本文开展了变压器油中溶解糠醛的定量分析研究, 基于实验室搭建的拉曼光谱液体检测平台, 对不同浓度糠醛含量的变压器油样进行光谱信号检测及预处理, 运用主成分分析法对光谱数据进行数据降维并结合支持向量机建立油中溶解糠醛含量拉曼光谱定量分析回归模型, 检测下限为0.625 mg/L。结果表明, 拉曼光谱技术结合支持向量机能有效地对变压器油中溶解糠醛进行定量分析, 为实现油中溶解老化特征物的准确、快速检测提供新的手段。

In this paper, the quantitative analysis of dissolved furfural in transformer oil was carried out.Different furfural content of transformer oil samples were detected based on the Raman spectroscopy liquid test platform.After the efficient pretreatment, spectrum data was used principal component analysis method for data dimensionality reduction, and the establishment of furfural content dissolved in oil analysis regression model was based on support vector machine, the detection limit is 0.625 mg/L.The results show that the Raman spectroscopy combined support vector machine for quantitative analysis of furfural dissolved in transformer oil effectively, and it provides a new method for realizing a fast and accurate detection of dissolved matter in oil ageing characteristics.