Artificial intelligence for sustainability: Challenges, opportunities, and a research agenda

International Journal of Information Management, Journal Year: 2020, Volume and Issue: 53, P. 102104 - 102104

Published: April 20, 2020

Language: Английский

---

Short-Term Photovoltaic Power Forecasting Using an LSTM Neural Network and Synthetic Weather Forecast

IEEE Access, Journal Year: 2020, Volume and Issue: 8, P. 172524 - 172533

Published: Jan. 1, 2020

In this paper, a forecasting algorithm is proposed to predict photovoltaic (PV) power generation using long short term memory (LSTM) neural network (NN). A synthetic weather forecast created for the targeted PV plant location by integrating statistical knowledge of historical solar irradiance data with publicly available type sky host city. To achieve this, K-means used classify into dynamic groups that vary from hour in same season. other words, types are defined each uniquely different levels based on day and This can mitigate performance limitations fixed categories translating them numerical data. The proved embed features data, which results significant improvement accuracy. model investigated intraday horizon lengths seasons. It shown up 33% accuracy comparison when an hourly categorical used, 44.6% daily used. highlights significance utilizing forecast, promote more efficient utilization reliable prediction. Moreover, superiority LSTM NN verified investigating machine learning engines, namely recurrent (RNN), generalized regression (GRNN) extreme (ELM).

Language: Английский

---

Advanced Methods for Photovoltaic Output Power Forecasting: A Review

Applied Sciences, Journal Year: 2020, Volume and Issue: 10(2), P. 487 - 487

Published: Jan. 9, 2020

Forecasting is a crucial task for successfully integrating photovoltaic (PV) output power into the grid. The design of accurate forecasters remains challenging issue, particularly multistep-ahead prediction. Accurate PV forecasting critical in number applications, such as micro-grids (MGs), energy optimization and management, integrated smart buildings, electrical vehicle chartering. Over last decade, vast literature has been produced on this topic, investigating numerical probabilistic methods, physical models, artificial intelligence (AI) techniques. This paper aims at providing complete review recent applications AI techniques; we will focus machine learning (ML), deep (DL), hybrid these branches are becoming increasingly attractive. Special attention be paid to development application DL, well future trends topic.

Language: Английский

---

Recent Advances in Machine Learning Research for Nanofluid-Based Heat Transfer in Renewable Energy System

Energy & Fuels, Journal Year: 2022, Volume and Issue: 36(13), P. 6626 - 6658

Published: June 13, 2022

Nanofluids have gained significant popularity in the field of sustainable and renewable energy systems. The heat transfer capacity working fluid has a huge impact on efficiency system. addition small amount high thermal conductivity solid nanoparticles to base improves transfer. Even though large research data is available literature, some results are contradictory. Many influencing factors, as well nonlinearity refutations, make nanofluid highly challenging obstruct its potentially valuable uses. On other hand, data-driven machine learning techniques would be very useful for forecasting thermophysical features rate, identifying most influential assessing efficiencies different primary aim this review study look at applications employed nanofluid-based system, reveal new developments research. A variety modern algorithms studies systems examined, along with their advantages disadvantages. Artificial neural networks-based model prediction using contemporary commercial software simple develop popular. prognostic may further improved by combining marine predator algorithm, genetic swarm intelligence optimization, intelligent optimization approaches. In well-known networks fuzzy- gene-based techniques, newer ensemble such Boosted regression K-means, K-nearest neighbor (KNN), CatBoost, XGBoost gaining due architectures adaptabilities diverse types. regularly used fuzzy-based mostly black-box methods, user having little or no understanding how they function. This reason concern, ethical artificial required.

Language: Английский

---

Day-ahead photovoltaic power forecasting approach based on deep convolutional neural networks and meta learning

International Journal of Electrical Power & Energy Systems, Journal Year: 2019, Volume and Issue: 118, P. 105790 - 105790

Published: Dec. 31, 2019

Language: Английский

---

A review and taxonomy of wind and solar energy forecasting methods based on deep learning

Energy and AI, Journal Year: 2021, Volume and Issue: 4, P. 100060 - 100060

Published: March 7, 2021

Renewable energy is essential for planet sustainability. output forecasting has a significant impact on making decisions related to operating and managing power systems. Accurate prediction of renewable vital ensure grid reliability permanency reduce the risk cost market Deep learning's recent success in many applications attracted researchers this field its promising potential manifested richness proposed methods increasing number publications. To facilitate further research development area, paper provides review deep learning-based solar wind published during last five years discussing extensively data datasets used reviewed works, pre-processing methods, deterministic probabilistic evaluation comparison methods. The core characteristics all works are summarised tabular forms enable methodological comparisons. current challenges future directions given. trends show that hybrid models most followed by Recurrent Neural Network including Long Short-Term Memory Gated Unit, third place Convolutional Networks. We also find multistep ahead gaining more attention. Moreover, we devise broad taxonomy using key insights gained from extensive review, believe will be understanding cutting-edge accelerating innovation field.

Language: Английский

---

Machine learning on sustainable energy: A review and outlook on renewable energy systems, catalysis, smart grid and energy storage

Process Safety and Environmental Protection, Journal Year: 2021, Volume and Issue: 174, P. 414 - 441

Published: Aug. 17, 2021

Language: Английский

---

Deep learning and wavelet transform integrated approach for short-term solar PV power prediction

Measurement, Journal Year: 2020, Volume and Issue: 166, P. 108250 - 108250

Published: July 20, 2020

Language: Английский

---

Load Forecasting Techniques for Power System: Research Challenges and Survey

IEEE Access, Journal Year: 2022, Volume and Issue: 10, P. 71054 - 71090

Published: Jan. 1, 2022

The main and pivot part of electric companies is the load forecasting. Decision-makers think tank power sectors should forecast future need electricity with large accuracy small error to give uninterrupted free shedding consumers. demand can be forecasted amicably by many Machine Learning (ML), Deep (DL) Artificial Intelligence (AI) techniques among which hybrid methods are most popular. present technologies forecasting work regarding combination various ML, DL AI algorithms reviewed in this paper. comprehensive review single models functions; advantages disadvantages discussed comparison between performance terms Mean Absolute Error (MAE), Root Squared (RMSE), Percentage (MAPE) values compared literature different support researchers select best model for prediction. This validates fact that will provide a more optimal solution.

Language: Английский

---

Artificial intelligence techniques for enabling Big Data services in distribution networks: A review

Renewable and Sustainable Energy Reviews, Journal Year: 2021, Volume and Issue: 150, P. 111459 - 111459

Published: July 16, 2021

Artificial intelligence techniques lead to data-driven energy services in distribution power systems by extracting value from the data generated deployed metering and sensing devices. This paper performs a holistic analysis of artificial applications networks, ranging operation, monitoring maintenance planning. The potential for system needed sources are identified classified. following networks analyzed: topology estimation, observability, fraud detection, predictive maintenance, non-technical losses forecasting, management systems, aggregated flexibility trading. A review methods implemented each these is conducted. Their interdependencies mapped, proving that multiple can be offered as single clustered service different stakeholders. Furthermore, dependencies between AI with identified. In recent years there has been significant rise deep learning time series prediction tasks. Another finding unsupervised mainly being applied customer segmentation, buildings efficiency clustering consumption profile grouping detection. Reinforcement widely design, although more testing real environments needed. Distribution network sensorization should enhanced increased order obtain larger amounts valuable data, enabling better outcomes. Finally, future opportunities challenges applying grids discussed.

Language: Английский

---