A transient solution costs an order of magnitude more than a steady-state one, so the decision has to be justified. Neither "let us run it transient just in case" nor "steady state is probably fine" is a justification [1] [2].
Steady state is sufficient
- If the boundary conditions do not change with time
- If the flow is largely attached and separation is limited
- If the quantity you want is an average: pressure drop, total heat flux, efficiency
Transient is mandatory
If the physics is inherently periodic. Vortex shedding behind a cylinder, blade passing frequency, sloshing. In these problems a steady-state solver either fails to converge or gets even the mean wrong.
If a boundary condition varies in time. Start-up, shutdown, duty cycle.
If the quantity you want is time-dependent. Oscillating force amplitude, fatigue loading, temperature fluctuation, acoustic source.
If thermal inertia matters. How long a device takes to warm up, or its cool-down curve, is by definition a transient question.
A middle road: pseudo-time
For problems that are genuinely steady but will not converge in steady state, advancing in pseudo-time works. The solution gives no physically meaningful time history, but it reaches the steady answer stably.
Choosing the time step
Picking a time step that looks "reasonable" is common and wrong. There are two criteria.
Courant number. The fluid should not cross more than one cell in one time step. Implicit solvers relax this, but in the region where separation and vortex transport are resolved it should stay near one.
The frequency you want to resolve. There should be at least twenty time steps in one period of the highest-frequency event you intend to capture. Sample blade passing with five steps and you lose the amplitude.
Time-step independence is as necessary as grid independence: halve the step, and if the monitored quantity changes it was not yet small enough.
References
- Ferziger, J. H., Perić, M. & Street, R. L. (2020). Computational Methods for Fluid Dynamics, 4th ed. Springer. DOI: 10.1007/978-3-319-99693-6
- Celik, I. B., Ghia, U., Roache, P. J., Freitas, C. J., Coleman, H. & Raad, P. E. (2008). Procedure for estimation and reporting of uncertainty due to discretization in CFD applications. Journal of Fluids Engineering, 130(7), 078001. DOI: 10.1115/1.2960953