Research on splicing method of temperature distribution curve of EHV–UHV long string porcelain insulators

Abstract

The infrared thermal image detection method has unique advantages in the field of low zero value diagnosis of porcelain insulators. However, since the length of the extra–ultra-high voltage (EHV–UHV) porcelain insulator string can reach more than 10 m, it is difficult for common industrial thermal imagers/airborne lenses to clearly capture the infrared spectrum of the full string of EHV–UHV porcelain insulators. In order to solve this problem, this paper proposes a two-segment temperature curve splicing method based on weighted data fusion. The method avoids the problem of missing targets existing in conventional infrared image splicing methods and has higher fitting accuracy than curve splicing methods such as rounding splicing and midpoint splicing. This paper takes a 1000 kV insulator string as an example to illustrate the basic algorithm and process of the proposed curve splicing method and takes a 500 kV insulator string as an example to carry out a comparative study of various curve splicing methods, which verifies the effectiveness and feasibility of the proposed method.