Fault detection and classification of VSC-HVDC transmission lines using
a deep intelligent algorithm
Abstract
Considering the sensitivity of HVDC-transmission system protection and
the difficulty in identifying high-resistance earth-faults,this paper
presents three methods for fault location and classification in VSC-HVDC
transmission lines.These methods are evaluated in terms of efficiency
and reliability.The current and voltage signals obtained from the
network are pre-processed by performing DWT,and then using feature
extraction methods,special and unique features are extracted for
different states of the signals,and then using these features and
proposed algorithms,network learning was performed to detect faults.In
addition,the effectiveness of the proposed plan has been confirmed for
different fault scenarios related to extensive changes in fault
resistance,fault starting angle and fault location.These algorithms were
also investigated in unwanted noise-conditions and the reliability of
these algorithms was confirmed.In this article methods,k nearest
neighbor(KNN),support vector machine(SVM) and deep-neural network(DNN)
have been investigated.The strength of this research is use of a new
method of extracting features from the fractal dimension,which has been
able to provide outstanding capabilities that can lead to improved
diagnosis with a small number of study data and different conditions.The
main advantages of proposed-method are higher speed and accuracy than
conventional methods.The test results show that the proposed method can
reliably and accurately identify and classify high-impedance faults.