Analysis of Artificial Neural Networks for Forecasting Photovoltaic Energy Generation with Solar Irradiance

Abstract

The growth in the use of solar energy has encouraged the development of techniques for shortterm prediction of solar photovoltaic energy generation (PSPEG). Machine learning with Artificial Neural Networks (ANNs) is the most widely used technique to solve this problem. However, comparative studies of these networks with distinct structural configurations, input parameters and prediction horizon, have not been observed in the literature. In this context, the aim of this study is to evaluate the prediction accuracy of the Global Horizontal Irradiance (GHI), which is often used in the PSPEG, generated by ANN models with different construction structures, sets of input meteorological variables and in three short-term prediction horizons, considering a unique database. The analyses were performed with controlled environment and experimental configuration. The results suggest that ANNs using the input GHI variable provide better accuracy (approximately 10%), and their absence increases error variability. No significant difference (p>0.05) was identified in the prediction error models trained with distinct meteorological input data sets. The prediction errors were similar for the same ANN model in the different prediction horizons, and ANNs with 30 and 60 neurons with one hidden layer demonstrated similar or higher accuracy than those with two hidden layers.

Solar Irradiance Forecasting to Short-Term PV Power: Accuracy Comparison of ANN and LSTM Models

“…This is also observed in the slightly lower R 2 absolute value of the LSTM model at the 60 min horizon (Figure 5b). The decrease in prediction accuracy as the prediction horizon increases is a known result from the PSPPG literature [11,25,27]. However, this study quantifies and statistically validates this result.…”

“…The results showed that temperature and cloud amount are determinant impact factors affecting solar irradiance. Finally, considering the characteristics of studies comparing the accuracy between ANN and LSTM models for short-term prediction, detailed in Table 1, there is a lack of a complete study that uses different exogenous meteorological input variables, some of them being determinants of solar irradiance [55], and also including different short-term prediction horizons that influence the model's accuracy [25,56].…”

“…Dataset split used in three-fold (3-fold) cross-validation for each ANN and LSTM model is presented in Table 5. [25,26]. With the intention of performing the comparative analysis between ANN and LSTM models, the settings of the number of hidden layers, cells and parameters of LSTM models were defined similarly to the previously mentioned ANN models.…”

“…It is highlighted that the motivation for this study originated from previous studies [25][26][27] and by the fact that, to date, no research has been identified in the scientific literature comparing ANN and LSTM models in the PSPPG context, specifically considering the following simultaneous characteristics of controlled experimental configuration, a single dataset with exogenous meteorological input variables and three short-term prediction horizons. Although works have compared these models in other application areas, no study evaluating both models with the exogenous meteorological and solar irradiance data used in this study was observed in the literature.…”

Machine Learning Approach on Time Series for PV-Solar Energy

“…ANN can handle both regression and classification. ANN will work as a nerve system, but transferable mathematical functions establish that system [21,22]. For regression, the process can choose only one, but the classification method can handle several classes.…”