the Raman spectrum detection platform and self-built aging equipment. Based on the sampling time of thermal aging samples, the thermal aging samples were divided into 6 categories, and the support vector machine was used to classify the original data, and the parameters of the support vector machine were optimized by Gridsearch, Particle Swarm Optimization (PSO), and Genetic Algorithm (GA). It takes a long time to optimize the parameters of the support vector machine due to the large number of samples and the high spectral dimension. In order to further improve the classification accuracy and speed, asymmetric reweighting penalty least squares, SG smoothing, PCA and other preprocessing methods are used for the original spectral data, which eliminate the influence of baseline shift and environmental noise, and reduce the classification time, and improve the classification accuracy. The research paves a way to effectively evaluate the aging state of transformer oil-barrier insulation.