AI-powered analysis of thermally magnetized EMHD Casson hybrid nanofluid

This study introduces an innovative approach of Artificial Neural Networks (ANNs) based on the Levenberg-Marquardt Backpropagation Scheme (ANNs-LMBPS) to simulate the EMHD Darcy-Forchheimer Flow of Casson Hybrid Nanofluid (EMHD-DFFCHNFs) over a stretching surface. The hybrid nanofluid comprises uranium dioxide and molybdenum disulfide nanoparticles suspended in a blood-based fluid. The aim of this investigation is to explore the influence of electro-osmosis, Joule heating and porous media interactions on Casson hybrid nanofluid under EMHD-DF flow conditions. The novelty of this study lies in combining ANNs with EMHD-DFFCHNF modelling in the presence of electro-osmosis and porous media effects, which offer a practical and efficient proxy model for biomedical applications. The governing equations are converted into nonlinear differential equations using an appropriate level of non-similarity transformations and solved numerically using the bvp4c scheme in MATLAB. Furthermore, the ANNs-LMBPS model is utilized to analyze variations in heat transport and skin friction coefficient across various dimensionless physical parameters. The performance of the trained model is evaluated in terms of regression, training state, mean square error and error histogram. Comparison tables provide excellent agreement of ANN predictions with numerical results. The key results show the method’s accuracy by a decrease in error from to for heat transfer and from to for the skin friction coefficient, relative to numerical and ANNs approaches, respectively. The effects of key parameters on the dynamics of bio-magnetic fluid flow are clearly demonstrated through well-structured tables and graphical illustrations, offering comprehensive visual insights into their role and significance. A rise in the magnetic number results in a decreased velocity profile. In contrast, higher electric field parameter, Darcy number and Helmholtz-Smoluchowski velocity contribute to an improvement in the hybrid nanofluid velocity. Furthermore, the results highlight a clear trade-off between thermal enhancement and hydrodynamic resistance. Hybrid nanofluids exhibit superior heat transfer performance, achieving Nusselt number improvements of approximately 3.9% to 5% across various parameter configurations, notably with increased magnetic, Ecker and Darcy numbers. In contrast, conventional nanofluids offer lower skin friction, with up to a 2% reduction in skin friction, making them preferable in flow-sensitive applications. The ANNs-LMBPS model, validated by minimal absolute errors, demonstrates both high predictive accuracy and computational efficiency. These insights support the selection of hybrid nanofluids for heat-intensive systems such as electronic or industrial cooling, while nanofluids remain advantageous where minimizing energy losses due to drag is critical. The findings of this study offer valuable insights for optimizing thermal regulation in biomedical applications, such as thermotherapy, material science, artificial organs, bioinformatics and targeted drug delivery.

Hali tarjima qilinmagan