[Bldg-sim] simulation software

Justin Spencer jspencer17 at gmail.com
Tue Dec 18 16:27:43 PST 2007


This is truly an interesting topic and I'll add my $0.02.

I know very little of TRACE.  I'm relatively inexperienced in simulation
compared to some folks in our midst.  I started using EnergyPlus 2.5 years
ago and I've since used DOE2.2 and eQuest a fair amount.  Personally, I like
eQuest, but if anybody wants a job done right, it's going to demand an
understanding of BDL.  Yes, the output from DOE2 is confusing.  But just
because you get simple answers from a tool doesn't mean they are right
answers.  Personally, I'd suggest using eQuest as a front end and learning
DOE2 BDL.  I guess I have a somewhat unique perspective among the younger
crowd of modelers because I started out writing my input files by hand, just
like the older guys did back in the 70s and 80s (or earlier?!)  But it's
critical that you understand what's going on behind the scenes if you really
want to get consistent higher quality results.

The fundamental problem here is one of the users judgment.  Nothing
demonstrates garbage-in, garbage-out like an easy-to-use energy simulation
tool.  Just because you hire a firm with lots of energy simulation
experience doesn't mean you're going to get high quality simulation
results.  You're going to get access to the engineering judgment of people
at that firm, which is worth a lot because they're going to be better at
smelling garbage.  But I don't believe you can really trust any big firm to
tell you with certainty whether a given building design uses 27% or 32% less
energy than a baseline building.  You could certainly trust the results more
if you hired one of these experienced DOE2 experts to do the modeling
themselves, but it's likely that their billing rates are too high and the
lower bid will come from the firm who is using younger, less experienced
modelers and using the more experienced people for oversight and
management.  That's just how it goes.

It's amazing how many projects choose a top-flight architecture firm and
then go after commodity engineering and energy expertise, especially when
the architects tout the energy efficiency of the buildings.  They drop $10
million on architectural features and then choose their energy consultant
based on a $10 thousand difference in bids.  It's embarassing how many
LEED-certified pieces of garbage are out there.

But where can we go with this energy simulation problem?  A big problem with
these projects is that they are more likely to use new technologies that
aren't simple to model and haven't yet seen enough use in simulation engines
to really work all of the bugs out.  I cringe every time I use the GSHP
model in any model.  So what can you do, except get deeper into the code?
Pretty soon you'll be reading FORTRAN, a step I've never taken, but come
quite close to taking.

Personally, I think there are way too many assumptions that may be
questionable in our modeling to be sure about absolute energy usage.  But
this isn't necessarily that important, really.  It seems that what we want
to be sure about is the relative difference in savings between a base case
and a case including a given measure.  Where we really need tool or process
improvement is in garbage sniffing.  I've been using DView with hourly
output for debugging purposes for a while.  It's a tremendous help in
ferreting out problems in your models.  But it seems we could get more
advanced in this sort of capability.  Just like you check the geometry in
eQuest, you should check the building operation to make sure it makes sense
-- do you see the lights coming on as scheduled?  Do you see the HVAC system
cycling on and off and behaving as it should?  Does that ice storage system
load when it's supposed to load and unload when it's supposed to?

While it might be difficult to calibrate an individual model of a new
building, we need calibration methods for making sure that a given
technology is performing as it should.  If this isn't possible, we should be
completely clear and candid with our clients and tell them this.  You're
misleading yourself if you think an uncalibrated hourly model in eQuest is
more accurate in depicting the performance of a novel system than an
uncalibrated Excel model is.  And calibration methods for these things....
These need to be standardized and improved.

In the long run, I think the models with the most resources behind them are
going to be the ones to use.  This probably means EnergyPlus, as soon as it
starts to acquire more users, provided that funding continues.

I'm not sure how the real pros in the industry can distinguish themselves
from more commodity type energy consultants going forward.  Maybe
performance contracting is the answer?  What if you hired a single firm as
designer, energy consultant, commissioner, and ESCO going forward?  That
would certainly provide the right kind of incentive.

Maybe LEED certification should require an independent audit and
recommissioning at the 2 year mark to determine what the real energy usage
should be?

There's just a lot out there, process-wise, that we can work as a community
to improve.  As fast as the industry is growing, there's the potential for
quality to go downhill and genuine simulation experts find themselves being
outbid by groups that do shoddy work.

Back to the task at hand, learn eQuest and learn BDL if you really want to
do high quality work.  Also be sure to practice looking at hourly output and
using your engineering judgment to verify that things are operating as
expected.  My experience is that there are guaranteed to be bugs in your
work -- you just need to develop a keen nose for sniffing them out.



On Dec 18, 2007 4:23 PM, Bernard W. Nelson <bnelson at mcw.com> wrote:

> Hi All,
>
> I agree with Mike Roberts.  I started my career 34 years ago as a sales
> engineer for the Trane Company.  One of my responsibilities was to teach
> consulting engineers how to use TRACE.  This was before the advent of time
> sharing on mainframes when a completed input book was mailed to La Crosse,
> Wisconsin for a two week turnaround.
>
> I became an Energy Consultant in 1979 and used TRACE extensively for many
> years in both new construction and the energy retrofit business. With the
> migration of TRACE and DOE-2 from running on time shared mainframes to
> running on the PC, the simulation world has really changed our business.
> Some runs used to take 24 hours to process on the first PC versions.  I
> incurred $20,000 in run time costs on a large project using time sharing
> on
> a mainframe. Now, if software costs $1,700 to purchase with an annual $300
> maintenance fee we think that it's cost prohibitive.  When a simulation
> program takes three minutes to make a run we start looking at our watches.
>
> In 1981 I learned how to read and write the Building Description Language
> (BDL) and input DOE2.1E the hard way by coding everything by hand
> (sometimes
> in Cartesian coordinates) and running it on a time shared mainframe.  I
> ran
> many DOE2.1E simulations using time sharing and then using Micro-DOE2 on
> an
> IBM 386 PC in the early 90's.  DOE-2 is a great program and provides the
> knowledgeable user with just about any kind of hourly output you would
> like
> to analyze.
>
> In the last number of years I have used both TRACE 700 and eQUEST.  I like
> using both programs depending on what I am trying to model.  If you are
> going to run eQUEST you had better plan on experiencing a steep learning
> curve so you can tweak the BDL as required instead of relying on the
> Wizards
> for your input and output flexibility.  The CAD import feature of eQUEST
> is
> a great tool.
>
> I have received excellent technical support from TRACE CDS over the years
> and very little technical support on DOE-2.  The quality of the technical
> support received depends many times on how you phrase your questions and
> whether or not you really understand what you are trying to accomplish.
>
> For the last 17 years I have specialized in conducting facility
> assessments,
> modeling new and existing building systems, preparing retrofit designs of
> chiller plants, steam and hot water boiler plants, air systems, pumping
> systems, direct digital controls systems and then providing "hands-on"
> commissioning of the building's Direct Digital Control systems to ensure
> the
> projects work.
>
> Building system modeling is going to be "garbage in - garbage out" if you
> expect pre-processing and post-processing software to do all the "heavy
> lifting" for you. To be an effective modeler one has to develop a thorough
> understanding of how HVAC systems work and how to design these systems
> that
> are being modeled.
>
>
> Bernard W. Nelson P.Eng., PE
> Project Manager
>
> MCW Consultants Ltd.
> 1400-1185 West Georgia St.
> Vancouver, BC V6E 4E6
> Phone 604-687-1821 Ext.126
> Fax 604-683-5681
> E-Mail: bnelson at mcw.com
>
>
>
> -----Original Message-----
> From: bldg-sim-bounces at lists.onebuilding.org
> [mailto:bldg-sim-bounces at lists.onebuilding.org] On Behalf Of Mike Roberts
> Sent: Tuesday, December 18, 2007 12:15 PM
> To: 'Karen Walkerman'; bldg-sim at lists.onebuilding.org
> Subject: Re: [Bldg-sim] simulation software
>
> Now this is the beginning of a very useful thread!
>
> Steve Playo pointed out some important things to consider in choosing an
> energy simulation program.  I also agree with Karen Walkerman in her
> defense
> of eQUEST.  I, too, would use it in preference to TRACE in most cases.
>
> What too many people do not realize is energy simulation is a very
> complicated and detailed process.  Far too many assume the eQUEST wizard
> or
> the TRACE entry method will allow them to simulate a building for LEED or
> whatever without going through the agony of learning what the building is
> doing or how the program models it.  It just isn't going to happen.
>
> I have been doing energy simulation for 30 years.  I started with Cal-ERDA
> 1.3 (predecessor to DOE-2.2, the eQUEST calculation engine) on a time
> shared
> main frame computer.  After a couple years of expensive time sharing, I
> helped my employer install DOE-2.1 on a DEC System 10 main frame.  In
> 1986,
> I left that employer to do energy simulation for my own company.  Over the
> years, I have used DOE, TRACE, BLAST, and a number of other simulation
> programs.  I have taken training courses in most of the ones I have used
> and
> assisted in training courses for eQUEST.  The point is that even now, I
> learn something new on every project.  I used to say that running DOE took
> the equivalent of a five hour college course.  Now, I think it is the
> equivalent of a college degree.
>
> Don't worry about how cheap the program is to obtain.  Whatever it costs
> will pale compared to the cost of learning how to use it.  Take the
> training
> for whatever program you are going to use and then expect to spend a lot
> more time on the learning curve.
>
> This is not to discourage using energy simulation.  Energy simulation is a
> wonderful tool.  It is easy to come up with more than enough energy and
> dollar savings to offset the additional engineering hours.  And it will
> only
> become more important in the future.  It's just not as easy as some people
> would like to think.
>
> Mike Roberts
>
>
> -----Original Message-----
> From: bldg-sim-bounces at lists.onebuilding.org
> [mailto:bldg-sim-bounces at lists.onebuilding.org] On Behalf Of Karen
> Walkerman
> Sent: 12/18/2007 11:49 AM
> To: bldg-sim at lists.onebuilding.org
> Subject: Re: [Bldg-sim] simulation software
>
>
> Hi All,
>
> I too have spent many frustrating nights trying to figure out what is
> wrong
> with my eQuest models.  After all the brow-beating, I have always come to
> the conclusion that something was wrong with my inputs, not with the
> software, and been able to remedy the issue.  Not true for Trace700.  I
> spent about a year with Trace, found a number of bugs (some of which the
> company is now fixing), got frustrated with the long (hour-plus)
> simulation
> times, and limitation to modeling 4 alternatives in one file.  Yes it's
> nice
> that there's someone to call when you have questions, but they don't
> always
> know the answer.  I had people telling me to check my fan energy inputs
> when
> the odd data I was getting had nothing to do with fans.
>
> In addition, I often have trouble with data corruption when editing
> templates.  Granted, templates are a great idea and I wish eQuest had a
> similar option, but when your inputs are corrupted and you have to go
> through and check room by room to see what remains...  not a fun process.
>
> Trace700 has no visual feedback for window, wall, room or roof geometry.
> This can be a pain.
>
> I also believe that their water source heat-pump system does not share
> energy in the loop properly.  I created a test-file, consisting of two
> 10X10X10 rooms.  One room had a large cooling load, the other a large
> heating mode.  Neither had any external surface areas.  The total energy
> required to condition the spaces should have been much lower for heat
> pumps
> than for fain coils... but, they were practically identical.  I sent the
> model to Trace, and they argued that this was not a fair test of their
> software????
>
> So... though eQuest can be frustrating at times, I gladly use it over
> Trace.
>
> ~Karen~
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
>
> http://lists.onebuilding.org/pipermail/bldg-sim-onebuilding.org/attachments/
> 20071218/f81778dd/attachment.htm
> _______________________________________________
> Bldg-sim mailing list
> Bldg-sim at lists.onebuilding.org
> http://lists.onebuilding.org/listinfo.cgi/bldg-sim-onebuilding.org
>
> _______________________________________________
> Bldg-sim mailing list
> Bldg-sim at lists.onebuilding.org
> http://lists.onebuilding.org/listinfo.cgi/bldg-sim-onebuilding.org
>
> No virus found in this incoming message.
> Checked by AVG Free Edition.
> Version: 7.5.503 / Virus Database: 269.17.4/1188 - Release Date:
> 12/17/2007
> 2:13 PM
>
>
> No virus found in this outgoing message.
> Checked by AVG Free Edition.
> Version: 7.5.503 / Virus Database: 269.17.4/1188 - Release Date:
> 12/17/2007
> 2:13 PM
>
>
> _______________________________________________
> Bldg-sim mailing list
> Bldg-sim at lists.onebuilding.org
> http://lists.onebuilding.org/listinfo.cgi/bldg-sim-onebuilding.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.onebuilding.org/pipermail/bldg-sim-onebuilding.org/attachments/20071218/56bd6f19/attachment.htm 



More information about the Bldg-sim mailing list