In this work, regression analysis is invoked to predict relationship between varistor microstructure response and temperature for degradation characterisation or assessment of metal oxide surge arresters (MOSA). Accelerated degradation test is conducted on varistor samples—having similar electrical and physical characteristics—at randomly selected temperature points for equal period of time. The reference voltage—measured before and after accelerated degradation test—is used as the degradation criterion for varistor samples. Linear intercept performed on Scanning Electron Microscopy (SEM) micrographs is relied upon to estimate varistor average grain size at each applied thermal stress. The coefficients obtained for the developed regression model are statistically validated using the one-way analysis of variance (ANOVA). Results show that for each 1 °C variation in the accelerating temperature, which corresponds to 52. 6 h of service time, varistor average grain size increases by 0.143 microns.

在这项工作中，回归分析被用来预测压敏电阻微观结构响应与温度之间的关系，以用于金属氧化物避雷器 (MOSA) 的退化表征或评估。加速退化测试是对具有相似电气和物理特性的压敏电阻样品在随机选择的温度点进行相同时间段的测试。在加速退化测试之前和之后测量的参考电压用作压敏电阻样品的退化标准。依靠在扫描电子显微镜 (SEM) 显微照片上执行的线性截距来估计每个施加热应力下的压敏电阻平均晶粒尺寸。使用单向方差分析 (ANOVA) 对开发的回归模型获得的系数进行统计验证。