APPLICATION OF DIFFERENT INDICES OF PERFORMANCE EVALUATION OF FORECAST FOR WIND SPEEDS
DOI:
https://doi.org/10.24084/repqj12.568Keywords:
forecast, performance evaluation, prediction error, numerical weather prediction modelAbstract
In this study, firstly, we´ll see an introduction about wind energy and its current use in Brazil, giving an overview of the development of the treatment of wind forecasting through two techniques for forecasting. One of the data sources available are predictions from numerical models, due to the large volume of information, it were saved only prediction´s images made. Worked these images you can retrieve a value of wind speed forecast for a specific city. We´ll show the used methodology, seeing how processed data were taken, we´ll also see the accumulator´s installations, which will contain for the same date, actual values and projected wind speeds, and we’ll conclude with the indexes’ calculations for evaluating performance working with three indexes: “MAPE”, “RMSE”, “MAE”, in the range of a year in four quarters and 45 samples in each quarter. Then we'll see according to several articles, the most used indexes evaluation, which are “MAPE", "RMSE" and "MAE", that´s the reason in this study we’re also going to work with these ones. Finally, we´ll proceed to the analysis of the results. In the rainy season and the dry season, we will see what was the numerical model that best approached to the reality.