To better forecast solar variability, spatio-temporal methods exploit spatially distributed solar time series, seeking to improve forecasting accuracy by including neighboring solar information. This review work is, to the authors’ understanding, the first to offer a compendium of references published since 2011 on such approaches for global horizontal irradiance and photovoltaic generation. The identified bibliography was categorized according to different parameters (method, data sources, baselines, performance metrics, forecasting horizon), and associated statistics were explored. Lastly, general findings are outlined, and suggestions for future research are provided based on the identification of less explored methods and data sources.

The availability of solar irradiance time series without missing data is an ideal scenario for researchers in the field. However, it is not achievable for a variety of reasons, such as measurement errors, sampling gaps, or other factors. Time series imputation methods can be a solution to the lack of data and, in this paper, we study the applicability of Bidirectional Encoder Representations from Transformers (BERT) as an irradiance time series imputation solution. In this regard, a BERT model was trained from scratch for the masked language modelling (MLM) task, and the quality of the imputation was evaluated according to the number of missing values and the position within the series. The experiments were conducted over a dataset of 165 stations, captured by meteorological stations distributed over the Spanish regions of Galicia, Castile, and León. In the evaluation process, an average coefficient of determination (R² score) of 0.89% was obtained, the maximum result being 0.95%.

Accurate solar resource forecasting is important because of the inherent variability associated with solar energy and its significant impact on the cost for energy producers. The traditional method applied in solar irradiance forecasting involves two main phases, related to (1) data selection and (2) model selection, training, and evaluation. In this study, we propose a novel end-to-end methodology for solar irradiance forecasting that starts with the search for the data and all of the preprocessing operations involved in obtaining a quality dataset, continuing by imputing missing data with the BERT (Bidirectional Encoder Representations from Transformers) model, and ending with obtaining and evaluating the predicted values. This novel methodology is based on three phases; namely, Phase_1, related to the acquisition and preparation of the data, Phase_2, related to the proposed imputation with a BERT model, and Phase_3, related to the training and prediction with new models based on deep learning. These phases of the proposed methodology can be applied in a disjointed manner, and were used on two public datasets accessible to the scientific community. Each of the proposed phases proved to be valuable for the workflow, and the application of the novel method delivered increases in performance of up to 3 percentage points (3%) when compared to the traditional approach.

In solar forecasting, it is common practice for solar data (be it irradiance or photovoltaic power) to be converted into a stationary index (e.g., clear-sky or clearness index) before being used as inputs for solar-forecasting models. However, its actual impact is rarely quantified. Thus, this paper aims to study the impact of including this processing step in the modeling workflow within the scope of very-short-term spatio-temporal forecasting. Several forecasting models are considered, and the observed impact is shown to be model-dependent. Persistence does not benefit from this for such short timescales; however, the statistical models achieve an additional 0.5 to 2.5 percentual points (PPs) in terms of the forecasting skill. Machine-learning (ML) models achieve 0.9 to 1.9 more PPs compared to a linear regression, indicating that stationarization reveals non-linear patterns in the data. The exception is Random Forest, which underperforms in comparison with the other models. Lastly, the inclusion of solar elevation and azimuth angles as inputs is tested since these are easy to compute and can inform the model on time-dependent patterns. Only the cases where the input is not made stationary, or the underperforming Random Forest model, seem to benefit from this. This indicates that the apparent Sun position data can compensate for the lack of stationarization in the solar inputs and can help the models to differentiate the daily and seasonal variability from the shorter-term, weather-driven variability.

Accurate solar forecasting lately relies on advances in the field of artificial intelligence and on the availability of databases with large amounts of information on meteorological variables. In this paper, we present the methodology applied to introduce a large-scale, public, and solar irradiance dataset, CyL-GHI, containing refined data from 37 stations found within the Spanish region of Castile and León (Spanish: Castilla y León, or CyL). In addition to the data cleaning steps, the procedure also features steps that enable the addition of meteorological and geographical variables that complement the value of the initial data. The proposed dataset, resulting from applying the processing methodology, is delivered both in raw format and with the quality processing applied, and continuously covers 18 years (the period from 1 January 2002 to 31 December 2019), with a temporal resolution of 30 min. CyL-GHI can result in great importance in studies focused on the spatial-temporal characteristics of solar irradiance data, due to the geographical information considered that enables a regional analysis of the phenomena (the 37 stations cover a land area larger than 94,226 km²). Afterwards, three popular artificial intelligence algorithms were optimised and tested on CyL-GHI, their performance values being offered as baselines to compare other forecasting implementations. Furthermore, the ERA5 values corresponding to the studied area were analysed and compared with performance values delivered by the trained models. The inclusion of previous observations of neighbours as input to an optimised Random Forest model (applying a spatio-temporal approach) improved the predictive capability of the machine learning models by almost 3%.

In this work, three new convolutional neural network models—spatio-temporal convolutional neural network versions 1 and 2 (ST_CNN_v1 and ST_CNN_v2), and the spatio-temporal dilated convolutional neural network (ST_Dilated_CNN)—are proposed for solar forecasting and processing global horizontal irradiance (GHI) data enriched with meteorological and astronomical variables. A comparative analysis of the proposed models with two traditional benchmark models shows that the proposed ST_Dilated_CNN model outperforms the rest in capturing long-range dependencies, achieving a mean absolute error of 31.12 W/m², a mean squared error of 54.07 W/m², and a forecast skill of 37.21%. The statistical analysis carried out on the test set suggested highly significant differences in performance (p-values lower than 0.001 for all metrics in all the considered scenarios), with the model with the lowest variability in performance being ST_CNN_v2. The statistical tests applied confirmed the robustness and reliability of the proposed models under different conditions. In addition, this work highlights the significant influence of astronomical variables on prediction performance. The study also highlights the intricate relationship between the proposed models and meteorological and astronomical input characteristics, providing important insights into the field of solar prediction and reaffirming the need for further research into variability factors that affect the performance of models.