Dear all,
I've recently played around quite a bit with E-quest to simulate the effect of dynamic fenestration on building energy consumption.
For these simulations I've been using a very simple office model with a skylight on the top as a test-bed for my calculations (with default system for HVAC and daylighting controls with dimming for lights) but, observing the values I'm obtaining in output, I see that the values I obtain for lighting, cooling and heating consumption make sense relatively, i.e. if compared to themselves in different conditions, but are sometimes of totally different order of magnitude if compared to each other (lighting/cooling loads are often much higher than heating, by orders of magnitude), especially when scaling up the model to bigger sizes. This way, it's very difficult to see the effect of a variable change on the total energy consumption.
I know the model I'm using is very idealized, but is there any variable that I can act on in order to obtain a model which has energy consumption values more similar to a real building? For example, is there a "suggested" size of the building that gives better results? Or maybe simulating a single room in a large building gives more realistic results than the single room alone?
My point is that I would like to have a model whose results in scale might be consistent when applied to bigger buildings.
Thanks for your hints and suggestions,
Stefano
--
Stefano Moret
The short answer is that every building is different, and there is no
"typical" model that will fit them all.
As you mentioned, the ratio between lighting/cooling/heating changes
depending on the size of the building. I'm willing to guess that this is
because the core zones that have very little need for heat get much larger
in relation to the perimeter zones. Also, the building energy needs will
change depending on orientation and global location.
I would suggest that you take a slightly different approach - try to come
up with values that are representative for different types of spaces. You
might do say:
1. A perimeter office space with XX% glazing (run for North, East, South
and West exposures)
2. A core office space with no skylights
3. A core office space with XX% of roof area as skylights
4. A core manufacturing space???
5,6,7....
Run each model in the applicable climate. Then, if you have an office
building that is 70% core and 30% perimeter space, you'll have a better
understanding of the building. This is still a very rough approximation,
but should get you farther than trying to model one "typical" building.
--
Karen
Dear Karen,
Thanks for your quick reply.
I've done plenty of simulation runs, which basically covered different types of buildings, of different sizes, in different locations.
Of course different parameters grow with different laws, and if lighting wattage is strictly related to the building area, heating/cooling loads are influenced by the lateral surfaces and perimeter zones. This makes perfectly sense, and it was somehow what I expected to see.
Said that, let me explain a little better my question: when I write about different results, I can probably show the point with an example. I take a 5x5x3m building with one skylight on the top (4% SRR) and I get certain values for lighting and HVAC loads. Then I simulate the same exact building with 50x50x6m dimensions, with 4% SRR for skylight, and I get lighting and cooling increasing by a factor of 100, heating by a factor of 10. I also have real data for this larger building, and I simulated it in Energyplus as well, and both these comparisons confirmed that heating and cooling should grow in a similar way in a temperate climate.
So my question is, since I want to keep the model simple, especially on the HVAC side, which are the most important variables in Equest which might lead me to more consistent results without bringing excessive complexity into the model. Of course I'm not looking for a unique building which could be valid for all cases, but I'm caring about model consistency and validation.
Thanks,
Stefano
...I see my response is arriving too late to the conversation but here are some further thoughts that might help Stefano and others:
I have some parallel experience from which I can build/add to Karen's suggestions.
I have performed studies with similar structure as Karen is suggesting (creating and adding "building blocks" to evaluate building massing options). However, for Stefano's purposes I would caution to carefully take stock of what questions you are trying to answer and whether an energy model is really the way to go about finding those answers.
Using an energy model to develop generic guidelines with "arbitrary" buildings/spaces for something inherently site/climate/building-geometry-specific (such as "ideal window wall ratio" or "ideal window properties") is actually a terribly complex endeavor, as I'm sure you are appreciating. You can ultimately come up with a set of guidelines/spreadsheets/principles to use, but applying such developed guidelines to an actual project for decision-making will always be fraught with accuracy disclaimers (perhaps that is not the ultimate goal - I'm only throwing this out for consideration).
Guidelines built around generic conditions are indeed useful and have their place (ASHRAE's advanced energy design guides are worth reviewing if you're at the outset of trying to find such results), but should only be relied upon in the context of "I don't want to invest in a site-specific study and the increased accuracy / differing suggestions that may produce." Evaluating how well any study applies to the project-at-hand is something to always keep in mind.
To cut myself off: I think my response drafted so far is falling away from Stefano's original question as he's more recently clarifying it, but I'm ultimately driving towards furthering Karen's first line: Build project-specific models where you need a project-specific answer.
To Stefano's issue of scaling - I am nodding my head vigorously as you're observing how spaces modeled in isolation do not always additively represent a group of spaces modeled together. It's possible for that to happen under certain circumstances, but the variables in play do not normally scale linearly between an isolated space and a whole building (consider the ratio of envelope surface area to conditioned volume, for one instance). A big part of "big picture" modeling differences when compared to "the smaller parts" is the tempering of various loads across spaces and the way that has energy saving effects on HVAC systems in particular .
I would not spend too much time deliberating on such (expected) differences, and would emphasize focusing on the results of isolated or "whole building" models as appropriate to the purpose of your study =).
~Nick
[cid:489575314 at 22072009-0ABB]
NICK CATON, P.E.
Dear Nick,
Thanks for your suggestions, I totally agree with your points, and I'm conscious of the issues about the "generic guidelines". I'm still surprised by some unexpected results in my models though: I'll work on my models to try to understand the reasons leading to results which are different from the measured ones.
Thanks, Stefano
I had all to similar issues with eQuest, and and realized that eQuest is
not really a reliable software. *I know the list is going to hate me.. The
thing is that eQuest is deceptively easy, but when you want to find
reliable results you basically have to find some "workaround"...
The deal breaker for me in using eQuest for anything more then exploring
early design options was:
I had several real buildings complete with Energy Star portfolio manager
data, as well as building audits, when I attempted to use eQuest discovered
that eQuest was not capable of coming close to reality, in fact it wanted
me to have something like zero insulation, and very poor U values to even
come close. Given Energy Plus is not going to give exact results either but
at least it comes close.
Watch the lists, eQuest is full of "how do I come up with a work around",
TRNSYS is mostly "how do I customize" and Energy Plus is mostly "how does
this feature work"
I do not mean to insult eQuest users but my prospective is that it makes
more since to use software that is harder to use, has better support and is
updated regularly for such professional and high paid work, rather then the
"workaround"
Why eQuest is so popular is because it is easy, but for something like your
project I suggest you use Energy Plus.
*Jeremiah D. Crossett*
Thanks all for your considerations, I much appreciate your recommendations.
The nice thing about E-quest is that it support some elements, e.g. skylights, with many different features, while in Energyplus they are usually treated as normal windows.
Thanks again, Cheers
Stefano
--
Stefano Moret
Thanks I will keep that in mind..
I kinda enjoy the eQuest interface, so it is good to know there are
modeling tasks that it shines in! It has a nice parametric skylight tool
and good libraries, so will now be my tool of choice when skylights are
involved..
*Jeremiah D. Crossett*
Good points Jeremiah. Yes - eQUEST is frustrating and it's a pain in the backside sometimes to try to make it do what you want it to. Maybe Energyplus is a better program to use for this.
However - I don't think workarounds are necessarily a bad thing. In my experience - if you are experimenting with workarounds, it enhances your understanding of the process you are trying to model, and as a result makes you more thoughtful and allows you apply critical thinking to the process.
Ultimately - most modeling is about that last item - critical thinking.
Vikram Sami, LEED AP BD+C
Canon vs. Nikon.
When you choose a photographer, do you ask to see their camera equipment? No, you look at their portfolio. There is no camera that reliably takes good pictures - you need a photographer that knows what they're doing.
The best energy modeling program is the one you are most familiar with.
~Bill
P.S. I can capture great shots with my Pentax K1000 35mm or my Nikon D40x digital SLR, but the Pentax collects dust because the Nikon is fast, inexpensive to use, and gives me immediate feedback. Sure, there are some qualities of film that are hard to reproduce with digital, but 9 times out of 10 you just work within the limitations.
[cid:image001.png at 01CD16F8.E6E525D0]
Classification: UNCLASSIFIED
Caveats: NONE
Shrews must eat 80-90 % of their own body weight in food daily. An elephant eats about 5% of their own body weight in food daily.
The surface area to mass ratios are vastly different. Heat loss, metabolism (internal loads) and fur are very different.
If an elephant had the fur of a mouse it would die from overheating.
You can't model a small building and try to extrapolate the results to a large building.
John Eurek
Hi Stefano,
John's right on point (especially the elephant part). The results likely
represent some sort of curve, and are probably multivariate. My suggestion
would be to model large, medium and small scenarios and do hourly reports of
the loads to see what is happening at different outside temperatures.
Shaun Martin