Numerical Simulation and Deep Neural Network Revealed Drag Reduction of Microstructured Three-Dimensional Square Cylinders at High Reynolds Numbers

Abstract

Square cylinders are widely used in various fields. For example, they are common structures in fishways. The flow around square cylinders has been a common problem in various fields. However, reducing the flow drag of the square cylinder is a problem that remains unexplored. Many previous studies have reported the drag reduction of 2D square cylinders, which failed to reflect the drag of real structures. Also, some studies focus on the drag force of the inner wall of the square cylinder modified by the microstructure. Achieving drag reduction by microstructuring the surface of the 3D square cylinder is a challenging problem. This study applied a 3D numerical simulation and deep neural network to study the drag reduction performance of the square cylinder under different patch sizes. We studied the drag reduction performance of protrusion and pit-patched square cylinders and tried to find the rule between drag reduction performance and patch configuration. The results show that the square cylinder has better drag reduction performance in some cases. However, its drag reduction performance is greatly affected by the protrusion structure. Also, too large protrusions will increase the drag force of the structure. When the surface protrusion accounts for 10% of the total area of the square cylinder, the drag reduction performance is the best (22.1%). The pit patch structure demonstrated an insignificant drag reduction performance and even increased the drag in most cases. The DNN prediction model demonstrated the robustness of the numerical simulation data.