Energy Infrastructure of the Future - Data-Driven Results from the...
37
Data-Driven Results from the Pecan Street Smart Grid Demonstration Project Joshua Rhodes Ph.D. Candidate, The University of Texas at Austin Department of Civil, Architectural and Environmental Engineering Webber Energy Group Pecan Street IAC Quarterly Technical Workshop March 06-07, 2013
Transcript of Energy Infrastructure of the Future - Data-Driven Results from the...
Data-Driven Results from the Pecan Street Smart Grid Demonstration Project
Joshua Rhodes Ph.D. Candidate, The University of Texas at Austin
Department of Civil, Architectural and Environmental Engineering Webber Energy Group
Pecan Street IAC Quarterly Technical Workshop
March 06-07, 2013
Joshua Rhodes Pecan Street IAC: Slide 2/23
March 06-07, 2013
Presentation Overview:
• What is Pecan Street?
• Energy audit analysis – Set the stage
• Measured energy retrofit results – Do the homework
• Residential energy use curves – Go where none have gone before…
• Future Work and Conclusions
Joshua Rhodes Pecan Street IAC: Slide 3/23
March 06-07, 2013
First smart grid project that is customer focused
• 1,000 home area networks, 100% voluntary
• Smart electricity, water, and gas
• High penetration of residential solar (170+)
• Smart appliances and home energy management
• Dynamic pricing and information feedback
Pecan Street is a One-of-a-Kind Smart Grid Test Bed
Joshua Rhodes Pecan Street IAC: Slide 4/23
March 06-07, 2013
The Mueller Smart Grid Test Bed is at the Old Austin Airport
Joshua Rhodes Pecan Street IAC: Slide 5/23
March 06-07, 2013
Mueller in Context to the Best of Austin
Joshua Rhodes Pecan Street IAC: Slide 6/23
March 06-07, 2013
We Will Use Different Data Sets for Our Analysis
Analysis
Electric meter
Electric sub circuits
Natural gas meter
Water meter
Energy audits
Surveys
Direct measurements (quantitative)
Indirect assessments (quantitative and qualitative)
Joshua Rhodes Pecan Street IAC: Slide 7/23
March 06-07, 2013
Austin’s (Semi-) Unique Energy Conservation Audit and Disclosure Ordinance (ECAD)
• Most homes > 10 years old must get energy audit – Use market forces to increase the energy
efficiency of buildings – Address the Austin Climate Protection Plan
• Reduce peak demand by 700 MW by 2020
• To date the ordinance has resulted in over 12,000 energy audits on file for buildings (commercial and residential) in Austin, TX
Joshua Rhodes Pecan Street IAC: Slide 8/23
March 06-07, 2013
Our Analytical Results of ECAD Data Reveal Opportunities to Reduce Peak Demand
• Inefficient A/C units – Excess peak power demand of 205 MW – 8% of Austin’s peak!
• Oversized A/C units – Excess peak power demand of 41 MW
• Duct leakage – Excess energy consumption of 18%
• Reduced A/C capacity – Excess energy consumption of 20%
Joshua Rhodes Pecan Street IAC: Slide 9/23
March 06-07, 2013
Many Installed Air-conditioners are Inefficient
0.
100.
200.
300.
400.
500.
600.
Up To 6 7 8 9 10 11 12 13 14
Num
ber o
f Hom
es
Installed EER
Recommendation 205 MW of excess peak
Joshua Rhodes Pecan Street IAC: Slide 10/23
March 06-07, 2013
Many Air-conditioners are Oversized
4
6
8
10
12
14
16
18
20
4 6 8 10 12 14 16 18 20
Act
ual I
nsta
lled
Cap
acity
(kW
)
Manual J Recommended Capacity (kW)
Installed Units 120% Manual J 100% Manual J 75% Manual J
oversized
undersized (~1MW)
41 MW excess peak
Joshua Rhodes Pecan Street IAC: Slide 11/23
March 06-07, 2013
0%
2%
4%
6%
8%
10%
0% 10% 20% 30% 40% 50% 60%
Perc
enta
ge o
f Sys
tem
s
Total Duct Leakage
Ducts are Notoriously Leaky
AE recommendation 18% excess energy consumption
Blasnik et. al, 1995 26% Cummings et. al, 1990 24% Proctor et. al, 1997 20% Modera & Jump 1995 19% Siegel et. al, 1996 16%
Joshua Rhodes Pecan Street IAC: Slide 12/23
March 06-07, 2013
Air-conditioners are Operating Under Capacity
0%
5%
10%
15%
20% 40% 60% 80% 100% 120%
Perc
enta
ge o
f Sys
tem
s
Sensible Capacity: Measured vs. Rated
Under capacity
20% excess energy consumption
Joshua Rhodes Pecan Street IAC: Slide 13/23
March 06-07, 2013
Measured results from residential energy retrofits – a couple regressions…
Joshua Rhodes Pecan Street IAC: Slide 14/23
March 06-07, 2013
Some residents chose to undertake retrofits as a result of energy audits
• Retrofits included: – Applying solar shading to windows – Air sealing the house – Increasing attic insulation – Replacing HVAC equipment – Upgrading to an Energy Star appliance – Replacing duct work
Joshua Rhodes Pecan Street IAC: Slide 15/23
March 06-07, 2013
A panel regression (both temporal and cross-sectional) revealed some significant results
• Air-sealing, upgrading HVAC equipment, and duct sealing resulted in significant energy savings
– Attic insulation – lower degree of certainty
• Weather effects normalized by cooling and heating degree days
• Regression results reinforce energy audit results
Joshua Rhodes Pecan Street IAC: Slide 16/23
March 06-07, 2013
A second regression into home characteristic data reveals some interesting results
• This model assessed the impact on total yearly energy use of:
– Year built – Home size – Number of kids – Number of adults – Thermostat set points – Energy and Water IQ scores – Income
Joshua Rhodes Pecan Street IAC: Slide 17/23
March 06-07, 2013
Some results were expected, some were more interesting
• Year built – negative impact (newer homes use less)
• Size, number of adults/kids – positive impact on energy use
• Income – negative impact on energy use
• Energy/Water IQ – negative impact – Knowledge is power! – or a power use
reduction?
Joshua Rhodes Pecan Street IAC: Slide 18/23
March 06-07, 2013
Representative residential energy use profiles
Joshua Rhodes Pecan Street IAC: Slide 19/23
March 06-07, 2013
Time of Day
Powe
r Dem
and
(kW
)
0
1
2
3
4
5
6
0 6 12 18 24
It is hard to get an intuitive feel for this amount of data
Joshua Rhodes Pecan Street IAC: Slide 20/23
March 06-07, 2013
Clustering techniques allow “groupings” of similar use profiles
• Real hourly, seasonal-average home use profiles (70-100) are normalized
• The values (0-1) at each hour are used to create a 24 valued vector
• Centers of groups are formed and iterated on until the Euclidean distance from each profile to a center is minimized
• Averages are then taken for each cluster
d(p, q) =
vuutnX
i=1
(qi � pi)2
Joshua Rhodes Pecan Street IAC: Slide 21/23
March 06-07, 2013
●
●●
● ●
●
●●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●
●
0 6 12 18 24
Nor
mal
ized
Use
0.2
0.4
0.6
0.8
1.0
Time of Day
●
●●
● ●
●
●●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●
●
0 6 12 18 24
●
●●
● ●
●
●●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●
●
0 6 12 18 24
Time of Day
●
●●
● ●
●
●●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●
●
0 6 12 18 24
●
●●
● ●
●
●●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●
●
0 6 12 18 24
Nor
mal
ized
Use
0.2
0.4
0.6
0.8
1.0
Time of Day
●
●●
● ●
●
●●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●
●
0 6 12 18 24
●
●●
● ●
●
●●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●
●
0 6 12 18 24
Time of Day
●
●●
● ●
●
●●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●
●
0 6 12 18 24
These techniques reveal that there might be some variation in profile types
Joshua Rhodes Pecan Street IAC: Slide 22/23
March 06-07, 2013
●
●
●● ●
●
●
●●
●
●
●●
●
●●
●
●
● ●
●
●
●
●
0 6 12 18 24
Nor
mal
ized
Use
0.2
0.4
0.6
0.8
1.0
Time of Day
●
●
●● ●
●
●
●●
●
●
●●
●
●●
●
●
● ●
●
●
●
●
0 6 12 18 24
●
●
●● ●
●
●
●●
●
●
●●
●
●●
●
●
● ●
●
●
●
●
0 6 12 18 24
Time of Day
●
●
●● ●
●
●
●●
●
●
●●
●
●●
●
●
● ●
●
●
●
●
0 6 12 18 24
●
●
●● ●
●
●
●●
●
●
●●
●
●●
●
●
● ●
●
●
●
●
0 6 12 18 24
Nor
mal
ized
Use
0.2
0.4
0.6
0.8
1.0
Time of Day
●
●
●● ●
●
●
●●
●
●
●●
●
●●
●
●
● ●
●
●
●
●
0 6 12 18 24
●
●
●● ●
●
●
●●
●
●
●●
●
●●
●
●
● ●
●
●
●
●
0 6 12 18 24
Time of Day
●
●
●● ●
●
●
●●
●
●
●●
●
●●
●
●
● ●
●
●
●
●
0 6 12 18 24
The same seems to be true for winter profiles
Joshua Rhodes Pecan Street IAC: Slide 23/23
March 06-07, 2013
Developing representative energy use profiles will allow for more realistic simulations
• Profiles can be randomly chosen to simulate groups of homes
– Some profiles are more common – Weighting factors can be determined
• Multipliers can be used to simulate realistic home draw patterns with magnitudes
• Some profiles might be more desired by utilities – Some are ‘cheaper’ to utilities
• Some profiles might be less carbon intense
Joshua Rhodes Pecan Street IAC: Slide 24/23
March 06-07, 2013
Home audit and homeowner survey data might correlate with certain patterns
• Static data might indicate what temporal, seasonal patterns look like
• Certain aspects (efficiency upgrades) might change profile shapes
• Might not just be reduction by constant amount
• Some technologies might be more effective with certain profiles
Joshua Rhodes Pecan Street IAC: Slide 25/23
March 06-07, 2013
Conclusions
• There is vast potential for residential power and energy reduction
• Measured results indicate what retrofits make significant differences
• Education seems to play a role in energy use
• Clustering might be a helpful data analysis tool
Joshua Rhodes Pecan Street IAC: Slide 26/23
March 06-07, 2013
Acknowledgements
• Pecan Street Inc.
• Doris Duke Foundation
• The University of Texas at Austin – Dr. Michael E. Webber – advisor – Charles Upshaw
• Texas Advanced Computing Center (TACC)
• Austin Energy
Joshua Rhodes Pecan Street IAC: Slide 27/23
March 06-07, 2013
Thank you - Questions?
Joshua Rhodes
Civil, Architectural and Environmental Engineering Cockrell School of Engineering University of Texas at Austin
Joshua Rhodes Pecan Street IAC: Slide 28/23
March 06-07, 2013
Extra slides
Joshua Rhodes Pecan Street IAC: Slide 29/23
March 06-07, 2013
From 1 to 2 to 3D and beyond!
(x)
(x, y) (x, y, z)
2D 3D
1D
Joshua Rhodes Pecan Street IAC: Slide 30/23
March 06-07, 2013
The main technique here is a distance between two “points”
1
2
a
b c
y
x
a2 + b2 = c2
c =p
a2 + b2d(p, q) =
vuutnX
i=1
(qi � pi)2
Joshua Rhodes Pecan Street IAC: Slide 31/23
March 06-07, 2013
●
●
● ●●●
●●● ●
●●●
●●● ●●●
●●
●●
●●●●● ●●
0 1 2 3 4
x
y
●●
●●●●
●● ●
●●●●●●
●
●●●●●● ●● ●
●●●●
●
0
1
2
0 1 2 3 4
This allows for the automatic grouping of items that are close to each other
Joshua Rhodes Pecan Street IAC: Slide 32/23
March 06-07, 2013
Residential demand increases drastically from minimum to peak demand
Spring 2010
Residential Commericial
48% 52%
Summer 2010
Residential Commericial
Total Grid load ~30.6 GW Total Grid load ~63.6 GW
Joshua Rhodes Pecan Street IAC: Slide 33/23
March 06-07, 2013
A view of the change in magnitude 0
1020
3040
5060
70
Spring Summer
0
10
20
30
40
50
60
70
Elec
tric
Grid
(ER
CO
T) L
oad
(GW
)
Residential LoadCommercial Load
Joshua Rhodes Pecan Street IAC: Slide 34/23
March 06-07, 2013
Residential summer average temporal profile shows non-AC use somewhat level
050
010
0015
0020
0025
00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
0
500
1000
1500
2000
2500
Non−AC electricity useAC electricity use
Elec
trici
ty (W
)
Hour of Day
Joshua Rhodes Pecan Street IAC: Slide 35/23
March 06-07, 2013
A more clear view of summer non-AC electricity use
050
010
0015
0020
0025
00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
0
500
1000
1500
2000
2500
Non−AC electricity use
Elec
trici
ty (W
)
Hour of Day
Joshua Rhodes Pecan Street IAC: Slide 36/23
March 06-07, 2013
Powe
r Dra
w (k
W)
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
−2
0
2
4
6
8 UseGenerationGrid
Example 1 min. Data for One Home for One Summer Day
Joshua Rhodes Pecan Street IAC: Slide 37/23
March 06-07, 2013
Powe
r Dra
w (k
W)
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
−2
0
2
4
6 UseGenerationGrid
Example 1 min. (Average) Data for Multiple Homes (14) for a Summer Day
