Extreme Learning Machine Approach for Transmission Line Protection using Wavelet Singular Entropy

This research work presents a novel approach to detect faults in transmission lines. The proposed technique combines two methods: wavelet singular entropy and extreme learning machine. After finding wavelet packet transform, singular value decomposition is applied. The extreme machine learning technique is applied to the entropy values. The proposed approach is evaluated on a microgrid system using MATLAB software. The results show that it can detect various types of faults. Furthermore, it is a simple technique that requires less complexity and can operate at a lower sampling rate.