Optimal Dispatch Including Wind Power Based On Chance-constrained Programming

With the development of economy and society, there is an ever-growingdemand for energy. The importance of wind power is geting more and moreattention as the reduction of fossil energy reserves and the seriousness ofenvironmental pollution. As wind power is renewable, non-polluting, low cost andhas a large reserve, it’s irreplaceable in new energy generation area. But thedistribution of wind power is influenced by the climate, it’s random, intermittent,fluctuant and uncertain, there are lots of adverse effects when massive wind powerconnected to the power system. To reduce the effects bringing to the system of windpower, the traditional optimizing dispatch method is no longer applicable, so weshould seek a more reasonable and effective method to study the optimal dispatchproblem including wind power.In this paper, we firstly analyse the wind power fluctuation characteristics.Decompose the measured data of wind power using Mallat decomposition algorithmto extract the average data and hour level fluctuation. By curve fitting, we get thefunction formula between the average data and wind power fluctuation, so the windpower fluctuation of each moment can be estimated by the average data, then thewind power fluctuation can be quantized.Secondly, we establish an optimal dispatch model based on chance-constrainedprogramming. Use chance-constrained programming to improve the traditionaldeterministic model, express the objective function and constraints with probabilityforms including confidence levels. Then use the function formula of the wind powerfluctuation in the chance-constrained model, so that the model is more in line withthe actual situation.Thirdly, solve the chance-constrained model proposed. As there are chanceconstrains in the model, the traditional method is no longer applicable, so we use thegenetic algorithm based on stochastic simulation to solve the model. Compare theimpacts to the system cost between wind power fluctuation and prediction error byan example, and verify the effectiness of the model and algorithm proposed in thispaper.Finally, establish frequency modulation models for single-area system anddouble-area system, input the dispatching results to the models and simulation,compare the impacts to the system frequency modulation effects between windpower fluctuation and prediction error. From the simulation results we can also seethat after the regional system interconnection, the frequency deviation of eachregion has been reduced accordingly. According to the research results, the model and algorithm proposed in thispaper are more in line with the actual situation when massive wind power connectedto the power system, it can reduce the system cost, improve the dispatchingeconomy, reduce the system frequency deviation and optimize the dispatch of thepower system. There is guiding significance to the making of dispatch instructionswhen massive wind power connected to the power system.This paper is supported by the National Key Basic Research and DevelopmentProgram of China (973program)<The Basic Research of the Security and EffectiveUse of Massive New Energy in the Smart Grid>(Sub-topic1: The DynamicCharacteristics and Modeling Theory of the Power System with New Energy)(NO.2012CB215201) and the National High Technology Research and Development ofChina (863Program)<Key Techniques of Intermittent Energy with HighPermeability in the Regional Power Grid>(NO.2011AA05A105).