MLFV: a novel machine learning feature vector method to predict characteristics of turbulent heat and fluid flow DOI
Iman Bashtani, Javad Abolfazli Esfahani

International Journal of Numerical Methods for Heat &amp Fluid Flow, Journal Year: 2024, Volume and Issue: 34(10), P. 3979 - 4006

Published: Aug. 22, 2024

Purpose This study aims to introduce a novel machine learning feature vector (MLFV) method bring overcome the time-consuming computational fluid dynamics (CFD) simulations for rapidly predicting turbulent flow characteristics with acceptable accuracy. Design/methodology/approach In this method, CFD snapshots are encoded in tensor as input training data. Then, MLFV learns relationship between data rod filter, which is named vector, learn features by defining functions on it. To demonstrate accuracy of MLFV, used predict velocity, temperature and kinetic energy fields passing over an innovative nature-inspired Dolphin turbulator based only ten Findings The results indicate that contours alongside scatter plots have good agreement predicted solved R 2 ≃ 1. Also, error percentage histograms reveal high precisions predictions MAPE = 7.90E-02, 1.45E-02, 7.32E-02 NRMSE 1.30E-04, 1.61E-03, 4.54E-05 prediction temperature, at Re 20,000, respectively. Practical implications can state-of-the-art applications wide range ability train small data, practical logical regarding number required tests. Originality/value paper introduces novel, super-fast address challenges associated traditional approach physics heat real time superiority

Language: Английский

Energy and exergy assessment of a photovoltaic-thermal (PVT) system cooled by single and hybrid nanofluids DOI Creative Commons
Mohammed Alktranee, Qudama Al-Yasiri,

Khwaiter Imam Rahama Mohammed

et al.

Energy Conversion and Management X, Journal Year: 2024, Volume and Issue: unknown, P. 100769 - 100769

Published: Oct. 1, 2024

Language: Английский

Citations

7
Research on mini-channel heat exchangers with honeycomb modular structure: Design principles and convective heat transfer DOI
Zhaoyuan Wang, S. T. Tan, Jiamin Zhu

et al.

International Journal of Thermal Sciences, Journal Year: 2025, Volume and Issue: 214, P. 109832 - 109832

Published: March 6, 2025

Language: Английский

Citations

0
Heat exchange improvement and drag force reduction around a heated square cylinder controlled by three partitions DOI
Youssef Admi,

Mohammed Amine Moussaoui,

Ahmed Mezrhab

et al.

The European Physical Journal Plus, Journal Year: 2025, Volume and Issue: 140(4)

Published: April 23, 2025

Language: Английский

Citations

0
MLFV: a novel machine learning feature vector method to predict characteristics of turbulent heat and fluid flow DOI
Iman Bashtani, Javad Abolfazli Esfahani

International Journal of Numerical Methods for Heat &amp Fluid Flow, Journal Year: 2024, Volume and Issue: 34(10), P. 3979 - 4006

Published: Aug. 22, 2024

Purpose This study aims to introduce a novel machine learning feature vector (MLFV) method bring overcome the time-consuming computational fluid dynamics (CFD) simulations for rapidly predicting turbulent flow characteristics with acceptable accuracy. Design/methodology/approach In this method, CFD snapshots are encoded in tensor as input training data. Then, MLFV learns relationship between data rod filter, which is named vector, learn features by defining functions on it. To demonstrate accuracy of MLFV, used predict velocity, temperature and kinetic energy fields passing over an innovative nature-inspired Dolphin turbulator based only ten Findings The results indicate that contours alongside scatter plots have good agreement predicted solved R 2 ≃ 1. Also, error percentage histograms reveal high precisions predictions MAPE = 7.90E-02, 1.45E-02, 7.32E-02 NRMSE 1.30E-04, 1.61E-03, 4.54E-05 prediction temperature, at Re 20,000, respectively. Practical implications can state-of-the-art applications wide range ability train small data, practical logical regarding number required tests. Originality/value paper introduces novel, super-fast address challenges associated traditional approach physics heat real time superiority

Language: Английский

Citations

1