Estimation of the electrical parameters of overhead transmission lines using Kalman Filtering with particle swarm optimization

Abstract

The parameter estimation for transmission systems is important to power flow analysis, planning the expansion of electric power systems, stability, dispatch and economic analysis. This type of task is developed through systems identification methods, being the least squares method and its variations the most common techniques to obtain the transmission line parameters. However, these techniques have some disadvantages, such as non-recursive parameter estimation or the availability of an ideally transposed line, in order to address a problem with symmetric matrices, which simplifies the estimation process. In this paper, a non-linear method (Extended Kalman Filter) is presented to obtain the states of the transmission line terminals jointly with the vectorized matrix of parameters; such approach is strongly affected by the initial conditions; these conditions are usually obtained manually, which requires a lot of time and effort. Therefore, an optimization method (Particle Swarm Optimization) is applied in order to improve the convergence of the EKF, which reduces the time for adjusting the hyper-parameters and improves the estimated results. The proposed method showed accurate results for non-transposed systems, and also in comparison with results obtained from the same EKF-based method without the proposed optimization technique.