Short-Term Forecasting of Wind Energy: A Comparison of Deep Learning Frameworks

Abstract

Wind energy has been recognized as the most promising and economical renewable energy source, attracting increasing attention in recent years. However, considering the variability and uncertainty of wind energy, accurate forecasting is crucial to propel high levels of wind energy penetration within electricity markets. In this paper, a comparative framework is proposed where a suite of long short-term memory (LSTM) recurrent neural networks (RNN) models, inclusive of standard, bidirectional, stacked, convolutional, and autoencoder architectures, are implemented to address the existing gaps and limitations of reported wind power forecasting methodologies. These integrated networks are implemented through an iterative process of varying hyperparameters to better assess their effect, and the overall performance of each architecture, when tackling one-hour to three-hours ahead wind power forecasting. The corresponding validation is carried out through hourly wind power data from the Spanish electricity market, collected between 2014 and 2020. The proposed comparative error analysis shows that, overall, the models tend to showcase low error variability and better performance when the networks are able to learn in weekly sequences. The model with the best performance in forecasting one-hour ahead wind power is the stacked LSTM, implemented with weekly learning input sequences, with an average MAPE improvement of roughly 6, 7, and 49%, when compared to standard, bidirectional, and convolutional LSTM models, respectively. In the case of two to three-hours ahead forecasting, the model with the best overall performance is the bidirectional LSTM implemented with weekly learning input sequences, showcasing an average improved MAPE performance from 2 to 23% when compared to the other LSTM architectures implemented.

Wind-Ramp Predictability

“…However, other more sophisticated methodologies, such as the Complete Stochastic Modeling Solution (CoSMos), could be applied for ramp forecasting [48]. Deep-learning technologies can also be applied to the post-processing of NWP results for wind-power applications [47,49]. Regarding atmospheric processes, to improve the prediction of wind ramps associated with storms, high-resolution numerical simulations are being carried out to evaluate the sensitivity of the parameterizations of physical processes (e.g., PBL, convection, etc.)…”

Robust Scheduling of Networked Microgrids for Economics and Resilience Improvement

“…For wind power, the hourahead forecast error could be made below 10%. However, the day-ahead forecast error is generally over 20% [28,29]. As to PV power forecasting, the problem is getting more difficult due to random cloud coverage and changing ambient temperature, both of which affect the PV generation significantly [30,31].…”

Robust Microgrid Scheduling Considering Unintentional Islanding Conditions

“…For wind power, the hour-ahead forecast error could be made below 10%. However, the day-ahead forecast error is generally over 20% [18], [19]. As to PV power forecasting, the problem is getting more difficult due to random cloud coverage and changing ambient temperature, both of which affect the PV generation significantly [20], [21].…”

“…Unlike stochastic optimization, robust optimization doesn't require the probability distributions and correlations of the stochastic variables. Without loss of generality, wind power P W wt , PV power P PV vt and load P L dt are assumed independent, symmetric and continuous random variables as in (19). Meanwhile, the grid-connection condition Z G t is assumed binary random variable.…”

“…To consider the uncertainties of both renewable generation and load, wind power W wt , PV power P PV vt and load P L dt are assumed independent, symmetric and bounded random variables as in (19). For robust optimization, we set γ IS = 0.25 and γ P = 0.5, and solve the robust optimization problem to determine the commitment status of DGs.…”