Upload
bethany-fox
View
216
Download
0
Tags:
Embed Size (px)
Citation preview
2013 STATEWIDE AC CYCLING PROGRAMS PROCESS EVALUATION
PG&E SMARTAC, SCE SUMMER DISCOUNT PLAN, SDG&E SUMMER SAVER PROGRAMS
Findings & Recommendations
November 12, 2014
Presentation Agenda
PY2013 AC Cycling Process Evaluation Findings 2
Overview of Evaluation Effort Key Observations Evaluation Tasks Program Descriptions Integrated Findings Recommendations
Overview of PY2013 AC Cycling Process Evaluation Effort
3
Primary research objectives include: Document the AC Cycling programs’ administration and
delivery strategies Assessing the effectiveness of the program administration Evaluating participant experience with the programs and their
DR events
Secondary research objectives included: Review marketing costs per enrolled customer Examine the range of appropriate costs for AC Cycling programs
This presentation focuses on providing highlights of these overarching topics across all utilities, their
programs, and their customers
PY2013 AC Cycling Process Evaluation Findings
Key Observations
4PY2013 AC Cycling Process Evaluation Findings
Key Observations
5
The IOUs are implementing the three programs as designed Program is triggered on days with peak temperatures or high demand
Overall customer satisfaction is high among all three programs
Although each program has marked differences in the sub-groups studied
Customers who are notified of events are more satisfied and take more action to reduce energy consumption
But there was no clear strategy for notifications
Satisfaction with the program incentives were consistent for all IOUs
Despite the fact that the incentive structures vary significantly
Technologies are generally as effective as other programs in the US
Devices could be upgraded and override capability is a benefit
Marketing costs are slightly higher than industry average
PY2013 AC Cycling Process Evaluation Findings
Evaluation Tasks
6PY2013 AC Cycling Process Evaluation Findings
Opinion Dynamics spoke with 1,160 active and lapsed customers to assess their experience with the program
7
Evaluation Task Description
Program Materials and Database Review
Analyzed the program databases to characterize and understand the participant population
Reviewed program marketing and outreach materials
Program Manager Interviews
Conducted telephone / in-person interviews with program and EM&V staff from PG&E (3), SCE (3) and SDG&E (2); and SCE account representatives (2)
Implementer/Contractor/ Aggregator Interviews
Conducted telephone interviews with GoodCents (PG&E), GoodCents and NRG (SCE) and Comverge (SDG&E)
Participant Survey
Fielded telephone survey with 880 participants ensuring representation by technology type, dual-enrollment, geography and duty cycle selection, where relevant
PG&E SmartAC (328) SCE Summer Discount (264) SDG&E Summer Saver (288)
Lapsed Customer Survey
Fielded telephone survey with 280 lapsed customers ensuring representation by geography, technology and duty cycle selection
PG&E SmartAC (68) SCE SDP (71) SDG&E Summer Saver (141: 96 residential and 45 nonresidential)
AC Cycling Process Evaluation Tasks
Program Descriptions
8PY2013 AC Cycling Process Evaluation Findings
Combined, the AC Cycling programs have half a million participants and installed load control devices
9
AC Cycling Program Participants and Load Control Devices(As of December 2013)
PY2013 AC Cycling Process Evaluation Findings
AC Cycling Programs
PG&E SmartAC
SCE Summer Discount Plan
SDG&E Summer Saver
Residential Participants
Yes X X
Small Nonres Close to new enrollment X XLarge Nonres Not available X
Technology Offerings
Adaptive Load Switch Programmable
Thermostat Direct Load Switch Direct Load Switch
Event Triggers
Emergencies Economic dispatch Sub-lap System testing
Emergencies Economic dispatch Sub-lap System testing
Emergencies Economic dispatch System testing
Duty Cycle Options
All: 50% Residential: 100% and 50% Nonres: 100%, 50% and 30%
Residential: 100% and 50%Nonres: 50% and 30%
ContractorsScheduling and conducting installations, maintenance GoodCents
Scheduling and conducting installations, maintenance GoodCents and NRG
Aggregator model Comverge
Number of 2013 Events
3 sub-lap events 1 test system-wide test
event (all participants) 1 test sub-lap event
12 residential events (11 during the summer and 1 in the winter)
4 summer nonres events
6 system wide events
Event Periods in 2013
3-4 hours in evening for sub-lap events;
Varied for tests, with 1.5 hours in duration
Varied: most residential in late afternoon, with most with 2 hours duration; all nonresidential 1 hour in duration
All events to all segments four hours in duration
Incentive Structure
One-time incentive payment at enrollment
Summer monthly credits that vary according to duty cycle selection, AC tonnage controlled, and whether customers choose the option to override events
Bill credit calculated based on duty cycle option in at the end of year
Participant’s Experience
11PY2013 AC Cycling Process Evaluation Findings
Despite limited knowledge of program design features and low awareness of AC Cycling events, satisfaction levels are high overall
12
While we present statistical significance when comparing within program results, no statistical tests were established for across program comparisons.
Residential Nonresidential
PG&E SmartAC 8.0 Not Applicable
SCE Summer Discount Plan
8.5 8.2
SDG&E Summer Saver 8.3 7.3
Overall Program Satisfaction Ratings (0 to 10 point scale)
Residential Participants
Dually Enrolled (PG&E)
Residential-Highest
Duty Cycles
Nonresidential Lowest Duty Cycle
Characteristics of Program Satisfaction
PY2013 AC Cycling Process Evaluation Findings
Effectiveness of Program Administration and Delivery Strategies
13PY2013 AC Cycling Process Evaluation Findings
The program technology – adaptive/traditional load control switches and programmable thermostats - are appropriate for set-it and forget-it design
Failure rates for load switches are in line with national ranges.
All program technologies are one-way communicating devices, which provide limitations on:
Visibility into failed devices and available load as well as ability for the program to evolve to provide ancillary service and increase costs of QA/QC.
The Effectiveness of Load Control Technology
14
Failure rates: Within range of
similar programs for load switches (7% to 12%)
PG&E discontinued programmable
thermostats due to paging errors that lead to high failure
rates (40%)
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
Load Switches (12%)
Programmable Thermostats
(40%)
Fai
lure
Rat
es
National Average
PY2013 AC Cycling Process Evaluation Findings
Program options are not always clearly explained, potentially reducing potential savings from program
15
Lack of awareness of multiple duty cycle options were more pronounced among 50% and 30% Duty Cycle participants
Awareness of Multiple Duty Cycle Options
SegmentClearly
Explained
Not Clearly Explained/ Don’t
know
SDG&E
Residential 54% 46%
Nonresidential
28% 72%
SCEResidential 69% 31%
Nonresidential 62% 38%
Program options were clearly explained
30% Duty Cycle
(n=26)
50% Duty Cycle
(n=40)
50% Duty Cycle
(n=47)
15%
13%
5%
21%
18%
6%
RES
NO
N R
ES
SDG&E SCE
50% Duty Cycle (n=45)
50% Duty Cycle (n=105)
23%
45%
56%
40%
PY2013 AC Cycling Process Evaluation Findings
Note: Might be a self-selected group; satisfaction may be due to SmartRate program characteristics
- It might lead to additional savings
- Participants might feel discomfort when notified of events and leave the program
PROS
CONS
Event notifications are either non-existent (PG&E) or are opt-in. Evidence indicates those who receive notifications are more satisfied
16
Those who receive notifications (dually-enrolled) are more satisfied (PG&E)
Although some of this may also be attributed to other SmartRate features
Those who receive notification take action (PG&E)
Participants report taking action upon receiving notification (from other DR programs)
There are pros and cons to notifications:
Turn off lights (21%)
Delay use of appliances
(26%)
Unplug devices (20%)
PY2013 AC Cycling Process Evaluation Findings
Some participants report that they are likely to engage in actions that lead to snapback had they been aware of AC Cycling events
17
PG&E SmartAC
SCE Summer Discount
Plan
SDG&E Summer Saver
Res Nonres
Change temp settings
23% 10% 23% 18%
18% 6% 4% 16%
TOTAL 41% 16% 27% 34%
Precool
15% 22% 10% 16%
11% 8% 22% 10%
TOTAL 26% 30% 32% 26%
Participant Self-Reported Behavior had they been aware of an AC Cycling Events
Always
Sometimes
PY2013 AC Cycling Process Evaluation Findings
Satisfaction ratings are similar across utilities irrespective of incentive structure
18
(*) $ /calculated AC ton per day
Segment CategoryPG&E
SmartAC
SCE Summer Discount Plan
(bill credit)
SDG&E Summer
Saver (bill credit)
Residential
Incentive
One-time $50 per load switch at installation
Maximum Savings (100%):• Standard: $0.36/AC
ton/day• Override: $0.18/AC
ton/dayMaximum Comfort (50%)• Standard: $0.18/AC
ton/day• Override: $0.09/AC
ton/day
• 100% duty cycle: $38/ton paid at year end
• 50% duty cycle: $11.50/ton paid at year end
Satisfaction with
Incentive7.6 7.9 7.5
Commercial
Incentive NA
• Maximum Savings (100%): $12.69/AC ton
• Good Value (50%): $4.44/AC ton
• Maximum Comfort (30%): $0.89/ AC ton
• (all per month)
• 50% duty cycle: $15/ton per month
• 30% duty cycle: $9/ton per month
Satisfaction with
IncentiveNA 7.7 6.8
PY2013 AC Cycling Process Evaluation Findings
Program Marketing
19PY2013 AC Cycling Process Evaluation Findings
In general, marketing costs are slightly higher than industry average practices
20
Marketing costs Interviewed 12 program managers of AC Cycling programs across the
country Marketing costs are slightly higher (average of $68.00-69.00 per
enrolled customer) than industry average practices (average of $61.00/per enrolled customer for print mail)
Vary by the stage of the program, with newer programs having much higher marketing costs than established programs that are focusing on replacement of attrition
Marketing effectiveness Program marketing campaigns have, for the most part, included
customers in the entire service territories. This results in high costs and potentially suboptimal customer reach.
Targeted marketing, with the support of data analytics to identify optimal participants, can both reduce marketing costs per enrolled customer by increasing conversion rates, while simultaneously increasing program overall impact performance.
PY2013 AC Cycling Process Evaluation Findings
Program Recommendations
21PY2013 AC Cycling Process Evaluation Findings
Leverage Smart Meter Data/Advanced Analytics
22
Use data analytics to actively target participants more likely to provide load reduction:
Target by geography and participant profile Emphasize warmer climates / avoid coastal areas Emphasize those with peaky load during likely event days (warm
afternoons) / Avoid those who do not Target areas of locational grid constraints/historically high Locational
Marginal Price (LMP)
Leverage data analytics to more readily identify failed load control devices:
Identify whether participants consistently do not respond to AC Cycling events (indication of failure device or a non-contributor)
This option may be limited where the events are mostly sub-Lap limiting participants’ exposure to multiple events
Will reduce costs of random inspections and unnecessary truck rolls and increase the reliability of program capacity
PY2013 AC Cycling Process Evaluation Findings
Upgrade technology for two-way communications to allow for provision of ancillary services (if desired)
23
Requires both an upgrade to the program technology and program design (to allow events outside of current afternoon windows). Options for upgrading the devices to allow for ancillary services include, where feasible:
Technology Upgrade Options
Short-term,Low cost
• Replace load switch one-way door with a two-way communication chip.
Medium-term,High cost
• Replace load devices with upgraded two-way communicating devices.
Long-term• Leverage the “internet-of-things” and allow for web-addressable technology-agnostic options driven by the market. Utilities move away from providing technology to enable demand response services.
PY2013 AC Cycling Process Evaluation Findings
Consider introducing widespread event notification
24
If the utilities choose to keep the program as currently designed and thus, not use AC Cycling programs for ancillary and other grid-regulation services:
Notify participants of AC Cycling events: this may increase satisfaction and load shifted (where participants may engage in additional energy-saving behaviors)
Consider testing event notification with a customer subset to establish whether notification yields additional savings and higher satisfaction, or lead participants to leave the program at a higher rate
Enhance education about program duty cycle options Where multiple duty cycle options offered, enhance materials explaining
options to allow participants to select a higher-level duty cycle: Account representatives should clearly explain these tradeoffs when
marketing the programs Call center representatives should be trained to encourage higher duty
cycle options provided the customer feels comfortable with that choice Consider introducing a higher duty cycle option (PG&E only).
PY2013 AC Cycling Process Evaluation Findings
Revisit incentive structure
25
The lag between event experience and the process surveys potentially introduces recollection biases in survey results.
Ideally, participants would receive a survey about a program event within a few days of experiencing the event.
The programs can use pre-programmed phone and/or internet surveys to collect participant responses immediately following an AC Cycling event.
To address false positives, surveys could also be administered in a typical AC Cycling day (hot afternoon temperature/humidity) where no events are called to normalize results.
Survey customers immediately following an AC Cycling event
Given similar satisfaction ratings across programs despite different incentive structures, it may be that multi-year incentives and/or incentive amounts could be revisited without a significant adverse effect on satisfaction.
We recognize that incentives needed to maintain capacity may be different from incentives needed to recruit participants.
PY2013 AC Cycling Process Evaluation Findings
Detailed Program Slides
26PY2013 AC Cycling Process Evaluation Findings
One-time Incentive
$50 (since 2012; $25
earlier)None
Duty Cycle 50% for residential and 33% for previously enrolled nonresidential
SmartAC Only74%
Dually Enrolled
26%(n=154,398)
Load Switches
81%
PCTs19% (n=181,976)
Enrollment Option
Participants
Load Switche
sPCTs
Total Devices
SmartAC Only 114,751 109,032 18,117 127,149
Dually Enrolled (SmartRate)
39,647 37,912 6,086 43,998
Total 154,398 146,944 24,203 171,147
Participan
ts
Load Switche
sPCTs
Total Devices
Total 5,762 1,001 9,828 10,829
Number of 2013 events
5 5
Scope of Events3 sub-Lap /2
test3 sub-Lap /2
test
Event triggersEmergency/
TestEmergency/
TestAlignment with system peaks
Aligned Aligned
TechnologiesAdaptive Load Switch
PCTAdaptive Load Switch
PCT
Override Option
Offered* Offered*
Event Notification
Not offered Not offered
*online or through SmartAC hotline
Residential Participants
Control Devices
Program Design Features Event Characteristics
PG&E SmartAC Program
Cycling Option
Participants
Control Devices
Standard
Incentive
Override Incentiv
e
Maximum Savings (100%)
268,424 273,022 $0.36 $0.18
Maximum Comfort (50%)
27,856 28,572 $0.18 $0.09
Total 296,280 301,594 Per calculated AC ton per day Cycling Option Participants
Control Devices
Incentive
Maximum Savings (100%) 6,034 7,309 $12.69
Good Value (50%) 1,775 2,417 $4.44
Maximum Comfort (30%) 694 913 $0.89
Total 8,503 10,639 Per
calculated AC ton per day
Number of 2013 events
12 4
Scope of Events Most sub-LapSystem-
wideMost common event triggers
Economic Economic
Alignment with system peaks
Data not provided
Data not provided
TechnologiesDirect Load
SwitchT-Stat Pilot *
Direct Load Switch
T-Stat Pilot *Override Option
With capable load switch
Not offered
Event Notification
Not offered Opt-In
*1,300 thermostats tested in 2013
Maximum Sav-ings 91%
Maximum Comfort 9%
(n=296,280)
Maximum Savings
71%
Good Value 21%
Maximum Com-fort 8%
(n=8,503)
SCE Summer Discount Plan
Program Design Features Event Characteristics
Duty Cycle Choices Among Residential Participants
Duty Cycle Choices Among Nonresidential
Participants
50% Duty Cycle(69%)
30% Duty Cycle(31%)
(n=4,620)
100% Duty Cycle(49%)
50% Duty Cycle(51%)
(n=23,085) Cycling
OptionParticipant
sControl Devices
IncentiveIncentive Limit
100% Duty Cycle 11,302 3,929 $38/tonup to $152
50% Duty Cycle 11,783 13,787 $11.50/tonup to $46
Total 23,085 27,716 Annual bill credit in December
Cycling Option
Participants
Control Devices
Incentive
Incentive Limit
50% Duty Cycle 3,201 7,511 $15/ton n/a
30% Duty Cycle 1,419 3,839 $9/ton n/a
Total 4,620 11,350 Annual bill credit in December
Number of 2013 events
6 6
Scope of Events System-wideSystem-
wideMost common event triggers
System loadSystem
load
Alignment with system peaks
Aligned Aligned
TechnologiesDirect Load
SwitchDirect Load
SwitchOverride Option
Not offered Not offered
Event Notification
Opt-in by phone Opt-In by phone
Duty Cycle Choices Among Residential Participants
Duty Cycle Choices Among Nonresidential Participants
Program Design Features Event Characteristics
SDG&E Summer Saver Program