Research On Uncertainty Modeling And Power Dispatching Method Of Multi-wind Farm Joint Output Considering High Dimensional Nonlinearity

Wind power forecasting is an important means to deal with the uncertainty of wind power and assist dispatching decision.However,due to the inaccuracy of wind speed forecasting and the imperfection of forecasting methods,the accuracy of current wind power forecasting methods is low and difficult to improve significantly.Therefore,for the wind power prediction error effectively and accurately describe the modeling,prediction error uncertainty,has important theoretical and practical significance for dispatch.In order to effectively characterize the uncertainty of prediction errors,this paper takes multi-wind farm cluster in the same area as the breakthrough point,analyses the statistical characteristics of typical wind farm data,and deepens the modeling based on Copula theory,proposes the modeling method and establishes static Copula model of single wind farm output,dynamic Copula model of single wind farm output and combined output uncertainty model of multiple wind farms.Finally,the day-ahead scheduling simulation is carried out in PJM-5 system,and the superiority and practicability of the model are proved by the scheduling results.Firstly,the statistical characteristics of wind power prediction data and prediction error data are analyzed quantitatively.Combining Copula theory and its research status,the types of Copula function,the topological structure of Vine Copula and the method of calculating correlation coefficients suitable for wind power prediction output uncertainty modeling are selected.Based on the idea of resolving the contradiction,this paper discusses the current technical level of wind power forecasting and the technical difficulties it faces,and explains the research content of this paper,that is,the necessity of establishing the uncertainty model of wind power forecasting output.Based on the research of domestic and foreign experts and scholars on the uncertainty modeling of Copula function in wind farm,it illustrates the introduction of Copula theory as a modeling tool.According to Sklar theorem,Pair-C opula theory and Vine C opula theory,the basic forms of C opula function are listed,and its solution method is clarified.Taking C Vine and D Vine as examples,the topological structure of Vine C opula function is illustrated,and the functional expressions of both are enumerated.The C opula theory and basic modeling method are introduced in general.Secondly,a dynamic Copula model for single wind farm output is proposed.Eight groups of static and dynamic Copula models are established.The advantages of introducing dynamic characteristics into the model are proved by a comparative example.In this paper,the statistical analysis of wind farm output prediction data and prediction error data is carried out,and the correlation between them is quantitatively analyzed,and the basis for joint modeling is explained.Based on the correlation between single wind farm output and prediction error,and the correlation between two wind farms in the same area and their dynamic characteristics,the dynamic Copula theory is introduced and the dynamic phase is proposed.Estimation method of relation number and evaluation method of model;8 sets of dynamic and static models are established based on measured data,and the goodness of fit of the model is compared,which shows the superiority of dynamic Copula model.Thirdly,the method of constructing high-dimensional dynamic Vine Copula model and uncertain model of combined output of multi-wind farms are proposed,and the highdimensional dynamic Vine Copula model and uncertain model of combined output of multiwind farms are established.The advantages of the above model are proved by comparison examples.Based on the correlation and dynamic characteristics of the predicted output of two wind farms in the same area,the high-dimensional dynamic Vine Copula theory is introduced,and the modeling method based on the high-dimensional dynamic Vine Copula is proposed.Combining with the discrete convolution method based on the Copula function,the modeling method of the uncertainties of the combined output of multi-wind farms is put forward.The output of Vine Copula model and dynamic Copula model are compared,which shows the superiority of the combined output uncertainty model of multi-wind farms.Finally,the model output of the uncertain model of combined output of multi-wind farms is taken as the constraint condition of rotating reserve,and the day-ahead optimal dispatching model is established taking into account the factors such as total generation cost and wind power absorption rate,and the superiority and practicability of the model are proved by a comparative example.In this paper,the model output of the combined output uncertainty model of multi-wind farms is taken as the basis of the rotating reserve constraints,the objective function is to minimize the total generation cost,and the wind power absorption rate is taken into account.The optimal dispatch of PJM-5 system is carried out before the day.The optimal dispatch results are compared with those of the normal error model,the Vine Copula model and the dynamic Copula model,which are simply superimposed on the model output.The practicability of uncertain model of field combined output.