The following geometry consists of a barrel with a conical convergence and a capillary positioned at the middle of the length of the cone, on its surface (Figure 1). The cylindrical capillary has a diagonal entrance. Post processing shows that the center line is 15-20% of the wall shear rate when it should be zero. In addition, the ANSYS calculated wall shear (~30,000 s-1) is 15% lower than the analytically calculated shear rate (shear rate= (4*flow rate)/(Pi*r_capillary^3) =34,500 s-1).

The way I have meshed this geometry was using the adaptive sizing function with the element size 0.023 mm, resulting in about 2.26 million elements. After decreasing the element size to 0.0225 mm, the computational power of my desktop was not enough to run the simulation. I was wondering if anyone has suggestions for improving the meshing of the geometry such the resulting center line shear rate is close to or equal to zero, as well as smoother, even contours on limited computational power.