Transient Simulations for AEC Applications

If the questions “how long” or “how fast” are asked during a design review meeting, then a transient simulation may be required. Transient analysis examines changes in CFD simulation results over time.

Steady-state simulations assume that all inputs have been running for a long time and the fluid and thermal results have settled out completely.  On average, steady-state runs account for roughly 90% of all simulations. 

With a transient analysis, changes in results over a finite time period can be reviewed in detail.  Here are a few AEC examples of transient simulations:

  • Determine the amount of cold air that enters a heated room when an outside door is opened for 10 seconds.
    • Steady-state assumes that the door has been open for many hours.
  • Find out the temperature rise in a conditioned room if the cooling system is shut down for 30 minutes.
    • Steady-state assumes that the system has been shut down for many hours.
  • Tracking smoke or fluid mixing over time in a fluid domain.  For example, building codes may require that exit signage remain visible for a certain amount of time after the onset of a fire.
    • Steady-state cannot process results over a short time period.

Transient Setup

When setting the Solution Mode to Transient in the Control tab of the Solver dialog, the input fields will change.  The additional transient inputs are detailed further in the following links: 

Solve dialog for steady-state (left) and transient simulations (right).

Transient solver parameters can vary widely based the specifics of the application and the frequency that data needs to be output (i.e. sampled), including:

  • Speed of the flow or how fast the inputs are changing.
  • Overall duration of the transient event (seconds, hours or days).
  • How often results need to be visualized.
    • For an event lasting 10 minutes, saving results every second produces 600 sets of data.  By comparison, saving results every 30 seconds produces only 20 sets of data.

TIP:  Only use output data that NEEDS to be visualized for sufficient design guidance.  For large AEC applications, each incremental set of data can be hundreds of megabytes in size.   Saving results that are not really needed slows down the solving process since extra data has to be written to the hard drive, and may even fill up the hard drive in a worst case scenario.

The following table lists two examples of possible transient parameter combinations.

  Case 1 Case 2
Length of fluid domain (ft) 100 1000
Average fluid velocity (ft/s) 0.1 200
Duration of Transient (s) 1000 5
Time Step Size (s) 10 .02
Time Steps to Run 100 50
Save Interval (s) 10 30

Note that these are examples; your specific transient requirements will dictate the proper settings.  A small enough time-step will be required to accurately resolve sudden changes in input conditions.  The time-step selection process is identical to the mesh convergence process detailed here in the Fundamentals section of Meshing module.  Too large of a time-step will impact the accuracy of the results; too small of a time-step results in simulation run times that are longer than needed.  To determine the optimal time-step size, follow the same procedure as mesh convergence:

  1. 1. Run model with reasonable initial time-step size
    • 1 second time step for 1000 steps equals 1000 second event duration
  2. Clone model
  3. Reduce time-step and run model again
    • 0.5 second time step for 2000 steps
  4. Compare results
    • If results do not change, initial time-step is OK
    • If results change, repeats steps 2-4 with a smaller time step (e.g. 0.25 seconds for 4000 steps)

TIP:  Transient simulations with saved results intervals can be animated in the interface to better visualize the changing results.  Setting up animations is described here Online Help: Animation.