How should I subdivide an atrium for a MacroFlo simulation?
In our experience, dividing a large space such as an atrium into 4 cells produces behaviour similar to that produced by CFD (which is of course a more rigorous basis for analysing this type of problem). I would advise making the vertical cut in such a way that it divides the room into zones that you expect to be at significantly different temperatures (for example along a plane parallel to a glazed facade).
I would also advise against using too many zones (four is good, eight borderline). If you have a lot of zones you will introduce artificial resistance into the flow field. It is important to understand that MacroFlo does not model stratification explicitly. MacroFlo deals with bulk air temperature and stratification is the result of plumes of warm air rising through the space driven by relatively small-scale spatial temperature variations. To account for the pluming effect, a proportion of convective gains occurring at low level from people, equipment or lights should be assigned to levels higher up.
As a rule of thumb we recommend that the gains should be apportioned in the ratio 1/3 to the level where they occur and 2/3 to higher levels. Here the lower level would normally represent the occupied space of approximately 1.8m height at the base of the space. If gains also occur higher up - eg from walkways within an atrium - the same principle should be applied there, with some of the gain being apportioned to the spaces above. The 2/3 ratio must be added as a watts figure as opposed to W/m2, as holes and windows in a floor reduce the calculated watts. Therefore a floor with a 100% hole (or window) will result in no watts from a W/m2 entry.
