Wind Power Optimal Dispatching Based On Chance Constrained Programming

With the development of economy and improvement of intelligent integration level, the social attention of wind power is also growing. Grid-connected operation is the main way of wind power, and the proportion of wind power in power system is becoming more and more. However, due to the inherent randomness of wind resource, the output of wind turbine generation is unstable. It is very adverse for power system which is very sensitive in deviation and requires for high accuracy as well as stability. In order to reduce the threat of gird-connected wind turbines to the stability of power system, on the one hand we have to reduce the randomness of wind turbine output; on the other hand we need consummate the power system dispatching plan, to make the effects of instability minimum. For wind power output which contained random variables, the traditional optimized dispatching method has too much limitation. Therefore, it is necessary to develop a more feasible dispatching method, adapt to the demand of wind energy integrated power system.First of all, we analyzed the credibility of wind predictive power as a standard of classification, to divide the predictive output into three levels. Laying the foundation of cooperation between wind power dispatching and power system control in different time scale, to reduce the predictive error affect the dispatching decision.Secondly, we analyzed the basic method of chance constrained programming and its commonly used equivalent form as deterministic model. We also analyzed the active power flow character of wind energy contained power system as well as the load character of wind turbines. And summed them as two objective function of chance constrained programming, with the restricts of various electric parameters in wind turbines and wind farm as the constrained condition, complete the wind power optimal dispatching chance constrained programming model based on flow and load character.Then, we set up a hybrid intelligent algorithm for solving the chance constrained programming, and analyzed the principle and character of three parts in the algorithm, stochastic simulation technology, neural network and genetic algorithm. We combined the advantages of the three parts, formed a genetic algorithm witch based on stochastic simulation, with neural network approaching the evaluation function. An Operation flow chart of the hybrid intelligent algorithm was provided.Finally, we take the actual data as an example, to operate those theoretical analyses, and then calculated the optimal dispatching plan about the wind farm.