Design And Implement Of Wind Farm's Data Acquisition Supervisory And Forecasting System

As an environmentally friendly, economical and conducive to social development energy production, wind power has attracted increasing national attention and the technology of wind power has gradually matured. With many domestic large-scale wind farm has been built and put into use, the instability of grid-connected wind power, caused by the intermittence and randomness of wind, has an adverse influence on the entire power grid. Recently the discrete control system (DCS) has been used in many wind farms and the extent of information integration is significantly improved, but for the geographical dispersion of wind farms, so it is still difficult to share the timely and effective information in the entire wind farm company. Under the circumstances, centralized power forecasting for wind farms is hard to achieve, which has became an element to restrict the market competitiveness of grid-connected wind power. Therefore, an effective and reliable platform of data transmission and storage for the wind farm should been built, and then construct the wind power forecasting system and data publishing system on the platform, which is conducive to ensuring the stable and secure running of the grid and to improving the utilization rate of wind power. Focusing on this actual demand, the wind farm's data acquisition supervisory and forecasting system is designed in this thesis and the following researches have been carried out.1. Data acquisition transmission and storage platform constructionFirst of all, the network topology is planned to collect and transmit the mass real-time information of the wind power generator and substation in verification mode to the remote monitoring center of the wind power. Then the PI server is picked up to store the real-time information of the wind farm. By chronically storing the fast refreshed and large capacity data in the form of prototype data with time tag, it has laid a strong foundation for wind power forecasting and information website.2. Wind farm's power forecasting system constructionAimed at wind farm's power forecasting, an ARMA model is built to forecast the following day's wind speed using the stationary time series. The series are gained by processing the non-stationary time series in the way, which is proposed in this thesis after analyzing the wind speed and power data. Then the following day's power could be forecasted by the neural network model using the wind speed and power data. Based on real-time and reliable running data, the forecast has high reliability and stability and could meet the accuracy requirements for the wind company. At the same time, as the guidance it is beneficial to scientifically allocating the wind power generators' start-up, stop and maintenance, and to improving the utilization rate of the wind power generator and economic benefits of the company.3. Data publishing platform constructionUpdataPanel control, Chartlet control, ReportViewer control and MATLAB engine are used to design the real-time information module, data curve module, data report module and power forecasting module of data publishing platform. The data publishing system of the wind farm based on B/S frame is developed by the method of IIS combined with ASP.NET and PI Server in Windows environment. The company information website is built up in this framework, to realize interactive, dynamic and efficient information management with network.