Background

Presently, the annual global wind energy is equivalent to about 1000 times of the annual energy generated through coal consumption, this is much higher than hydropower, also larger than the sum energy of solid and liquid fuels. The most important advantage of wind power is that it does not contain any pollutant and is a kind of clean raw material, not any pollutant will be generated in the wind power production process, it will also not cause excessive internal energy consumption.

Meanwhile, wind power is natural resource, it exists at anywhere and anytime, thus the development cost is extremely economical. Therefore, wind power is a type of energy-saving, clean and cheap energy. Wind power is the most important form in wind energy application, it is also the most mature one in today’s new energy development and application field, it is the most promising area in commercial development. However, uncertain factors also exist in wind power, for example, wind energy of the wind farms may vary depending on the season, climate and local weather changes, therefore, the daily power output of each wind farm may changes significantly, if the power dispatch system fails to accuracy predict the power generating capacity of each wind farm, it will be quite difficult to achieve balanced wind power supply, thus great resource waste and benefits losses may be caused.

To address this problem, relying on the numerical weather forecast technology, through forecasting the relevant wind power number in national wind farm regions, the New Energy Research Institute of China Electric Power Research Institute hoped to improve the forecast accuracy of the wind power forecasting system in the state grid.

Solution

Through close cooperation with the New Energy Research Institute of China Electric Power Research Institute, Sugon successfully constructed a high-performance numerical prediction platform for the client. Through three phases of construction, the peak generation of this platform was up to 67.58 TFLOPS, this numbered the 41st in 2011 National High-performance Computers Top 100. This wind power numerical predication and science research business system has extremely high requirement for computing resources, meanwhile, it proposed tough requirements for the storage system capacity and aggregated bandwidth, all these were firmly guaranteed by the ParaStor distributed parallel storage system provided by Sugon. In addition, Sugon CloudBASE C1000 refrigeration cooling and heating system greatly improved the refrigeration efficiency of the equipments and perfectly solved the computer room environment problem for the New Energy Research Institute.