Posters

Abstract

In the last 3 years, Vietnam has recorded the booming development of renewable energy as the result of competitive Feed in Tariff price. At the end of 2021, there are 83 wind power plants with the installed capacity of about 4000 MW, which is about 5% of Vietnamese generation capacity. Most of operating wind farms are onshore or nearshore with seasonal wind characteristics. The generated wind power variation throughout the day is pretty big. EVN National Load Dispatch Centre (EVNNLDC) has recorded the wind power of only 15 MW out of the total installed wind capacity (4000MW). These uncertainties impose enormous burden on the short-term wind power forecasting. Currently, EVNNLDC is having four power forecast sources, which are: two third-party suppliers from international providers, from the renewable power plant, and in-house forecasting by EVNNLDC. Each source has a certain degree of forecasting error and no source guarantees the error within the allowable range for all predicted records. Therefore, EVNNLDC has implemented a forecasting combination system aiming to combining these 4 forecasting sources to obtain better accuracy using Least Squares Error (LSE) method which is making the total of the square of the errors as small as possible. This application has shown a clear effectiveness in lowering the error of the final selection forecast results for actual operation, especially in the intraday forecast, effectively supporting System & Market Operators.