[Bldg-sim] eQuest - Model verification

Nick Caton ncaton at smithboucher.com
Fri Apr 29 07:41:51 PDT 2011


I know I have less experience calibrating models than some other
respondents, but I'll share my perspectives as well:

 

It is easy to adjust occupancy schedules / lighting loads in eQuest, but
it requires the end-user knows the program.  If the end-user is
inexperienced with eQuest or any other simulation program, then there is
no "easily modifiable" model you can provide.

 

1)      Even if I manage to put in daily occupant and fan schedules
based on actual use, will the model be accurate enough to take me within
1%?

 

With regard to historical modeling, showing any deadband of accuracy to
known costs/consumptions is *technically* possible given enough
information and time...  That said, falling within 1% of any holistic
measure, even on an annual basis, before "refining the unknowns" would
probably happen only by very lucky/happy coincidence.  It's fair to say
that in a relative sense, some buildings operate much more regularly and
predictably than others, and for these cases you or a client might
reasonably expect a tighter deadband.  Asking what is "accurate enough"
is like asking "what's reasonable," and the answer is that it depends on
what info/time you have. There are always "unknowable" variables that
can play a big role in your end-results (example:  what is the actual
leakiness of a 35 year old envelope?).

 

2)      We have been asked to remove plug loads (both for consumption
purposes and cooling loads).  Will this be too big of an assumption to
get any accuracy out of the model in the future?

 

Depends on your building, but for most buildings plug loads are a
significant fraction of consumption/internal loads.  

 

3)      How do I modify the weather input file?

 

I saved the attached email from David Reddy for the next time I need to
tackle "actual weather."  My attempts thus far haven't been terribly
successful, so I have no "go-to" procedures to share just yet.  I have
heard reference from those who appear in the know to macro/VBA-heavy
excel spreadsheets which can automate the workflow to one extent or
another, but I've not yet seen one with my own eyes.  I have yet to hear
of any program that can "easily" incorporate recently recorded weather
data.  

 

4)       Are there any other ideas I'm missing that will help us
validate a model as an absolute measure of actual building energy
consumption?

 

Generally:  the only "absolute measure of actual building consumption"
will be the utility bills.  Something about the 1% number tells me you
or your client might be trying to hammer a nail with a screwdriver...
Energy models are great for looking to the past to make decisions that
acknowledge the unknowns of the future, but if the real goal is to know
the actual energy usage on an hourly basis - why not set up some
sub-meters and track them?  While meters/bills measure energy usage and
energy models can use that information to make predictions, models don't
determine bills - it doesn't work in reverse!

 

PS:  For all the work you might put into calibrating a model with
historical weather and other known data, keep perspective that your
model will only ever be that accurate for behaviour in that year.  The
moment anyone wants  to use this model predictively (for the future),
you probably should swap back to TMY weather data and embrace the
inherent unknowns.

 

~Nick

 

 

NICK CATON, E.I.T.

PROJECT ENGINEER

Smith & Boucher Engineers

25501 west valley parkway

olathe ks 66061

direct 913 344.0036

fax 913 345.0617

www.smithboucher.com 

 

From: bldg-sim-bounces at lists.onebuilding.org
[mailto:bldg-sim-bounces at lists.onebuilding.org] On Behalf Of James V
Dirkes II, PE
Sent: Thursday, April 28, 2011 7:12 PM
To: 'Karen Walkerman'; 'Carol Gardner'
Cc: bldg-sim at lists.onebuilding.org
Subject: Re: [Bldg-sim] eQuest - Model verification

 

Weighing in late on this....

If I get >80% R-squared correlation, I think I'm doing well!  

There are so many unknowns, including the normal situation that the
client doesn't know their own operating schedules, occupancy patterns,
or maintenance / calibration procedures....

 

The Building Performance Team
James V. Dirkes II, P.E., LEED AP, BEMP
1631 Acacia Drive NW
Grand Rapids, MI 49504
616 450 8653

 

From: bldg-sim-bounces at lists.onebuilding.org
[mailto:bldg-sim-bounces at lists.onebuilding.org] On Behalf Of Karen
Walkerman
Sent: Thursday, April 28, 2011 3:31 PM
To: Carol Gardner
Cc: bldg-sim at lists.onebuilding.org
Subject: Re: [Bldg-sim] eQuest - Model verification

 

Also, a model that is calibrated to actual building use and weather can
help to highlight building operation issues.  I have had models which
were off from actual operating energy use that I was able to determine
how the building was operating inefficiently.

 

In one case, I created a model of an existing building.  For this model,
I made an actual count of lighting fixtures and computers.  For most of
the year, there were few other electrical loads.  I based on their
occupancy and use schedules, I was not quite able to calibrate the
electrical energy use.  When I met with the client, I informed him that
my model showed somewhat lower electrical use, equivalent to 1/3 of the
computers remaining on 24-hours per day.  As it turned out, about 1/3 of
the employees DID leave their computers on at all times so that they
could log in remotely.

 

... I guess what I'm saying is that maybe the goal is not to get the
model within 1% of the building energy use, but get the building within
1% of the model.

 

--

Karen

 

On Thu, Apr 28, 2011 at 3:20 PM, Carol Gardner <cmg750 at gmail.com> wrote:

Bill summed it up really well, Katherine. Sometimes a client will think
a calibrated model must match the real building by an unrealistically
high number, like 1%. I really can't imagine an instance when that would
be desirable. It certainly would be costly to achieve. 

 

Usually the plan is to use the calibrated model as a tool to examine
various energy saving measures/opportunities that the client may be
interested in implementing. It is important to remember, in this case
that you are looking at the usage between the calibrated model and the
ECM model, so the 1% deviation is consistent in both and not a worry.

 

Carol

On Thu, Apr 28, 2011 at 12:03 PM, Bishop, Bill <
wbishop at pathfinder-ea.com> wrote:

	Katherine,

	 

	The questions I would have for you (or your client) are:

	*         Is this an existing building, or something in the
design phase?

	*         What is the model going to be used for?

	*         Is there something magical about the 1% number?

	*         1% of what? Annual electrical consumption? Annual gas
consumption? Both? Monthly electric/gas consumption? Monthly/annual
electrical demand?

	*         Why remove the plug loads? Even if everything else is
modeled perfectly, plug loads are typically 10% or more of annual
electrical consumption but can be all over the place.

	*         "Easily adjusted" by whom? Does your client want you
to hand them a model and let them make all subsequent adjustments?

	 

	Even if you are an expert energy modeler, and there is monitored
energy data available for the building down to the individual electrical
panels or broken out by end use (lighting, plug loads, individual HVAC
units etc.) you will be hard-pressed to get to within 1% on a
month-by-month basis. Even 5% month-by-month is difficult. Sure, you can
tweak to your heart's content to get a model to match annual electric
consumption within 1% based on prior year's utility bills but to
maintain a calibrated model to that tolerance after that would require a
lot of work.

	 

	I suggest searching the bldg-sim and equest-users archives for
information about weather files.

	 

	Oh, one more question - how much money does your client have?

	 

	Regards,

	Bill

	 

	William Bishop, PE, BEMP, LEED(r) AP | Pathfinder Engineers &
Architects LLP

	Mechanical Engineer

	 

	Error! Filename not specified.134 South Fitzhugh Street
Rochester, NY 14608
	T: (585) 325-6004 <tel:%28585%29%20325-6004>  Ext. 114
F: (585) 325-6005 <tel:%28585%29%20325-6005> 

	wbishop at pathfinder-ea.com           www.pathfinder-ea.com
<http://www.pathfinder-ea.com/> 

	P   Sustainability - the forest AND the trees. P    

	From: bldg-sim-bounces at lists.onebuilding.org [mailto:
bldg-sim-bounces at lists.onebuilding.org] On Behalf Of Katherine
Louman-Gardiner
	Sent: Thursday, April 28, 2011 1:49 PM
	To: bldg-sim at lists.onebuilding.org
	Subject: [Bldg-sim] eQuest - Model verification

	 

	Hi All,

	Our client would like us to produce an eQuest model that

	a)      Can be easily adjusted to include current and recently
recorded weather data

	b)      Can be easily adjusted to reflect space occupancies and
lighting loads

	c)       Will reflect the actual building energy usage to within
1%

	Do any of you have experience trying to validate models like
this?  

	My concern is that models are approximations, and will not
account for small changes in occupancy, outdoor weather, etc.  

	I guess my questions are these:

	1)      Even if I manage to put in daily occupant and fan
schedules based on actual use, will the model be accurate enough to take
me within 1%?

	2)      We have been asked to remove plug loads (both for
consumption purposes and cooling loads).  Will this be too big of an
assumption to get any accuracy out of the model in the future?

	3)      How do I modify the weather input file?

	4)      Are there any other ideas I'm missing that will help us
validate a model as an absolute measure of actual building energy
consumption?

	Thank you for your help,
	Katherine

	 

	Katherine Louman-Gardiner, EIT

	Mechanical Engineer

	 

	 

	_______________________________________________
	Bldg-sim mailing list
	
http://lists.onebuilding.org/listinfo.cgi/bldg-sim-onebuilding.org
	To unsubscribe from this mailing list send  a blank message to 
BLDG-SIM-UNSUBSCRIBE at ONEBUILDING.ORG




-- 
Carol Gardner PE


_______________________________________________
Bldg-sim mailing list
http://lists.onebuilding.org/listinfo.cgi/bldg-sim-onebuilding.org
To unsubscribe from this mailing list send  a blank message to 
BLDG-SIM-UNSUBSCRIBE at ONEBUILDING.ORG

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.onebuilding.org/pipermail/bldg-sim-onebuilding.org/attachments/20110429/c6b4e742/attachment-0002.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/jpeg
Size: 1459 bytes
Desc: image001.jpg
URL: <http://lists.onebuilding.org/pipermail/bldg-sim-onebuilding.org/attachments/20110429/c6b4e742/attachment-0002.jpeg>
-------------- next part --------------
An embedded message was scrubbed...
From: "David Reddy" <david.j.reddy1 at gmail.com>
Subject: Re: [Bldg-sim] historic weather files for model calibration
Date: Tue, 30 Nov 2010 01:09:48 -0500
Size: 25745
URL: <http://lists.onebuilding.org/pipermail/bldg-sim-onebuilding.org/attachments/20110429/c6b4e742/attachment-0002.eml>


More information about the Bldg-sim mailing list