| In view of the shortage of traditional fossil fuels and environmental pollution,the development of renewable energy has become an important measure to solve energy and environmental problems.Distributed wind power as a widely distributed,renewable,clean and other characteristics of the power generation method has been vigorously developed.However,due to the strong volatility and intermittency of the output power of distributed wind turbines,the ability of the distribution network to accept distributed wind power is severely restricted,resulting in a serious waste of distributed wind power energy.The gradual evolution of the traditional passive distribution network to the multi-source controllable active distribution network and the mature development of energy storage technology are considered to be an effective way to solve large-scale wind power consumption.Active distribution networks can use source-network-load coordination to optimize a variety of active management measures to improve the capacity of the distribution network to accommodate distributed wind power,thereby increasing the absorption rate of large-scale distributed wind power;In addition,the combination of energy storage devices and distributed wind turbines to form the wind-ESS combined system,which can improve distributed wind power consumption and distribution network operational safety by smoothing distributed wind power output.Therefore,the specific research contents and research results are as follows:Firstly,in view of the uncertainty of distributed wind energy output and load demand,and the need to comprehensively consider the long-term planning of the evolution of large-scale and massive scenarios in the optimal allocation of wind-ESS,this paper adopts a method based on extensional distance K-means clustering to solve the uncertainty of distributed wind power output and load demand.The extensional distance K-mean clustering algorithm overcomes the disadvantages of the conventional K-mean clustering algorithm due to the randomness of the initial cluster centroid selection,which often leads to unstable clustering results and low clustering accuracy.Secondly,this paper combines source-network-load collaborative optimization and demand response multiple active management measures with the optimal allocation of wind-ESS combined system.With the objective of maximising the wind-ESS combined system investor’s earnings and the wind power local consumption rate,a multi-objective optimal allocation model for the wind-ESS combined system is constructed,taking the installed capacity and location of the wind-ESS combined system as well as the amount of active management strategies implemented by the source-grid-load triad as decision variables.Finally,this paper introduces the idea of multi-core parallel computing environment and sine function in the conventional differential evolution(DE)algorithm,and proposes the parallel multi-objective sinusoidal differential evolution(PMOSDE)algorithm to solve the optimal allocation model for the wind-ESS combined systeml efficiently;The PMOSDE algorithm uses multi-core parallel computing technology and sinusoidal function characteristics to adjust the variation scale factor and cross probability factor dynamically and scale periodically,and construct a new sinusoidal differential variation mechanism.The PMOSDE algorithm can not only maintain the diversity of individuals in the evolutionary process,but also achieve periodic traversal search,to overcome the shortcomings of conventional DE algorithms in terms of slow merit-seeking speed and insufficient population diversity in the late evolutionary stage.In order to verify the superiority of the proposed wind-ESS combined system optimal allocation method and the PMOSDE algorithm.A modified IEEE 33-bus system including the optimal allocation of wind-ESS is taken as an example to verify the feasibility and superiority of proposed method. |
