Numerical and machine learning investigation of bioconvection with dual-zone magnetic field in a curved cavity
Journal of Thermal Analysis and Calorimetry, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Basım Tarihi: 2026
- Doi Numarası: 10.1007/s10973-026-15786-9
- Dergi Adı: Journal of Thermal Analysis and Calorimetry
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aerospace Database, Chemical Abstracts Core, Chimica, Compendex, Index Islamicus, INSPEC, Academic Search Ultimate (EBSCO), Engineering Source (EBSCO), Materials Science & Engineering Collection (ProQuest), Technology Collection (ProQuest)
- Anahtar Kelimeler: Curvy boundary, Fe3O4-Water, Gaussian process regression, Induced magnetic field, Magnetotactic bacteria, Neural networks
- Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
- TED Üniversitesi Adresli: Evet
Özet
In this paper, the two dimensional, time independent bioconvective magnetohydrodynamic (MHD) flow is numerically investigated inside a curved cavity including magnetotactic bacteria and iron (Fe) concentration. Radial basis function collocation method is carried out for space derivative approximations in the governing dimensionless equations. The system is concerned under the effect of two different uniform magnetic field (MF) applied separately to the lower (Ha1) and upper (Ha2) portions of the cavity, with varying zone length (lb) of Ha1. The fluid flow, heat, mass and bacteria density transport are examined both in contours and quantitative analysis. Results reveal that the MF applied through the lower part of cavity has a slightly stronger suppression on convective heat, mass and bacteria density transport than the upper part, however, over the full parameter space, Ha2 retains marginally greater global influence on all transport characteristics, as confirmed by the machine learning analysis. The rise in lower zone length causes more suppression on all transport characteristics. The increase in curvy part reduces the cavity area which results in more retarding influence on fluid flow. In a collected set of numerical data, the trained models by neural networks (NN) and Gaussian Process Regression (GPR) are also compared, and the superiority of NN modeling on GPR is found while uncertainty is inherently quantified by GPR.