A Fast and Accurate Fault Location Technique for High Voltage Direct Current (HVDC) Systems Une technique rapide et précise de localisation des défauts pour les systèmes de courant continu à haute tension (CCHT)

To minimize the outage time and costs associated with faults on high voltage direct current (HVdc) transmission lines it is critical to locate faults in an accurate and sufficiently fast manner. Current fault location techniques based on artificial intelligence (AI) are accurate but require fault data from rectifying and inverting ends. This necessitates a communications system and incurs high computational burdens. Therefore, a novel fault location technique is proposed that requires fault data only from one end, eliminating the need for a communication system. It employs support vector machine (SVM) algorithms to reduce the time needed to locate faults through fault classification. After classification, Gaussian process regression (GPR) is used for location identification. The proposed technique is tested under real time simulation conditions. The test results show the SVM can classify different fault types with an accuracy of 99.7%, while the GPR is able to locate faults within 0.5197 s with a root mean square error (RMSE) value of 6.52e−5%. The performance of the technique is further investigated under varying fault impedance levels. The results show the proposed technique is robust, even under high impedance fault conditions.