Study On The Model And The Application Of Wind Power Prediction

In recent years, wind power prediction has achieved a series of breakthroughs, the technology is one of the most effective and economical means that can enhance the peaking capacity, enhance the acceptance of wind power for power grid capacity, improving the safety and economical operation of the power system.Purpose and significance of the thesis begins from the wind power forecasting, the prospect of the current domestic and foreign wind power forecasting technology development status and future trends. Then the paper briefly discusses the classification and prediction method of wind electric power, mainly divided into statistical method, physical method and comprehensive method. Research on paragraph in the physical method, focuses on the analysis of the atmospheric motion equations, obtains the influence factors for the prediction of wind power in atmospheric physics data; at the same time in this paragraph is the detailed analysis of the impact of bureau ground effect with the wind power forecasting, roughly divided into roughness change, changes in terrain to elaborate; and the wake effect are outlined. Research on paragraph in the statistical methods, carried out a detailed analysis of the definition, structure, algorithm of BP neural network, the learning rate of these two aspects to the BP neural network algorithm was optimized and improved the appropriate from a superposition of momentum method and optimization; at the same time the paragraph also focus from mathematical equations analyzed factors affecting the output power wind turbines, and discusses the model of energy conversion of wind power. Then, the paper from the input data selection, data normalization, wind power prediction of neural network structure selection and other aspects discussed the establishment of prediction model. Combined with the actual work and study, the paper discussed in detail the NSF3100 wind power Nari Technology Development Limited Company independent research and development of the forecasting system, the system includes a data monitoring, power prediction, software platform of three parts. Data monitoring is the base of prediction, data monitoring includes monitoring of meteorological information and operational status of wind farm. Power system short-term prediction can be realized, the ultra short term prediction function, meet the wind power business for different forecast demand. The software platform of monitoring and forecasting data results in an intuitive way to display and analysis.Finally, on the basis of this set of wind power forecasting system in the actual operation situation of a provincial Power Grid Corp dispatching system, the thesis makes a detailed display function and state prediction system in practice.