Intra-Day Solar Power Forecasting Strategy for Managing Virtual Power Plants

Abstract

Solar energy penetration has been on the rise worldwide during the past decade, attracting a growing interest in solar power forecasting over short time horizons. The increasing integration of these resources without accurate power forecasts hinders the grid operation and discourages the use of this renewable resource. To overcome this problem, Virtual Power Plants (VPPs) provide a solution to centralize the management of several installations to minimize the forecasting error. This paper introduces a method to efficiently produce intra-day accurate Photovoltaic (PV) power forecasts at different locations, by using free and available information. Prediction intervals, which are based on the Mean Absolute Error (MAE), account for the forecast uncertainty which provides additional information about the VPP node power generation. The performance of the forecasting strategy has been verified against the power generated by a real PV installation, and a set of ground-based meteorological stations in geographical proximity have been used to emulate a VPP. The forecasting approach is based on a Long Short-Term Memory (LSTM) network and shows similar errors to those obtained with other deep learning methods published in the literature, offering a MAE performance of 44.19 W/m2 under different lead times and launch times. By applying this technique to 8 VPP nodes, the global error is reduced by 12.37% in terms of the MAE, showing huge potential in this environment.

Hour-Ahead Photovoltaic Power Prediction Combining BiLSTM and Bayesian Optimization Algorithm, with Bootstrap Resampling for Interval Predictions

“…These devices record crucial variables such as solar irradiance, wind speed and direction, relative humidity, ambient temperature, and panel temperature. Additionally, measurements of the generated PV power at the facility are available [27,28]. It is noteworthy that the data's temporal resolution is set at 15-min intervals, covering only the time span from 1 January 2021, to 31 December 2022.…”

A Review of State-of-the-Art and Short-Term Forecasting Models for Solar PV Power Generation

“…Figure 2 presents the distribution of the five database input sources, of which the models using meteorological records [81][82][83][84] or numerical weather prediction (NWP) [85][86][87] are dominant, accounting for 49% and 25%, respectively. In several studies, 15% of the power generation information was shared from nearby PV power plants [56,59,88], 6% of the studies used satellite images as the input source data [89,90], and some studies combined with sky images have been very promising. Such studies account for 5% of all studies, although further work is needed to correctly identify cloud layers [72,[91][92][93].…”

“…It is easy to fall into local optimal solutions, thereby reducing the prediction accuracy. Deep learning (DL) networks are neural networks with many hidden layers, which can actively and comprehensively grasp the abstract features of samples by using layer-by-layer training and learning methods to form a feature space [86,89]. It overcomes the shortcomings of BP neural networks and SVM, thereby effectively improving the prediction accuracy.…”

Solar Prediction Strategy for Managing Virtual Power Stations

“…The power prediction at PV and the selected prediction interval for PV facilities in the VPP model can be replicated for real PV and can improve high accuracy with both MAE and RMSE models with values of 12.37% and 11.84% respectively (Moreno et al, 2021). Technology in the past has had an impact on the current electricity network, it is necessary to carry out a fundamental restructuring of fossil energy to serve the needs of renewable energy in the future (Rangu et al, 2020).…”