Non-Destructive Testing of Crystallinity and Breakdown Performance Assessment of XLPE Based on Near-Infrared Spectroscopy

With the rapid development of China's urban power grids, the problem of power cable insulation aging is becoming more and more prominent. Traditional testing methods have limitations such as insufficient sampling and long cycle time, and new testing techniques are urgently needed. Near-infrared (NIR) spectroscopy, as a non-destructive, rapid, and efficient testing method, has significant advantages in assessing the insulation status of power cables. In this paper, the correlation between AC breakdown field strength and crystallinity of XLPE is investigated, and its Pearson correlation coefficient is 0.813, which is an extremely strong correlation. A principal component regression (PCR) algorithm was used to establish a predictive model for the crystallinity of XLPE cable insulation based on NIR spectroscopy, enabling non-destructive prediction of crystallinity and subsequent assessment of its breakdown performance. By applying the random frog (RF) algorithm for wavelength selection in the NIR spectrum, the accuracy of the model was further enhanced. The RF-PCR model achieved a coefficient of determination of 0.97 and a root mean square error of 0.45 %. Validation on the test dataset showed that the percentage error was only 0.78 %, demonstrating high reliability.