The Study Of Estimating Degradation Of Rubber Insulation Low-voltage Cable Based On Frequency Dielectric Spectroscopy

The low-voltage cables play an important role in transporting energy in power system; the degradation of insulation directly affects the safety of electrical equipment and maintains stable operation of power system, so that it has great application value and widespread attention for the research of detection degradation cable insulation. In this paper, combined with the characteristics of frequency domain spectroscopy, a new diagnostic method called nondestructive detection was proposed based on the analysis of single-factor and multi-factor accelerated thermal aging test for low-voltage cables. The aging equation of insulation was built by this method, and the degradation of low-voltage cables was analyzed by using neural network. This nondestructive method could apply to detect the condition of low-voltage cables, which had practical significance for protecting safety.In this paper, the aging mechanism of low-power cables had been studied in detail, and the typical aging factors were pointed out in the process of aged insulation. Contrary to aging characteristics of insulation materials, the advantages and focus of the commonest diagnostic methods were discussed. Based on the equivalent structural model and aging model, the electrical parameters within different models were calculated. Meanwhile, this paper introduced the principle and characteristics of frequency dielectric spectroscopy, and analyzed the measurement characteristics and main influencing factors of frequency dielectric spectroscopy.The cable samples were designed into the accelerated thermal aging and thermal-water aging test, and the samples at different aging condition were obtained, respectively. The frequency domain spectroscopy was mainly measured and compared with mechanical and chemical curves to find out internal relation, and the sensitive measurement range went hand in hand with degradation was obtained. The sensitive range of low frequency was determined from10-2Hz to1Hz, and the integral characteristics of frequency domain spectroscopy were analyzed. The results showed that the integral values of frequency domain spectroscopy had obviously temperature characteristics which were in accordance with the regulation of mechanical properties. Through by thermal-water aging testing, the relationship between degradation and frequency domain spectroscopy was estimated. The results showed that the integral values of frequency domain spectroscopy were better responded the degradation and moisture content of insulation lay a foundation for the thermal aging model foundation of frequency domain spectroscopy.By analyzing the experimental data of thermal aging testing, the integral values were chosen as the characteristics of aging equation. After the reference temperature was selected, the temperature conversion method was obtained by deducing the time multiplicative shift factor with the same characteristics value at different aging temperatures; combined with integral equation of the deterioration tendency under reference temperature, the aging model of integral of frequency domain spectroscopy was derived by comprehensively analyzing with temperature conversion. And then, the mutual influence function of multi-factor was derived and the aging equation was fixed. Theoretical analysis and measurement results showed that the method of dielectric spectroscopy in low frequency was accurate and had good prediction, and the effectiveness of aging equation was checked by the actual operation data.Due to the environment of low-voltage cables, complex environment factors were analyzed by using artificial neural network with good prediction, and the effection of multi-aging-factor of accelerated aging model and conductor temperature were built. The feed forward neural network with back propagation error of three-layer was proposed, and the predicted data was measured and analyzed by Matlab on the basis of measurement. According to measurement errors, the effectiveness of aging characteristics and the veracity of conductor temperature were deduced. The results proved that the frequency domain spectroscopy which was trained by Artificial Neural Network effectively reflected the degradation and accurately predicted the conductor temperature.The method was based on frequency dielectric spectroscopy at low frequency range to assess the degradation of low-voltage cables, can effectively reflect the aging characteristics, and quantitatively analyze the aging condition of low-voltage cables with rubber insulation, so that this nondestructive detection method can be applied to estimate low-voltage cables in field.