Photovoltaic (PV) power intermittence impacts electrical grid security and operation. Precise PV power and solar irradiation forecasts have been investigated as significant reducers of such impacts. Predicting solar irradiation involves uncertainties related to the characteristics of time series and their high volatility due to the dependence on many weather conditions. We propose a systematic review of PV power and solar resource forecasting, considering technical aspects related to each applied methodology. Our review covers the performance analysis of various physical, statistical, and machine learning models. These methodologies should contribute to decision-making, being applicable to different sites and climatic conditions. About 42% of the analyzed articles developed hybrid approaches, 83% performed short-term prediction, and more than 78% had, as forecast goal, PV power, solar irradiance, and solar irradiation. Considering spatial forecast scale, 66% predicted in a single field. As a trend for the coming years, we highlight the use of hybridized methodologies, especially those that optimize input and method parameters without loss of precision and postprocessing methodologies aiming at improvements in individualized applications.

Solar energy is one of the main renewable energy sources capable of contributing to global energy demand. However, the solar resource is intermittent, making its integration into the electrical system a difficult task. Here, we present and compare two machine learning techniques, deep learning (DL) and support vector regression (SVR), to verify their behavior for solar forecasting. Our testing from Spain showed that the mean absolute percentage error for predictions using DL and SVR is 7.9% and 8.52%, respectively. The DL achieved the best results for solar energy forecast, but it is worth mentioning that the SVR also obtained satisfactory results.

Power generation from decentralised renewable energy (RE) sources has been increasingly used worldwide. The authors apply portfolio theory (PT) to solar and wind resource forecast, combining those two intermittent RE sources in different percentages to investigate the resultant effects on the prediction error with the use of a proposed impact factor. They use solar and wind data from a weather station in Brazil's Northeast region. The use of PT to improve resource forecast of the specific solar and wind conditions found in that Brazilian region is a pioneer project and an original contribution of their research. Traditionally, PT has been used in the finance sector to reduce investment risks by diversifying applications. Considering predictability, the efficient frontier indicates an optimum portfolio for the period under investigation composed by 30% solar and 70% wind resource, obtained by the smallest calculated standard deviation. The obtained average forecast error for wind speed was −1.54% and for solar irradiance was 3.16%; the average forecast error resulting from the integration of 30% solar and 70% wind was −0.13%. This study innovates by using PT to solar and wind forecast in the planning phase, before the installation of wind and solar plants.

An adaptation of the portfolio theory (PT) is proposed in this article, denoted as PrevPT, “Previsão” (in Portuguese) by PT, to integrate the three artificial neural networks, namely multilayer perceptron (MLP) backpropagation, radial basis function (RBF), and self-organizing map (SOM), based forecasting techniques, aiming to analyze the impact of wind speed forecasting errors and achieve more accurate results. In its first use, the PT goal was to maximize a financial return, at any risk, through the diversification of securities or investments that are not positively correlated. Based on the development of PrevPT, which was used until this work only for solar forecasting, the proposed technique is applied in this paper to integrate and improve the results of individual wind forecasts. Four-year wind speed data (January 2007 to December 2010) from two different locations (Algeciras, Spain and Petrolina, Brazil) were used. Our methodology develops a topology that integrates the forecasts obtained by MLP, RBF, and SOM aiming to obtain smaller forecast errors. By diversifying the forecasted asset, when one of the assets has negative prediction errors, another compensates for them and, thus, the total or partial cancellation of the errors occurs. PrevPT obtains a mean absolute percentage error of 1.13% for Spain and 2.35% for Brazil. PrevPT surpassed the results obtained by the three techniques applied individually in the two locations. The main innovations of the methodology are the significant reduction of errors and optimization of resource planning, and the beneficial features compared to other predictor integration techniques.

Photovoltaic (PV) solar generation is gaining an increasing attention due to technological advances such as higher efficiency and life of PV cells and cost reduction. Due to its vast territory, Brazil is composed of regions that can explore renewable energy sources for electricity generation, and the solar resource is found satisfactorily in several areas of the country. This article presents a solar irradiance prediction mechanism developed using data collected in Fortaleza-CE, Brazil. Due to the fact of its characteristic of unpredictability for this resource, many researchers look for several methods to take the generation of this type of energy. The predictions were performed using a Radial Basis Function (RBF) a computational model based on the human nervous system, it is a technical and effective for time series forecasting, which is a relatively complex problem, Artificial Neural Network (ANN) with the advancement of 1 hour. In the ANN performance, a total of 34.4% forecasts underestimated solar energy availability, 7% of the forecasts obtained error 0 and 58.6% of forecasts overestimated the solar resource. A total of 62.33% of forecasts was between -10% and 10% of forecast error. The prediction mean error was 5.93% and the Mean Absolute Percentage Error (MAPE) was 11.43%.