Transcript Slide 1

Cost-Benefit Analysis of Smart
Metering and Smart Pricing
Ahmad Faruqui, Ph. D.
NARUC Annual Convention
Miami Beach, Florida
November 14, 2006
A framework for quantifying costs and benefits
• Identify and measure costs
►
►
Deploying advanced metering infrastructure (AMI)
■ Advanced meters
■ Two-way communication links
■ Meter data management system
■ Billing system
Offering dynamic pricing signals
■ Administrative costs
■ Marketing costs
2
Framework (concluded)
• Identify and measure benefits
►
►
Operational benefits of AMI
Demand response (DR) benefits of dynamic pricing
• Operational benefits
►
►
►
Avoided meter reading costs
Faster outage detection
Remotely connect/disconnect service
• DR benefits
►
See next several slides
• Develop a present value of net benefits
3
Quantifying DR benefits
• Primary benefits = Quantity of DR (MW) * Value of
avoided MW
• Quantity of DR = kW reduction per participant * Number
of participants
• Value of avoided load = Cost of peaking capacity net of
energy profits
• Secondary benefits = Reduction in wholesale prices
+ increased system reliability + reduced
planning reserves + customer choice of rates
4
Will customers exhibit DR?
• Even a mild time-of-use (TOU) rate caused peak loads to drop by
5 % in Puget Sound
• Additional evidence is beginning to emerge from other pilots
►
►
►
►
►
►
►
►
►
AmerenUE, Missouri
Anaheim Public Utilities, California
BC Hydro, British Columbia, Canada
Commonwealth Edison, Illinois
Hawaiian Electric, Hawaii
Idaho Power, Idaho
Ontario, Canada
Pepco, Washington, D.C.
Public Service Electric & Gas, New Jersey
5
The most comprehensive evidence comes from California
• Two state commissions and three investor-owned utilities
conducted a scientifically designed experiment with 2,500
residential and small commercial and industrial customers in
2003-05
• Impacts were estimated for standard time-of-use (TOU) and
dynamic critical peak pricing (CPP) rates
• Customers on TOU rates dropped peak loads by 5 %, when
prices doubled
• Customers on CPP rates dropped loads by 13 %, when prices
quintupled
►
30% of the customers accounted for 80% of the impact
6
Price responsiveness varies by customer characteristics
C
N
o
H
ig C
h AC
I
Lo nco
w
m
In e
P
ov c o
er me
ty
Li
n
C e
o
N lle
o
ge
C
S
in oll
gl eg
e
e
F
M
a
ul m
ily
ti
Fa
m
ily
C
A
A
vg
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
Elasticity of Substitution
7
Enabling technologies boost the drop in critical peak
loads
Type of technology
60
50
40
30
20
10
0
Percent drop in critical peak load
Smart Meter
Smart
Thermostat
Gateway
Systems
8
Weighted
Average
Dynamic prices have a substantial impact in a hot climate
(Central Valley)
Figure 11
Hourly Load Shape - Complex Daily Share Model - Zone 4
2.5
kW Load
2.0
1.5
1.0
<-Peak Period->
0.5
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour
Control 4
Treatment 4
9
Dynamic prices even have an impact in a mild climate
(San Francisco)
Figure 8
Hourly Load Shape - Complex Daily Share Model - Zone 1
0.9
0.8
0.7
kW Load
0.6
0.5
0.4
0.3
<-Peak Period->
0.2
0.1
0.0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour
Control 1
Treatment 1
10
California’s utilities are developing advanced metering
infrastructure (AMI) business cases
• PG&E’s $1.7 billion AMI filing was unanimously approved by the
CPUC in July
►
►
►
Almost 90% of the benefits come from operational savings
By 2011, the utility projects more than 500 MW of demand response if
a third of its customers with central air conditioning adopt dynamic
pricing tariffs
It is proceeding to deploy five million electric meters and four million gas
meters
• SDG&E’s AMI filing is currently in hearings before the CPUC
• SCE has filed a Phase I feasibility report
►
It plans to file an application next year
11
Can others make use of the California results?
• Magnitude of response is driven by several factors
►
►
►
►
►
Existing rate design
New dynamic rate design
Existing load shape
Saturation of central air conditioning
Weather conditions
• Once these “initial conditions” are specified, the California
pricing model can be used to make preliminary forecasts of
dynamic pricing impacts in other regions
• Responses may be more transferable across regions than is
generally believed
►
In the mid-1980s, EPRI pooled data from five pricing experiments and
showed that customer response patterns were consistent across
California, Connecticut, North Carolina and Wisconsin
12
Percent drop in critical-peak load will vary with price and
climate
Figure 1-2
Percent Reduction in Peak-Period Energy Use on Critical Days
Average Summer, 2003/04
18
% Impact (kWh/Hour)
16
14
12
10
8
6
4
2
0
0.00
0.10
0.20
0.30
0.60
0.50
0.40
0.70
Peak Price ($/kWh)
Zone 1
Zone 2
Zone 3
13
Zone 4
State-wide
0.80
0.90
1.00
Putting it all together in five easy steps
• 1: Develop a dynamic pricing rate and estimate its impact per
customer
►
Ball park estimate: 10-30 % per participant
• 2: Identify the number of participants and associated marketing
costs
►
Ball park estimate: 10 – 30 % of the target market
• 3: Compute aggregate DR impact
►
Ball park estimate: 1 to 9 % of peak demand
• 4: Estimate value of avoided costs
►
Ball park estimate: $52 – 85 /kW-yr
• 5: Estimate the present value of benefits with the present value
of costs and derive an estimate of net benefits
14
Additional reading
• Ahmad Faruqui, “2050: A pricing odyssey,” The Electricity
Journal, October 2006
• Roger Levy, “A vision of demand response: 2016,” The
Electricity Journal, October 2006
• Plexus Research, Inc., Deciding on Smart Meters, Edison
Electric Institute, September 2006
• FERC, Demand Response and Advanced Metering, Staff Report,
August 2006
• Robert Earle and Ahmad Faruqui, “Toward a new paradigm for
valuing demand response,” The Electricity Journal, May 2006
• US Department of Energy, Benefits of Demand Response in
Electricity Markets, February 2006
15
Contact information
Ahmad Faruqui, Ph. D.
Principal
The Brattle Group
353 Sacramento Street, Suite 1140
San Francisco, CA 94111
Voice: 415.217.1026
Fax: 415.217.1099
Cell: 925.408.0149
Email: [email protected]
16