A shunt Compensation Impact on UHVTL Distance Relay using Machine Learning in Discrete Wavelet Classifier

الملخص

This paper presents an intelligent data mining approach for the Machine Learning-Adaptive Distance Relay (ML-ADR) fault classification model development, using hybrid discrete wavelet multiresolution analyses and machine learning (DWMRA-ML) algorithm on extracted 1-cycle short circuit transient voltage and current signals to discover the hidden useful knowledge that is deployed in the modification of existing ADR. The transmission line system, where the zone-3 element is protecting against the far-end fault is run with and without an integrated shunt compensating device midpoint along the line. The uniquely extracted 29 features across 2,560 fault sources from both faulty and healthy protected lines to build a historical fault database that is deployed for ML-ADR fault classification model development for effective short circuit fault detection, classification, and trip decision time reduction of the zone-3 protective element. The prior result from the Mat lab model of the adaptive numerical distance relay connected on midpoint- shunt compensator integrated transmission line system does indeed establish that the existence of the under-reach effect for zone-3 far-end short circuit fault causes wrong impedance estimation.

The ML-ADR provides the best integrated fault classifier model with the lowest mean absolute error value of 0.0009, this model satisfied and finally meets the objectives of the desired ADR.