Hi everyone,
So, a few months ago, I implemented a dynamic class so I could use the TRT model of Irina Ginzburg (which is a two parameter dynamic, with serializer and unserializer, in case it is of any matter), and another one for advection-diffusion, also a two parameter one, with a SecondBaseDescriptor. [code=“cpp”]
periodicity.toggle
works fine, so I didn't investigated further at the beginning, but then I had to use inflow and outflow, and I saw the following problems:
– I get allocation problems (the code still run however, I see stripes on the upper wall of the output .gif, and a diagonal one on the right wall) for the advection-diffusion scheme, while[code="cpp"]
periodicity.toggle(0,true);
periodicity.toggle(1,false);
give me exact mass conservation
– when using the flow scheme, the[code=“cpp”]
periodicity.toggle(0,true);
periodicity.toggle(1,false);
I get exact mass conservation, when I use the[code="cpp"]
OnLatticeBoundaryCondition2D
only on the top and bottom walls, and periodicity for the right and left walls, the mass conservation goes off by 1*10⁻⁶ in 50000 iterations, and when I put a boundary condition on every walls, with inflow and outflow, I lose 0.32% of my fluid in the same number of iterations.
On both cases, parallelism works fine.
Does anyone have any idea about what is causing this problem, and how to solve it?