A Review of Solar Forecasting Techniques and the Role of Artificial Intelligence

Solar, Journal Year: 2024, Volume and Issue: 4(1), P. 99 - 135

Published: Feb. 22, 2024

Solar energy forecasting is essential for the effective integration of solar power into electricity grids and optimal management renewable resources. Distinguishing itself from existing literature, this review study provides a nuanced contribution by centering on advancements in techniques. While preceding reviews have examined factors such as meteorological input parameters, time horizons, preprocessing methodology, optimization, sample size, our uniquely delves diverse spectrum spanning ultrashort intervals (1 min to 1 h) more extended durations (up 24 h). This temporal diversity equips decision makers sector with tools enhanced resource allocation refined operational planning. Our investigation highlights prominence Artificial Intelligence (AI) techniques, specifically focusing Neural Networks forecasting, we supervised learning, regression, ensembles, physics-based methods. showcases multifaceted approach address intricate challenges associated predictions. The Satellite Imagery, weather predictions, historical data further augments precision forecasting. In assessing models, describes various error metrics. literature discusses importance metrics, emphasis lies significance standardized datasets benchmark methods ensure accurate evaluations facilitate meaningful comparisons naive forecasts. stands significant advancement field, fostering development models crucial planning emphasizing imperative standardization, thus addressing key gaps research landscape.

Language: Английский

---

Open-source sky image datasets for solar forecasting with deep learning: A comprehensive survey

Renewable and Sustainable Energy Reviews, Journal Year: 2023, Volume and Issue: 189, P. 113977 - 113977

Published: Nov. 3, 2023

Language: Английский

---

Sky image-based solar forecasting using deep learning with heterogeneous multi-location data: Dataset fusion versus transfer learning

Applied Energy, Journal Year: 2024, Volume and Issue: 369, P. 123467 - 123467

Published: June 5, 2024

Solar forecasting from ground-based sky images has shown great promise in reducing the uncertainty solar power generation. With more and image datasets available recent years, development of accurate reliable deep learning-based methods using diverse multi-location data seen a huge growth potential. From that perspective, joint utilization these heterogeneous – such as captured with different camera setups, sensor measurements (i.e., irradiance versus photovoltaic output) varying scale distribution is both unique opportunity critical challenge. This study explores ways to cope heterogeneity compares three strategies for training models based on collected across continents by research groups. Specifically, location interest, we compare performance (1) local trained individually single dataset standard methodology literature); (2) global jointly fusion multiple datasets; (3) locally fine-tuned via transfer learning pre-trained model. The results suggest that, current modeling strategy, work well when deployed locally, but significant errors are observed applied offsite. model, proper normalization prediction targets, can adapt individual locations at cost potential increase efforts. Pre-training large diversified source transferring target generally achieves superior over other two strategies. 80% less data, model performs similarly baseline entire dataset. Overall, algorithms built have be accurate, robust, faster new than location.

Language: Английский

---

Probabilistic load forecasting for integrated energy systems using attentive quantile regression temporal convolutional network

Advances in Applied Energy, Journal Year: 2024, Volume and Issue: 14, P. 100165 - 100165

Published: Feb. 19, 2024

The burgeoning proliferation of integrated energy systems has fostered an unprecedented degree coupling among various streams, thereby elevating the necessity for unified multi-energy forecasting (MEF). Prior approaches predominantly relied on independent predictions heterogeneous load demands, overlooking synergy embedded within dataset. two principal challenges in MEF are extracting intricate correlations diverse loads and accurately capturing inherent uncertainties associated with each type load. This study proposes attentive quantile regression temporal convolutional network (QTCN) as a probabilistic framework MEF, featuring end-to-end predictor intervals electrical, thermal, cooling loads. leverages attention layer to extract between Subsequently, QTCN is implemented retain characteristics data gauge type. multi-task learning deployed facilitate simultaneous quantiles, expediting training progression model. proposed model validated using realistic meteorological from Arizona State University metabolic system National Oceanic Atmospheric Administration respectively, results indicate superior performance greater economic benefits compared baselines existing literature.

Language: Английский

---

Skill‐Driven Data Sampling and Deep Learning Framework for Minute‐Scale Solar Forecasting with Sky Images

Solar RRL, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 10, 2025

Accurate very short‐term solar irradiance forecasting is crucial for optimizing the integration of energy into power systems. Herein, an image‐based deep learning framework minute‐scale prediction presented. The locally developed model benchmarked against two commercial solutions deployed at same experimental site, demonstrating superior accuracy and adaptability. A key contribution introduction a skill‐driven sampling algorithm based on clear sky index persistence error, which optimizes training dataset by excluding low‐utility samples while retaining essential physical features like zenith azimuth angles. This enables exclusion up to 30% original data, resulting in ≈16% savings computational resources without affecting forecast validated using test set 324 991 observations. achieves skill score 7.63%, significantly outperforming models, exhibit negative scores under conditions.

Language: Английский

---

Review on green resources and AI for biogenic solar power

Deleted Journal, Journal Year: 2024, Volume and Issue: 2(1), P. 457 - 457

Published: Feb. 25, 2024

The need for clean and renewable energy has grown dramatically over the past few years. As potential candidates producing green in this region, photovoltaic bio-solar technologies have arisen. This review presents a novel approach designing developing photovoltaics cells using eco-friendly materials artificial intelligence (AI) techniques. An intriguing architecture is outlined cell that fuses electronics with photosynthetic organisms. A recyclable thin-film solar serves as basis of our system. To further maximize effectiveness device, we use AI algorithms. According to statistical calculations, proposed can produce sizable amount electricity while being ecologically sound. paper outlines significant advances nanomaterials AI, which provide exciting improving harvesting capacity. also an overview effects commercialization strategy, its social environmental benefits, pitfalls.

Language: Английский

---

SkyGPT: Probabilistic ultra-short-term solar forecasting using synthetic sky images from physics-constrained VideoGPT

Advances in Applied Energy, Journal Year: 2024, Volume and Issue: 14, P. 100172 - 100172

Published: April 10, 2024

The variability of solar photovoltaic (PV) power output, driven by rapidly changing cloud dynamics, hinders the transition to reliable renewable energy systems. Information on future sky conditions, especially coverage, hold promise for improving PV output forecasting. Leveraging recent advances in generative artificial intelligence (AI), we introduce SkyGPT, a physics-constrained stochastic video prediction model, which predicts plausible images using historical images. We show that SkyGPT can accurately capture producing highly realistic and diverse further demonstrate its efficacy 15-minute-ahead probabilistic forecasting real-world generation data from 30-kW rooftop system. By coupling with U-Net-based observe superior reliability sharpness compared several benchmark methods. propose approach achieves continuous ranked probability score (CRPS) 2.81 kW, outperforming classic convolutional neural network (CNN) baseline 13% smart persistence model 23%. findings this research could aid efficient resilient management electricity generation, particularly as renewable-heavy grids. study also provides valuable insights into modeling broad community, encompassing fields such meteorology atmospheric sciences.

Language: Английский

---

Efficient electricity generation forecasting from solar power plants using technology: Integration, benefits and prospects

Вісник Черкаського державного технологічного університету, Journal Year: 2024, Volume and Issue: 29(1), P. 73 - 85

Published: Feb. 17, 2024

Accurate prediction of electricity generation from renewable sources is an essential element to ensure the stability systems and transition more sustainable energy production. The study aims optimise operation Ukrainian power through introduction required share reliability system. To accuracy forecasting by photovoltaic plants in Ukraine, data analysis, a review existing models methods, comparative analysis using satellite images meteorological observations were used. Low output feature sources, which explained random nature related conditions. In problem qualitative becoming relevant. importance finding effective methods for Ukraine has increased with emergence market. This addresses issue day ahead conditions As part study, issues legislation regarding requirements consequences their failure considered. also reviewed modern explored new “forecasting system market” Ukraine. presents accepted metrics that allow estimating errors comparing effectiveness different methods. Considering dependence on parameters, was carried out. will determine material presented determining model generation, thus increasing efficiency companies reduce negative impact sector environment contribute efficient stable future

Language: Английский

---

A Review of Recent Hardware and Software Advances in GPU-Accelerated Edge-Computing Single-Board Computers (SBCs) for Computer Vision

Sensors, Journal Year: 2024, Volume and Issue: 24(15), P. 4830 - 4830

Published: July 25, 2024

Computer Vision (CV) has become increasingly important for Single-Board Computers (SBCs) due to their widespread deployment in addressing real-world problems. Specifically, the context of smart cities, there is an emerging trend developing end-to-end video analytics solutions designed address urban challenges such as traffic management, disaster response, and waste management. However, deploying CV on SBCs presents several pressing (e.g., limited computation power, inefficient energy real-time processing needs) hindering use at scale. Graphical Processing Units (GPUs) software-level developments have emerged recently these enable elevated performance SBCs; however, it still active area research. There a gap literature comprehensive review recent rapidly evolving advancements both software hardware fronts. The presented provides detailed overview existing GPU-accelerated edge-computing including algorithm optimization techniques, packages, development frameworks, specific packages. This subjective comparative analysis based critical factors help applied Artificial Intelligence (AI) researchers demonstrating state art selecting best suited combinations use-case. At end, paper also discusses potential limitations highlights future research directions this domain.

Language: Английский

---

A radiant shift: Attention-embedded CNNs for accurate solar irradiance forecasting and prediction from sky images

Renewable Energy, Journal Year: 2024, Volume and Issue: 234, P. 121133 - 121133

Published: Aug. 5, 2024

Language: Английский

---