ABSTRACT A numerical study examined buoyancy‐driven flow in a square cavity containing an inclined porous cylinder using a carbon nanotube (CNT)–water nanofluid. CNT–water nanofluid was selected as the working fluid, with its effective thermal and physical properties evaluated according to Thang et al. (2015). The effects of Darcy number ( Da ), Rayleigh number ( Ra ), and nanoparticle volume fraction ( ϕ ) were analyzed to examined the influence of key parameters 10 −6 ≤ Da ≤ 10 −2 ,10 3 ≤ Ra ≤ 10 6 , and0 ≤ ϕ ≤ 0.05, using finite element simulation. Additionally, machine learning (ML) models, random forest (RF), decision tree (DT), XGBoost, and support vector regression (SVR) were trained on the numerical data to predict Nu avg . Results showed that higher Da and Ra enhanced flow intensity and heat transfer. ML models predicted average Nusselt number, with RF achieving the best accuracy ( R 2 = 0.9959). SHAP analysis identified Da as the dominant factor influencing convective heat transfer performance.
Sinha et al. (2026) studied this question.