Performance evaluation of a new transmission line directional module using field data

This paper presents a neural network-based technique to identify the direction of a fault on a power transmission line. A novel network using a subset of voltage and current information as its inputs is presented. The network uses samples of voltage and current signals to learn the hidden relationship existing in the input patterns. The network performance is tested with real world fault data. Performance studies results show that the proposed network is able to detect the transmission network fault direction correctly and rapidly. Some of these test results are presented in the paper.