Reactive power optimizational configuration of wind farms on the distribution network based on improved random black hole particle swarm optimization algorithm

Abstract

At present, the reactive power compensation devices of the distribution network can’t compensate reactive power smoothly. To solve this problem, this paper studys reactive power optimizational generation of wind farms on the distribution network. First of all, in order to reflect the wind property,We adopt Weibull wind farm probabilistic model.Thus, the mathematical expected power of the model is the active power of double-fed induction wind turbine(DFIG) .In addition, we take the DFIG reactive power limit into account. Last but no least, we present a improved random black hole particle swarm optimization algorithm(IRBHPSO) which is faster convergence, better global searching. Using this algorithm, we can easily solve the optimal installation location and the optimal reactive power generation capacity of wind farms. We do a simulation on the IEEE33 system which has the same wind conditions on every node and compare IRBHPSO with other four algorithms. Finally, the result shows that the proposed algorithm is effective and feasible, and the installation of wind farms can enhance the power quality of the distribution network. Introduction Among the distrubuted generations, wind power generation has the advantages of its high efficiency, good controllability, appropriate cost and so on. Unlike the traditional reactive power compensation devices by switching capacitors, wind power generation can generate reactive power smoothly. Besides, the grid loss of the distribution network has a large proportion on the power system. Therefore, in order to decrease the grid loss and improve the voltage level effectively, it’s necessary to install the wind farms on the distribution network.However, the wind is random and intermittent. So the output of wind power generation turbine is random and uncertain. Thus, the DFIGs on the grid have a serious impact on the grid flow, node voltage level, grid loss and power supply reliability[1]. Swarm intelligence algorithms are widely used in solving nonlinear, discrete, non-convex problems, such as the distribution network wind farm optimization configuration problem[2]. Swarm intelligence algorithms which include particle swarm optimization, fish swarm algorithm, bacterial foraging algorithm, etc are always premature. And they always fall into local optimum but global optimal solution. the results are random. Based on a random black hole particle swarm algorithm, this paper solve the optimal installation location and the optimal reactive power generation capacity of wind farm on the distribution network. Considering the randomness of the wind and taking full advantages of the wind farm reactive power optimization capabilities, the distribution network optimizes the reactive power compensation by istalling wind farms. Finally, we do a simulation of IEEE33 system and it’s result that the proposed mathod and algorithm is effective and feasible. International Conference on Automation, Mechanical Control and Computational Engineering (AMCCE 2015) © 2015. The authors Published by Atlantis Press 457 Power properties of DFIG Active power properties of DFIG Active power of DFIG is related to wind speed. Active power formula of a 1.5MW double-fed induction wind turbine is Eq. (1) as follows[3]: