Fault Location Method Using Local Incipient Current Features for Submarine DC Power Systems

Due to the high fault occurrence probability of submarine cable caused by extreme working environment, accurate fault detection and location is very essential for submarine dc power systems. Electrical faults are usually accompanied by the failures of optical fiber communication, resulting in the poor reliability of the existing methods based on cabled communication. Therefore, we propose a communication-free fault location method using multiple incipient current features. First, the fault modeling and current loop decoupling of submarine dc power systems are carried out to obtain the time-domain description of fault line current. According to the varying trends and rates of incipient fault current with fault distance and fault resistance, multiple time-domain features are utilized to reflect the fault distance meanwhile weakening the adversity of transition resistance. Then, the segmented mapping model between selected features and fault distance is established by using the least-squares support vector machine (SVM) regression (LSSVR) algorithm. Via the integrated fault features and segmented location models, the proposed method can cover the fault location of whole line and offers the advantages of fast speed and high accuracy. Finally, power systems computer aided design/electromagnetic transients including DC (PSCAD) simulation and hardware-in-loop (HIL) experiment results validate the effectiveness of the proposed method under various fault conditions.