Artificial Intelligence-Based Improvement of Empirical Methods for Accurate Global Solar Radiation Forecast: Development and Comparative Analysis

Energies, Journal Year: 2024, Volume and Issue: 17(17), P. 4302 - 4302

Published: Aug. 28, 2024

Artificial intelligence (AI) technology has expanded its potential in environmental and renewable energy applications, particularly the use of artificial neural networks (ANNs) as most widely used technique. To address shortage solar measurement various places worldwide, several radiation methods have been developed to forecast global (GSR). With this consideration, study aims develop temperature-based GSR models using a commonly utilized approach machine learning techniques, ANNs, predict just temperature data. It also compares performance these empirical Additionally, it develops precise for five new sites entire region, which currently lacks AI-based despite presence proposed plants area. The examines impact varying lengths validation datasets on models’ prediction accuracy, received little attention. Furthermore, investigates different ANN architectures estimation introduces comprehensive comparative study. findings indicate that advanced both accurately GSR, with coefficient determination, R2, values ranging from 96% 98%. Moreover, local general formulas model exhibit comparable at non-coastal sites. Conversely, ANN-based perform almost identically, high ability any location, even during winter months. fewer neurons their single hidden layer generally outperform those more. efficacy precision models, ones, are minimally impacted by size data sets. This reveals was significantly influenced weather conditions such clouds rain, especially coastal In contrast, were less variations, approximately 7% better than ones best-developed thus highly recommended. They enable rapid is useful design evaluation continuously easily recorded purposes.

Language: Английский

---

Anomaly detection in all-sky images: An approach using robust ensemble modeling of cloud cover fraction and prediction bounds

Engineering Applications of Artificial Intelligence, Journal Year: 2025, Volume and Issue: 143, P. 110003 - 110003

Published: Jan. 18, 2025

Language: Английский

---

A weather station selection method based on the simulated annealing algorithm for electric load forecasting

Neural Computing and Applications, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 23, 2025

Language: Английский

---

Algorithms for big data mining of hub patent transactions based on decision trees

EPJ Web of Conferences, Journal Year: 2025, Volume and Issue: 318, P. 04013 - 04013

Published: Jan. 1, 2025

Dysfunctions of the patent supply and demand market have a negative impact on sustainability national innovation system, which stimulates growth prices for knowledge-intensive products. It is necessary to establish relationship between fiscal decisions regarding transactions prospects development commercialization results intellectual activity. One most promising methods improving accuracy system analysis big semi-structured transaction data use decision trees. Existing based error backpropagation method are quite slow, their accelerated versions lose in training accuracy. To effectively solve problem forecasting cost hub transactions, parametric algorithms been developed response bias with additional predicative structures model successive geometric transformations. The optimal number tree predicates has established taking into account computational efforts transactions. Based evolutionary computing, values parameters mining established.

Language: Английский

---

A hybrid Gaussian process-integrated deep learning model for retrofitted building energy optimization in smart city ecosystems

Applied Energy, Journal Year: 2025, Volume and Issue: 388, P. 125643 - 125643

Published: March 19, 2025

Language: Английский

---

Sunspot number-based neural network model for global solar radiation estimation in Ghardaïa

STUDIES IN ENGINEERING AND EXACT SCIENCES, Journal Year: 2024, Volume and Issue: 5(2), P. e7156 - e7156

Published: Aug. 27, 2024

In this investigation, the estimation of global solar radiation was meticulously carried out within Ghardaïa city, a region situated in Southern Algeria, utilizing sophisticated multilayer perceptron (MLP) neural network architecture. This research primarily concentrated on developing predictive model based singular input parameter, specifically, sunspot numbers, to forecast levels. The model's formulation rooted empirical data collected over an extensive period from 1984 2000, which used for training network. To assess accuracy and robustness, years 2001 2004 were employed validation purposes. outcomes study highly satisfactory, indicating that MLP-based possesses significant capability Diffuse Global Solar Radiation (DGSR). is substantiated by robust statistical metrics, including normalized Root Mean Square Error (nRMSE) 0.076, reflecting prediction, correlation coefficient (R) 93.16%, denoting strong between predicted observed values. These results underscore efficacy potential application accurately estimating specified region.

Language: Английский

---

Long-term scheduling strategy of hydro-wind-solar complementary system based on chaotic elite selection differential evolution algorithm with death penalty function

Engineering Applications of Artificial Intelligence, Journal Year: 2024, Volume and Issue: 142, P. 109878 - 109878

Published: Dec. 21, 2024

Language: Английский

---

Boosted Equilibrium Optimizer Using New Adaptive Search and Update Strategies for Solving Global Optimization Problems

Electronics, Journal Year: 2024, Volume and Issue: 13(24), P. 5061 - 5061

Published: Dec. 23, 2024

The Equilibrium Optimizer (EO) is an optimization algorithm inspired by a physical law called mass balance, which represents the amount of entering, leaving, and being produced in control volume. Although EO well-accepted successful literature, it needs improvements search, exploration, exploitation phases. Its main problems include low convergence, getting stuck local minima, imbalance between exploration This paper introduces Boosted (BEO) algorithm, where are proposed to solve these improve performance algorithm. New methods for three important phases algorithm: initial population, candidate pool generation, updating. In phase strengthened using uniformly distributed random population instead traditional versatile concentration strategy. Furthermore, balance improved with two new approaches updating phase. These novel enhance algorithm’s more effectively balancing exploitation. tested total 23 standard test functions, including unimodal, multimodal, fixed-size multimodal. results supported numerical values graphs. addition, BEO applied real-world engineering design problems. outperforms original on all

Language: Английский

---

Development of an hourly-based solar radiation prediction model with ANFIS and Coati optimization: a comparative analysis

Theoretical and Applied Climatology, Journal Year: 2024, Volume and Issue: 155(12), P. 9847 - 9869

Published: Oct. 16, 2024

Language: Английский

---