Computational intelligence of Bayesian regularization backpropagation neural networks to study the thermo-bioconvection flow of hybrid-nanofluid with thermal Radiation: Biotechnological perspectives

This study aims to develop a deep neural network that utilizes Bayesian regularization to investigate the performance of gyrotactic and oxytactic microbes in hybrid nanofluid flow over a sheet, taking into account local thermal non-equilibrium effects and thermal radiation. Two different activation...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	Journal of Radiation Research and Applied Sciences Ročník 18; číslo 3; s. 101777
Hlavní autoři:	Abbas, Munawar, Medani, Mohamed, Rajab, Adnan Burhan, Elaissi, Samira, Zayani, Hafedh Mahmoud, Faiz, Zeshan, Khan, Ilyas, Ben Khedher, Nidhal
Médium:	Journal Article
Jazyk:	angličtina
Vydáno:	Elsevier B.V 01.09.2025
Témata:	Bayesian regularization algorithm Hybrid nanofluid Local thermal non-equilibrium conditions Oxytactic and gyrotactic microbe Thermophoretic particle deposition Oxytactic and gyrotactic microbe Thermophoretic particle deposition Bayesian regularization algorithm Local thermal non-equilibrium conditions Hybrid nanofluid
ISSN:	1687-8507
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Popis
Shrnutí:	This study aims to develop a deep neural network that utilizes Bayesian regularization to investigate the performance of gyrotactic and oxytactic microbes in hybrid nanofluid flow over a sheet, taking into account local thermal non-equilibrium effects and thermal radiation. Two different activation functions, namely radial basis and log-sigmoid utilized in the designed network. The model improves the efficiency of pollutant removal by optimizing the dynamics of gyrotactic and oxytactic microbes under local thermal non-equilibrium conditions. The enhanced microbial activity for the degradation of organic waste and the breakdown of pollutants are made easier by the hybrid nanofluid's improved thermal and physicochemical characteristics. 80 % of the data fixed for the training of the deep neural network and 20 % data employed for the testing purpose. The base fluid water is coupled with titanium dioxide (TiO2) and iron oxide (Fe3O4) nanoparticles to produce a hybrid nanofluid. The input dataset of the network generated using the bvp4c command. To examine how the various parameters on the suggested model varied, three distinct examples were created. Various statistical metrics are taken into consideration to assess the network's accuracy and precision. [Display omitted]
ISSN:	1687-8507
DOI:	10.1016/j.jrras.2025.101777