Research On Wide Area Load Optimal Scheduling With Electric Vehicle And Energy Storage For Wind Power Consumption

Wind power consumption is the core issue in wind power utilization.Wind power has the characteristics of volatility and anti-peak regulation,which hinders the consumption of wind power.Although certain effects can be achieved through Longdistance delivery and conventional unit adjustment.However,with the development of smart grids,research on wind power consumption has expanded from the supply side to the active response on demand side.Therefore,research on wind power consumption should not only consider the adjustment of the supply side,but also develop the potential of the demand side,making it an important line of defense for consumption.Based on this,this research regards wind power as a kind of "negative" load,with the goal of reducing the fluctuation of equivalent load,starting from the perspective of wide-area load including electric vehicles and electric heating.Carried out wide-area load optimization dispatch research including electric vehicles and wind power storage.Firstly,the characteristics of anti-peak regulation of wind power forecasted output are analyzed,and the demand side has a larger space to promote wind power consumption,In addition,the fluctuation characteristics of wind power forecast output are analyzed,and it is concluded that the smoothness of the equivalent load curve has not changed significantly compared with the original load,which provides the possibility for wide-area load coordinated wind power consumption.Secondly,a charging control strategy based on reinforcement learning is proposed.This strategy uses the perception and learning ability of the deep deterministic policy gradient algorithm to explore the deep relationship between the uncertainty of perceiving environmental factors and the learning charging law,And from the point of view of neural network fitting function,an instant reward mechanism for actions acting on the environment is constructed.Combining single-step update and experience replay,a deep reinforcement learning algorithm is constructed,and a convergence experiment is performed on the reinforcement learning algorithm that expresses the reward function in mathematical form.Finally,the trained agent is verified by an example.Research shows that the charging control strategy based on deep reinforcement learning can optimize the charging process of electric vehicles under many uncertain factors.Thirdly,the dispatching characteristics and influencing factors of electric heating load are analyzed,and the identification method of electric heating load type based on multi-layer perception network is designed,thereby establishing a multi-agent control model.And the multi-agent deep deterministic policy gradient algorithm is used as the algorithm to solve the model,which verifies the effectiveness of the optimal dispatch strategy of electric heating load for wind power consumption.Research has shown that wind power after adding electric heating load can reduce the fluctuation of equivalent load under the electric heating load control based on the deep deterministic policy gradient algorithm.Finally,the research work is summarized and the outlook is put forward.By studying the optimization of wide-area load dispatching including electric vehicles and energy storage for wind power consumption,it is obtained that the wide-area load is included in the category of wind power consumption.And the method of actively responding to wind power consumption by controlling the demand side has also preliminarily verified the effectiveness of the method.We can further study the balance of reactive power to optimize the dispatching of wide-area loads.From the perspective of maintaining voltage stability,we can also study the coupling optimization problem of wind power and wide-area loads,or study other forms of loads in a wide area,and Other forms of energy storage that convert wind energy into biomass or mechanical energy.