Deep learning models for solar irradiance forecasting: A comprehensive review

Journal of Cleaner Production, Journal Year: 2021, Volume and Issue: 318, P. 128566 - 128566

Published: Aug. 11, 2021

Language: Английский

---

An Edge-AI Based Forecasting Approach for Improving Smart Microgrid Efficiency

IEEE Transactions on Industrial Informatics, Journal Year: 2022, Volume and Issue: 18(11), P. 7946 - 7954

Published: March 29, 2022

Smart Grid 2.0 is the energy Internet based on advanced metering infrastructure and distributed systems that require an instantaneous two-way flow of information. Edge computing benefits from its proximity to servers edge nodes smart grid systems, which can provide efficient low latency information transmission grid. With massive number Things being used, amount real-time power usage generated by represents a huge challenge for computing. To improve efficiency processing in this article combines different deep learning algorithms with analyze process renewable generation consumer data microgrid. Experiments two real-world datasets China Belgium show proposed framework obtain satisfactory prediction accuracy compared existing approaches.

Language: Английский

---

Photovoltaic power forecasting: A hybrid deep learning model incorporating transfer learning strategy

Renewable and Sustainable Energy Reviews, Journal Year: 2022, Volume and Issue: 162, P. 112473 - 112473

Published: April 21, 2022

Language: Английский

---

Hourly stepwise forecasting for solar irradiance using integrated hybrid models CNN-LSTM-MLP combined with error correction and VMD

Energy Conversion and Management, Journal Year: 2023, Volume and Issue: 280, P. 116804 - 116804

Published: Feb. 20, 2023

Language: Английский

---

Short-Term Solar Power Predicting Model Based on Multi-Step CNN Stacked LSTM Technique

Energies, Journal Year: 2022, Volume and Issue: 15(6), P. 2150 - 2150

Published: March 15, 2022

Variability in solar irradiance has an impact on the stability of systems and grid’s safety. With decreasing cost panels recent advancements energy conversion technology, precise forecasting is critical for system integration. Despite extensive research, there still potential advancement prediction accuracy, especially global horizontal irradiance. Global Horizontal Irradiance (GHI) (unit: KWh/m2) Plane Of Array (POA) W/m2) were used as objectives this a hybrid short-term model called modified multi-step Convolutional Neural Network (CNN)-stacked Long-Short-Term-Memory network (LSTM) with drop-out was proposed. The real data from Sweihan Photovoltaic Independent Power Project Abu Dhabi, UAE preprocessed, features extracted using CNN layers. output result to predict targets stacked LSTM efficiency proved by comparing statistical performance measures terms Root Mean Square Error (RMSE), Absolute Percentage (MAPE), Squared (MAE), R2 scores, other contemporary machine learning deep-learning-based models. proposed offered best RMSE values 0.36 0.98 61.24 0.96 POA prediction, which also showed better compared published works literature.

Language: Английский

---

Stacked LSTM Sequence-to-Sequence Autoencoder with Feature Selection for Daily Solar Radiation Prediction: A Review and New Modeling Results

Energies, Journal Year: 2022, Volume and Issue: 15(3), P. 1061 - 1061

Published: Jan. 31, 2022

We review the latest modeling techniques and propose new hybrid SAELSTM framework based on Deep Learning (DL) to construct prediction intervals for daily Global Solar Radiation (GSR) using Manta Ray Foraging Optimization (MRFO) feature selection select model parameters. Features are employed as potential inputs Long Short-Term Memory a seq2seq autoencoder system in final GSR prediction. Six solar energy farms Queensland, Australia considered evaluate method with predictors from Climate Models ground-based observation. Comparisons carried out among DL models (i.e., Neural Network) conventional Machine algorithms Gradient Boosting Regression, Random Forest Extremely Randomized Trees, Adaptive Regression). The hyperparameters deduced grid search, simulations demonstrate that is accurate compared other well persistence methods. obtains quality high coverage probability low interval errors. modelling results utilising an deep learning show our approach acceptable predict radiation, therefore useful monitoring systems capture stochastic variations power generation due cloud cover, aerosols, ozone changes, atmospheric attenuation factors.

Language: Английский

---

Short-term wind speed forecasting using an optimized three-phase convolutional neural network fused with bidirectional long short-term memory network model

Applied Energy, Journal Year: 2024, Volume and Issue: 359, P. 122624 - 122624

Published: Jan. 24, 2024

Wind energy is an environment friendly, low-carbon, and cost-effective renewable source. It is, however, difficult to integrate wind into a mixed grid due its high volatility intermittency. For conversion systems be reliable efficient, accurate speed (WS) forecasting fundamental. This study cascades convolutional neural network (CNN) with bidirectional long short-term memory (BiLSTM) in order obtain model for hourly WS by utilizing several meteorological variables as inputs their effects on predicted WS. input selection, the mutation grey wolf optimizer (TMGWO) used. efficient optimization of CBiLSTM hyperparameters, hybrid Bayesian Optimization HyperBand (BOHB) algorithm The combined usage TMGWO, BOHB, leads three-phase (i.e., 3P-CBiLSTM). performance 3P-CBiLSTM benchmarked against standalone BiLSTMs, LSTMs, gradient boosting (GBRs), random forest (RFRs), decision tree regressors (DTRs). statistical analysis forecasted reveals that highly effective over other benchmark methods. objective also registers highest percentage errors (≈ 53.4 – 81.8%) within smallest error range ≤ |0.25| ms−1 amongst all tested sites. Despite remarkable results achieved, cannot generally understood, so eXplainable Artificial Intelligence (xAI) technique was used explaining local global outputs, based Local Interpretable Model-Agnostic Explanations (LIME) SHapley Additive exPlanations (SHAP). Both xAI methods determined antecedent most significant predictor forecasting. Therefore, we aver proposed can employed help farm operators making quality decisions maximizing power integration reduced

Language: Английский

---

Power prediction of a wind farm cluster based on spatiotemporal correlations

Applied Energy, Journal Year: 2021, Volume and Issue: 302, P. 117568 - 117568

Published: Aug. 15, 2021

Language: Английский

---

Short-term solar irradiance forecasting based on a novel Bayesian optimized deep Long Short-Term Memory neural network

Applied Energy, Journal Year: 2022, Volume and Issue: 324, P. 119727 - 119727

Published: July 30, 2022

Language: Английский

---

Interpretable deep learning models for hourly solar radiation prediction based on graph neural network and attention

Applied Energy, Journal Year: 2022, Volume and Issue: 321, P. 119288 - 119288

Published: May 31, 2022

Language: Английский

---