K check interval

perforate.k_update_interval Unit: iterations Step 05 · Boundary Conditions

For a perforated tube with the grazing correction: how many iterations of the final solver phase lie between two checks of $K$ . At each check $K$ is re-evaluated from the sampled wall velocities; the solver is restarted with the new $K$ only when it moved by more than 2 %.

Last updated 2026-11-03

Theory

$K$ is evaluated first at the end of the upwind start-up phase and written into the restart the scheme switch makes anyway. While further checks remain, the final phase writes its fields and the wall samples at every multiple of this interval; each new sample is evaluated while the solver keeps running. A relaxed $K$ that moves by more than 2 % on some wall stops the solver at once (writeNow), the new $K$ goes into the saved pressure field and the solver restarts from there; the first check within 2 % is recorded and ends the checks, and the normal write interval returns.

With several perforated tubes one schedule is used for all: the shortest interval among them.

Equations

i_k = \left(\left\lfloor\frac{i_0}{N}\right\rfloor + k\right) N, \qquad k = 1,\ \dots,\ n_{max} - 1, \qquad i_k < i_{end}

(1)

\text{restart if } \max_{walls}\left\lvert\frac{K_{new}}{K_{old}} - 1\right\rvert > 0.02

(2)

where	meaning	unit
$i_k$	Iteration of the $k$ -th check	-
$i_0$	Last iteration of the upwind start-up phase (the first evaluation)	-
$N$	Check interval (this setting)	iterations
$n_{max}$	Maximum number of evaluations (Maximum K evaluations)	-
$i_{end}$	Last iteration of the run	-
$K_{new}$	Relaxed new loss coefficient	-

Default and rule

Default: 100Measured
Rule: 100 iterations.
Source: The 2 % threshold and the check scheme are based on a validation run of a two-baffle test muffler. Every restart disturbs SIMPLE for about 40 iterations ( $\Delta p_0$ +17 %, pressure residual about 0.4) whether $K$ changed or not, so the solver is restarted only for a real change. In a verification run on the same muffler the check at iteration 300 moved $K$ by 0.09 % and caused no restart. The interval of 100 iterations leaves the field time to respond to the previous $K$ ; it is the program's default and has not been varied in a study.

Valid range 10-5000

Values

10-5000 iterations.

Too short

A check soon after a restart sees a field that has not yet responded to the last $K$ ; keep the interval well above the 40 or so iterations a restart disturbs the solution.

Steady only

Transient runs make no checks.

Theory

Equations

Default and rule

Valid range 10-5000

Related 4