Study On The Optimizing Operation Of Power System With Large-scale Wind Farms

With the increasing depletion of fossil fuels such as coal, oil and natural gas, and its serious impact of pollution on the environment, wind power as the most mature alternative energy in generation technology has enjoyed large-scale development. However, as the penetration rate of wind power increases in power grid, characters of strong volatility and randomness will impact grid dispatch and operation severely. As a result, traditional power system dispatching and operation method are no longer applicable and corresponding new model considering characteristics of wind power should be established and analyzed, as well as its solving method and assessment method. In this paper, the history and status quo of three key issues in terms of unit commitment, economic dispatch and load flow calcula-tion is reviewed first. Based on that, considering the different characteristics of wind power in these three problems respectively, we do further research in unit commitment considering wind power ramping events, economic dispatch of power grid with wind power integration considering unit adjustment ability and probabilistic power flow based on wind power data samples. The main work includes the following three parts:(1) In the issue of unit commitment issue with wind power integration, due to the fact that large ramping events usually occur, unit commitment model considering wind power ramping event has been established. Firstly the stochastic nonlinear mixed integer pro-gramming model has been turned into stochastic linear mixed integer programming model using precise linearization technology, and then the wind power is depicted by the predicted value and the upper and lower limit of the interval prediction. Large scale of calculation can be done by commercial software. The feasibility and effectiveness of this method is vali-dated through simulation results, and that for general cases of wind power ramping events, the constraints has little effect while for large wind power ramp variation, the consideration of ramping event constraints can better insure the system security.(2) In the issue of economical dispatch of wind power integrated power grid, due to the fact that wind power is hard to predict accurately, so on the base of predicted wind power, the upper and lower limit of wind power prediction error has been taking into con-sideration to build a robust economic dispatching model of wind power integrated power grid. The error caused by imprecise wind power prediction can get linear optimal compen-sation through adjustment ability of conventional units and the stochastic problem can be turned into deterministic problem through robust optimization theory and then the calcula- tion can be done through commercial software and the secutity of power system can be en-sured, providing a new effective method for solving the economical dispatch problem in large wind power integrated power grid.(3) In the issue of probabilistic power flow in wind power integrated grid, due to the fact that wind power distribution characteristics can hardly be fitted using common prob-ability density function, a point estimation method based on wind power data samples is proposed, with the Cholesky decomposition technique adopted to handle the correlation of input variables and Gram-Charlier series expansion used to get the cumulative distribution of random output variables. The impact of whether taking correlation into consideration or not on the simulation result has also been analyzed in detail. Simulation results show that this method needs small amount of calculation of high accuracy, and that the correlation of input variables should be taken into consideration in real engineering application to avoid relative large error.