Research On Visual Monitoring Technology Of Windfarm Based On WEB

Wind energy, which is regarded as a clean, safe and renewable green energy, has been developed rapidly in recent years. With the implementation of our country’s Twelfth Five Year Plan on wind power, the industry of wind power has obtained outstanding achievement and become a front-runner in the new energy fields. However, owing to the high randomness and unpredictability properties of wind speed, wind power is fluctuant and intermittent. Thus, the wind power integration and control of voltage are descending to be difficult. According to the data, the wind power will bring serious harm to the security and stability of the grid when its penetrating power exceeds 8 percent, to a certain extent, which could limit the rapid development of the wind power industry. Therefore, in terms of improving the performance of the power system, the rational wind speed prediction in ultra-short-term is fairly significant.This project is creatively to put forward the combined theory of similarity in the aspect of wind speed forecasting theory, which is used to calculate the degree of similarity between the wind series, and then establish the ultra-short-term combined model of wind speed prediction; Moreover, with the three-dimensional graphics technology, this paper finally realizes the remote visual monitoring of wind farm. Firstly, this paper briefly describes the present status of the research about wind power prediction at homeland and abroad, and introduces the related characteristics and the utilization of wind energy; Secondly, several main methods used for wind speed prediction are introduced, and then the basic principles of the methods and their advantages and disadvantages are analyzed. The combined model of wind speed prediction based on the spatial correlation method is proposed which is based on the previous analysis. This model is to search the optimal correlation point of the host based on the spatial correlation method, and then calculate the optimal time lag between two points based on the combined theory of similarity which is proposed in this paper, and choose the training samples according to the time lag calculated in the previous step, then establish the combined model of RBF neural network and LS-SVM, and assign the weight by the method of minimum variance. Because the wind speed time series is doped with all kinds of noises during the transmission process, in order to reduce the noise and improve the prediction accuracy of the model, wavelet analysis is picked in this paper to do the multilayer decomposition, refinement and reconfiguration on the wind speed signal. Finally, a practical and reliable monitoring system on wind farm based on Web application is designed and developed, which mainly realizes the display of the wind turbines information based on Web GIS, the monitoring of running data, the analysis and statistic, the query of historical wind speed, the multi-step prediction of the real-time wind speed and power, and the 3D graphic display of the real-time wind farm data in the horizontal and the vertical through the JOGL, and then the effect diagram is embeded in web pages and an intuitive, friendly, efficiently operation interface is provided to the users.