Analyses of different approaches for detecting, classifying and locating faults in a three-terminal VSC-HVDC system

This paper presents three intelligent approaches for fault detection, classification and location in a three-terminal VSC-HVDC system. The proposed approaches are all based on ANNs, although each one has a specific pre-processing step to process the DC current samples. The approaches were implemented in Matlab, while the adopted three-terminal VSC-HVDC system was modelled in PSCAD. A large number of fault cases differing with respect to fault location, fault type and fault resistance were specifically generated to develop, test and compare the solutions presented here. The observed results encourage the application of the proposed approaches in practical situations, as a good performance was verified regardless the fault case characteristics. In terms of fault detection and classification functions the response time is always within 2 ms, while for fault location function the average accuracy is 1.7%.