The impact of input data resolution on neural network forecasting models for wind and photovoltaic energy generation using time series data

Abstract

Accurate photovoltaic (PV) and wind energy forecasting are crucial for grid stability and energy security. There are various modeling techniques and methods to design forecasting models, each leading to different accuracy. In this research, datasets were collected from a 546 kWp grid-connected PV farm and a 2 MW wind turbine for one full year. These data were used to train and test artificial neural network models to forecast day-ahead PV and wind energy utilizing time-series input data with 15-, 30-, and 60-min resolutions. The models were able to forecast the PV energy accurately, while the same models trained for wind showed poor performance. Higher input data resolutions lead to slightly better forecasting performance for the PV farm. Utilizing data with higher resolution can improve the forecast by 1%-5%. While for wind energy forecasting, the resolution has very minor effects, although the 30-min resolution shows a slightly better forecasting performance.