Question-and-Answer Resource for the Building Energy Modeling Community
 | 1 | initial version |
Let's get back to the basics of energy modeling for a little bit.
According to Wikipedia (and like I always say, "if it's on wikipedia it must be true"TM)
Building energy simulation, also called building energy modeling (or energy modeling in context), is the use of software to predict the energy use of a building.
There's also a short infographic video from energy-models.com here that's pretty good to inform your family and friends about what is it you do with your computer all day long at work, provided they speak any english.
Anyways, one good thing about building energy modeling (BEM) is that it allows you to calculate, to predict, the energy use of a building. You can do this with an abacus, a spreadsheet, or a simulation engine. And trying out different HVAC scenarios is a very common task in BEM.
One good thing about an energy modeling engine such as EnergyPlus is that it's much more detailed - that can also be a bit of a bad thing at first because it looks complicated - than anything you could probably do with a spreadsheet. After all, there are many smart people that have been developing the source code for a pretty long time: development of EnergyPlus itself started circa 1996, but it reused a code base from ealier projects. (History of Building Energy Modeling , BEMBOOK)
While it might be very tedious to handle interactivity correctly and to carry lots of performance curves etc. using a spreadsheet, an engine like EnergyPlus is perfectly capable of handling a VRV, and to handle a rooftop system.
So you would start by modeling your building, with its envelope, loads and internal gains. Then you would assign one of the two HVAC system, simulate this for an entire year of typical weather. Then you would create another model with the other HVAC system, simulate this for an entire year of typical weather.
By looking at the difference in consumption between your two options, you'd be capable of saying which options appears to be better than the other.
Let me use this occasion to mention something that is more an opinion than a fact. You'd also be able to just say stuff like 'VRV will save 23.0312% compared to a rooftop solution", which is genuinely what everyone is expecting from you. I've used 4 digits because it's generally too much and will instinctively trigger something in most people: they'll think about "precision", or more importantly "accuracy". Short of actually defining prior uncertainties and propagating them through your model, which you won't do, you're basically not quantifying uncertainties at all. But you can do two things at the very least:
