Fault Detection, Classification, and Location in VSC-HVDC Systems Using Metaheuristic Optimized ANFIS and Hilbert Huang Transform Method

Identifying, classifying, and locating various faults in HVDC networks can significantly reduce maintenance costs in these systems. For this reason, intelligent methods for detecting and locating faults with high accuracy and speed have recently been the focus of system planners and operators. This paper proposes an intelligent fault detection, classification, and location scheme based on the adaptive neuro-fuzzy inference system (ANFIS), whose training and testing data are generated using the Hilbert-Huang (HH) transform-based feature extraction method. DC link current is measured as a fault signal, and its statistical characteristics are extracted employing HH transform. Furthermore, for training and testing ANFIS, a meta-heuristic algorithm with dynamic search capability and high convergence speed called the Modified Cuckoo Search (MCS) Algorithm is utilized. In addition, two conventional training methods, i.e., least square estimation (LSE) and gradient descent (GD), are applied to train ANFIS, and its performance is compared with the MCS algorithm. The numerical analysis results validate the accuracy of 97.37% in classification and 98.39% in location using the proposed framework. Additionally, the comparative study confirms a 12.18% reduction in mean squared error (MSE) value compared with the best result obtained in the literature.