油纸绝缘老化状态判别的拉曼光谱特征研究

获取能准确反映油纸绝缘老化的有效特征量,对实现油纸绝缘老化的准确诊断具有重要意义.表面增强拉曼光谱技术在油纸绝缘老化诊断领域已表现出一定的应用潜力.对25#矿物变压器油和普通牛皮纸构成的油纸绝缘试品进行了加速热老化实验,结合共聚焦拉曼光谱检测平台和银纳米片表面增强基底获取其拉曼光谱信号,从多种角度提取了拉曼光谱特征量....

Study on the Raman Spectral Characteristics in Ageing Condition Discrimination of Oil-Paper Insulation

Obtaining effective characteristics that can accurately reflect the ageing condition of oil-paper insulation is of great significance to the accurate diagnosis of oil-paper insulation systems. Surface-enhanced Raman spectroscopy (SERS) has shown application potential in oil-paper insulation ageing diagnosis. This paper performed accelerated thermal ageing test for oil-paper samples composed of conventional 25# mineral oil and ordinary Kraft paper. The Raman spectral signal was obtained by the Confocal laser Raman spectroscopy (CLRS) platform and the silver nano-plates. Various methods extracted the features of Raman spectrum. The competitive adaptive reweighting algorithm was used to extract the key variables of the spectrum. The results correspond to the main characteristic peaks of aging characteristics of oil-paper insulation. The Voigt function is used to analyze the spectrum. A correlation between the profile parameters of the analytical peak and the aging degree of oil-paper insulation was obtained. The aging degree of the samples was classified according to the polymerization degree of the insulating paper. The first eight principal components and their loadings were correlated with aging characteristics and aging degree and can accurately classified the samples. Finally, wavelet packet energy entropy analysis was carried out to analyze the energy change of Raman spectrum in the aging process of oil-paper insulation. This study provides a basis for the application of surface enhanced Raman spectroscopy in the diagnosis for the ageing condition of oil-paper insulation. It lays the foundation for the rapid and non-contact comprehensive diagnosis for the fault and ageing condition of oil-paper insulation equipment in the field.