Adaline for symmetrical components detection in High Voltage transmission line faults

We present in this paper an ADAptive-LInear-NEuron (Adaline) method for symmetrical components identification in High Voltage (HV) transmission line faults. This method uses a current transformations in a Parks reference frame where the direct and inverse current components are linearly separated. Four Adalines are built in order to learn DQ currents. After the learning process, the Adaline weights are stabilized and allow identifying the RMS values and the phase angles of the direct and inverse currents. The weights are updated on line and track the power grid parameters evolution. This neural approach is compared with the three phase PLL. Simulation results show that our method is fast and efficient for transmission line faults detection and it is able to improve the response capabilities of the protection relay.