Short-term Load Forecasting And Nonlinear Economic Dispatch In Hydrothermal Integrated System

As an important economic and energy infrastructure system,power system is of great significance to the development of society.Ensuring the safe and stable economic operation of the power system is not only conducive to improving energy efficiency,optimizing the energy structure,but also accelerating the process of social modernization.The short-term load forecasting of power system is the basic work of power enterprise planning,scheduling,terminal power use and other departments.It provides data support and effective reference for arranging unit output,setting up unit operating plan,safe and economic dispatch and so on.However,most of the prediction models are traditional deterministic load forecasting models.It is difficult to describe the uncertainty risk of load in the future.Therefore,it is urgent to study the interval prediction method under certain confidence level to describe the scope of load fluctuation and provide more comprehensive load forecasting information.At the same time,the accurate and comprehensive load forecasting information helps to improve the economic efficiency of the short-term economic dispatch model.According to the comprehensive information,the influence of different load levels on the economic dispatch model can be quantified.For this reason,this paper studies the three dimensions of the short-term load point prediction,the load interval prediction under certain confidence level and the economic dispatch of the hydrothermal power system considering the load interval.The main research contents include:(1)For the ultra-short-term load forecast model,an algorithm based on similarity selection of linear difference degree is proposed on reducing the complexity of the algorithm and increase the operation rate.At the same time,around the short-term load forecast problem,the SVR algorithm was used to build prediction models based on different influence factors,and the model prediction results considering relevant factors were compared with model prediction results that did not consider related factors.Examples indicate extreme weather conditions.The prediction model that considers relevant factors can achieve higher accuracy.(2)Focusing on the demand for more comprehensive load forecasting information for power grid dispatch,Copula theory is introduced.Based on the correlation between the historical actual load value and the historical point prediction value,the semi-parametric method based on empirical distribution was applied to approximate the actual load value and point.The Copula function parameter values of the inter-predicted structure are selected,the optimal Copula function is evaluated and selected,and the interval prediction model based on the Copula function is constructed.The calculation is based on the joint probability distribution between the actual load and the point prediction value.The load forecast interval at a certain level of confidence.The example of network load in Hubei Province shows that the model is effective for the prediction of short-term load intervals.(3)In order to quantitatively analyze the influence of load uncertainty on power grid dispatching,the nonlinear economic scheduling of hydrothermal power system under the uncertainty of load is studied,and the load uncertainty is described in the way of a certain confidence level under the load interval.The generation cost of the power generating unit is incorporated into the constraint conditions,and the power generation cost of the thermal power unit,which considers the threshold effect,is brought into the objective function.A method of solving the nonlinear economic scheduling model considering the load interval is proposed,and optimistically pessimistic scheduling optimization model is established respectively,and the dual theory of nonlinear programming is introduced to solve the corresponding optimization.In this model,the interval solution of the economic dispatch model of the load interval is considered,and the total cost of electricity generation and the scheduling scheme under the two extreme conditions of optimistic and pessimistic conditions are obtained.The proposed method solves the problem of solving the two-layer nested nonlinear economic scheduling model in the actual operation of the power system,and can quantitatively analyze the impact of the load uncertainty on the scheduling,and provide a reference for the scheduling decision-makers.Finally,this part summarizes the research content of this article and puts forward the prospect of further elaboration.Some of the research results of this paper have been applied in the Dispatch Department of the Central China Branch of the State Grid.