Unexpected Custom SHGC Results

16 posts / 0 new
Last post

Hello,

I have an "L" shaped building with the point of the "L" facing North. The
inside of the "L" has both NE and NW surfaces, that include a high amount of
glass, which heats up the perimeter building spaces considerably during the
summer. Glazing is single pane tinted.

The customer wants to install some Verisol SilverScreen shades in these
windows. According to the manufacturer, the SHGC will reduce by about 33%. I
modeled in eQUEST, window properties in these windows to have an SHGC of 0.67
and ran an EEM reducing SHGC to 0.34, and got an increase in cooling load and
fan load year round, even in the summer months.

Am I seeing this wrong? I can't figure out how I could possible get results
like this?

Thank you,
Pete

Peter Baumstark's picture
Offline
Joined: 2012-07-14
Reputation: 0

Hi Pete,
I don't have a ton of experience with detailed modelling of shades in eQUEST, specifically, but two things that could be at play:
- The claim from the manufacturer sounds like it's not intended to universal in absolute terms. Also, it would be more conservative to reduce the SHGC by 33 percent than 33 percentage points (as you did) if you're going to take this simplified approach. Therefore, it would be closer to SHGC=0.44. Subtle but significant. You could try using software that specializes in window/shade performance like LBNL Window or Parasol to try to characterize the performance of your specific shade-glazing combination- Depending on the operating conditions and construction of the building, there's a chance your results aren't ridiculous. If shades intercept transmitted solar radiation, then a lot of that energy will almost immediately transfer to the air via convection. If you have thermally massive interior surfaces, there's a chance your building could actually perform better without those shades because the air conditioning won't kick in till later.
Liam
Date: Sat, 14 Jul 2012 11:51:02 -0700
From: pbaumstark at sbcglobal.net
To: equest-users at lists.onebuilding.org
Subject: [Equest-users] Unexpected Custom SHGC Results

Hello,

I have an "L" shaped building with the point of the "L" facing North. The inside of the "L" has both NE and NW surfaces, that include a high amount of glass, which heats up the perimeter building spaces considerably during the summer. Glazing is single pane tinted.

The customer wants to install some Verisol SilverScreen shades in these windows. According to the manufacturer, the SHGC will reduce by about 33%. I modeled in eQUEST, window properties in these windows to have an SHGC of 0.67 and ran an EEM reducing SHGC to 0.34, and got an increase in cooling load and fan load year round, even in the summer months.

Am I seeing this wrong? I can't figure out how I could possible get results like this?

Thank you,
Pete

Liam O'Brien's picture
Joined: 2011-10-02
Reputation: 0

Just admit it- eQuest is flawed, you don't have to make up things to
protect it-

If it is a mistake to use the percent points rather then percent reduced
from the abrataty eQuest assumption from 1999 window specs than the it
should have reduced solar heat gain by more then his product even provided-
Using the 33 percentage points but used the 33% should have provided him
over stated cooling reduction, (and extra added heating consumption
tradeoff)

O- and FYI LBL window does glass U-value not shading, ware-as LBL optics
can be used for film coefficients and used to create a custom glass type in
window, but do not do any calculations for "shading"

I say you try the same model in Energy Plus or TRNSYS and see if the
results differ.

*Jeremiah D. Crossett*

CleanTech Analytics's picture
Joined: 2012-02-09
Reputation: -1

I can think of a number of reasons why you might see an increase in
cooling and fan "load". Without spending some time looking at the
input/output, it is pretty hard for me to offer any definite
opinion. I know for myself, it is generally an input assumption I
hadn't considered, not a flaw on the modelling program.

For doing analysis on an existing building, you want to make sure
your inputs are not defaulting and that the model somewhat matches
the building.

>> Christopher Jones, P.Eng.

Chris Jones's picture
Offline
Joined: 2011-09-30
Reputation: 0

I've found this comment to be mildly offensive as well as showing a lack
of understanding about how heat gains ultimately get translated to
cooling energy consumption. It's extremely hard for me to believe that
eQUEST or DOE-2 has been flawed for 25 years in modeling something as
fundamental as solar heat gain through windows. In all the decades I've
used DOE-2 to analyze window performance for DOE's EnergyStar Program as
well as numerous other projects, whenever the results did not match or
ran counter to first-principle expectations, it was always because there
was some other factor that have been overlooked or ignored, chief among
them being the size of the HVAC system, its configuration, and control
strategy. Locations with mild cooling loads, such as San Jose, are
particularly sensitive to such system interactions. Were both runs done
using "autosizing"? What
kind of a system was modeled - VAV or CAV ? Did the model have an
economizer? What were the HEAT-CONTROL and COOL-CONTROL strategies ?
etc. It's far too early to lay blame on the DOE-2 algorithms.

Joe

Joe Huang's picture
Offline
Joined: 2011-09-30
Reputation: 406

Good points Joe- Counter-intuitive but your opinion carries the weight of
many years of experience I do not have, and did not think of. My opinion is
that there are issues with eQuest that have been left un-checked, but
no-one knows what they are- so my bad for assuming that this is one of them.

Best regards

*Jeremiah D. Crossett*

CleanTech Analytics's picture
Joined: 2012-02-09
Reputation: -1

Joe:

I agree. I find the comment to be more than mildly offensive.

I have the same experience that funky modeling results usually are the result of
flawed inputs or depending on too many eQuest defaults.

Paul Diglio, CEM, CBCP

Paul Diglio's picture
Offline
Joined: 2011-09-30
Reputation: 400

It's an interesting problem. With this model, no matter what I do, if I
manually change the SHGC or SC (if I use the U-Value method), or if I select
different glazing types from the library (with different SHGC values), no matter
what face of the building I change it, I get an increase in overall monthly
energy use with a lower SHGC.

I built the model through a zone by activity area method that pretty closely
matches the various zones of the building. It's a VAV with terminal reheat
system (1995 Trane Intellipaks), and I entered actual economizer, static
pressure and SAT settings. Airflow matches as-built design drawings.

I tried using other eQUEST models I've built for other customers using similar
methods, changed the location to San Jose, and ran window cases and results were
as expected.

I'm coming to believe that one issue with the building in question is the RTUs
seem over sized relative to the use patterns and internal heat gains. This
building previously had various lab areas, then was purchased by another
customer with lower internal heat load rates, but they kept the same RTUs.

Could it be possible that the lower heat gains from better fenestration products
could place the RTUs at a more inefficient spot on its performance curve? I've
ran into similar issues with chilled water systems, but never looked at DOE-2
performance curves for DX units.

Pete

Peter Baumstark's picture
Offline
Joined: 2012-07-14
Reputation: 0

So if the curve theory is correct I wonder if this would happen in other
modeling software or more importantly in the real world>? --reduce
internal gain and increase cooling consumption-- or if this phenomena is
only possible in the virtual world of eQuest..,,

Joe-Liam-Paul what do you think >?<

*Jeremiah D. Crossett*

CleanTech Analytics's picture
Joined: 2012-02-09
Reputation: -1

You might want to verify with Compare-it! or another file difference software that no other changes were inadvertently made.

Assuming none, also verify that the changes are actually from the loads and not something in the system configuration by comparing the zone component load reports for the two cases. This would help you identify an AHU controllability issue separate from the magnitude of the zone loads.

One thing you mention in the original message is that the spaces with the large windows "get heated up" - did the model accurately reflect the actual and proposed IEQ for the spaces?

DSE Mobile

David Eldridge's picture
Offline
Joined: 2012-05-08
Reputation: 1

An HTML attachment was scrubbed...
URL:

Bruce Easterbrook's picture
Joined: 2011-09-30
Reputation: 0

Pete,

The performance curves are one item to check out if the RTUs are greatly oversized.

A couple other things come to mind:

* If they originally had lab areas with high make up air requirements, the RTUs could be greatly oversized and operating at a low enough load with the reduced SHGC to bring on hot gas by-pass (within the simulation).

* Check the minimum % cfm of the VAV boxes. If it's a high % like 60%-70% which I've seen in the past, you could be bringing on reheat, which will increase cooling load in addition to reheat energy.

Mike Busman

Busman, Michael R's picture
Joined: 2012-07-12
Reputation: 0

I totally agree. The suggestion that eQUEST is flawed in modeling such a basic heat transfer process as solar gain through windows is extremely unlikely. My experiences in doing such parametrics with DOE-2 over many years if not decades is not that the program is wrong, but that it's considering many other factors often overlooked in a simplistic view of things. Chief among these are the sizing of the HVAC system, its configuration, and operations. At the bare minimum, you should make sure that you input the same system size in both runs. Furthermore, in a mild climate like San Jose, something as innocuous sounding as COOL-CONTROL=WARMEST can cause havoc in the expected savings. In all the runs I've looked at, there was always an explanation when the results do not meet or run counter to the original expectations. In fact, I regard those as the most valuable instances for using building energy simulations. After all, if all you want is confirmation of your back-of-the-envelope estimates, why do a simulation at all? Just do a degree-day calculation instead.

Joe Huang

Joe Huang's picture
Offline
Joined: 2011-09-30
Reputation: 406

I firmly believe that interfaces that do the work for you end up
crippling your understanding of the underlying program if you don't
pay attention.

My first modelling assignment:
I was given a copy of the DOE2.1e Basics, a clipboard, rolling ruler
and tape measure. I was sent to an existing building and was
instructed to find as much design documentation as was available and
measure the rest. I found some floor plans and elevations and the
monthly utility bills.

From that I was to construct the DOE2 model using a text editor and
the command line to run the model.

I constructed the model using full X,Y,Z and calibrated to the utility bills.

With that approach, I was forced to RTFM, something that seems to
escape most novices.

Unfortunately, that has left me with a good understanding of the
underlying engine but I am terribly slow at constructing models via
the interface. I am much more comfortable using DOE22 directly. I
just update the USRLIB.dat file to contain the eQ_lib.dat
information. The results are the same but the models run faster.

Chris Jones, P.Eng.

Chris Jones's picture
Offline
Joined: 2011-09-30
Reputation: 0

I apologize for not getting to this sooner but I wanted by make sure
my analysis was "real".

The PVAV systems are set with COOL-CONTROL = CONSTANT. That is
driving up both the cooling energy and the reheat energy.

The zone air flows are self-sized and the minimum flow is too low for
heating. When you change the glazing in some of the windows, the
zone air flows change, skewing the results. I am pretty sure that a
"real world" project would have fixed air flows to the zones. In
order to gain some understanding of how the change in glazing affects
the building you would want to program the zone air flows to be the
same between the three runs - Base Case, EEM 1, EEM 2.

There is no central heating coil. As a result, air is delivered to
the zones at much lower temperatures than required to meet the
heating loads. Hence the thousands of hours underheated.

First I fixed the zone air flow rates to the values reported in the
Base Case run. The results are more "real":
Cooling energy decreases in each EEM run, heating energy increases as
expected - less heat in the windows, more heating required.
The fan energy also increases as expected, more run hours at higher
zone air flows as this is a VAV system after all.

Interestingly though, the total energy remains exactly the same
between all three runs.

Next I took the bold step of setting COOL-CONTROL = WARMEST,
MAX-COOL-REHEAT-T = 70.
The base case total energy consumption dropped by 8.5%.

The two EEM runs again save cooling energy, the fan energy and
heating energy increase. The result is that EEM 1 saves 0.17% total
energy. EEM 2 saves 0.02% total energy consumption.

I hope that provides something a bit more real.

Chris Jones's picture
Offline
Joined: 2011-09-30
Reputation: 0

Jeremiah,

I'm not quite sure what you're driving at, but performance curves are
used throughout DOE-2 and EnergyPlus to model the performance of HVAC
equipment. In fact, if my memory is correct, the chiller model in
EnergyPlus is taken directly from DOE-2.
The performance curves are not a "theory", but empirically derived
equations that map performance as functions of the ambient air
conditions and the part-load-ratio at that time step based typically on
manufacturers' test data. so yes, these curves do reflect
real world conditions. Take out these curves, i.e., assume that all
equipment perform the same under all conditions, and there wouldn't be
much point to simulating the system at all.

But getting back to Peter's hypothesis, I don't think the increase in
cooling energy is due only to the shape of the equipment curves. I think
it's a combination of that plus the zone air flows, and the
COOL-CONTROL. This last item is particularly tricky in a moderate
cooling climate. We have to remember that air-handling systems are not
ideal - if you reduce the cooling load in a zone that's not the warmest
without changing the relative zone air flows, you may just overcool that
zone and in the worse case increase reheat energy. The federal
government learned that back in the early 70's when in response to the
Arab Oil Embargo
they asked that all government offices raise their cooling setpoint
temperatures as an energy saving measure.

Joe

Joe Huang's picture
Offline
Joined: 2011-09-30
Reputation: 406