Abstract Recent technological advances and increased participation of energy systems based on photovoltaic solar energy place this renewable energy source in a prominent position in the current scenario. With the increase in the share of solar photovoltaic systems, the impact of power fluctuations in these sources has worsened, which can affect the quality of electrical energy and the reliability of the electrical power system. Therefore, with the use of energy storage together with control algorithms based on artificial intelligence, it is possible to control and perform power smoothing. In this context, the study presents a technical feasibility study on the use of artificial neural network (ANN) to perform the power smoothing of the photovoltaic system connected to the network. Being studied the performance of a real photovoltaic system operating in conjunction with an ideal energy storage for comparative analysis of the performance of the artificial neural network when the numbers of neurons and layers are modified for different real operating conditions considered as temperature variation, humidity, irradiation, pressure and wind speed, which are considered to be ANN input data. The results obtained point to the feasibility of using ANN, with acceptable precision, for power smoothing. According to the analyzes carried out, it is clear that ANN's with few neurons, the smoothing profile tends to be more accurate when compared to larger amounts of neurons. In the current state of the study, it was not possible to determine a relationship between the variations in the number of neurons with the most accurate results, it is important to note that the development of the curve pointed by the neural network can be influenced by the database. It should be noted that, when ANN exceeds or does not reach the optimal smoothing curve, the storage system compensates for the lack or excess of power, and there is a need for other mechanisms to optimize power smoothing.

Abstract The growth in the use of solar energy has encouraged the development of techniques for short-term 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.