[Baliztech]

Hello,

I am simulating a pressure swirl nozzle. I use RSM model. I achieve successful converge with 2D simulations using third order muscl. When I do 3D simulations, I am not able to converge the model neither with second order upwind nor muscl. My mesh has around 10 million polyhedral elements. Min. orthogonal quality is 0.3. Max. aspect ratio is 25. My goal is to have the solution with at least second order upwind, preferably with muscl.

At the moment, I am using first order upwind in all spatial discretization schemes. If I switch to third order in momentum, I achieve convergence. Mainly, reynolds stresses does not converge using higher order. So how do I achieve at least second order accurate result? Reducing time step is not working. On the other hand, in terms of numerical accuracy, having the solution with momentum in third order and other schemes first order is advantageous or it may converge to a worse point than all first order upwind?

Thanks in advance.