Artificial Neural Network-based Fault Location in Terminal-hybrid High Voltage Direct Current Transmission Line

Abstract

In this article, a fault location technique based on artificial neural networks (ANN) for Terminal-Hybrid LCC-VSC-HVDC has been assessed and scrutinized. As is known, in conventional HVDC systems (LCC-based and VSC-based HVDCs), the same type of filter is used on both sides due to the use of similar converters in both sender and receiver terminals. In this article, it is concluded that due to the use of two different types of converters at the both ends of the utilized Terminal-hybrid LCC-VSC-HVDC system, and the use of different DC filters on both sides, fault location using positive and negative pole currents of the rectifier side has much better results than the rest of input signals. Therefore, it will be finalized that by increasing and designing suitable DC filters on the transmission line of HVDC systems, fault localization matter will be remarkably and surprisingly facilitated. Nowadays, the fault location of HVDC transmission lines with a value of more than 1% is generally discussed in most articles. In this research, the fault location with a value of 0.0045%, i.e., a distance of 22.5 meters from the fault point in the most satisfactory case is obtained, which shows the absolute feasibility of the ANN along with the wavelet transform. To validate the proposed method, a ±100 KV, Terminal-hybrid LCC-VSC-HVDC system is simulated via MATLAB. The outcomes verify that the proposed technique works perfectly under various fault locations, resistances, and fault types.