Neural network analysis of electro-osmotic Sutterby nanofluidics in complex porous microsystem

Purpose The purpose of this study is to investigate nonlinear electro-osmotic transport of Sutterby nanofluid in a tapered porous microchannel saturated with a non-Darcy medium. This work analyzes the combined effects of electro-hydrodynamic forces, heat transfer, viscous dissipation, Joule heating and motile microorganism-induced bioconvection. A novel hybrid framework based on the homotopy perturbation method and artificial neural network is developed specifically for this problem to provide accurate analytical interpretation and fast predictive modeling for advanced microfluidic and biomedical applications. Methodology The governing nonlinear equations are formulated using Debye–Hückel linearization for the electric double layer and lubrication theory to simplify the flow model. The homotopy perturbation method is used to obtain analytical solutions. An artificial neural network model is then developed in TensorFlow using homotopy perturbation method (HPM)-generated data, ReLU activation and the Adam optimizer for accurate prediction. Findings The results of this study show that electro-osmotic effects enhance the pressure gradient, while the modified Darcy parameter reduces flow resistance. The Forchheimer parameter increases fluid temperature because of inertial resistance, whereas the Darcy parameter supports cooling and bolus formation. Higher Peclet numbers suppress microorganism transport and weaken bioconvection. The artificial neural network (ANN) model achieves excellent prediction accuracy with $R2 > 0.999$. Originality/value The originality of this study lies in presenting a novel hybrid HPM–ANN framework for nonlinear electro-osmotic Sutterby nanofluid flow with bioconvection in a tapered microchannel saturated with a modified Darcy–Forchheimer porous medium. Unlike previous studies, the model incorporates modified Darcy–Forchheimer dissipation to accurately capture thermal behavior, along with velocity slip wall conditions for a more realistic microchannel flow description. The coupling of analytical HPM solutions with TensorFlow-based ANN prediction provides a fast, reliable and highly accurate strategy for complex electro-biothermal microfluidic transport systems.