Prediction and optimization of unsteady forced convection around a rounded cornered square cylinder in the range of Re

An unsteady two-dimensional laminar forced convection heat transfer around a square cylinder with the rounded corner edge is numerically investigated for Re = 80–180 and non-dimensional corner radius, r = 0.50–0.71 at Pr = 0.71 (Air). A structured non-uniform mesh is used for the computational domain discretization, and the finite-volume-method-based commercial code FLUENT is used for numerical simulation. The heat transfer characteristics over the rounded corner square cylinder are analyzed with average Nusselt number (Nu _avg) at various Re and various corner radii. The heat transfer characteristic is predicted by gene expression programming (GEP), and the GEP-generated explicit equation of Nu _avg is utilized in particle swarm optimization to optimize the corner radii for maximum heat transfer rate. The data required for the training the GEP model have been collected from the authors’ recent published article (Neural Comput Appl, 2015. doi:10.1007/s00521-015-2023-8). It is found that the heat transfer rate of a circular cylinder can be enhanced 12 % by introducing a new cylinder geometry of corner radius r = 0.51.