Multi-step prediction of carbon emissions based on a secondary decomposition framework coupled with stacking ensemble strategy

Environmental Science and Pollution Research, Journal Year: 2023, Volume and Issue: 30(27), P. 71063 - 71087

Published: May 9, 2023

Language: Английский

---

Short-Term Load Forecasting of Electricity Demand for the Residential Sector Based on Modelling Techniques: A Systematic Review

Energies, Journal Year: 2023, Volume and Issue: 16(10), P. 4098 - 4098

Published: May 15, 2023

In this paper, a systematic literature review is presented, through survey of the main digital databases, regarding modelling methods for Short-Term Load Forecasting (STLF) hourly electricity demand residential and to realize performance evolution impact Artificial Intelligence (AI) in STLF. With these specific objectives, conceptual framework on subject was developed, along with based scientific publications high bibliometric study directed towards production AI The research articles over 10-year period, which took place between 2012 2022, used Preferred Reporting Items Systematic Meta-Analyses (PRISMA) method. This resulted more than 300 articles, available four databases: Web Science, IEEE Xplore, Scopus, Science Direct. organized around three central themes, were defined following keywords: STLF, Electricity, Residential, their corresponding synonyms. total, 334 analyzed, year publication, journal, author, geography by continent country, area application identified. Of 335 documents found initial after applying inclusion/exclusion criteria, allowed delimiting addressed topics interest analysis, 38 (thirty-eight) English (26 journal 12 conference papers). results point diversity techniques associated algorithms. measured different metrics and, therefore, cannot be compared directly. Hence, it desirable have unified dataset, together set benchmarks well-defined clear comparison all

Language: Английский

---

A reinforcement learning-based online learning strategy for real-time short-term load forecasting

Energy, Journal Year: 2024, Volume and Issue: 305, P. 132344 - 132344

Published: July 9, 2024

Real-time Short-Term Load Forecasting (STLF) is crucial for energy management and power system operations. Conventional Machine Learning (ML) methodologies STLF are often challenged by the inherent variability in demand. To tackle challenge associated with variability, this paper presents a novel Reinforcement (RL)-enhanced method. Different from conventional methods, our method dynamically improves model selecting optimal training data to capture recent usage trends possible variations demand patterns. By doing so, can significantly reduce impact of unforeseen fluctuations real-time forecasting. In addition RL-enhanced method, we propose comprehensive evaluation framework, encompassing three key dimensions: accuracy, runtime efficiency, robustness. Tested on distinct real-world datasets, demonstrates superior forecasting performance across metrics achieving accurate robust predictions under varying scenarios. Furthermore, approach provides uncertainty bounds practical applications. These results underscore significant advancements made RL-based precision, We have algorithm openly accessible online promote continued development advancement methods.

Language: Английский

---

Short-term power load forecasting based on sparrow search algorithm-variational mode decomposition and attention-long short-term memory

International Journal of Low-Carbon Technologies, Journal Year: 2024, Volume and Issue: 19, P. 1089 - 1097

Published: Jan. 1, 2024

Abstract To improve the forecasting accuracy of power load, model based on sparrow search algorithm (SSA), variational mode decomposition (VMD), attention mechanism and long short-term memory (LSTM) was proposed. Firstly, SSA is used to optimize number penalty factor in VMD realize operation initial data. Then, LSTM predict each component, this basis, feature temporal mechanisms are introduced. Feature introduced calculate contribution rate relevant input features real time, weights modified avoid limitations traditional methods relying threshold expert experience association rules. Temporal applied extract historical key moments stability time series prediction effect. Finally, final result obtained by superimposing results component complete load prediction. Practical examples show that, compared with other methods, proposed achieves highest accuracy, an RMSE 1.23, MAE 0.99 MAPE 11.62%.

Language: Английский

---

Cakformer: Transformer Model for Long-Term Heat Load Forecasting Based on Cauto-Correlation and KAN

Energy, Journal Year: 2025, Volume and Issue: unknown, P. 135460 - 135460

Published: March 1, 2025

Language: Английский

---