Predictive analysis of flow profiles of Ag-Au/blood hybrid nanofluid using ANN with bvp4c for the combined impact of dissipative heat and convective condition

• The predictive simulation of hybrid nanofluid shows increasingly significant in biomedical and energy transport systems due to their superior heat transfer properties. • Particularly, blood-based hybrid nanofluids with metallic nanoparticles have shown remarkable potential in targeted drug delivery, therapeutic cooling, hyperthermia treatment, etc. • The present model focuses on the predictive analysis of Ag-Au/blood hybrid nanofluid via a permeable substrate affected by transverse magnetization and dissipative heat impact. • Furthermore, the convective thermal boundary condition with incorporating thermal radiation enriches the thermophysical efficiency of the hybrid configuration, thereby promoting superior heat transfer performance. • The implementation of Joule and Darcy dissipation intensifies the temperature gradient within the system. • The mathematical model formulated for the interaction of proposed assumptions is transformed into their standard dimensionless form by utilizing similarity rules and subsequently handled by using bvp4c numerical algorithm with a predictive design of data-driven neural architecture simultaneously. The predictive simulation of hybrid nanofluid shows increasingly significant in biomedical and energy transport systems due to their superior heat transfer properties. Particularly, blood-based hybrid nanofluids with metallic nanoparticles have shown remarkable potential in targeted drug delivery, therapeutic cooling, hyperthermia treatment, etc. The present model focuses on the predictive analysis of Ag-Au/blood hybrid nanofluid via a permeable substrate affected by transverse magnetization and dissipative heat impact. Furthermore, the convective thermal boundary condition with incorporating thermal radiation enriches the thermophysical efficiency of the hybrid configuration, thereby promoting superior heat transfer performance. The implementation of Joule and Darcy dissipation intensifies the temperature gradient within the system. The mathematical model formulated for the interaction of proposed assumptions is transformed into their standard dimensionless form by utilizing similarity rules and subsequently handled by using bvp4c numerical algorithm with a predictive design of data-driven neural architecture simultaneously. The model is developed and trained using the computed data to forecast both the comparative representations of velocity and thermal outlines are provided to demonstrate the impact of distinct flow-controlling parameters on the system’s behavior. The ANN employs a Levenberg backpropagation algorithm achieving high prediction accuracy with minimal mean square error. The ANN model, precisely predicted the axial, transverse, and temperature profiles with very low validation errors of 1.0947×10⁻⁷, 9.4038×10⁻¹¹, and 2.0535×10⁻⁸, respectively. The gradient values of 1.4505, 9.227×10⁻⁸, and 9.8642×10⁻⁸ demonstrate fast convergence and strong model stability. The ANN outputs closely align with the numerical results (R ≈ 1), confirming the model’s high accuracy and robustness.