应用红外光谱和混沌学分析变压器油的老化特性

Analyze Aging Characteristics of Transformer Oil Based on Infrared Spectroscopy and Chaos Theory

为监测诊断变压器油的老化状态,采用开口杯劣化法在120°C恒温下对变压器油进行老化试验,通过红外光谱监测变压器油老化过程中官能团的变化趋势,从微观角度分析变压器油老化机理,试验结果表明变压器油在老化过程中有醛类和酸类化合物生成。并对不同老化程度的变压器油采用尖-板电极局部放电,用混沌-分形理论分析了油中局部放电脉冲波形,通过计算波形的关联维数D(m)和Kolmogorov熵(KE),分析波形的频谱和功率谱密度。油质老化初期,关联维数呈下降趋势,放电信号的K熵值呈上升趋势,油中局放信号进入稳定期;随着老化加剧,关联维数迅速增大,局放波形的复杂性增强。运用混沌特性的变化来识别变压器油的老化状态取得了很好的效果,并为变压器油绝缘老化状态的诊断提供了一种可行的方案。

Aiming at discriminating the aging conditions of transformer oil ,the oil was aged at 120℃ by an openedcup deterioration method. Then infrared spectroscopy was used to analyze thermal aging of transformer oil, and absorption peaks, absorption strength and changes of absorption figures were used to judge the change trend of functional group during aging of oil. The results show that the aldehydes and the acids compound are produced in aging process. The PD pulse waveform of aging transformer oil is analyzed through Chaos - fractal theory. The results of calculating the correlation dimension （D（m）） and Kolmogorov Entropy（KE） and comparing the amplitude spectrum and Power Spectral Density（PSD） show that the correlation dimension presents declining trend, Kolmogorov Entropy of discharge signal presents increasing trend during initial period of oil aging, and the PD signal enters the stabilization period. Along with the aging of transformer oil, the correlation dimension increases rapidly and PD waveform becomes more complex. The chaos characteristic is used to identify condition of transformer oil and presents obvious effect, providing a feasible diagnosis method for the monitor of transformer oil.