Research On Calculation Method Of Optimal Power Flow Including Wind Farm

Wind power is not only environmentally friendly and has characteristics of high reliability and mature technology, but is also beneficial for scale and cost-effectiveness.But it is no doubt that unpredictability, stochastic volatility and uncontrollable factors of wind speed result in the fluctuating properties of the output of the wind turbine, which will do harm to the system when connected to it on the trend of the power grid and net loss, transient and voltage stability of the power grid and the protection parameters. In order to estimate accurately the influence of the wind farm on the power flow of the present power system and find out the optimal solution as soon as possible, it’s necessary to seek for optimal power flow calculation method of the electric power systems.The main work of this article is as follows:Firstly, it introduces the basic principle of wind poweras well as the working principle and characteristics of the asynchronous wind turbine and doubly-fed wind generators, it also analyzes briefly the influences when wind power farm is connected to the electric power system.Secondly, it includes the basic principle and characteristics of all kinds of neural network prediction and especially analyzes the differences of the application characteristics between BP neural network and RBF neural network. Record instruments will contain some unreasonable wind speed datas on wind farms named abnormal datas because of weather factors and the failure of machine. The abnormal datas regarded the nerve forecast datas samples will debase the forecast accuracy. On contrary, accurate predictions on wind power output can do a lot to avoid or reduce the adverse influences of the power system when the wind farms are connected to the power grid, and increase the competitiveness of wind power in the electricity market.Therefore, this article makes use of abnormal data eliminate technology to analyze datas and exclude the abnormal datas.It predicts the active power output on the basis of RBF neural network and the case analysis results indicate that the method can improve the prediction accuracy of the neural network.Thirdly, it shows the basic principle and characteristics of Artificial Fish-swarm Algorithm and analyzes its disadvantages. Basing on the conclusion of various previous calculating methods of wind farm power flow and analyses of the disadvantages of the basic artificial fish swarm algorithm, the article makes some adaptive adjustments in the artificial fish’s field of view and step method of choosing. What’s more, it improves from the random generation method to the evenly in distribution method for producing in a solution space in terms of the methods of the initial population. Besides, it put individual fish that swim in random places originally in feasible space.Last but not least, terminate iterative conditions depend on the method of the biggest algebra ranges mixed with the minimum reserve algebras instead of the Maximum iterations. As a result, it comes up with the improved artificial fish swarm algorithm.Finally, examples analysis show that the improved artificial fish swarm algorithm is quite remarkable. The distribution of initial fish can cover the global solution area. The improvement of view, step and pattern of actions not only help increase the search for global optimal solution speed and avoid to fall into searching the local optimal solution but also enhance the search abilities on the global optimal solution. In addition,improvements on the termination of the iteration conditions increase the convergence rate. Therefore, the improvements make a great difference and have practical significance of the reference value of engineering applications.