Question-and-Answer Resource for the Building Energy Modeling Community
Get started with the Help page
Ask Your Question

Has anyone validated the DOE Reference Building Measure?

asked 2019-03-07 07:47:19 -0500

Vaclav's avatar

I used the OpenStudio measure "Create DOE Prototype Building" for a medium and large office buildings, but found that the site EUIs vary widely from the EUI summary table on DOE's website. Specifically, I ran both buildings for the 90.1-2013 standard and 4A climate zone without any further modifications and got a 35% lower EUI for the medium office building and 240% higher EUI for the large office building, compared to the DOE summary table EUIs.

Has anyone else come across this issue? Are there any best-practice suggestions for using this measure to avoid such an issue? Has this measure been validated or is it still under active development?

edit retag flag offensive close merge delete

3 Answers

Sort by » oldest newest most voted

answered 2019-03-07 11:10:32 -0500

See the answer to this question about the DOE Prototype buildings.

The link you reference is for ASHRAE 90.1-2004, Post-1980, and Pre-1980, not ASHRAE 90.1-2013. Also, PNNL updated the models from the version you referenced: PNNL, Enhancements to ASHRAE Standard 90.1 Prototype Building Models, April 2014. Most importantly, the updated included adding a data center to the Large Office model which explains the much higher EUI. The lower medium office building EUI is explained by the new code version.

Both the PNNL reference buildings and theCreate DOE Prototype Building measure are under active development (e.g. recently switching to ASHRAE 169-2013 instead of ASHRAE 169-2006 for design days), but should be very close.

edit flag offensive delete link more

answered 2019-05-11 19:20:43 -0500

Using the "create DOE Prototype Building" measure for the existing post-1980 large office in San Francisco I get an EUI of 55 kbtu/ft2, about 17% higher than the summary table for the original reference model in this location. (Also, just running the file in E+ v 9.1, after updating from v7.2, I get a 10% lower EUI than in DOE's summary table).

So it does not seem that this measure is currently working too well.

edit flag offensive delete link more


Are you using the newer model with the detailed space plan? Or the older model that has only zone level loads?

__AmirRoth__'s avatar __AmirRoth__  ( 2019-05-12 12:49:01 -0500 )edit

HI Amir, do you mean the 90.1 models?

I just checked, and the measure does match well for those.

I was referring to Version 1.4_7.2 updated November 13, 2012

Hayes Zirnhelt's avatar Hayes Zirnhelt  ( 2019-05-12 13:09:29 -0500 )edit

The newer large office model contains a data center, which mostly explains the higher EUI.

mdahlhausen's avatar mdahlhausen  ( 2019-05-13 15:58:45 -0500 )edit

Ok, that's sort of confusing. So when you run the 'create DOE prototype building' for post-1980 it adds a data center similar to the newer PNNL 90.1 reference models, even though there isnt one in the 'post-1980' DOE reference model.

Hayes Zirnhelt's avatar Hayes Zirnhelt  ( 2019-05-13 16:07:21 -0500 )edit

I also noticed this measure is creating natural gas powered elevators, which is interesting!

Hayes Zirnhelt's avatar Hayes Zirnhelt  ( 2019-05-16 16:59:49 -0500 )edit

answered 2019-06-27 15:07:05 -0500

mpigman's avatar

I’m also comparing the results of the PNNL models (90.1 2013 and pre-1980) with the OpenStudio "Create DOE Prototype Building” measure. I have the most recent measure from the BCL, and they’re all CZ 3C models run with the same SFO weather file, but there are still some pretty big differences:

  • Pre-1980 large office: The energy consumption of the OS model is 16% higher than the PNNL model. This is driven primarily by the cooling and fans, which help make the OS HVAC twice as large as the PNNL HVAC. The interior equipment is virtually identical between the 2 models, so it’s not because of a difference in data center.
  • 2013 secondary school: The energy consumption of the OS model is 16% higher than the PNNL model. This is driven primarily by the water heating being 4 times as large in the OS model. This is moderated a bit by the cooling being almost 3 times as large in the PNNL model.
  • 2013 warehouse: The energy consumption of the OS model is 16% higher than the PNNL model. This is driven primarily by the heating and interior lights being about three times as much in the OS model.
  • Strip mall: Overall differences aren’t very big (0 and 6%), but the heating is substantially more in the OS models than the PNNL ones.

Ideas about what might be going on?

edit flag offensive delete link more


cf. mdahlhausen's answer: "Both the PNNL reference buildings and theCreate DOE Prototype Buildingmeasure are under active development"

Jeremy's avatar Jeremy  ( 2019-06-27 15:18:02 -0500 )edit

OK. If I'm going to choose one set of results to report, which one should it be? I'm guessing I should use the PNNL models because they're more established, but I'd love some advice.

mpigman's avatar mpigman  ( 2019-06-27 15:33:24 -0500 )edit

Use the measure; it represents the most recent version of development. And if you are trying to do an analysis with reference buildings instead of a representative sample of buildings or actual buildings, you aren't going to be that accurate in representing the building stock anyways.

mdahlhausen's avatar mdahlhausen  ( 2019-06-28 10:37:41 -0500 )edit

@mdahlhausen - How do I access the measure in the OpenStudio Application? It used to show up, and now I cant access the BCL Library to add measures

helpneeder's avatar helpneeder  ( 2022-02-11 15:49:12 -0500 )edit

@helpneeder you can download the measure on bcl here:

mdahlhausen's avatar mdahlhausen  ( 2022-02-17 22:14:32 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Training Workshops


Question Tools



Asked: 2019-03-07 07:47:19 -0500

Seen: 1,132 times

Last updated: Jun 27 '19