Real-time Load Modeling Of Wind Power Connected To Substation

With the rapid development of renewable energy power generation technology,large-scale wind power is connected to the grid,so that wind turbines have gradually become an important part of the bus load model.Existing load modeling considering wind power connection can only realize parameter identification in a very short time,and cannot realize continuous parameter identification in a long period of time,and the model structure is fixed and single,and cannot change with the change of wind power permeability.Therefore,the paper conducts research from the following three aspects:First of all,for the problem of missing values in the actual measured data of the daily load of the busbar in the substation,the supplementary ensemble empirical mode decomposition and the algorithm combining the long and short-term memory neural network(CEEMD-LSTM)are used to fill it;A clustering method(SCK-SHAPE)that combines the contour coefficient and the K-SHAPE algorithm is proposed to make up for the clustering algorithm’s inability to objectively determine the optimal number of clusters;through the actual measurement of the 110kV side busbar of a 330kV substation in a certain province The simulation analysis of the data verifies that the CEEMD-LSTM algorithm can accurately predict the missing values,and the SCK-SHAPE algorithm has a better clustering effect than the K-means algorithm.Then,the method of real-time load modeling under the entire daily load curve is studied.The actual daily load curve closest to the typical daily load curve is selected as the research object of load modeling and parameter identification,and a generalized load model including wind turbine connection is established under it;15 minutes of SCADA data is used as a period,and the period is selected Multiple sets of PMU data suitable for identification of small disturbance parameters,use genetic algorithm to perform continuous parameter identification on multiple sets of PMU data;calculate the average value of multiple sets of parameter identification results and model output power,and compare with the actual measured daily load curve fitting;The simulation results verify the feasibility of the method.Finally,research is carried out on the influence of wind power penetration rate on the accuracy of the bus load model.Based on the parameter identification results in Chapter 3,the model accuracy evaluation method based on numerical value and trend similarity is used to evaluate the results,introduce wind power permeability,and describe the relationship between wind power permeability and model accuracy;The wind power permeability value at the point where the accuracy of the model drops as a demarcation point,two load models under different wind power permeability are established.One is the generalized load model that takes into account the connection of wind turbines,and the other is the traditional model that ignores the connection of wind turbines.Load model;the variable structure load model is identified again under the same daily load curve for continuous parameter identification;the simulation results show that the wind power penetration rate and the model accuracy generally obey the inverse relationship,and the traditional load model has a higher wind power penetration rate.The load characteristics of the bus can be described well even when it is low.The thesis provides a reference for the real-time load modeling of wind power connected to substations and the study of variable bus load structure under different wind power penetration rates.