[hconel]

I'm solving the laminar flow around a custom-shaped cylinder at low Reynolds numbers (around 100) using pisoFoam and pimpleFoam. I am familiar with the algorithmic differences between these two solvers, but still surprised that they yield slightly different results of shedding frequency and flow field. The main differences in case setup are:

pisoFoam:

• Adjustable deltaT subject to Courant number max: 0.99
• 1 outer iteration

pimpleFoam:

• Fixed deltaT corresponding to roughly average Courant number of 10
• 50 outer iterations, every timestep converges with 10^-5 residual tolerance

Everything else is identical (I'm not going to write all of them down, but they are generally accepted values). Results are:

.............. Strouhal ..... Drag coeff.
piso ....... 0.1746 ......... 1.7123
pimple ... 0.1831 ......... 1.7812

Anybody else delved into the source of this difference? Results are close to each other and to other studies. But, in which result should we be more confident (feels like pimple for me)?