FAULT LOCATION ESTIMATION FOR A BIPOLAR HVDC TRANSMISSION LINE USING DIFFERENT BACK PROPAGATION ALGORITHMS OF ANN

In electrical transmission system uninterrupted supply is a primary demand from system. But every system under go some type of situation where supply gets disconnected for some time based on severity of problem. In HVDC transmission system main problem arises due to line to ground fault. Now the first step to resolve this situation is to locate the exact location of fault occurrence. During this situation for a line long enough, it is very difficult to manually check the whole because that will take long time. Hence a technique is to be derived which has the ability to find the location so as to do the required maintenance as soon as possible and continue the supply again. In this work an attempt is made to resolve this problem with the help of artificial neural network. Two different algorithms of ANN are utilized in this work to calculate fault location of Bipolar HVDC transmission line. This a bipolar HVDC transmission line is simulated for fault at a step of 1 km in PSCAD/EMTDC software and data of both sending end and receiving end is collected. This data is input data for Neural Network. In the present study a line of DC voltage of 500kV 816 km is taken. This line is a prototype of India first bipolar line i.e. Rihand-Dadri HVDC line. After collection of data further modelling of neural network model is done in neural network toolbox in MATLAB software. The results show that BR will give more accurate results than LM thus proving to be more efficient method.