A Data-Driven Approach on Bioconvection Flow

Pekmen B., Gürbüz Çaldağ M.

25th International Conference on Computational Science and Its Applications, ICCSA 2025, İstanbul, Türkiye, 30 Haziran - 03 Temmuz 2025, cilt.15650 LNCS, ss.356-365, (Tam Metin Bildiri) identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası: 15650 LNCS
  • Doi Numarası: 10.1007/978-3-031-96962-1_25
  • Basıldığı Şehir: İstanbul
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.356-365
  • Anahtar Kelimeler: Bioconvection Flow, Feature Importance, Neural Network, Oxytactic Bacteria, Radial Basis Function
  • TED Üniversitesi Adresli: Evet

Özet

In this study, machine learning modeling is utilized on a bioconvection flow problem. The two dimensional, time independent natural convection flow problem in the presence of oxytactic bacteria is considered inside a unit square cavity. The dimensionless governing equations in terms of stream function, temperature, oxygen concentration, bacteria density and vorticity are numerically solved by radial basis function collocation method. In variation of Rayleigh, bioconvection Rayleigh, Peclet, Lewis and buoyancy ratio parameter, average Nusselt, Sherwood and bacteria density along the heated wall as well as average bacteria density throughout the entire cavity are calculated, and organized into a matrix to form a dataset. This dataset is used for neural network modeling to model the outcomes as averages. The obtained mean squared error metric results show the good fit even by one layer neural network. A feature importance analysis also approves the importance of Rayleigh number on average Nusselt number, and Peclet number on all other averages.