Parallel lines are one of the main parts of the power systems due to their capability of increasing the transmitted power and reliability of the power network. However, due to their complexity, fault location in these transmission lines has always been a potential problem. Furthermore, when the wind farms are connected to the power networks, a new challenge is created for the fault location algorithms. The reason is that the impedance changes continuously in these wind power plants. Moreover, the Static Var Compensator (SVC), which is used for compensation in transmission lines, can cause problems in accurate calculation of the fault location in power transmission lines. Accordingly, this paper presents a new method based on Long Short-Term Memory (LSTM) networks for source impedance and other network parameters estimation, which are key parameters to calculate the fault location accurately. The simulations are accomplished using PSCAD and Python software on a sample two-circuit network. This network consists of two parallel lines, a wind farm, and a synchronous generator to supply the network's power. The double-fed induction generator (DFIG) model is considered as the wind farm in the studied network. The results obtained from the tests clearly illustrate the applicability of the proposed method to estimate fault location and source impedance with high accuracy. • A fast and accurate fault location algorithm is achieved using Long Short-Term Memory (LSTM) networks. • Determining the location of different fault types in two circuit transmission lines. • Accurate performance for the location of different fault types under nosy conditions. • Estimation of the source impedance with a much smaller error value.
Read full abstract