eduzhai > Applied Sciences > Engineering >

Combination of WRF Model and LSTM Network for Solar Radiation Forecasting—Timor Leste Case Study

  • KanKan
  • (0) Download
  • 20210301
  • Save

... pages left unread,continue reading

Document pages: 37 pages

Abstract: A study of a combination of Weather Research andForecasting (WRF) model and Long Short Term Memory (LSTM) network for locationin Dili Timor Leste is introduced in this paper. One calendar year’s results ofsolar radiation from January to December 2014 are used as input data toestimate future forecasting of solar radiation using the LSTM network for threemonths period. The WRF model version 3.9.1 is used to simulate one year’s solarradiation in horizontal resolution low scale for nesting domain 1 × 1 km. It is done by applying 6-hourly interval 1º × 1º NCEP FNL analysis data used as Global ForecastSystem (GFS). LSTM network is applied for forecasting in numerous learningproblems for solar radiation forecasting. LSTM network uses two-layer LSTMarchitecture of 512 hidden neurons coupled with a dense output layer withlinear as the model activation to predict with time steps are configured to 50and the number of features is 1. The maximum epoch is set to 325 with batchsize 300 and the validation split is 0.09. The results demonstrate that thecombination of these two methods can successfully predict solar radiation wherefour error metrics of mean bias error (MBE), root mean square error (RMSE),normalized MBE (nMBE), and normalized RMSE (nRMSE) perform small errordistribution and percentage in three months prediction where the errorpercentage is obtained below the 20 for nMBE and nRMSE. Meanwhile, the errordistribution of RMSE is obtained below 200 W m2 and maximum biaserror is 0.07. Finally, the values of MBE, RMSE, nMBE, and nRMSE conclude thatthe good performance of the combination of two methods in this study can beapplied to simulate any other weather variable for local necessary.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...