[araujo]

Hello all,

in the case of my research, I'm using direct numerical simulation to compute the flow around a cylinder for the laminar case (Reynolds around ~ 100) through high order finite difference scheme (6th order). The cylinder is subjected to one-degree-of-freedom vortex-induced vibration (VIV) in the transverse direction to the flow (y axis).

Right now, I compute the aerodynamic/hydrodynamic coefficients (drag and lift coefficient) through the conservation of momentum equation which uses a control volume (0.75D distanced to each side of the cylinder). Due to the cylinder displacement, there is a lot of noise in the time signal of the coefficients. I'm using a cartesian mesh which isn't conformed to the cylinder.

Is this the most efficient way that minimizes numerical error? For example, a paper by Shiels (2001) uses a vorticity scheme (referred on link). I suppose that integrating along the surface of the cylinder would increase the error.

Thanks in advance for any highlights