Accuracy in St number for Flow over Cylinder

[airwarriorg91]

Hello,
I am simulating flow over a cylinder. I tried various set of parameters for simulation. I reduced the error in the St number for my case (Re = 300) to 6.16% from 22% with modifications in mesh-size and domain-size. I am not able to reduce the error further lower than 6.16%. Any suggestions on how accurate we can get the St number with cartesian mesh.

Here are the parameters of the simulations I performed with the St number I got and the cooresponding error.

S. No. | Domain Size | Mesh Size | Time Step | Number of Steps | Simulation St Number | Error
1 | 20x5x3 | 129x65x16 | 0.0075 | 60000 | 0.222 | 20
2 | 20x5x3 | 257x80x16 | 0.0075 | 60000 | 0.226 | 22.16
3 | 12x8x3 | 201x120x16 | 0.0025 | 60000 | 0.213 | 15.14
4 | 12x8x3 | 301x201x16 | 0.0025 | 60000 | 0.213 | 15.14
5 | 16x16x3 | 301x301x8 | 0.0025 | 60000 | 0.2 | 8.11
6 | 20x20x3 | 301x301x8 | 0.0025 | 60000 | 0.199 | 7.57
7 | 30x30x3 | 1041x1040x20 | 0.0001 | 600000 | 0.1964 | 6.16
8 | 30x30x3 | 1281x1280x20 | 0.0005 | 300000 | 0.2 | 8.11
9 | 60x60x6 | 2081x2080x20 | 0.0005 | 150000 | 0.2 | 8.11

In the last three simulations (S.No. 7,8 and 9), I noticed that increasing the time-step increased the error (7 and 8) whereas increasing the mesh size and domain size proportionally without changing the time-step, didn't change the error (8 and 9).