A statistical algorithm for predicting the energy storage capacity for baseload wind power generation in the future electric grids

• A statistical algorithm is presented for predicting storage capacity for wind energy.
• The algorithm can be utilized at different stages of the wind energy industry.
• Introduced methods are based on parametric and nonparametric statistical models.
• The algorithm contributes toward the goal of baseload wind power in the future grids.

We propose a statistical algorithm for sizing the energy storage system required for delivering baseload electricity to a selected confidence level for a wind farm. The proposed algorithm can be utilized by utilities to assess wind integration and to investigate better capacity credits for wind farms connected to the grid, by wind farm operators to potentially increase their return on investment by designing a baseload wind farm to a selected confidence level, and by financial institutions to calculate the confidence level for baseload wind farm projects. Methods introduced are based on parametric and nonparametric statistical models using wind resource assessment data and available wind turbine information that reflect different stages of a wind farm project—from site selection to operational status. To study the performance of each method, we apply these to a North America operational wind farm data set. We use averaged 10-min and hourly data to calculate and compare the firm capacity of the wind turbine for each proposed method. The results show that for different stages of the wind farm development, and depending on the available information, the proposed algorithm can properly estimate the energy storage capacity required to deliver constant power to a user selected confidence level.