I feel compelled to add further comments because I think there's been some misunderstanding or misperceptions about comparing results between two programs, in this case, specifically EnergyPlus and eQUEST or DOE-2:
1) Several people (some even in private e-mail) have mentioned looking at the BESTTEST or SP-140 test runs that are repeated everytime there's a new EnergyPlus release. Those simulations are extremely simplified test cases, mostly of one-zone models designed to test fundamental heat transfer algorithms - the impact of solar gain, wall heat loss, thermal mass - across various programs. These test cases
do not begin to address the dynamic conditions found in any real building, particularly not a large commercial building. Unfortunately, it's
these interactions between the transient loads and the HVAC system response that will color or dominate the resultant energy numbers.
In my opinion, a far more revealing test suite would be either the CEC Title 24 certification test suite or even the ten DOE-2 sample runs,
where the models are of realistic buildings with typical systems and operating controls (for more details on the Title-24 certification test suite, please see my SimBuild 2006 paper, available for download at http://www.whiteboxtechnologies.com/downloads/SimBuild2006_EPlusDOE2_translatorF.pdf ). My point is that the BESTTEST/SP140 results may show that the basic heat transfer algorithms in the programs are (reasonably) consistent, but that doesn't mean the programs will show similar results when used to model actual buildings, particularly not commercial buildings with intermittent operations and complex HVAC systems and controls.
2) Several people sounded incredulous that the differences could be as much as 50%. I can only say that if you try it, i.e., model the same building (preferably not a shoebox with a window on one side :-)) with two programs, you will quickly appreciate the difficulty. Actually, from my vantage point, 50% does not sound exaggerated at all. I just got back from a simulation workshop in China, where differences of several hundred percent were reported comparing programs such as DOE-2.1E, TRNSYS, and DeST, a Chinese program, even though they were looking only at residential apartment buildings. The dirty secret in our field is that even if you asked two reasonably competent modelers to model the same building with the same program, they're apt to come up with quite different results. Doing it with two different programs just makes that match exponentially more difficult.
3) Several people (including me) have alluded to the difficulty in making the inputs consistent. I cannot emphasize this enough. All programs have hundreds of hidden defaults or assumptions that aren't necessarily transferable. My paper highlights nine of the most prominent ones found in the CEC project, e.g., DOE-2 models drapes as a SHADING-FRACTION, EnergyPlus models them as internal shading surface requiring 10 inputs; DOE-2 allows a distribution loss in the water loop, EnergyPlus didn't; DOE-2 degrades the wind speed in its infiltration calculation, EnergyPlus didn't; DOE-2 allows a thermostat throttling range; EnergyPlus didn't; etc., etc. (some of these limitations in EnergyPlus may have been corrected since 2007). Since I had access to both the DOE-2 and EnergyPlus source code at the time, I believe I did the most thorough job possible in matching inputs, but even so I had to "punt" in numerous places (no drapes, infiltration at raw wind speed, etc.) for the sake of an "apples-to-apples" comparison. Yet, even under these ground rules, I was still seeing up to 40% differences in heating loads in some cases.
4) So which are the correct, or better, results? I really can't say. When you look into the codes, there is room for questioning the solution techniques in both programs. Just by dint of having seen many more DOE-2 results, I tended to regard them as the "conventional wisdom", but then conventional wisdom could always be wrong. Maybe we have been overestimating the heating energies in California all these years. After reading through this litany of problems, some may want to throw up their hands and say either, "it's hopeless" or even, "who cares?". That to me is a wrong and dangerous response, because it ultimately damages the credibility of simulations. Going back to China, many experts there have become disenchanted and distrustful of simulations, and are calling for it NOT to be used for compliance calculations. The same can happen here, as well. So, I think we need to spend more effort to have a better sense of the relative performance of different programs, and what is the "ground truth". Doing a parallel set of calibrated simulations against good monitored data would be a good start.
Joe
Joe Huang White Box Technologies, Inc. 346 Rheem Blvd., Suite 108D Moraga CA 94556 yjhuang@xxxxxxxxxxxxxxxxxxxxxxxx www.whiteboxtechnologies.com (o) (925)388-0265 (c) (510)928-2683 "building energy simulations at your fingertips"
Paul Diglio wrote:
Mizra:
Thank you for the time it took you to send the results of your study to the forum. Some of the top simulation firms in my area use Energy Plus and I was confused when other people claimed a higher discrepancy rate than you modeled.
Paul Diglio
From: Mirza Sajjal <Mirza.Sajjal@xxxxxxxxxxxxxxx>
To: Pasha Korber-Gonzalez <pasha.pkconsulting@xxxxxxxxx>; deepika khowal <deepika.khowal@xxxxxxxxx>
Cc: equest-users@xxxxxxxxxxxxxxxxxxxxx; energyplus_support <EnergyPlus_Support@xxxxxxxxxxxxxxx>
Sent: Thu, April 28, 2011 5:55:09 PM
Subject: Re: [Equest-users] equest and energy plus outputs
This message has been scanned by MailController - portal1.mailcontroller.co.uk |This was a little study I did out of curiosity last year, and the results between eQUEST (v3.63) and EnergyPlus (v4.0) matched quite well (~1.4% difference). All the schedules, zone areas, surface areas were exactly the same (I checked), but I also chose a simple system; PTAC units for the comparison. I believe when I was doing this I had checked the performance curves that were being used in EnergyPlus and they matched the eQUEST curves.
I assume the results will begin to differ when the systems become more complicated, but even then I believe if both models are calibrated to match as much as possible the results shouldn’t vary significantly (assuming all systems are native to the programs and we’re not creating work-arounds i.e. DOAS in eQUEST using dummy zones or the like).
(FYI, the geometry was ported over to EnergyPlus through a tool I wrote using Excel and VBA, but it’s a messy process)
Following are some of the results and graphical outputs:
____________________
Mirza Sajjal
Engineer
Buro Happold
100 Broadway
New York, NY 10005Tel: +1 212 334 2025
Direct: +1 212 616 0380
From: equest-users-bounces@xxxxxxxxxxxxxxxxxxxxx [mailto:equest-users-bounces@xxxxxxxxxxxxxxxxxxxxx] On Behalf Of Pasha Korber-Gonzalez
Sent: Thursday, April 28, 2011 4:46 PM
To: deepika khowal
Cc: equest-users@xxxxxxxxxxxxxxxxxxxxx; energyplus_support
Subject: Re: [Equest-users] equest and energy plus outputs
As I understand it--Energy+ is free-ware in it's raw form. I guess I was wrong to make an assumption that Deepika is using it without an interface. If this is not the case it might explain such a variance in results. When I model in DOE-2 raw form---it is so much harder for me to manage my data and inputs in my head and such----I was not born to be a programmer, and it gives me nightmares from struggling to pass FORTRAN so many years ago...ugh.
Therefore, I haven't looked at E+ myself for a very long time and purchasing or putting out the cost for the user-interface programs is not as desireable as using the eQuest free-ware. or the future CANQuest free-ware (future SI version).
BUT-- I know that E+ is supposed to have some great capabilities in which we are limited with DOE-2 (to an extent.) Has anyone else had the time and desire to compare these program engines more closely recently? Also--why is it so difficult to "match" systems in eQuest and E+....forgive my blase'-ness, but isn't a pkgd system a pkgd system...a pkgd system??? :)
Pasha
On Thu, Apr 28, 2011 at 11:39 AM, deepika khowal <deepika.khowal@xxxxxxxxx> wrote:
So based on my analysis, there is a difference of about 10% in results of equest and E+. I would believe that because there are certain parameters which are very difficult to match.
this seems reasonable to me.
On Thu, Apr 28, 2011 at 9:09 AM, John Aulbach <jra_sac@xxxxxxxxx> wrote:
I am sure EnergyPlus is a fine and well thought out program (consider the sources and fine people involved). However, since I am in production mode (and don't have time to create a model in two different programs-who pays for THAT), I chose to migrate to eQuest after DOE-2.1 A, B, C, D, and E. There are plenty of resources to ask questions of, the program started out in Windows (didn't need "add ons"), and gives me a 3-D rendering on the building immediately.
I dealt with 20 years of "raw" DOE-2, where I didn't know what my building really looked like (until Joe Huang came along with BDL Draw..). So I would need at age (you guess..) to not relearn an entirely new program and stick with what I had learned over the past 25+ years.
Senior Energy Engineer
Partner Energy
1990 E. Grand Avenue, El Segundo, CA 90245
W: 888-826-1216, X254| D: 310-765-7295 | F: 310-817-2745
From: deepika khowal <deepika.khowal@xxxxxxxxx>
To: Paul Diglio <paul.diglio@xxxxxxxxxxxxx>
Cc: equest-users@xxxxxxxxxxxxxxxxxxxxx; energyplus_support <EnergyPlus_Support@xxxxxxxxxxxxxxx>
Sent: Thu, April 28, 2011 8:56:34 AM
Subject: Re: [Equest-users] equest and energy plus outputs
As Joe and others mentioned, its really difficult to create all parameters same in both softwares.I am still working on it .
For ex, which system should I use in equest as equivalent to unitary system in E+?
I understand its difficult to match every input and hence, getting same results is very tricky.
Thanks all for their inputs
On Thu, Apr 28, 2011 at 4:54 AM, Paul Diglio <paul.diglio@xxxxxxxxxxxxx> wrote:
I thought E+ is a free program? I would be interested to hear from anyone who has used Google SketchUp and the Open Studio Plug-in to generate a 3D view in E+.
I would like to hear more about the discrepancies between eQuest and E+ from those who use both programs.
Paul Diglio
From: Pasha Korber-Gonzalez <pasha.pkconsulting@xxxxxxxxx>
To: deepika khowal <deepika.khowal@xxxxxxxxx>
Cc: energyplus_support <EnergyPlus_Support@xxxxxxxxxxxxxxx>; equest-users@xxxxxxxxxxxxxxxxxxxxx
Sent: Thu, April 28, 2011 1:27:44 AM
Subject: Re: [Equest-users] equest and energy plus outputs
Hmmm- an energy plus model showed 50% less energy use (EUI) than eQuest? Did you use the same weather files? (i.e. convert the EPW you used in E+ to .bin and use the same weather file in eQuest?)
If this is truly the case, this is unsettling as a simulator. Wouldn't it be safer for our clients to error on the conservative side and give the eQuest results instead of the E+ results?
Also--if this is the case, then what is the market advantage to spending thousands of dollars on E+ software rather than use the FREE-ware eQuest program??
I'd appreciate any commentary to help me "see the light" of this topic. And if Deepika is willing to share a visual of his energy results output, I'm super curious to see what it is showing...
Good question/good info...thanks,
Pasha
Korber Energy Consultants
On Wed, Apr 27, 2011 at 7:11 PM, deepika khowal <deepika.khowal@xxxxxxxxx> wrote:
Thanks Joe
Even I realized the same thing. the total energy use in Energy plus was almost 50% less than equest.
If this is the case, who would you know that you model is working fine?
On Wed, Apr 27, 2011 at 6:06 PM, Joe Huang <yjhuang@xxxxxxxxxxxxxxxxxxxxxxxx> wrote:
I was involved in a two year project 2005-2007 to convert the Calif. Energy Commission's Title-24 certification suite of building tests from DOE-2.1E to EnergyPlus. There are some areas where it's difficult to get comparable inputs due to differences or limitations in the models. The differences between the two programs varied a lot depending on the building, weather, and HVAC system. For the CEC certification suite of 160 runs, cooling results were more consistent, within 10% in most cases, with EnergyPlus almost always on the high side; for heating, the differences were much greater, sometimes with EnergyPlus being 40-60% lower than DOE-2.1E. I have a 120-page report on this comparison, but haven't bothered to put it on the Web.
Joe
Joe Huang
White Box Technologies, Inc.
346 Rheem Blvd., Suite 108D
Moraga CA 94556
yjhuang@xxxxxxxxxxxxxxxxxxxxxxxx
www.whiteboxtechnologies.com
(o) (925)388-0265
(c) (510)928-2683
"building energy simulations at your fingertips"
deepika khowal wrote:HI All
I am trying to create same model in equest and energy plus to see whether i see same results and just to validate my simulation files.
has anyone done this before?
I would like to know what is the % difference in both software outputs?
Thanks
Deepika------------------------------------------------------------------------
_______________________________________________
Equest-users mailing list
http://lists.onebuilding.org/listinfo.cgi/equest-users-onebuilding.org
To unsubscribe from this mailing list send a blank message to EQUEST-USERS-UNSUBSCRIBE@xxxxxxxxxxxxxxx
_______________________________________________
Equest-users mailing list
http://lists.onebuilding.org/listinfo.cgi/equest-users-onebuilding.org
To unsubscribe from this mailing list send a blank message to EQUEST-USERS-UNSUBSCRIBE@xxxxxxxxxxxxxxx
_______________________________________________
Equest-users mailing list
http://lists.onebuilding.org/listinfo.cgi/equest-users-onebuilding.org
To unsubscribe from this mailing list send a blank message to EQUEST-USERS-UNSUBSCRIBE@xxxxxxxxxxxxxxx
This message has been scanned by MailController - portal1.mailcontroller.co.uk |
This message has been scanned by MailController.
This message has been scanned by MailController.
_______________________________________________ Equest-users mailing list http://lists.onebuilding.org/listinfo.cgi/equest-users-onebuilding.org To unsubscribe from this mailing list send a blank message to EQUEST-USERS-UNSUBSCRIBE@xxxxxxxxxxxxxxx