Evaluation Of Wind Power Acceptance Capacity Of Power Grid Based On Wind Power Interval Forecastion

With the rapid development of wind power technology in China,the installed capacity has reached a high level,and wind farms in many regions have formed scale.However,the inherent characteristics of wind power and the limitation of the spatial distance between source and load have led to the existence of a large number of “abandoned wind power restrictions”.The wind power acceptance capacity of the power grid is to determine the extreme penetration ability of the power grid for wind power.There are many factors affecting the wind power acceptance capacity,and the wind power forecastion interval is an indispensable item.After summarizing the limiting factors of wind power acceptance capacity,this paper proposes a wind power interval forecastion method and probabilistic load flow algorithm,and applies it to the evaluation of wind power acceptance capacity based on the constraints of the upper and lower limits of wind power prediction.The main contents of this article include:(1)The characteristics of wind power output,the impact of wind power on the power grid,and the limiting factors of wind power capacity are summarized and analyzed.Based on this,the constraints for the wind power capacity assessment model are summarized.(2)The classification and definition of wind power interval forecastion are introduced,and the extreme learning machine network algorithm is used as a point forecastion model.Based on the Copula correlation theory,and considering the correlation between power fluctuation and forecastion error,a wind power interval forecastion model is established.Evaluation indicators validate models and methods.(3)The advantages and disadvantages of the parametric and non-parametric methods of wind power probabilistic modeling in current wind power grid-connected probabilistic power load calculation methods are analyzed and summarized,and a non-parametric kernel density estimation considering boundary correction is used to establish wind power in probabilistic power load calculation Probabilistic model,and references the Grid Search method with Cross Validation(CV)in machine learning to optimize the bandwidth parameters of kernel density estimation.The IEEE 39-bus power system improved after adding wind power was used for power load simulation analysis.(4)Establish a wind power acceptance capacity evaluation model based on the wind power forecast interval.First,take the wind power forecast interval information and load information as inputs,and combine the conventional unit power generation model to optimize the start and stop status of the conventional units,and give the optimal unit combination and acceptable The upper and lower limits of wind power;based on this,the abandoned wind rate and the amount of cost change before and after wind power access are used as evaluation indicators to compare the results obtained at different confidence forecastion intervals with the single-point forecastion values as model inputs to analyze and obtain the final conclusion.