The Time Step Size is always in seconds. The correct choice of time step depends on the time scale of the analysis. For non-motion flow analyses, the time step size is a fraction of the mean flow velocity, and should be at least a tenth of the time needed to traverse the length of the device. In many cases a much smaller time step size will be required to adequately resolve the flow.
For non-motion heat transfer analyses, the time scale is usually much larger, so a larger time step size can be used. The time step should not exceed one tenth of the expected heat-up time. For solar heating analyses, a much larger time step can be used because the time scale is typically a day or more. A time step for a typical solar heating analysis can be on the order of 100 seconds or more.
If Intelligent Solution Control is enabled, Autodesk Simulation CFD automatically calculates a time step size based on convergence progression and the mesh. This time step size is usually quite small, and often a larger step size can be used effectively ( ).
For Rotating analyses, a time step size ranging from individual blade passages to complete revolutions can be used effectively. Smaller time step sizes are recommended for devices with many blades to resolve the interaction between the blades and surrounding, non-rotating geometry.
To facilitate this, a time step calculator computes the time step size based on either a prescribed number of degrees per time step or the number of blades. Open the dialog by clicking the pop-out button on the Time Step Size line. This is only available when a exists in the model.
Select either the Degrees per Time Step or the Number of Blades, and enter the appropriate value. The time step will be computed based on the rotational speed specified as part of the Rotating Region. If the number of blades is specified, the time step size will be computed using a single time step per blade passage.
The time step size for moving solids analyses is computed based on the specified motion parameters and the mesh size. When the Solve dialog is first opened after assigning Motion parameters, the time step size is computed automatically. If changes are made to the flow or motion velocities, click the button to recalculate the default time step.
Enter a specific time (in seconds) in the Stop Time field to indicate when the solution should stop. This is a very useful way to end certain transient analyses in which Intelligent Solution Control is enabled. An example is the simulation of flow-driven motion because it is not known how many time steps will be required to complete a certain amount of time. If it is not desired to stop the analysis at a certain time, enter “-1” in the Stop Time field, and be sure to specify the number of time steps to run.
Enter the number of steps to run in the Time Steps To Run field. After completing the indicated number of time steps, the solution will stop. This is a recommended way to run transient analyses whose time step size will not likely change. If the number of time steps to run is not important (only reaching the stop time is), then enter “-1” as the number of time steps to run, and be sure to specify a Stop Time.
If both a Stop Time and the number of Time Steps To Run are specified, then the first of the two that is met will cause the analysis to stop. For example: the user wants to run a transient for 3 seconds, but doesn’t want to exceed a total number of time steps of 1000. The user would set the Stop Time as 3, and the Number of Time Steps to 1000. If 1000 time steps are calculated, but only 2.5 seconds have passed, the solution will stop. Alternatively, the solution will stop if 3 seconds is reached in only 450 time steps.
Because Autodesk Simulation CFD uses an implicit method to discretize the transient terms in the governing equations, the calculation has to be iterated at each time step. This transient inner iteration is similar to a global steady state iteration. The governing equations are solved at each inner iteration as they are at each global iteration in a steady state analysis. The difference is that far fewer inner iterations are needed in a transient time-step because the transient equations are much more numerically stable.
Typically, 5-10 inner iterations per time step are sufficient for a transient analysis. If the convergence monitor indicates that this is not enough (the convergence plot does not flatten), this number can be increased. If the convergence monitor shows that this is too many inner iterations (curves are flat for several iterations), you can decrease this number.