Some numerical works in the literature applied the semi-empirical correlations for estimating temperature-dependent properties of hybrid nanofluids (HyNf). Nevertheless, these correlations (developed based on the regression analysis of an extensive variety of experimental data), were originally proposed for mono NFs. No one has examined the accuracy of properties data of mono/hybrid NF obtained by using semi-empirical correlations compared to the relevant experimental data. In this study, the SST k-ω model in the single-phase framework based on OpenFOAM is applied for the turbulent forced convection heat transfer inside a double forward-facing step (DFFS) channel. The effects of converging/diverging bottom adiabatic wall and dispersing hybrid CNT-TiO2, ND-Ni, and mono Ni NPs in the water (base fluid) are investigated considering various imposed velocities at the inlet. The NFs and pure fluid properties are temperature-dependent, and various performance evaluation criteria (PEC) are evaluated. Compared to available experimental data, one well-known empirical model underestimates the value of FOM and does not generate the temperature-dependent variation of FOM while the other model overpredicts it. While replacing the canonical case with the converging channel improved the average Nu number considerably, the augmentation of the pressure drop exceeds it. The diverging channel is found preferable compared to the canonical case irrespective of the kind of working fluid. The thermal efficiency of less effective NFs in separating flow is affected strongly not only by the deflection angle of the bottom wall but also by the imposed velocity. Based on PECnf, using TiO2-CNT/water NF with ϕ = 0.001, 0.002 gives the highest thermo-hydraulic performance compared to pure water, and it is effective for converging/diverging channels at various imposed velocities.
