Photovoltaic Power Prediction Using Artificial Neural Networks and Numerical Weather Data

Abstract

The monitoring of power generation installations is key for modelling and predicting their future behaviour. Many renewable energy generation systems, such as photovoltaic panels and wind turbines, strongly depend on weather conditions. However, in situ measurements of relevant weather variables are not always taken into account when designing monitoring systems, and only power output is available. This paper aims to combine data from a Numerical Weather Prediction model with machine learning tools in order to accurately predict the power generation from a photovoltaic system. An Artificial Neural Network (ANN) model is used to predict power outputs from a real installation located in Puglia (southern Italy) using temperature and solar irradiation data taken from the Global Data Assimilation System (GDAS) sflux model outputs. Power outputs and weather monitoring data from the PV installation are used as a reference dataset. Three training and testing scenarios are designed. In the first one, weather data monitoring is used to both train the ANN model and predict power outputs. In the second one, training is done with monitoring data, but GDAS data is used to predict the results. In the last set, both training and result prediction are done by feeding GDAS weather data into the ANN model. The results show that the tested numerical weather model can be combined with machine learning tools to model the output of PV systems with less than 10% error, even when in situ weather measurements are not available.

Conceptual Design of a Semi-Automatic Process Line for Recycling Photovoltaic Panels as a Way to Ecological Sustainable Production

“…Photovoltaic installations are very popular because they are available to most people, and their installation does not require as much capital as in the case of installations using other renewable energy sources which is confirmed by many authors in their works [1][2][3][4][5]. In addition, along with the development of technology, they are becoming cheaper and more effective way to obtain electricity [6][7][8]. Currently, little attention is paid to the problem of recycling waste generated as a result of the development of the photovoltaic industry, but in the coming years it may turn out to be a big mistake.…”

Photovoltaic power electricity generation nowcasting combining sky camera images and learning supervised algorithms in the Southern Spain

“…In this paper, it was shown that adding time series inputs, in the form of common weather variables, does not improve the current ANN model. Gomes et al [18] forecasted the solar resource in a 4-h time horizon, using meteorological variables as inputs to the ANN.…”

Short-Term Forecasting and Uncertainty Analysis of Photovoltaic Power Based on the FCM-WOA-BILSTM Model

“…For a long time, domestic and foreign scholars have done a lot of research on the short-term prediction of optical volt power in the direct method to solve the problems of photovoltaic grid connection to maintain the stability of the power system. Researchers have successively proposed support vector machine (SVM) (Mayer and Gróf, 2021), Markov chain (Hu and Zhang, 2018), limit learning machine (Wang, 2018), artificial neural network (ANN) (López Gómez et al, 2020), time series prediction and other methods (Zhu et al, 2019;Singh et al, 2021). Traditional ANN achieves power prediction by establishing a mapping between input data and output data, and the lack of consideration of time correlation in the data series makes it impossible for neural network models to capture the relationship between data and time, which limits its application in time series forecasting methods.…”