[JamesR]

Hello,

I am trying to conduct a transient simulation of a external flow over a motorcycle with rotating wheels accelerating constantly from rest to 30 m/s over a period of 7 seconds. I am using overset meshing and I have managed to set the ground and inlet boundaries to time dependent velocities using a uniformFixedValue table of timesteps and corresponding vectors. I have tried to move the internal domain with the inlet and lowerWall boundaries with an expression for the internalField:

internalField uniform #eval "(30/7*time())*vector(0,1,0)";

Upon execution of decomposePar, no errors or warnings are given but looking at the individual processor/0 directories it reverts the internalField back to uniform (0 0 0).

I then altered the processor directories to contain the expression and ran the simulation for a few time steps and it is again set back to uniform (0 0 0). Is there a way to give the internalField some sort of time depency with a function or a table/dataFile.

Currently the domain has zero velocity internally and only the cells close to the ground and inlet and moving wheels have the desired velocity at the corresponding time step which is not an accurate representation of a motorcycle moving through fluid.

I have done this easily at steady state as one can just set the internalField to a uniform vector of the velocity.

Thank you for any help!

[olesen]

Using expressions, tables or whatever to change your internalField dependent on time sounds really problematic - I think you need to rethink what you are doing. If you change the internalField during the calculation, it will usually become decoupled from your boundary fields. This is what the correctBoundaryConditions() partly helps with.


If you are doing this type of advanced manipulation (and you know what you are doing), I think that a coded functionObject or some other type of user coding is what you will need.