Evaluation Of The Insulation Aging Of 35kV And Below XLPE Power Cables

Cables that have been in service for over 20 years in Shanghai, a city with wet weather condition and abundant surface water, failed more frequently and induced different cable accidents. This necessitates researches on the insulation aging state of low and medium voltage cables working in special circumstances.Various insulation tests were performed to study the characteristics and detection methods of cable insulation aging in this paper, including water tree investigation, tensile test, dielectric spectroscopy test, thermogravimetric analysis(TGA), fourier transform infrared spectroscopy(FTIR), and electrical aging test. In order to reflect various possible aging phenomena, many test parameters were proposed for characterization of water treeing degree, thermo-oxidative aging degree, degradation level, electrical aging resistance ability, thermal decomposition difficulty and aging state of cable insulation respectively, including water tree content and maximum water tree length, high frequency dielectric loss peak value and carbonyl index, tensile strength and elongation at break, accumulative breakdown strength, activation energy. A new parameter ln(t/τT) was also proposed to characterize the insulation aging state of field aged cables based on equivalent thermal history parameters. The results of analysis of variance(ANOVA) performed on the insulation test data of cable samples with different aging states show that all the proposed test parameters can indentify field aged cables of different aging degree. In addition, a large number of typical bowtie water trees were found in field aged cables, indicating that water treeing is a typical aging characteristic of in-service low and medium voltage cables in shanghai. The results of FTIR analysis on the insulation of different locations along cable radial direction show that oxygen is the main factor causing the thermo-oxidative aging of cable insulation. As shown in the chemical reaction kinetics analysis along with regression analysis, 60% can be taken as the criteria of weight loss percentage for activation energy calculation using Ozawa or Toop method and the activation energy is able to be calculated quickly from the TG curve at the heating rate of 20℃/min by Coast-redfern method(S9 mechanism). The results of accelerated water treeing test show the water treeing of cables tends to be accelerated as the degree of in-service aging increases.The tests mentioned above were performed with 466 samples from about 300 cable lines in Shanghai. The correlation analysis between every two test parameters indicate that: both water tree investigation and tensile test are effective and reliable; the severer the water treeing, the lower the mechanical properties; the rule that the increase of microscopic imperfections in solid dielectrics improves the withstanding voltage is verified; the severer the aging of cable insulation, the higher the degradation rate; the thermo-oxidative aging can prevent water trees from growing along the cable radial direction; crosslinked network of XLPE is gradually broken up during the thermal oxidative aging process; ln(t/τT) can be used as an effective test parameter to characterize the aging of field aged cable insulation. The results of two-sample t-Test and the one way ANOVA of each parameter indicate that: there are significant differences in the carbonyl index and tensile strength for different laying methods, the duct cables are more sensitive to thermo-oxidative aging than direct-buried cables, but the insulation degradation is severer in direct-buried cables than duct cables; for different sampling sections and cable-laying areas, there are significant differences only in tensile strength and breakdown strength, with the same results obtained by ANOVA analysis, which confirms the reliability of the two functional parameters for insulation diagnosis of cables.The characteristic parameters were selected from the ones which have good results in statistical analysis, including water tree content, tensile strength, accumulative breakdown strength and activation energy. The data of test parameters showing strong correlation in regression analysis was combined into a two-dimensional data set for fuzzy clustering. The obtained clustering centers were taken as the thresholds of membership functions. On the basis of the results of fuzzy clustering and experience of experts, the rules of fuzzy diagnosis were established. Using intensity transfer method, a model for insulation diagnosis of cables was presented. It is shown by case study that the model is applicable, and that there is no correlation between the cable service time and the insulation aging state. According to the insulation test data and the diagnosis model, a piece of software for cable maintenance was developed. Using the software, the insulation diagnosis was performed on all the cable samples. The statistical conclusion of the diagnosis results of cables is same as that of insulation test data, which proves that the model is reliable.For the same in-service aging condition, the aging time is found to have significant multiple linear regression relationship with the water tree content, accumulative breakdown strength and activation energy by step-wise regression. With the aid of the multiple linear regression analysis, a multivariable aging model of cable insulation was established. The model parameters depend on the specific field condition. The case study proves that the aging model is applicable.