Competitive adaptive reweighting key data extraction method for Raman spectrum analysis of insulating oil

Abstract

A competitive adaptive reweighting key data extraction method for Raman spectrum analysis of insulating oil is characterized by comprising the following steps: 1, obtaining spectral data by actual measurement; 2, processing all original data, and extracting a PLS subset from actually measured spectral data through a partial least square method; 3, screening out singular samples by using Monte Carlo interaction verification; and 4, screening variables of the Raman spectrum of the insulating oil for multiple times by adopting a CARS method. According to the Raman spectrum key variables screenedby the method, unnecessary wavelength points in the Raman spectrum can be removed, and the colinearity of Raman spectrum data is removed. Due to limit by the Raman signal detection limit of an aging characteristic substance in the oil, characteristic peaks of all aging characteristic substances are difficult to observe by directly detecting the insulating oil, redundant information is removed by utilizing a variable screening means, key variables are obtained, and the relationship between the key variables and the characteristic peaks is analyzed, and thus, effective support can be provided for extraction and discrimination of spectral aging characteristic quantity.

Description

technical field [0001] The invention belongs to the field of on-line monitoring and fault diagnosis of electrical equipment insulation, and in particular relates to an analysis method of Raman spectrum of insulating oil. Background technique [0002] After years of research and development of testing technology, Raman spectroscopy has been proven to be effective in quantitative and qualitative analysis of non-destructive testing in various fields. However, due to the very high detection resolution of the spectrometer currently used to build the spectrum detection platform, the number of wavelength points in the original oil-paper insulation Raman spectrum is very large, often far more than the number of samples. On the one hand, not all wavelength points are useful when analyzing oil-paper insulation samples, and may even cause unnecessary interference; on the other hand, there is serious collinearity among a large number of wavelength data points in the spectrum. Due to th...

AI Technical Summary

Problems solved by technology

[0006] ③Due to the complex composition of insulating oil, various aging characteristics will be produced during the aging process, and its Raman spectral data usually has serious peak overlapping phenomenon, and the sensitivity of aging information contained in some wavelengths in the spectrum is relatively poor;

[0007] ④There are usually multiple correlations (ie, multicollinearity) in Raman spectral information, and the wavelength range detected by the spectrometer is wide, resulting in a large amount of redundant information in the spectrum;

[0008] ⑤In the process of spectral detection, due to the existence of fluorescent substances and fluorescence effects in oil, as well as the influence of spectral instrument noise and cosmic rays, the signal-to-noise ratio (SNR) of the spectrum in certain bands is low, resulting in the poor overall quality;

[0009] ⑥Many external interference factors (such as temperature, humidity, incident angle, etc.) have a great influence on the Raman characteristics of aging characteristics in oil at certain wavelengths, and the concentration of dissolved aging characteristics in oil is very low. The situation of information detection is very unfavorable, so the robustness of the model will be reduced; due to the increase of the number of wavelength points, the estimated variance of the model will expand with the increase of complexity, so it will not only waste system resources, but also reduce the prediction accuracy of the model;

Images