3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
Comparison of Software Models for
Energy Savings from Cool Roofs
Joshua Newa, William A. Millera,
Yu (Joe) Huangb, Ronnen Levinsonc
a Oak
Ridge National Laboratory (United States)
b White Box Technologies (United States)
c Lawrence Berkeley National Laboratory (United States)
1
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
The Roof Savings Calculator (RSC) is a webbased tool to estimate cool-roof energy savings
2014-08-20:
The Roof Savings
Calculator (RSC) is
undergoing revision and
validation. Results from
the current version of RSC
(beta release 0.92) may be
inaccurate and should not
be cited.
RSC = doe2attic + web interface
doe2attic = DOE2.1E + AtticSim
2
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
RSC uses doe2attic, a modified version of
DOE-2.1E coupled with AtticSim
• DOE-2.1E is used as the whole-building simulation program to take
the ceiling heat flows and duct losses calculated by AtticSim and
simulate their effects on the building’s heating and cooling energy
use
• AtticSim is a detailed simulation program to model attic heat
transfer, developed by ORNL starting in the 1990s
• AtticSim is an ASTM-certified procedure for estimating heat gain
and loss through ceilings under attics (ASTM C1340/C1340M-10)
• AtticSim has been calibrated against measured data
• AtticSim runtime for two simulations of a two-zone home
is 58 seconds with ducts in the attic, and 14 seconds without
• DOE-2 requires less than 5 seconds for the two annual building
simulations, thus accounting for only 7-25% of total RSC runtime
3
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
Early benchmarking efforts validated AtticSim,
doe2attic against southern California homes
Attic air temperature
Attic floor heat flux
•
•
•
•
single-family home
direct-to-deck cool tile roof
RSI-6.7 (R-38) batt insulation on attic floor
tested August 2008 in Fort Irwin, CA
4
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
RSC results diverged from earlier LBNL studies
RSC v0.92 (2010)
vs.
DOE2.1E (Akbari & Konopacki 2005):
• cooling savings ≈ 1.2X
• heating penalty ≈ 12X
source energy savings (kBTU/y∙ft2 attic)
20
RSC v0.92 (2010)
Old office w/built-up-roof
y=x
15
Phoenix, AZ
10
Miami, FL
New Orleans, LA
Los Angeles, CA
Houston, TX
Fort Worth, TX
5
Atlanta, GA
0
-10
-5
0
5
-5
Baltimore, MD
New York City, NY
Philadelphia, PA
Chicago, IL
10
15
20
-10
DOE 2.1E (Akbari & Konopacki 2005)
RSC v0.92 (2010)
vs.
MICROPAS (Akbari et al. 2006):
• cooling savings ≈ 1.4X
• heating penalty ≈ 11X
source energy savings (kBTU/y∙ft2 attic)
5
4
RSC v0.92 (2010)
New home w/asphalt shingle roof
y=x
3
2
1
Daggett, CA (CZ 14)
0
-5
-4
-3
-2
-1
-1
-2
0
1
2
3
4
5
Fresno, CA (CZ 13)
Los Angeles, CA (CZ 9)
-3
-4
-5
Santa Maria, CA (CZ 5)
Arcata, CA (CZ 1)
MICROPAS (Akbari et al. 2006)
5
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
Adding radiant heat transfer to attic model
should amplify annual source energy savings
Old office w/built-up roof, unventilated attic, no ducts in attic
doe2attic (2014-04-28)
vs.
DOE2.1E:
• cooling savings ≈ 1.73X
• heating penalty ≈ 1.59X
source energy savings ≈ 1.78X
doe2attic (2014-04)
source energy savings (MBTU/y)
30
y = 1.78x
R² = 0.98
25
20
i
e
15
c
10
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
(k)
b f
d
j
5
a
gh
k
0
0
5
10
y=x
15
IL_Chicago
CA_Los-Angeles
TX_Fort-Worth
TX_Houston
FL_Miami
LA_New-Orleans
NY_New-York
PA_Philadelphia
AZ_Phoenix
GA_Atlanta
MD_Baltimore
20
25
30
DOE-2.1E
6
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
doe2attic, EnergyPlus each predict % heating
penalties, % cooling savings >> DOE-2.1E
% Heating Penalties
80
y=x
doe2attic (blue) and EnergyPlus (red)
doe2attic (blue) and EnergyPlus (red)
50
% Cooling Savings
40
30
20
doe2attic
EnergyPlus
10
y=x
60
40
doe2attic
EnergyPlus
20
0
0
0
10
20
30
DOE-2.1E unmodified
40
50
0
20
40
60
80
DOE-2.1E unmodified
7
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
Team has identified several issues
(to date) with RSC 0.92
Symptom
Attic temperatures did
not vary with amount of
ceiling insulation
Cause
Mix-up between Conduction
Transfer Functions (CTF) in
AtticSim and Response
Factors (RF) in DOE-2
Resolution
DOE-2 RFs
converted to CTFs
when passed to
AtticSim
AC undersized in office
building
Order of zones in SYSTEM
did not match that in LOADS,
Roof bottom, attic,
causing heat flows calculated Input file correction
ceiling temperatures and by AtticSim to be added to the
heat flows incorrect in
wrong zone
office
Unusual heat flow
timing in top of attic
Arrays for duct elements in
AtticSim too small, causing
overwriting of heat flows
Arrays for ducts
expanded in AtticSim
8
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
Undersized AC in office building diminished
cooling energy savings in RSC v0.92
RSC v0.92 (2010), Fresno
doe2attic (2014-03-30), Fresno — fixed
9
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
Office building's roof bottom, attic, ceiling
temperatures and heat flows were incorrect
RSC v0.92 (2010), Fresno
doe2attic (2014-03-30), Fresno — fixed
10
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
Home's attic-floor heat flux peaks at 10:00 LST,
about 4 to 5 hours earlier than expected
RSC v0.92 (2010), Fresno
Issue appears to be fixed
in latest runs (September 2014).
11
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
doe2attic is being validated against measurements
from Fresno cool roof study (Rosado et al. 2014)
60
60
Cool house (cool tile roof)
Attic air temperature (°C)
Attic air temperature (°C)
Standard house (warm asphalt shingle roof)
measured
doe2attic (2014-09-18)
50
40
30
20
measured
doe2attic (2014-09-18)
50
40
30
20
10
10
0
6
12
18
0
24
12
18
24
12
18
24
20
20
Cool house (cool tile roof)
Attic floor heat flux (W/m²)
Standard house (warm asphalt shingle roof)
Attic floor heat flux (W/m²)
6
measured
doe2attic (2014-09-18)
15
10
5
0
-5
measured
doe2attic (2014-09-18)
15
10
5
0
-5
0
6
12
18
Hour of day, 1 August 2012
24
0
6
Hour of day, 1 August 2012
12
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
We seek to repair and validate RSC this fall,
release update in 2015
• All identified problems have been resolved
• Continue review of code (e.g., for reasonableness of local heat
transfer coefficients)
• Evaluate magnitudes, timing of temperatures, heat fluxes,
energy uses, energy savings predicted by doe2attic
• Validate doe2attic against Fresno home measurements
• Release RSC v1.0 in 2015
13
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
Thank you
14
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
SUPPLEMENTAL SLIDES
15
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
Why doe2attic (DOE-2.1E + AtticSim)
instead of EnergyPlus (i/ii)
• AtticSim is an ASTM standard (C1340 / C1340M-10)
• AtticSim has better algorithms for modeling the attic and ducts
• DOE-2 runs up to 80 times faster in calculating the building
loads, thus minimizing RSC runtime
• The authors are intimately familiar with AtticSim and DOE-2
and able to modify them as needed
16
3rd International Conference on Countermeasures to Urban Heat Island – Venice, October 13-15, 2014
Why doe2attic (DOE-2.1E + AtticSim)
instead of EnergyPlus (ii/ii)
Heat
AtticSim
EnergyPlus
transfer
method of calculation method of calculation
mechanism
Radiation
View factors
Calculated internally
based on fixed
geometry shape
Ventilation
The ventilation air is a
combination of stack
and wind.
Heat transfer includes
multiple surfaces in
contact with
ventilation air.
Recommended changes to
EnergyPlus
Approximated method or view
factor inputs
E+ will use Radiance to calculate
view factors. This configuration
algebra approach requires fixed
The approximated method based geometry.
on area ratios may be too simple.
Mean Radiant Temperature may
be better than the approximated
Inputs may be used to provide
method as a simplified approach
more accurate view factors.
Have multiple ventilation objects E+ may adopt this approach.
to calculate ventilation rate as a
Authors are unsure if EnergyPlus
function of temperature, wind
includes heat transfer between
speed, and stack effect.
ventilated air and surfaces.
No heat transfer along flow
pathway between surfaces and
ventilation air.
17