Modeling of uncertainty in distribution network reconfiguration using Gaussian Quadrature based approximation method

Abstract

Distribution feeder reconfiguration is a cost-effective scheme to improve the system operating condition. Making decisions on the status of switches along the feeder with high penetration of renewable resources requires a tractable and accurate input data analysis. This paper presents a method using Gaussian quadrature to approximate probability distribution function of input variables with a probability mass function. The method is applied to Weibull wind speed and Beta solar irradiance probability distribution functions. These approximated input variables are fed into an energy loss minimization problem to find the best configuration of 33-bus Baran test system in a typical summer day. Group search optimizer, a swarm intelligence optimization method, is applied to solve the problem. In order to analyze the accuracy and efficiency of the proposed method, Monte Carlo simulation and two other approximation methods-bracket midpoint and bracket mean-are used for comparison. Results show the superiority of the method in computational time and its adequate precision.