In DOE-2, operable windows are modeled as natural ventilation in SYSTEM using the Sherman-Grimsrud Infiltration Model (when I asked the author Max Sherman whether this was an acceptable use of his model, Max said it was actually better suited for large bulk flows than through small cracks and leaks!). S-G is a single-zone model, so there's no capability to model air flow from zone to zone, as is possible with network models. However, at the single-zone level it uses similar physics with a wind-driven and a temperature driven component. For many years, natural ventilation was allowed only with the RESYS (residential) system, but starting with 2.1E it was also made available to other air systems. The amount of window openings is entered as a FRAC-VENT-AREA, and controlled using either the temperature or enthalpy difference between indoors and out. One thing I've like about the control is that it first tests if natural ventilation can hold the setpoint, and if not, it then models the windows as closed and simulates the A/C.
There are a lot of good answers here for specific tools. Maybe my question was too broad to ask for "any simulation tool". It makes it hard to choose a correct/best answer.
To recast the question a little bit, given that many tools have some way to model operable windows--explicit capability or workaround--what assumptions do you use for the inputs and controls? I have not looked at all of the tools and capabilities mentioned so far, but I'm guessing they all take vaguely similar inputs for setting threshold temperatures (when its open/closed), maybe flow rates or ACH in some cases, etc.
@Peter Ellis I have noticed that may questions ask "How do I model XYZ." I recommend editing these questions to either one of "What feature do I use to model XYZ" and/or "What modeling assumptions should be made when modeling XYZ" Does this make sense? I don't know how to suggest changes to a question yet. Some people (like you) put this into the detailed description, but I think it should go into the question title itself for clarity on the main list of questions.