A Comparison of New Methods Based on ANNs for Detecting and Locating Faults in MTDC Systems

This work presents a protection scheme based on Artificial Neural Networks (ANNs) for fault detection and location in Multi Terminal Direct Current (MTDC) systems. To perform these tasks only local signals are used and communication links are not required. Three different schemes based on ANNs are evaluated (DC current samples, DC current spectrum and DC current wavelet coefficients), allowing a comparison between the advantages and disadvantages of each one. The proposed method has been tested for different fault locations and fault resistances in a four terminal MTDC system with three DC lines. The adopted MTDC system was modeled in ATP-Draw, while the proposed algorithm was developed in Matlab. The results show that the proposed scheme is able to properly detect and locate a fault within a few milliseconds.