A Data Clustering Based Probabilistic Power Flow Method for AC/VSC-MTDC

Abstract

This paper investigates the role of modified clustering methods to account correlated uncertainties in the probabilistic power flow problem for alternate current/voltage source control-multiple terminal direct current (ac/VSC-MTdc). Such uncertainties are introduced by stochastic renewable generation units or variable loads. The wind farm is a type of stochastic generation units that grows dramatically in power networks. Due to the increase of penetration level of renewable generations, number of uncertainty sources is increased. Clustering methods are suitable tools for handling stochastic power system problems, because these methods provide acceptable results with less computation in comparison to other common methods such as Monte Carlo simulations. In this paper, two modified methods are used for clustering uncertainty sets. These methods include modified K-means algorithm, and modified self-organized map neural network. After clustering, the power flow of ac/VSC-MTdc is carried out for cluster centers of each method. Results of the probabilistic power flow for case study that is based on the IEEE 24-bus and IEEE 118-bus test systems are presented and compared with different clustering techniques, Monte Carlo simulation, and Latin hypercube sampling (LHS) methods. Such results show that power flow calculation with clustering methods is more accurate than LHS method and is faster than Monte Carlo simulation method.