Abstract: Due to the random fluctuation of wind power having impact on the real-time operation state of power system， reasonable optimization of both location and capacity of grid-connected wind farm under the classical operation mode should be considered. It is benefitial for both reducing the static safety risk and improving the economic operation of the grid. Considering the uncertainty of both power-source side and load side， the correlation of different wind speed and the correlation of different loads， a probability power flow calculation method considering the correlation of random variables is studied in this paper， which generates the correlation samples by combing point estimation and inverse Nataf transform. Moreover， through embedding the point-estimation based algorithm for probabilistic power flow， which considers the random correlation， into the adaptive particle swarm optimization algorithm， and avoiding the optimism on the constraints caused by using the point estimation method to solve the probabilistic power flow， taking the minimization of both the active power loss rate of the grid and the average deviation of bus voltages as the objective， the method for reasonably planning both the location and capacity of grid-connected wind farm is presented. Finally， the effectiveness of the proposed optimizing method is validated by the simulation carried out on the IEEE 57-node system. At the same time， the simulation result also shows the necessity of improving the reliability of the optimizing result by considering the correlation of the random variables.

Key words: wind power integration, point estimation, probabilistic load flow, probability correlation, particle swarm optimization