Research Of Wind Power Forecast Methods Based On Cloud Computing And Artificial Intelligence Algorithm

As the energy problem is more serious, the renewable energy in China's energy structure is gradually increasing. As a kind of clean energy, wind power will become the main body of the future energy structure. Wind power equipment and technology has been quite mature, but the wind power generation is not stable directly affect the wind power grid, but also caused a lot of wind power waste. So wind power forecasting can improve the utilization of wind power, and become the premise of wind power grid.The wind power forecasting is more complex than the general forecasting problem,and there are many factors that affect the wind power generation. In order to be able to carry out effective wind power forecasting input parameters must be more than the average power forecasting problem. Artificial neural network is widely used in the field of wind power forecasting, but it is easy to fall into local optimum, so it can be improved by using intelligent optimization algorithm. The PSO algorithm, genetic algorithm and other algorithms can solve the problem of neural network into local optimum. But when the input parameter dimension is increased, the performance of the algorithm is very low.It is difficult to predict the wind power effectively. The algorithm is implemented in Spark cloud platform, improve the operation rate of the whole algorithm. This article mainly carries on the following several aspects of the work.(1) the problem of the traditional wind power forecast is analyzed, and the advantages and disadvantages of different forecasting methods are studied. The performance of several intelligent algorithms is compared. According to the special nature of the wind power forecasting, a suitable algorithm is selected to improve the existing algorithms.(2) the characteristics of the study and analysis of the principle of the optimization algorithm are analyzed, and the relative other optimization algorithms are more advantages than other optimization algorithms. A novel neural network prediction method, which is based on artificial neural network, is proposed. The neural network weights and threshold values are defined as a vector, which is used as the optimization algorithm of the bacterial population.(3) a swarm neural network algorithm based on cloud computing is proposed, and the performance of the neural network is evaluated by training data set in each Spark node. The performance index of the neural network is passed to the Spark main node, and the weights are given to each node according to the weight of the weights, and the final results are obtained.(4) an experimental test and an example analysis. The real wind power parameters of a wind farm in Inner Mongolia are selected and the performance of the proposedalgorithm is tested on the 9 nodes in the laboratory. The experimental results show that the proposed algorithm has better performance than the existing algorithms, which can provide effective basis for wind power load forecasting, and has better parallel performance.