Timely detection and discovery of abnormal temperatures is the key to improving the service life of cables. In this paper, the thermal-force coupling field (temperature field and stress field) of cable joints is mathematically modeled to quantify cable energy changes. The infrared detection technology is used to measure the temperature of the cable and generate a thermogram to investigate the trend of the cable’s thermal state. The machine learning algorithm “Support Vector Machine” (SVM) and Sparrow Search Algorithm (SSA) are integrated to identify the local thermal characteristics of cables from three aspects: feature extraction, parameter optimization and pattern recognition. Simulation experiments are conducted to test the quality of the proposed detection technique. The results show that when the cable has localized thermal aging, the frequency response of the channel in this part are less than 0dB, which can not transmit the signal normally, and need to be maintained in time. The technology in this paper can realize the effective detection of cable thermal characteristics and reduce the risk of cable faults.
