This is where the duration of the simulation in real time is specified. It is used to determine when to finish the transient run. It can be set to one of:

The Timesteps list is where you set the actual real time intervals at which the CFX-Solver solves for the flow field. The timesteps you select need to be based on the time scale of the transient behavior that you want to resolve in your flow simulation. The timesteps can be:

When not entering the list as an expression, you can also specify a comma separated list of timestep values with multiple timesteps per item in the list. For example, specifying 5*0.1, 2*0.5, 10*1, with the units [s] specified separately, would result in the following CCL:

0.1 [s], 0.1 [s], 0.1 [s], 0.1 [s], 0.1 [s], 0.5 [s], 0.5 [s], 1 [s], 1 [s], 1 [s], 1 [s], 1 [s], 1 [s], 1 [s], 1 [s], 1 [s], 1 [s]

If you accidentally enter 5*0.1 [s], 2*0.5 [s], 10*1 [s] as an expression, the multiplication would be carried out, and the corresponding CCL that would be generated would be:

0.5 [s], 1.0 [s], 10.0 [s]

If a single value or expression is set, this timestep size is used to move forward in real time from the initial time to the end of the transient run. The end of the transient run is determined by the Time Duration: Option setting (see Time Duration).

If a list is set, each timestep size is used in turn to move forward in real time from the initial time to the end of the transient run. The end of the transient run is still determined by the Time Duration: Option setting. The end of the run can be reached before all timesteps in the list have been used, in which case the remaining timesteps are ignored. If all timesteps in the list are used before the end of the run is reached, then the last timestep in the list continues to be used repeatedly until the end of the transient run.

When restarting a transient run, the time completed in previous runs is considered. For example, if a list is specified and the first three timesteps in the list were completed in the initial run, then the run will restart at the fourth timestep in the list.

A simple timestep dependence study is useful to determine the effect of the timestep size on the accuracy of the results. You should initially choose a timestep size of the order of the time scales of interest, and then running a fixed but small period of the simulation in physical time. The same period is then re-run using double and half this timestep size, and the results compared. Significant differences between the initial and smaller timestep mean that an even smaller timestep is required, and this process can be repeated until the similarity in results is within some acceptable solution tolerance. Even if your timestep selection for the run is not the limiting case, this kind of analysis provides a good feel for the level of error introduced by your selection of timestep size. For details, see Max. Iter. Per Step.

This is the same as the Timesteps list described above, except that when restarting a run, the time completed in previous runs is ignored. When CCL is used to set both Timesteps and Timesteps for the Run lists, the Timesteps list takes precedence (see Timesteps: Timesteps List).

Adaptive timestepping automates the process of adapting the timestep size according to one of the following three options:

In addition, each condition requires specification of following parameters:

If adaptive timestepping is used, the final simulation time may slightly exceed the specified time duration.

The Initial Time for a transient simulation corresponds to the time before beginning the first timestep in the current simulation. The following options can be set to control the initial time:

If the Initial Time is read from an initial values file, it is set to the last timestep value read from the results file. Information on specifying this option in CFX-Pre is specific.

At each timestep in a transient simulation, the CFX-Solver performs several coefficient iterations or loops, either to a specified maximum number or to the predefined residual tolerance. These settings are controlled on the Solver Control panel in CFX-Pre. The maximum number of iterations per timestep may not always be reached if the residual target level is achieved first.

Using a large number of coefficient iterations for transient runs is not recommended. Improved accuracy is much more efficiently achieved by reducing the timestep size. If the solution is not converging within each timestep, which is required to maintain a conservative discretization, it may signify that the timestep is too large for good transient accuracy. Decreasing the timestep size typically makes convergence within a timestep easier to achieve, and also improves the transient accuracy of the solution. Actual behavior is, of course, problem dependent. For your particular application, you need to make a compromise between timestep size, number of coefficient iterations and residual tolerance to get the most cost effective solution.

A value of 3 iterations per timestep should be sufficient for most single phase simulations, and values higher than 5 are unlikely to improve accuracy. In multiphase cases, the default value of 10 iterations per timestep may be more appropriate. Initially, you should use these values and then adjust the timesteps to control the balance between accuracy and solution time.

Long transient simulations can obviously end up being very computationally expensive. It is easy to define a problem that could potentially take weeks to compute, even with the coupled solution capability of Ansys CFX. It is very important, therefore, to balance the need for accuracy against run-time cost when performing these types of analyses.

The transient scheme defines the discretization algorithm for the transient term.

For related theory, see Transient Term in the CFX-Solver Theory Guide.