Fluids

Fluids

Time step relevance for particle tracking.

    • Jeroen
      Subscriber

      I have been trying to model the change of mass for a multicomponent particle using a user defined function of the type DEFINE_DPM_HEAT_MASS. However, it has come to my attention that no change in particle mass can be observed unless the particle tracking time step is very small (see figures below, no reduction degree = conversion; this is related to the change in mass of the particle). The temperature profile for the particle does not show this type of behaviour.

      One would expect some sort of converging behaviour related to the timestep, however I cannot explain the huge difference that these time steps of 10^-4 and 10^-5 seconds show. Also, when using the particle surface reactions model this behaviour is not there and a timestep of 10^-2 gives reasonable results. Therefore, it seems to me like a threshold value is present.

      Using a time step of 10^-5 seconds is too expensive in my model when tracking many particle and when considering interaction between the discrete and continuous phase.

      Can someone please explain why this behaviour / threshold value is there?

    • Rob
      Ansys Employee
      I'm not aware of any issues, and do run models with a range of time steps. What exactly does the UDF do? n
    • Jeroen
      Subscriber
      The UDF calculates the change in mass of 3 components. The sum of these component masses is not constant and decreases over time. The UDF calculates the rate of change for each of the components in kg/s and returns this (as dydt in DEFINE_DPM_HEAT_MASS) to fluent.n
    • Rob
      Ansys Employee
      If you divide the UDF result by 10 (ie slow down the material transfer) what happens if you run at 1e-4s? n
Viewing 3 reply threads
  • You must be logged in to reply to this topic.