United States
                 Environmental Protection
                 Agency
                           Air and
                           Radiation
                           (6204-J)
EPA430/B-95-012
Jurse 1995
 -4 '•->
&EPA  Conservation
          Verification Protocols
          Version 2.0
          A Guidance Document for Electric Utilities
          Affected by the Acid Rain Program of the
          Clean Air Act Amendments of 1990
EPA
430
B
95
012
c.2
                                     ACID
         RAIN
                                           PROGRAM

-------
   If you are interested in obtaining more information on the Acid Rain Program,
or in receiving a copy of the User's Guide to the Conservation Verification Protocols.
            >: which includes a reporting form and instructions, call:

                           ACID RAIN HOTLINE
                                202-233:9620
       Monday through Friday, 9:00 a.m. to 5 p.m.. Eastern Standard lime

                             :'  or write to:

                     i     Acid Rain Division (6204J)
                                 U.S. EPA
                             401 M Street, S.W.
                           Washington, D.C 20460

-------
                                    ,-* • s , r
                              H) :  j'. .'_>;,£ >i;> ;'^_i

                              ^v:nctc;i,bc'' 20460
 CONSERVATION  VERIFICATION PROTOCOLS
                       Version 2.0
            V
            A Guidance Document for Electric Utilities
            Affected by the Acid Rain Program of the '
              Clean Air Act Amendments of 1990

                          by
  <                   Barry D. Solomon*
                      Alan K. Meier**
                     &EPA
                     ACID RAIN DIVISION
              U.S. ENVIRONMENTAL PROTECTION AGENCY
* U.S. Environmental Protection Agency
** Lawrence Berkeley Laboratory

-------

-------
ACKNOWLEDGMENTS
*"i^^i-^« -v^»»     — —
The Conservation Verification Protocols (CVPs) were developed as a team effort!  Instrumental was
the support provided by Alan Meier and his staff at the Lawrence Berkeley Laboratory (L6L). Also
instrumental was the Conservation Verification Subcommittee of EPA's Acid Rain Advisory Commit-
tee, ably chaired by Stan Hulett, which provided valuable guidance, advice, and direction for the
content of the CVPs. Membership of the Subcommittee included Tom Buckley, Cary Bullock, John
Fox, Larry Frimennan, Bill Harding, Jeff Harris (of .the Northwest Power Planning Council), Liz Hicks,
Eric Hirst, Stan Hulett, Marty Kushler, Steve Nadel, Bob San Martin, Vince Schueler, Sam Swanson,
and Steve Wiel (now with LBL). Other people who have given important input to the CVPs include
Joe Etc, Chuck Goldman, Jeff Harris, Barbara Litt, and Ed Vine of LBL, and Dan Blank, Marilyn
Brown, Paul Centolella, Pat Quran, Bill Gavelis, Phil Hanser, John Hoffman, Marvin Horowitz, John
Hughes, Phil Hummel, Ken Keating, Joe Kruger, Anne Gumerlock Lee, Brian McLean, Bill Miller,
Rick Morgan, Steve Offutt, Gil Peach, Ralph Prahl, Renee Rico, Judy Tracy, Uoy d Wright, Roger Wright,

and Cathy Zoi.

-------

-------
                      Table of Contents
Executive Summary....	.'	1

Purpose and Philosophy Behind the Approach	5

Section 1. Linkage With the Acid Rain Program	9
   Background	•	•	9
   Conservation and Renewable Energy Reserve (CRER)	;..	-9
   Reduced Utilization (RU) Provision	;	10
   Criteria for Deferring Conservation Verification to States	11
   Verification by EPA	12
   Forms and Data Requirements for Users	:	12
       First Full-Year Savings	13
Section 2. Verifying Energy Savings	..15
   Introduction	'..	,	15
   Path I: Monitored Energy Use	.'.	.15
       Reference Case	...»	15
       Service Adjusted Energy Use and Savings	;	16
       Gross and Net Savings Distinctions	:	17
       Confidence in Estimated Savings: Satisfying a 75% Hypothesis Test	17
       Acceptable Measurement Procedures	1B
       Duration of Metering...	19
   Path II: Stipulated Savings	19
       Stipulated Savings for Specific Measures	19
       Best-Practice Engineering Estimates	:	20
   Subsequent-Year Savings	>	20
       Option 1: Monitoring Subsequent-Year Savings	22
       Option 2: Inspection for Presence and Operation of Subsequent-Year Savings...	22
       Option 3: Default Subsequent-Year Savings	22
   Supply-Side Efficiency Improvements	23

Appendix A
   Definitions	;	25

Appendix B                            ;
   Service-Adjusted Savings and Hypothesis Test	27

Appendix C
   t-Statistics for 75% Confidence One-Tailed Hypothesis Testing	31

Appendix D                                               -
   Stipulated Savings, Procedures, Lifetimes, and Conversion Factors	33
       Engineering Estimates of Energy Savings	33
       Reduced Transmission and Distribution Losses Due to Demand-Side
       Efficiency Improvements	:	•	-35

-------
         Table of Contents (Cont.)
Energy-Efficient Refrigerators...	
Residential Water Heating Conservation Measures..	
Ground Source Heat Pumps For Homes (New and Retrofit)	
Higher Efficiency Lights in Office Buildings	,	
De-Lamping in Commercial Buildings	
Exit Sign Light Replacement	
Higher Efficiency Street Lights	'.	
Higher Efficiency Motors for Constant Load Applications	
Amorphous Metal Distribution Transformers	
Stipulated Lifetimes for Common Conservation Measures	
Useful Lives of Residential Energy Conservation Measures	
Useful Lives of Commercial and Industrial Energy Conservation Measures.
Stipulated Net-to-Gross Factors for Stipulated Measures	
.36
.37
.39
.40
.42
.43
.44
.45
.46
.47
.48
.49
.50

-------
                    Tables and  Figures
Figure 1.   Overview of Verification Options	r	2
Figure 2.   Relationship Between EPA's CVPs and Acid Rain Program Allowances	6
Table 1.   Qualifying Criteria and Applicability for Electric Utilities Using Energy.
           Efficiency Incentives Under Title IV of the CAAA	10
Table 2..   Bonus Allowances Awarded from the Conservation and Renewable Energy
           Reserve	'.	13
Figure 3.   Steps to Calculate Service-Adjusted Savings	16
Figure 4.   Stipulated Savings Path	'.	20
Figure 5.   Subsequent-Year Verification Options	21

-------

-------
                      Executive Summary
The Conservation Verification Protocols (CVPs)
.have been prepared by EPA as part of its mis-
sion to implement the Acid Rain Program of the
dean Air Act Amendments of 1990. The Acid
Rain Program has a goal of reducing SO2 emis-
sions by 10 million tons from the 1980 level. Be-
ginning in the year 2000, most electric utilities'
that bum fossil fuels will have to hold enough
SO2 emission allowances to cover their emissions.
    Energy efficiency programs can be an impor-
tant component of utility acid rain compliance
strategies because they can help meet customer
demand for electricity with reduced emissions
of SO2. In addition, the Acid Rain Program has
two explicit conservation incentives, which are
linked through opportunities to earn or save a
new tradable commodity, the emission .allow-
ance.  The verification of energy efficiency sav-
ings under these incentives is thus essential to
the credibility of the market approach. Most in-
vestor-owned utilities that qualify for these in-
centives will probably have their energy savings
verified by procedures specified by their State
Public Utility Commission (PUC). The primary
users of the CVPs under the Acid Rain Program
are expected to be public power utilities.
    The CVPs are designed to be rigorous with-
out being burdensome on the utility or the regu-
lator. The CVPs help to ensure the cost-effective-
ness of utility conservation programs and SO2
emission reduction measures, as well as the reli-
ability of energy savings from those measures.

Conservation Incentives in the
Acid Rain Program

The Conservation and Renewable  Energy Re-
serve (CRER) is a special pool of 300,000 emis-
sion allowances that is available to utilities that
meet electric demands with either conservation
or renewable energy generation.  To qualify for
the CRER, a utility must have at least one affected
unit, a least-cost plan, and net income neutral
rated for conservation measures (to be certified
by the U.S. Department of Energy). In addition,
the utility can only receive credit for demand-
side energy conservation measures.
   The Reduced Utilization provision applies to
the 110 Phase I power plants that must lower their
emissions in 1995-1999. In contrast to the CRER,
this provision allows for supply-side efficiency
measures and does not require least-cost plan-
ning or net income neutrality. There is no pre-
scribed limit to the number of allowances that
these utilities can save through conservation. •
Verifying Energy Savings:
The CVP Approach

The CVPs allow for two general savings paths
(see Figure 1): Monitored Energy Use or Stipu-
lated Savings. The Monitored Energy Use Path
is the preferred verification approach, and its
goal is to measure energy use in such a way as to
infer net energy savings, i.e.. the savings attrib-
utable to the utility conservation program.  The
Stipulated Savings Path includes procedures for
estimating savings as well as simple equations
and standard values for estimating stipulated en-
ergy savings from a limited number of corner-'
vation measures for which expected energy sav-
ings are well understood. This path also includes
criteria for developing program-specific engi-
neering estimates that may be used by a utility
in limited cases.  Finally, the CVPs also include
guidelines for verifying the persistence of energy
savings from conservation measures.

Path I: Monitored Energy Use

The preferred verification approach is  to infer
energy savings through the measurement of en-
ergy use. The key features of this verification
path include:

• Specifying a reference case.

-------
          • Adjusting for differences in factors such as
           weather,  operating hours,  and production
           rates.            ,      .        •
          • Determining net energy savings.
          • Establishing the appropriate statistical confi-
           dence in the savings.

          The reference case is the set of conditions and
          levels of service from which the energy savings
          are to be estimated.  To develop the reference
          case, the energy use of a group not participating
          in the conservation program is often used.
             Raw energy savings, i.e.. simple differences
          in metered energy use, nearly always require ad-
    Figure 1. Overview of Verification Options
               Utility Conservation Programs
      Stipulated Savings
  Standard
Equations and
  Values
                Monitored Energy Use
Engineering
 Estimates
                              Service-Adjusted Savings
                                  Gross
                                 Savings
    Net-to-Gross Discounting
                                   Net Savings
                         FirsMfear Savings
                       Subsequent-Year Savings
justments to account for differences in metering
periods, weather, indoor temperature, and lev-
els of service. The CVPs require these adjust-
ments for both the reference case and the cus-
tomers  participating in the conservation pro-
gram.
    The CVPs require estimation of net energy
savings, 1&, the savings attributable to the util-
ity conservation program. A standard method
for estimating net savings involves comparing
the gross savings of customers participating in
the program to the change in energy use of a simi-
lar group not participating in the program (the
"comparison group"), some of whom may have
installed similar conservation measures on their
own initiative. The comparison group will nor-
mally be considered the reference case.  If this
approach is not followed, "the CVPs allow  for
other ways to develop a reference case and to
derive net savings. For example, net savings may
be estimated indirectly through market research,
surveys, and inspections of non-participants.
    There will be uncertainty about the magni-
tude of energy savings from utility conservation
programs because of the many factors that can
bias estimates of energy savings. An objective
of the CVPs is to award allowances for savings
that occurred with reasonable certainty. The
CVPs require that the savings be expressed in
terms of the utility's confidence, based on statis-
tical analysis, that the true savings were equal to
or greater than those for which it applied. Based
on the state of the art of energy conservation
evaluation, a 75% confidence (or 1-tailed hypoth-
esis test) is required.
    Energy savings that have been adjusted for
levels of service and that satisfy the 75% hypoth-
esis test are considered "first-year savings" if the
reference case captures net savings (e.g.. through
use of a comparison group as discussed above).

Subsequent-Year Savings

A major uncertainty in utility conservation pro-
grams is how long the energy savings persist over
time.  Three options are available for verifying
savings in subsequent years: monitoring and
inspection, inspection only, and a default.

-------
    If monitoring and inspection is used to verify
persistence of savings, a utility must continue to
meet the 75% hypothesis test used to estimate
the first-year savings. The utility must measure
energy use in the 1st and 3rd subsequent years.
After the third subsequent year, the utility may
continue to receive full credit for its savings for
the lifetime of the measure.
    If the inspection only option is used, a util-
ity can receive credit for at least 75% of the en-
ergy savings in subsequent years,for up  to half
of the measure's lifetime. In the case of "passive
conservation measures" (gsg., wall insulation), a
utility may receive credit for 90% of the savings
for the full lifetime of the measure.
    If the utility discontinues any form of moni-
toring or inspection, the default option provides
for 50% of the first-year savings to continue in
subsequent years for half the lifetime of the en-
ergy conservation measure.

Path II:  Stipulated Savings

While the CVPs are oriented toward measure-
ment of energy savings, in some cases utilities
may use simple values or algorithms that have
been provided for a limited number of technolo-
gies. In other cases, they may develop their own
engineering estimates.  In such cases, a utility
may convert gross savings into net savings
through simple multiplication by a default fac-
tor, which varies between 0.50 and 0.95, or by
documenting an actual "net-to-gross" ratio. Sub-
sequent year savings for this path are estimated
by the same procedure used in the Monitored
Energy Use Path (see previous discussion).
    The rationale for the Stipulated Savings ap-
proach is that the performance of some measures.
is well understood and may not be cost-effective
to monitor. Use of stipulated savings in these
limited circumstances should prove very attrac-
tive to small utilities and ones with new conser-
vation programs. The CVP contains 15 measures
for  which there are stipulated savings: energy?
efficient refrigerators, pick-up of old refrigera-
tors, ground source heat pumps for homes, of-
fice building lighting upgrades, de-lamping in
commercial buildings, exit sign lights, street
lights, water heater insulation blankets, anti-con-
vection valves, pipe  insulation, low-flow
showerheads, water-faucet aerators, heat pump
water heaters, constant load motors, and amor-
phous metal core transformers (a supply-side
measure and thus ineligible for CRER allow-
ances).  Finally, Stipulated Energy Savings are
also provided for a utility to take credit for trans-
mission and distribution energy losses that are
avoided by demand-side energy conservation
measures.
   ' In certain circumstances in which extensive
measurement and analysis are not cost-effective
or feasible, an engineering estimate of energy
savings will be acceptable. These circumstances
include any of the following situations:
• Measurement costs would exceed 10% of pro-
  gram cost.
• No customer accounts for more than 20% of
  total savings.
• Energy savings are expected to be less than 5%
  of use of the smallest isolatable circuit (e.g..
  residential lighting efficiency improvements).

-------

-------
                 Purpose and Philosophy
                                    the Approach
Purpose
The U.S. Environmental Protection Agency (EPA)
has prepared the Conservation Verification Pro-
tocols (CVPs) as part of its mission to implement
the Acid Rain Program authorized by Title IV of
the Clean Air Act Amendments (CAAA) of 1990.
Two cornerstones of the Acid Rain Program are
SO2 emission allowance trading and energy effi-
ciency improvements as part of compliance strat-
egies of affected electric utilities. These corner-
stones are linked through two explicit incentives
for energy conservation that are described in
more detail in the Acid Rain rules: the Conser-
vation and Renewable Energy Reserve (CRER)
discussed in the Allowance System Rule (40 CFR
Part 73); and the Reduced Utilization (RU) pro-
vision for Phase I affected units discussed in the
Permit Rule (40 CFR Part 72). These two incen-
tive programs allow electric utilities to earn or
save allowances through qualified conservation
measures.
   The main goal of tile CVPs is to credit elec-
tric utility conservation programs for energy sav-
ings that they generated as a part of an SO2 emis-
sions reduction strategy. An ancillary goal is to
describe consistent procedures for measuring
and verifying these savings.
   The CVPs are intended  to describe good
practice for impact evaluators of electric utility
energy conservation programs. They are not in-
tended to be prescriptive. While the formal fed-
eral verification requirements will end when the
CRER and RU programs expire, users may want
to continue to follow the CVPs or to use them in
other applications such as State-level conserva-
tion programs.
   A good verification program should be rig-
orous without being burdensome on the utility
or the regulator. The CVPs will have the benefit
of helping to'ensure the cost-effectiveness of util-
ity conservation programs and SO2 emission re-
duction measures, as well as the reliability of
actual energy savings from these programs.
Eventually, energy savings should be as reliable
an option for providing energy services to util-
ity customers as energy production is currently.
Incentive for
Measurement
The CVPs described herein are intended to pro-
vide stronger motivation for monitoring energy
savings from conservation measures and pro-
grams. The CVPs embody an incentive struc-
ture designed to encourage more and improved
measurement and less reliance on estimation of
energy savings, while allowing for simple veri-
fication procedures through engineering esti-
mates and stipulated savings (i.e.. standard al-
gorithms for calculating energy savings from
specific measures) when those approaches best
fit the circumstances of the utility. The stipulated
savings approach, while generally much less ex-
pensive and time consuming to implement, is de-
signed to be conservative and will often result in
less energy savings credit than will a conven-
tional verification procedure. As a result, utili-
ties will more typically analyze, representative
customer groups to determine the energy sav-
ings.              -
   The CVPs have been designed to create a
single set of protocols with sufficient flexibility
to accommodate Acid Rain Program provisions
as well as future applications by other federal or
State agencies. Thus, the CVPs are self-standing
and independent of the individual regulatory
programs. This necessitates language linking the
individual allowance provisions and the CVPs.
Section I specifies the administrative details
unique to each regulatory program, such as tun-

-------
   Figure 2. Relationship Between EPA's CVPs
       and Acid Rain Program Allowances
   I   EPA'sAcId Rain Program    |
                                     State Energy
                                     Conservation
                                      Program
              EPA's Conservation
              Verification Protocols
                     Revise
  List
   °f
Stipulated
 Energy
 Savings
Update
        Technical Review
         • Committee
         ing, reporting and auditing, etc. Figure 2 illus-
         trates the relationship between the allowance
         schemes and the CVPs.  '
         The Need for Flexibility

         Energy conservation is a diverse activity.  No
         general protocols for verifying energy conserva-
         tion savings can anticipate every kind of conser-
        . vation technology, program, or activity that will
         be undertaken by utilities. Procedures for veri-
         fying the energy savings must therefore be flex-
                                              ible enough to accommodate verification of the
                                              common conservation measures as well as new
                                              efficiency options. For example, in the past many
                                              utility programs have focused on commercial and
                                              residential buildings. Evaluation and verifica-
                                              tion techniques developed for the residential and
                                              commercial sectors, however, will not be directly
                                              transferable to the industrial sectors. For these
                                              reasons, the CVPs will give general guidelines
                                              for verifying energy savings rather man specify-
                                              ing the verification procedure for each kind of
                                              measure.
                                                 The CVPs have been written to be accessible
                                              to all electric utilities: large and small, public and
                                              private. They can be used by utilities well versed
                                              in evaluation of energy conservation programs
                                              as well as by those who are new to this field.
                                                        Cost-Effective Evaluation
The cost of savings verification will depend on
the kind of energy savings program and verifi-
cation procedures used by the utility.  If the an-
ticipated extra cost of evaluation reverses the
cost-effectiveness of the conservation program,
a less stringent evaluation will be permitted. For
purposes of discussion, it is generally expected
that 5-10% of the program cost will be devoted
to savings verification. No specific measurement
technology is required; a utility can use what-
ever approach will achieve the verification re-
quirements at lowest cost. It is anticipated, how-
ever, that most utility verification programs will
rely on a mixture of utility billing data, inspec-
tion of  conservation  measures,  simple
submetering, and stipulated savings (to be ex-
plained below). In limited cases engineering es-
timates of energy savings can be used by a util-
ity.
   A rapidly increasing body of literature quan-
tifying the field energy savings from various con-
servation measures is emerging. In some cases,
the savings are sufficiently reliable that a com-
prehensive verification program may not be cost-
effective.  A list of these measures, with stipu-

-------
lated energy savings or a stipulated energy sav-
ings formula, has been developed by EPA and
appears in Appendix D.
    While the list of measures with stipulated
savings is not comprehensive, it will reduce the
monitoring and verification burden and allow
utilities to focus measurements on demand-side
management (DSM) programs where impacts are
less predictable. The  provision for stipulation
also offers some energy savings credit to conser-
vation programs that were undertaken without
adequate verification plans.  The list and, the
stipulated savings will be subject to periodic, pro-
fessional reviews and revisions.
    This document uses considerable specialized
vocabulary, some of which has not gained gen-
eral acceptance. A list of definitions of the most
important verification terms is presented in Ap-
pendix A.         „     (
Outlook for Verification

The practice of verification of energy conserva-
tion savings based on program evaluations is rap-
idly evolving. While an increasing number of
State governments are instituting Verification
standards for their investor-owned utilities, even
the leading States have been reassessing their
protocols. Section 111 of the Energy Policy Act
of 1992 will accelerate these activities, since it
amends the Public Utility Regulatory Policies Act
of 1978 and requires that integrated resource
planning include the verification of energy sav-
ings achieved through conservation. The Agency
strongly believes that the CVPs need to remain
flexible in order to allow utilities ample freedom.
to evaluate their energy conservation savings in
the most cost-effective manner.,

-------

-------
                                                                                     Section
      Linkage with the Acid  Rain  Program
Background
The purpose of this section is to provide back-
ground and context for the CVPs by linking them
with EFA's Acid Rain Permits and Allowance
System rules, which were promulgated on Janu-
ary 11,1993 (40 CFR §§ 72.43 and 72.91 and Part
73, SubpartF, respectively). Application require-
ments and applicability criteria will be discussed,
along with the differences between the Conser-
vation and Renewable Energy Reserve (CRER)
and the Reduced Utilization (RU) provision and
who is responsible for verification of energy sav-
ings. For a more complete discussion of these
issues, refer to 40 CFR §§ 72.43  and 72,91 and
Part 73, Subpart F.
    The overall goal of the Acid Rain Program is
to achieve significant environmental benefits
through reductions in emissions of SO2 and NOX,
the primary causes of acid rain. To achieve this
goal at the lowest cost to society, the Program
employs both traditional and innovative, mar-
ket-based approaches for controlling air pollu-
tion. In addition, the Program encourages en-
ergy efficiency and pollution prevention.
   Title IV of the CAAA sets as its primary goal
the reduction of annual SO, emissions by 10 mil-
lion tons below the 1980 level. To achieve these
reductions, the law requires a two-phase tight-
ening of the restrictions placed on fossil-fuel-fired
power plants. Phase I begins in 1995 and affects
110 mostly coal-burning electric utility plants lo-
cated in 21 eastern and midwestem States. Phase
n, which begins in the year 2000, tightens the
annual emissions limits imposed on these large,
high-emitting plants and,sets restrictions on
smaller, cleaner plants fired by coal, oil, and natu-
ral gas. The Program affects existing plants with
an output capacity of 25  megawatts or greater
and new utility units under 25 megawatts that
use fuel with a sulfur content greater than 0.05%.
   The Acid Rain Program creates a new trad-
able commodity, the SO2 emission allowance. The
verification of energy efficiency savings under
the CRER and RU programs is essential to the
credibility of the market approach. Once these
programs end, EPA's requirements for continu-
ous emissions monitoring of SO2 will provide
automatic verification of emissions reduction (40
CFR Part 75).
Conservation and
Renewable Energy
Reserve (CRER)

The CRER is a special pool of 300,000 allowances
taken from the Phase n allocations. The allow-
ances in the pool are available to utilities that
meet electric demands with either conservation
or renewable energy resources. Congress estab-
lished this Reserve to provide an early "jump
start" to energy efficiency and renewable energy
strategies for reducing SO, emissions. There are
several criteria that an electric utility must meet
in order to qualify for a share of the bonus al-
lowances in the CRER, which are issued on a first-
come, first-served basis (see Table 1):
• A utility must own or operate one or more af-
  fected power plant units or be the partial owner
  or operator of an affected unit (see 40 CFR
  § 73.82(a) and 40 CFR § 73.10 for a listing of
  these affected units and EPA's yearly base al-
  lowance allocations).
• Beginning July 1,1993, a utility with at least
  one Phase I affected unit may apply to the
  CRER for allowances for verified energy sav- '
  ings (post-hoc) that occur from January 1,1992
  to December 31,1994, or  until all the allow-
  ances in the Reserve are allocated, whichever
  occurs first (40 CFR §§ 73.82(g)(l), 73.80(b)).
• Beginning July 1,1993, a utility with only a
  Phase n affected unit (or units) may apply to
  the CRER for allowances  for verified energy
  savings (post-hoc) that occur from January 1,
                                                                                     9

-------
      1992 to December 31,1999, or until all the al-
      lowances in the Reserve are allocated, which-
      ever occurs first (40 CFR §§ 73.82(g)(l),
      73.80(b)).
     1 A utility must have a "least-cost plan" or plan-
      ning process for meeting future electric needs,
      which may consider social and environmental
      externality costs (40 CFR § 73.82(a)(4)).
     1 Investor-owned utilities must be subject to a
      rate-making process that ensures net income
      neutrality, including a provision allowing the
      utility's net income to be compensated in full
      for lost sales attributable to its conservation
      program (40 CFR § 73.82(a)(9)). The U.S. De-
      partment of Energy began certifying the net in-
      come neutrality of State Public Utility Com-
      missions' (PUCs1) electric rate-making proce-
      dures on January 1,1993 (40 CFR § 73.82(g)(2)).
      Bonus allowances may be awarded condition-
      ally pending  DOE certification (40 CFR
      §73.84(c)).
Table 1.
Qualifying Criteria and Applicability for Electric Utilities
Using Energy Efficiency Incentives
Under Title IV of the CAAA
FEATURE
Who May Apply?
* Utilities Must Have a
Least-Cost Plan and
Net Income
Neutrality?
Type of Efficiency
Measures That
Qualify
Emission Rate at
Which Allowances
Earned/Saved
How Many
Allowances
Available?
Timing of Program
CRER BONUS
ALLOWANCES
Owner/Operator of Any
Affected Unit
Yes
Demand-Side Measures
Only
0.004 Ibs/kWh
(0.4 Ibs/mmBtu)
300,000 Total
1992-94 (Phase I Utilities)
1992-99 (Phase II Utilities)
REDUCED
UTILIZATION
Owner/Operator
of Phase I
Affected Units
No
Both Demand-
and Supply-
Side Measures
Average Emission
Rate of
Phase II Units
During Phase I Year
No Limit
1995-1999
(Phase I Utilities)
• Utilities applying to the CRER must use quali-
  fied demand-side energy conservation mea-
  sures (listed in Appendix A(l) of 40 CFR Part
  73, Subpart F) that were installed on or after
  January 1,1992, or measures that do not ap-
  pear on the list but that qualify because they
  meet the requirements of 40 CFR § 73.81.
• Qualified measures must be consistent with a
  utility's least-cost plan or least-cost planning
  process; must be funded in whole or in part by
  the utility; must increase customer end-use ef-
  ficiency; must not increase the use of any other
  'fuels (other than renewables, industrial waste
  heat, or industrial waste gases); and must not
  be programs that are solely informational or
  educational (40 CFR §§ 73.81(a)(2)(i), 73.81(b),
  73.82(a)(3)).
A qualified electric utility will earn allowances
from die CRER based on a formula that assumes
an emissions rate of 0.4 pounds of SO2/mmBtu
(or .004 pounds/kWh). Verified conservation
savings of 500 MWh, for example, would earn
one allowance.
                                                   Reduced Utilization
                                                   (RU) Provision

                                                   In establishing the Acid Rain Program, Congress
                                                   recognized that during Phase I utilities might cir-
                                                   cumvent SO2 emission reductions by shifting load
                                                   to units not regulated in Phase I. To avoid this
                                                   potential problem, the rules implementing the
                                                   CAAA require that during Phase I, a unit whose
                                                   utilization is reduced below its 1985-87 baseline
                                                   level must surrender allowances representing a
                                                   shift of emissions to unregulated units, unless
                                                   such under-utilization is accounted for in one of
                                                   several ways. One way to account for under-
                                                   utilization is through energy conservation. The
                                                   RU conservation option may be especially valu-
                                                   able to some utilities, since it allows them to ac-
                                                   count for reduced load at Phase I units (and avoid
                                                   surrendering allowances) by receiving credit for
                                                   system-wide energy savings. Unlike the CRER
                                                   program, under RU a Phase I affected unit would
                                                   effectively receive credit for conservation at the
10


-------
SOj emission rate of the units to which load
would otherwise have been shifted (primarily
Phase U units). An affected utility could thereby
potentially save allowances at a much greater
rate than could be earned based on the f onnula-
basedCRER.
   The criteria for a utility to use energy con-
servation under the RU provision (40 CFR
§ 72.43) are less restrictive than the criteria for
using the CRER (see table 1). The RU provision
is applicable only to electric utilities with Phase
I affected units that file RU plans with EPA pur-
suant to § 72.43. Such utilities may use both de-
mand-side and supply-side (isfi., power genera-
tion, transmission, or distribution efficiency)
measures that result in verified energy savings
to reduce utilization of Phase I units below 1985-
1987 baseline levels.  The supply-side conserva-
tion measures are listed in Appendix A(2) of 40
CFR Part 73, Subpart F; supply- and demand-
side measures installed as early as January 1,1988
may qualify, as long as the energy savings that
occur during the RU period of 1995-1999 because
of those measures are verified.
   \ferificationofenergysavingsisrequiredfor
each year in which energy conservation measures
contribute to the RU. The savings from any con-
servation measures (other than improved unit ef-
ficiency measures) instituted within a Phase I
unit's utility system may be attributed  to the
Phase I unit.1  An RU plan must be filed with
EPA by November 1 of the year in which RU oc-
curs. An initial estimate of energy conservation
savings must be filed in the annual compliance
certification report by March 1 of the following
year. The utility unit's designated representative
must submit the verification results as part of a
confirmation report by July 1 of each year for
which an affected unit's annual compliance cer-
tification report is submitted to EPA.  Thus the
first possible verifications would take place in
the first 6 months of 1996 for energy savings that
occur in 1995, following submittal of an RU plan
by November 1995. The EPA Administrator may
grant, for good cause shown, an extension of the
time to file the confirmation report, however. If
the energy savings do not result in utilization of
Phase I units below baseline levels, such savings
 do not have to be verified.
    In contrast to the CRER program, a utility
 using the RU provision does not have to meet
 the least-cost planning or net income neutrality
 requirements, and since it is freeing up its own
 allowances, there is no .limit to the amount of
 conservation for which it may receive credit. Fi-
 nally, unlike the demand-side measures, which
 are installed at the customer side of the meter,
 the supply-side conservation measures at the
 power plant or in the transmission and distribu-
 tion system will result in energy savings that are
 all directly caused by the electric utility. Conse-
 quently, "net" energy savings from supply-side
 conservation measures can be assumed by util-
 ity applicants to equal "gross" energy savings, as
 discussed in the next section. Separate applica-
 tion forms for RU plans will be available at a fu-
. hire date.
 Criteria for Deferring
 Conservation
 Verification to States

 The primary users of the CVPs are expected to
 be public-power utilities. Most investor-owned
 utilities in States that qualify for the CRER will
 probably have their energy conservation savings
 verified by procedures specified by the PUC, and
 EPA recognizes that it would be burdensome to
 require these utilities to use two separate verifi-
 cation procedures. Therefore EPA will defer to
 the  conservation verification requirements of
 PUCs based on criteria discussed in the Allow-
 ance System Rule (40 CFR Part 73, Subpart F)
 and summarized below.
     EPA will require State PUC verification of
 energy savings claims of electric utilities if the
 FUC uses periodic evaluation of energy savings
 to determine any type of performance-based rate
 adjustment for conservation programs (40 CFR
 § 73.82(c)(l)). For RU conservation  (40 CFR
 § 72.43), use of qualifying State PUC verification
 is optional.  EPA believes  that the verification
 performed by State regulators is likely to be fairly
 rigorous when the evaluation of energy savings
' If certain conditions are met, a unit that is part of a holding company system may be attributed savings from conservation
  within any part of the holding company system. See40CFR§.72.43(b)(2){iv).
                                                                                           11

-------
    will be used to determine the cost of the conser-
    vation incentives to the ratepayers. In the case
    of both "shared savings mechanisms" and "lost
    revenue adjustment mechanisms," regulators
    must evaluate energy savings achieved by util-
    ity conservation programs in order to determine
    the proper adjustment to a utility's rates neces-
    sary to recover lost revenues or to provide a fi-
    nancial award.
        Under this approach for deferring conserva-
    tion verification to a State, the use of engineer-
    ing estimates of energy savings is allowed. The
    requirement to periodically modify performance'
    based rate adjustments through the evaluation
    of net energy savings, however, will help to en-
    sure that a State is committed to good evalua-
    tion practices. EPA recognizes that evaluation
    of energy savings is a young and rapidly evolv-
    ing field and that many States and utilities begin
    their evaluation programs with engineering es-.
    timates. Most evaluation experts agree that while
    engineering estimates are useful, they should be
    supplemented with statistically valid measure-
    ments of the decrease in energy use  resulting
    from utility energy conservation programs (Ls^
    net savings). These measurements should reflect
    both the actual number of measures installed and
    an ex post estimate of the energy savings
    achieved per measure.
        EPA believes that this approach toward con-
    servation verification and deferral to State PUCs
    is consistent with the approach of the Acid Rain
    Program, which encourages flexibility but insists
    upon appropriate measurement  standards.
    While a minimum standard for deferring con-
    servation verification to States is being estab-
    lished, EPA is not prescribing standard verifica-
    tion methods or techniques. States have the flex-
    ibility to design their own verification protocols,
    as many have. Because of the diversity of evalu-
    ation methods and the ongoing evaluation ac-
    tivities of  State agencies, electric utilities, and
    other professionals, EPA believes that this ap-
    proach will be more effective in ensuring good
    practice than would mandating specific methods
    or requiring universal use of the CVPs.
Verification by EPA

The CVPs are a guidance document for electric
utilities, and they may be revised periodically.
Moreover, as explained earlier in this report,
many utility applicants to the CRER will not have
to use the CVPs. Consequently, use ol the CVPs
is not a requirement, though the reader should
note that under EPA's Allowance System Rule
applicants to the CRER must demonstrate to the
satisfaction of the EPA Administrator that in the
case of qualified energy conservation measures,
energy savings have actually been achieved. EPA
will consider the CVPs as a source of guidance
for reasonable conservation verification proce-
dures. The rule allows, but does not require, that
applicants that do not qualify to have verifica-
tion performed by their State public utility com-
mission verify savings through use of the CVPs.
Forms and Data
Requirements for Users

Users of the CVPs should summarize their veri-
fication results on the CVPs reporting forms for
the CRER and RU. These forms and instructions'
are included in the User's Guide to the CVPs.2 Ap-
plicants to the CRER should send the appropri-
ate forms and any attachments (or a State PUC-
approved conservation verification, where appli-
cable), to EPA at the address provided at the end
of the form instructions. Applicants should call
the Acid Rain Hotline (202-233-9620) to receive
a copy of the reporting forms.
   EPA began its review of the verified conser-
vation savings claims when the CRER applica-
tion period opened on July 1,1993, on a first-
come, first-served basis. EPA reviews applica-
tions to determine whether they meet the require-
ments of Subpart F of 40 CFR Part 73. Varifica-
tion claims may be based on the CVPs, a State
PUC-sanctioned method, or other methods.  If
other methods are used, however, the time pe-
    2U.S. Environmental Protection Agency, The User's Guide to the Conservation Verification Protocols, EPA 430-B-92-002, Washington,
      D.C., April 1993.
12

-------
riod required lor EPA's verification of an appli-
cation is likely to be longer than if the CVPs or a
State PUC-sanctioned method is used, because
EPA would need to evaluate the alternative
method as well as the application claim. Defi-
cient verification claims will lose their place in
line for bonus allowances if they must be sent
back to the applicant for revision (see 40 CFR §
73.84(b)). Nonetheless, there is no deadline for
sending in applications to the CRER (although
the program has a limited duration), nor is there
a prescribed reporting period for conservation
results. The CRER allowances awarded to date
are shown'in Table 2.

                  Table 2.
       Bonus Allowances Awarded From the
    Conservation and Renewable Energy Reserve
    Utility
No. of Allowances
    Centerior Energy
       Cleveland Electric                   2
       Toledo Edison                      4
    Cincinnati Gas & Electric                 11
    City of Austin (Texas)   ,                97
    Dayton Power & Light                    4
    ESI Energy/Florida Power & Light         109*
    Long Island Lighting                   535
    Minnesota Power                       8
    New England Power
       Granite State Electric                24*
       Massachusetts Electric             436*
       Narragansett Electric             '  129*
    New York State Electric & Gas            142
    Niagara Mohawk           ,          --177
    Northeast Utilities
       Connecticut Light & Power           173
       Western Massachusetts Electric       30
    Orange & Rockland              -       46
    Otter Tail Power                        42*
    Portland  General Electric                277
    PSI Energy                           41
    Puget Sound Power & Light            1,002
    Rochester Gas & Electric                  7
    Sierra Pacific                        635*
    United Illuminating                     47
    Wisconsin Public Power                  3
   Total
    ' Some or all of total is for renewables.
        4,161
    A detailed evaluation of an energy conser-
vation program requires a large amount of data
and statistical analyses. In an attempt to mini-
mize the reporting burden for verification, EPA
is requesting applicants to submit general sum-
mary data (as is indicated on the CVPs Report-
ing Forms). The stipulated measures require fur-
ther collection of information, which is not re-
quired to be submitted to EPA but which should
be maintained in a utility's files in case an audit
is initiated by the Agency (40 CFR § 73.82(c)(3)).
An audit of a conservation verification might be
used to determine whether an incorrect number
of allowances was disbursed to a utility  or to
collect more detailed information about the per-
formance of energy conservation programs and
SQj emissions offsets or reductions.  Adequate
verification programs by utilities should mini-
mize the number of audits conducted by EPA. If
such an audit determines that an applicant's en-
ergy savings claims were overstated during the
verification, EPA may reduce the allocation of al-
lowances or require the surrender of allowances
by the applicant (40 CFR § 73.82(c)(4)).

First Full-Year Savings

A special accounting problem may arise during
the first calendar year in which a conservation
measure is installed and operated. The energy
savings during the first calendar year is typically
for only a part of one year rather than a full year,
since the conservation measures typically will not
all be installed on January 1. This creates some
confusion about the savings during the first year.
For reporting to EPA, it is necessary to report the
actual, partial year savings that coincide with the
calendar year upon which the CRER allowance
award will be based.  (The procedure for calcu-
lating the part-year savings is described in The
User's Guide to the CVPs.3) After the program's
first year, the'actual annual savings and calen-
dar-year savings  will be synchronized for the
measures that were installed in the previous year
and whose savings persist without attrition. The
CVPs, however, use the concept of the first-year
JU.S. Environmental Protection Agency, The User's Guide to the Conservation Verification Protocols, EPA430-B-92-002, Washington,
  D.C, April 1993.                                          .

                                                                                                13

-------
    savings for calculations of subsequent-year sav-
    ings. For example, the default subsequent-year
    verification option (see Figure 5) allows 50% of
    the first-year savings. The utility may thus need
    to calculate a first full-year savings to provide a
basis for estimation of subsequent-year savings.
The utility can have the option, however, of us-
ing the full calendar-year savings (which may be
the program's second year) as the basis for cal-
culations of subsequent-year savings.
14

-------
                                                                                      Section
                 Verifying Energy Savings
 Introduction

 As indicated earlier, the philosophy of the CVPs
 is to present general verification procedures
 rather than specific requirements. This approach
 offers utilities the maximum amount of flexibil-
 ity in developing verification methods. Never-
 theless, some consistency of method is needed
 in order to permit quantification of the energy
 savings and objective evaluation of energy sav-
 ings claims.  Two general savings verification
 paths are described: Monitored Energy Use and
 Stipulated Savings.  The two paths were illus-
 trated earlier in Figure 1 and  are discussed in
 detail below. This section also discusses require-
 ments for determining the persistence of energy
 savings and procedures for quantifying the im-
 pact of supply-side efficiency improvements.
Path I:  Monitored
Energy Use

The preferred verification approach is to mea-
sure energy use in such a way as to infer energy
savings. The following sections describe the key
steps in developing a method for converting en-
ergy consumption measurements into energy
savings. These steps are:
                 \
* Specifying a reference case against which to
  measure savings.
• Adjusting for several factors, such as weather,
  indoor temperature, and industrial production
  levels.
* Determining net energy savings, i.e.. the en-
  ergy savings resulting from the conservation
  program that would not have happened with-
  out the utility intervention.
• Establishing confidence in the energy savings.
The discussion concludes with an overview of
acceptable measurement procedures and the ap-
propriate duration of metering.  No particular
measurement technique is prescribed, however.
The intent of the CVPs is to suggest what may
generally constitute good-practice verification
procedures. The discussion begins with specifi-
cation of the reference case from which to mea-
sure the energy savings. Subsequent steps de-
scribe the adjustments and other methodologi-
cal issues.

Reference Case

The "reference case" is the set of conditions and
levels of service from which the energy savings
are to be estimated. The principal element of the
reference case is measured energy use. By defi-
nition, the reference case consists of the untreated
(pre-retrofit) group's energy use adjusted to ser-
vice levels from the treated group. This differs
from many definitions of reference cases, where
the untreated energy use and service levels es-
tablish the reference case. In the CVPs, service
levels from the treated case are preferred as the
reference because service levels often change as
a result of the retrofit and are more likely to pre-
vail for the subsequent years. (A formulation of
this definition is given in Appendix B.)
   In simple situations such as a conservation
program in single-family homes, the reference
case is the energy use of the pre-retrofit condi-
tion or comparison group.  The reference case
also includes levels of service that can affect the
level of energy use (and hence estimates of en-
ergy savings) such as weather, operating hours,
and industrial output. For an insulation program
of electric-heated homes, the reference case might
consist of the space heating energy use of the
homes prior to retrofit, adjusted to the long-term
average number of degree-days. More complex
situations will rely on data from surveys or in-
spections. For example, the reference case for a
building lighting retrofit program might consist
of a similar group of buildings unaffected by the
program. In addition, the operating hours and
illumination levels should be surveyed in order
to adjust for yearly variation.
                                                                                       15

-------
                The specification of the reference case will
            also determine if the utility is estimating gross
            or net energy savings. If the reference case cap-
            tures the impact of the conservation program
            (rather than only the technology), then the util-
            ity may follow the Net Savings path as shown in
            Figure 1. For example, if the reference case is a
            similar group of customers that did not partici-
            pate in the program, then the utility is estimat-
            ing net energy savings.

            Service-Adjusted Energy
            Use and Savings

            Raw energy savings, that is, simple differences
            in metered energy use, will nearly always require
Figure 3. Steps to Calculate Service-Adjusted Savings
   Monitoring Path.

^
r

energy use for treated and untreated cases (through monitoring) 1
            Step 2. Specify reference case from which to measure
              energy savings (energy use and levels of service)
               Step 3. Calculate service-adjusted energy
               	use tor treated case	
              Step 4. Estimate Service-Adjusted Savings (SAS)
             satisfying hypothesis test at the 75% confidence level
                          Does the
                        reference case
                         capture net
                          savings'
 adjustments to account for differences in meter-
 ing periods, weather, and levels of service. These
 adjustments to derive the "service-adjusted en-
 ergy use" (SAEU) and "service-adjusted savings"
 (SAS) are described below, while the general ap-
 proach is shown in Figure 3. The major steps
 involve calculating the SAEUs for the reference
.case and for the treated case. The difference be-
 tween the SAEU of the reference case and the
 SAEU of the treated case is the SAS.
     The energy consumption of many activities
 depends on the level of service required. For ex-
 ample, a house's space heating energy use de-
 pends on, among other factors, the outside and
 inside temperatures; a factory's energy use de-
 pends on the number of products manufactured
 in mat period; and a restaurant's energy use de-
 pends on the number of customers served. Some
 of the most important service factors are:
 • Severity of heating or cooling season.
 • Indoor temperature.
 • Operating hours.
 • Conditioned floor'area.
 • Production rate (or mix).
 • Illumination levels.
     A relationship between the level  of service
 and energy  use can usually be developed. For
 example, space heating energy use is a function
 of the heating degree-days, and lighting energy
 use depends on the number of operating hours
 and the lighted floor area. These relationships
 can often be derived through simple regressions
 of energy use and the level of service determined
 through measurements, surveys, and other data.
 For example, RCG/Hagler, Bailly, Inc.4 presents
 one formula quantifying the relationship be-
 tween energy use and various factors determin-
 ing the level of service, but other functions may
 be needed for industrial or special commercial
 applications. Although there is some uncertainty,
 these relationships may be used to normalize
 energy use in two periods with different levels
 of service. A step-by-step description of how to
 determine service-adjusted savings appears in
 Appendix B.
            4RCG/Hagler, Bailly, Inc., Impact Evaluation of Demand-Side Management Programs, EPRICU-7179, Electric Power Research
             Institute, Palo Alto, CA, February 1991.

-------
     When estimating energy savings, energy use
 should be adjusted to the treated ("post-retrofit")
 levels of service In all cases except where there
 are weather variables or where a long-term av-
 erage differs from the measurement period. This
 practice accounts for the fact that energy-effi-
 ciency improvements are often made in build-
 ings at the same time that service levels are im-
 proved.               .

 Gross and Net Savings Distinctions

 The intent of the CVFs is to estimate the net en-
 ergy savings from conservation programs. The
 utility may specify the reference case such that it
 captures the net savings. For example, the refer-
 ence case may be the energy use of an untreated
 group of similar customers. In this case, the net
 energy savings for the first year ("first-year sav-
 ings") emerge directly from the SAS calculation.
     The net effects can also be estimated indi-
 rectly through market research, surveys and in-
 spections of non-participants.  Based on these
 methods, a utility can determine a "net-to-gross"
 conversion  factor. (Multiplying the gross sav-
 ings by the "net-to-gross" conversion factor yields
 net savings.)
     A special case of net savings arises when the
' treated group of customers is effectively the en-
 tire population (leaving no untreated group
 available for comparison). This might occur when
 a utility targets the sole factory manufacturing a
 certain product. In this situation, the utility may
 still receive net savings credits by constructing a
 plausible scenario of what the customer would
 have done without  intervention by the utility
 program. This scenario might be based on mar-
 ket research, surveys, economic models, or in-
 dexes of investment activity.
    The net-to-gross factor is usually less than
 one, g.g.. if a certain fraction of "free riders" ac-
 cept the utility incentives even though they in-
 tended to do the measures anyway, at the same
 time or in the future; program customers may
 also increase their energy use (the snapback ef-
 fect). In some circumstances, however, the net-
 to-gross factor may be greater than one. These
 situations may involve customers that implement
 the conservation measures as a consequence of
 the utility program, even though they receive no
 direct benefits. Another example is when pro-
 gram participants adopt additional energy effi-
 ciency measures. For example, a utility rebate
 program to encourage consumers to buy more
 efficient refrigerators might cause the manufac-
 turer to reduce shipments of ineligible units to
. that utility's service area. Thus, the average effi-
 ciency of all refrigerators available to customers
 is increased, even if a customer chooses not to
 participate in the utility rebate program.
    This phenomenon (sometimes called  "free
 drivership," "spillover," or "market transforma-
 tion") 5 may amplify  the impact of a utility's re-
 bate program and may even raise the net-to-gross
 factor above one. Values greater than one can be
 used if the utility documents and quantifies the
 basis of the net savings estimate. See chapters in
 the report by Hirst and Reed6 (especially that by
 Saxonis) for a more detailed discussion. The
 documentation might include analyses of retailer
 inventory patterns before and after rebates were
 implemented, customer surveys, or other indi-
 cators of shifts in the market caused by the util-
 ity rebates.

 Confidence in  Estimated Savings:
 Satisfying a 75% Hypothesis Test

 There is always uncertainty in the savings asso-
 ciated with energy conservation measures. Un-
 certainty occurs over  the reliability of energy sav-
 ings estimates, the interaction effect of multiple
 conservation measures on total energy savings,
 and the persistence of energy savings over time.'
 For purposes of the  emissions allowances, the
 objective is to award  allowances for savings that
 occurred with reasonable certainty. The CVPs
 require that the savings be expressed in terms of
 the utility's confidence that the true savings were
 equal to, or greater than, those for which it ap-
 plied. In other words, in order to qualify for 100
 allowances using the CVPs,  a utility must be rea-
 sonably confident that the true savings equaled
 or  exceeded 100 allowances. The utility must
 5 Xenergy, final Report: Spillover Scoping Study, prepared by Xenergy, Inc. and Easton Consultants, Burlington, MA, January 1995.
 'E. Hirst and J. Reed, eds.. Handbook of Evaluation of Utility DSM Programs, ORNL/CON-336, Oak Ridge National Laboratory,
  Oak Ridge/TN, 1991.
                                                                                              17



-------
demonstrate that the true savings are equal to or
greater than the requested allowances with a 75%
confidence. Sampling procedures to calculate the
savings with the 75% confidence test are pre-
sented in Appendix B, while the critical values
for the t-statistic are shown in Appendix C..
    The procedure described in Appendix B,
while differing from that usually employed by
electricity rate regulators (who typically require
confidence plus or minus precision), offers utili-
ties more flexibility, the ability to use smaller
sample sizes, and the opportunity to claim some
legitimate savings even when the evaluation it-
self was not as successful as planned.7 The CVPs
take this approach because while it is not based
on usual statistical standards, it reflects the state-
of-the-art for reasonable impact evaluation of
savings from utility conservation programs. The
Service-Adjusted Savings satisfying the 75% hy-
pothesis are considered the "first-year savings"
if the reference case captured the net effects (see
Figure 4). If only gross savings were measured,
then a further adjustment is needed (see above
for a discussion of gross and net savings distinc-
tions).
           j*                    '
Acceptable Measurement Procedures

Energy savings must be derived from the differ-
ence between a reference case and a treated con-
dition. A variety of different methods are accept-
able for measuring this difference. For whatever
method is chosen, a utility  may use a variety of
analytical techniques, such as billing analysis,
econometric modeling, submetering, or some
combination of techniques. .Several common
methods are:
• Measurement of a treated group's energy use
  pre-/post-installation of conservation mea-
  sures compared to a comparison group's
  pre-/post-installation.
• Measurement of a treated group's energy use
  pre-/post-installation of conservation mea-
  sures (Lp.. the treated group is its own com-
  parison group).
 • Post-installation measurement of the treated
  group's energy use compared .to the untreated
  group's energy use fiue.. cross-sectional analy-
  sis).
 • Flip-flop of measures (old device is still in place
  and can be switched on for a short period).
    The conventional verification procedure for
 a retrofit project involves measuring energy use
 before and after the conservation retrofit. In gen-
 eral, however, if this procedure does not include
 analysis of a comparison group, it yields gross
 energy savings (rather than net) because it does
 not control for what would have happened to
 energy use if the retrofit had not occurred. Post-
 treated versus post-untreated groups might be
 used to estimate savings from a conservation pro-
 gram applied to new building construction. This
 procedure could also supplement a before/after
 measurement to provide net energy savings es-
 timates. Alternatively, building simulations of
 energy use are often used for these conservation
 programs (see Appendix D discussion  of Engi-
 neering Estimates).
    A flip-flop procedure allows for alternating
 measurements of the original equipment arid the
 replacement unit. A good candidate for flip-flop
 measurement is replacement of a resistance wa-
 ter heater with a heat pump water heater. A heat
pump water heater can be easily switched from
heat pump to resistance heating mode at weekly
 or monthly intervals until an adequate monitor-
ing period has been reached. Certain energy
management and control systems (from simple
 timeswitches to sophisticated Energy Manage-
 ment  and Control Systems) can also be tested
with a flip-flop procedure, that is, with the con-
 trols disabled. External controls would periodi-
cally switch from one system to another, thus
permitting estimates of energy savings during
only one heating season. A flip-flop procedure
is generally best applied when only one or two
technologies are involved.
    Combinations of these procedures are also
acceptable (and may even be necessary  for esti-
mating net energy savings). For  example, a net
' For a comparison of this procedure with other determinations of statistical confidence, see P. Hanser and D. Violette, "DSM
 Program Evaluation Precision: What Can You Expect? What Do You Want?" Proceedings, Fourth National Confirmee on
 Integrated Resource Planning. National Association of Regulatory Utility Commissioners, Washington, D.C., September 13,1992,
 pp. 299-313.

-------
savings analysis of a retrofit program will require
measurement of the energy use of participants
before and after the treatment, in addition to the
energy use of a group of non-participants.
    There are many different methods to com-
bine monitored data and survey data to estimate
energy savings from utility efficiency programs.
One popular method is the use of "Statistically
Adjusted Engineering estimates (SAE)." The
principles (and some of the variants) are de-
scribed by RCG/Hagler, Bailly, Inc8  The SAE
(and related, hybrid methods) is an acceptable
method for estimating savings as long as the
analysis includes energy use in the pre- and post-
retrofit periods and reflects the reference service
levels.

Duration of Metering

Metering is necessary to determine both the first-
year energy savings and the extent to which those
savings persist over the measure's lifetime Very
different considerations apply to these periods.
    First-year energy savings calculations must
be based on a sufficiently long period to cover
any natural cycle of activities that influences en-
ergy savings. For example, the "natural cycle"
for a space heating retrofit is one winter. Briefer
metering periods are acceptable under the CVPs,
however, if they can be reasonably extrapolated
to the whole year. For conservation measures
applied to constant loads (or those automatically
controlled), instantaneous power determinations
are satisfactory.
    Once all of these procedures are completed
by the utility, the result is the calculation of first-
year energy savings.
Path II:  Stipulated
Savings
The CVPs are oriented towards the estimation
of savings based on measurement of energy use.
In certain cases, however, the utility may use the
 simple Stipulated Savings methods given in Ap-
 pendix D to earn credit for the applicable con-
 servation measures instead of relying on a moni-
 toring program. These Stipulated Savings meth-
 ods are of two different types: (1) algorithms for
 calculating energy savings for specific measures;
 and (2) a set of criteria for using best-engineer-
 ing practices.

 Stipulated Savings for
 Specific Measures

 The rationale for the list is mat the performance
 of some conservation measures is well under-
 stood and may not be cost-effective  to monitor.
 Use of Stipulated Savings in these limited cir-
 cumstances should prove especially attractive to
 small utilities and ones with new conservation
 programs (although the energy savings assump-
. tions are often conservative). The Stipulated Sav-
 ings list contains the following kinds of measures:
 • Common measures whose savings are  well-
   documented in the literature (e.g.. energy-effi-
   cient refrigerators).
 • Measures where instantaneous metering is an
   acceptable proxy for first-year savings (g,g..
   exit sign lights).
 • Measures where short-term metering is an ac-
   ceptable proxy for first-year savings  (e.g..
   transformers).

 In some cases the laboratory test, "label," or other
 pre-specified efficiency parameter may be a reli-
 able indicator of energy use or savings. For stipu-
 lated measures, the utility may calculate energy
 use and savings based on these values rather than
 through metering. For example, in a refrigera-
 tor rebate program, the utility could calculate en-
 ergy savings by calculating the differences be-
 tween the maximum energy use dictated by the
 federal standard and the labeled consumptions
 of the refrigerators purchased in  its high-effi-
 ciency program. This verification  procedure
 would be much less costly than metering because
 the utility would need only to tabulate labeled
 values and count installations.
8 RCG/Hagler, Bailly, Inc., Impact Evaluation of Demand-Side Management Programs, EPRICU-7179, Electric Power Research
 Institute, Palo Alto, CA, February 1991.
                                                                                            19

-------
   Figure 4. Stipulated Savings Path
     Use stipulated fii
     utility-researched
     net-to-gross factor
      List of stipulated
    net-to-gross factors
       (Appendix 0}

ed
tors
)


Utility-researched 1
net-to-cross factor 1
(subject to approval)]


  First-year savings calculated [
Subsequent-year sayings calculated
1
        The approach for calculations of stipulated
    energy savings is outlined in Figure 4. The cal-
    culations in the stipulated measures result in
    gross energy savings.  The utility may convert
    the gross savings into net savings through simple
    multiplication by a default factor  (the "net-to-
    gross conversion factor") given in the list of stipu-
    lated net-to-gross factors in Appendix D; alter-
    natively it can use its  own adjustment factor
    based on its own research. The stipulated net-
    to-gross factors range from 0.50 - 0.95 and are
    based on experience with utility conservation
    programs. It should be noted, however, that net
    energy savings estimates for similar conserva-
    tion programs can vary widely, and if a utility
develops its own estimate through internal re-
search it may be able to claim larger net savings
than if it uses the stipulated value. The utility
may use its own estimate if it can demonstrate
the legitimacy of its analysis to EPA: Subsequent-
year savings for stipulated measures are esti-
mated using the same procedure as is used in
the Monitored Energy Use path (see the discus-
sion earlier in this section).

Best-Practice Engineering Estimates

In certain circumstances, measurement of energy
savings is not feasible or practical. These circum-
stances include any of the following situations:
            <
• No single customer accounts for more than 20%
  of total savings.
* There is an extremely small anticipated energy
  savings per customer, spread over many cus-
  tomers.  (An example might be a residential
  program  involving   distribution   of
  compact fluorescents.)  Such savings may be
  difficult to observe in a billing analysis  or
  submetering.
• The cost of monitoring and verifying energy
  savings is high. A utility may still want to ap-
 • ply for verified energy savings even though
  verification through monitoring would cost
  significantly more than 10% of the total pro-
  gram costs.
Determination of energy savings through energy
monitoring is the preferred means of verification.
When this is not feasible, in limited circumstances
engineering estimates of savings may be substi-
tuted, with an adjustment factor to account for
the actual realization rate of projected energy sav-
ings. The precise conditions in which engineer-
ing estimates can be used are specified in the list
of Stipulated Savings in Appendix D.


Subsequent-Year Savings

The persistence of energy conservation savings
over time is uncertain. This uncertainty stems
from both the eventual deterioration in a conser-
vation measure's technical performance and be-

-------
 havioral factors such as the energy user's shift-
 ing preferences and maintenance practices. The
 limited research that has been conducted on per-
 sistence of energy savings indicates that conser-
 vation programs often overestimate measure per-
 sistence, particularly in the commercial sector
 and for residential compact fluorescent lamps.9
    Three options are available for verifying sub-
 sequent-year energy savings: monitoring, inspec-
 tion, and a default.  Figure  5 illustrates these
 options, which are applicable to both the Stipu-
 lated Savings and the Monitored Energy Use
 paths. Various discounting factors have been in-
 serted to make monitoring attractive. By moni-
 toring, a utility can obtain credit for a greater frac-
 tion of the savings and for a longer period. In
 contrast, the default option greatly restricts the
 allowable savings.  This reflects decreased cer-
 tainty that savings persist when  they are not
 monitored by the utility.
    For all three options, the net-to-gross conver-
 sion  factor is calculated once for the first-year
 savings. The same conversion factor is used in
 all subsequent years to convert estimates of gross
 savings. For example, if a utility through exami-
 nation of a comparison group (or other accepted
 technique) determines that 15% of the treated
 group would  have implemented  the measure
 even without a utility incentive (a net-to-gross
 conversion factor of 0.85), then this factor car-
 ries to all subsequent-year savings.  Thus it is not
 necessary to follow the comparison group be-
 yond the first year.
    All three options make certain assumptions
 about the lifetimes of specific conservation mea-
. sures. The lifetime of a conservation measure is
 the median length of time (in years) that a mea-
 sure produces energy savings (see Appendix D).
 Stipulated lifetimes are provided  for the most
 common conservation measures. When a stipu-
 lated lifetime is not given, the utility must esti-
 mate a. lifetime. If significant degradation in per-
 formance occurs, then the effective lifetime is
     Figures. Subsequent-Year Verification Options
First-year I
savings |
ring
on for
eand
(ton
jit

100% of observed
hypothesis test
/ measure >^
fr/ raipJmadfvo \

Biennial verification In
subsequent years 1 and 3
(including inspection)
' Savings for remainder of
physical lifetime are
average of last two
• measurements
No
'"\. JSrtS^ V^ C"8881**
XraHwroa^r measure)
| Y|S

(Active
measure)
• '
^ 50% of first-year savtnos
physical libdiiM

90% of first-year
IfBSme

75% of ftstyear savings
fof units present and
operating tor rtaS of
physfcaJBeame
(bJenNaltapBcbom)

tor hall ot I
_\
shortened. Ameasure may be physically present
for many years beyond the time it is saving en-
ergy at its designed level.
    Utilities may shift from one verification path
to another in subsequent years.  It is expected
that these shifts will be from the monitoring op-
tions to less extensive verification procedures.
These shifts are permissible; once a shift is un-
 9See, £,&, E. Hirst and I. Reed, eds., Handbook of Evaluation of Utility DSM Programs, ORNL/CON-336, Oak Ridge National
  Laboratory, Oak Ridge, IN, 1991; K. M. Keating, "Persistence of Energy Savings," in E. Hirst and ]. Reed, eds.. Handbook of
  EvtAvaiion of Utility DSM Programs, ORNL/CON-336, Oak Ridge National Laboratory, Oak Ridge, TN, 1991, pp. 8949;
  Northeast Utilities, A Survey of Other Utilities for Measure-Life and Persistence-of-Savings Issues, Conservation and Load Manage-
  ment Department, Northeast Utilities, Hartford, CT, March 1992; L. A. Skumatz, et al. "Bonneville Measure Life Study: Effect of
  Commercial Building Changes on Energy Using Equipment," SRC No. 7619-R2, Synergic Resources Corporation, Seattle, WA,
  December 1991; E. LJ Vine, "Persistence of Energy Savings: What Do We Know and How Can It Be Ensured?" Energy—The
  International Journal, Vol. 17, No. 11,199Z PP-
                                                                                                  21

-------
    dertaken, however, a utility cannot reverse itself
    and return to a stricter option and receive con-
    tinuous credit from it.

    Option 1: Monitoring
    Subsequent-Year Savings

    The monitored savings verification option, which
    includes a requirement for inspection of installed
    measures, credits the utilities with the greatest
    possible savings. This recognizes the greater con-
    fidence in the existence and accurate estimate of
    the savings derived from monitoring.  In addi-
    tion, the utility receives credit for,savings mat
    occur after it terminates monitoring as long as
    the utility completes two subsequent-year veri-
    fications.
        The monitoring option requires that the veri-
    fied savings estimated by the utility continue to
    meet the 75% hypothesis test used for the first-
    year savings calculation. As mentioned previ-
    ously, conversion of gross savings into net sav-
    ings can be done through the first-year default
    conversion factor or by the utility-researched fac-
    tor, and the first-year values can be used in all
    subsequent-year savings.  A utility may, if it
    wishes, re-evaluate the net-to-gross conversion
    factor and use updated values.
        Verifications should be conducted every
    other year (biennially) after the first-year verifi-
    cation for a total of three verifications after the
    retrofit. Savings in off-years (i.e.. when no verk
    fication is performed) will be calculated at the
    inspection or default rate for no subsequent-year
    savings verification. When the next monitoring
    verification is performed, however, the utility
    may recoup uncredited savings that occurred in
    the off-year, based on the average of the last two
    measurements.  More frequent verifications are
    also acceptable.  The savings from the fourth year
    onwards are calculated as the average of the two
    subsequent-year verifications until the end of the
    measure's lifetime.

    Option 2: Inspection for Presence and
    Operation of Subsequent-Year Savings

    Surveys of the treated group (or representative
    samples) must be undertaken to confirm that the
conservation measures are both present and ac-
tively operating in subsequent years. A measure
is considered to be inoperative if its performance
is found to be significantly impaired, that is,
about half of the first-year savings are not occur-
ring. For example, clogged coils in a commer-
cial refrigeration system and broken time docks
for a thermostat setback control are reasonable
evidence of performance deterioration beyond
an acceptable limit.
   The first-year savings must be reduced by
the fraction of sites found to have non-function-
ing measures or where measures were removed.
For example, if an "active conservation measure"
was determined to have been removed from 10%
of the sites  inspected, then the subsequent-year
verified savings for that year must be reduced
by 10%.  Following this adjustment, where ap-
plicable, the utility may claim 75% of first-year
savings for half of the measure's physical life-
time (see Figure 5).
    Certain conservation measures do not need
regular maintenance and are unlikely to be re-
moved once they are installed. In mis sense, they
are "passive conservation measures" whose en-
ergy savings are likely to continue with a high
degree of certainty (assuming correct installa-
tion). For calculation of subsequent-year savings,
passive measures will receive 90% of first-year
energy savings without further inspection. Utili-
ties will receive these savings for the measure's
physical lifetime.
   A utility may use the passive measure path
shown in Figure 5 if it demonstrates that the
measure can be reasonably expected to remain
in place and perform to first-year levels without
requiring active maintenance or operation. Ex-
amples of passive measures include:  building
shell insulation, pipe insulation, and window
improvements.

Option 3: Default
Subsequent-Year Savings

If the utility decides not to continue any moni-
toring or inspection, it is still eligible for the de-
fault savings.  These savings consist of 50% of
first-year savings but are limited to half of the
measure's lifetime (as discussed previously).
22




-------
                                                                             REVISED P. 23
Line Loss Adjustment

Verified   savings   from  demand-side   energy
conservation measures or programs may include
savings from  transmission  and distribution line
losses mat are avoided because of improved end-use
efficiency. The utility may provide documentation
of line losses that are avoided by a conservation
program, or it may utilize the stipulation procedure
outlined in Appendix  D.  Such savings may be
eligible for either CRER allowances  or reduced
utilization credit.
Supply-Side Efficiency
Improvements

Supply-side efficiency improvements can also be
verified with these protocols. Although supply-side
measures are not eligible for CRER allowances, they
are eligible for credit with  respect to  annual
reconciliation for reduced utilization of Phase  I
generating units  (see 40  CFR 72.90).  Thus,
documenting supply-side efficiency improvements
can provide substantial value to a utility that has
underutilized Phase 1 units in a particular year.
Common supply-side improvements include boiler
and turbine improvements, heat rate improvement
programs, high efficiency transformer switchouts,
and other types of transmission and distribution
(T&D) efficiency improvements.
       Under   these   protocols,  supply-side
efficiency measures are divided into two categories.
The first of these is transmission and distribution
efficiency improvements, where both the inputs and
outputs are electricity. In general, energy savings
from T&D measures or programs should be verified
using the same procedures outlined  in  earlier
sections.  A reference  case must be specified
utilizing a statistically valid sample of installations,
with careful attention to the levels of service for the
equipment.
The second category is efficiency improvements in
generating facilities, where fuel or heat is converted
into  electricity.  The lack of a large sample of
similar generation efficiency measures precludes the
use of statistical tests applied to other measures
under  these  protocols.     Instead,   efficiency
improvements at a generating unit can be measured
in a straightforward manner by determining the
improvement in conversion efficiency, or heat rate,
through a systematic analysis of savings year fuel
inputs and electricity outputs relative to a base
year.1 If the efficiency improvement attributable to
a particular measure or program cannot be precisely
determined, measures or programs that collectively
contribute to verified heat rate improvements may
be grouped into a generic "heat rate improvement
program" category for each generating unit.
       Although generation  efficiency can  be
defined and measured in a number of ways, the
"nominal" heat rate is the appropriate measure for
purposes of these protocols. Nominal heat rate is
the ratio  of the unit's total fuel input to its total
generation  output during  the  relevant  year,
consistent with data reported on Form EIA-767. The
maximum credit which  may be earned for  all
measures at any generating unit is the reduction in
fuel  input attributable to the change in heat rate
since the base year  at  utilization  levels which
occurred during the savings year.
       Further information on credit for supply-
side  efficiency measures with regard to reduced
utilization of Phase I generating units can be found
in  The  User's Guide to  the  Conservation,
Verification Protocols  (Version 2.0)  and the
Energy  Conservation  and  Improved  Unit
Efficiency   Report   (EPA  Form  EC1   and
instructions).
                                                            For purposes of reduced utilization
                                                    submittals, the base year is 1987. ,

-------

-------
                                 Appendix A
                                      Definitions
 The following definitions are used in the CVPs.
 Several types of energy savings are used in this
 document, all of which are assumed to be elec-
 tricity. Since each type of energy savings has a
 different meaning and application, the.defini-
 tions are listed together below. Definitions for
 other terms are listed alphabetically following the
 energy savings terms.


 Energy Savings

' Raw, metered energy savings: The simple dif-
 ference in metered energy use between the
 treated customers and the comparison group
 during the relevant periods. Essentially no pro-
 cessing has been undertaken, so the periods need.
 not be similar and no normalizations have been
 performed. (The "savings" can even be negative.)
 This term is rarely used in this document but is
 defined to emphasize the distinctions in the defi-
 nitions of gross and net energy savings (below).

 Gross energy savings: The difference in energy
 use after adjusting for changes in levels of ser-
 vice and periods of measurement. Typical ser-
 vice adjustments include differences in weather
 (degree-days), operating hours, factory output,
 customers, illumination levels, conditioned floor
 area, inside temperature, etc. Periods of measure-
 ment for the treated and comparison groups will
 often differ and need to be normalized to a com-
 mon length (typically a year). Gross energy sav-
 ings seek to quantify the savings from the tech-
 nologies applied, so the comparison groups con-
 sist entirely of similar installations mat did not
 receive the conservation measures.

 Net energy savings: The energy savings attrib-
 utable to the energy conservation program. Al-
 ternatively net savings equals gross savings mi-
 nus free ridership and increased energy use from
 the snapback effect, plus any free drivership sav-
 ings. A standard method for estimating net en-
 ergy savings involves comparing the gross en-
ergy savings for the treated group to the change
in energy use of a similar group not participat-
ing in the program, some of whom may have
implemented similar conserving (or increasing)
measures on their own initiative.

First full-year savings: The energy savings ac-
tually achieved by conservation measures after
their installation for a calendar year or another
12 month period, based on measurement, esti-
mation techniques, or bom.

First-year energy savings: A specific period of
a conservation program's energy savings. Soon
after installation, the conservation measures be-
gin saving energy.' For purposes of verification,
however, the savings are divided into the "first-
year" and "subsequent-year" savings. The "first-
year" begins after the conservation program is
being deployed in a routine manner and the con-
servation measures are operating in their normal
modes.
                    «
Service-adjusted savings (SAS): The reduction
in energy use from the reference case due to the
conservation measure. The SAS include adjust-
ments for changes in service that may have oc-
curred between the untreated, and treated
groups.

Subsequent-year energy savings: The energy
savings occurring in the annual periods follow-
ing the first year.

Stipulated Savings: Energy savings for speci-
fied measures (or procedures) that a utility may
use instead of savings determined through moni-
toring.


Other Terms

Active conservation measure: A conservation
measure that requires regular maintenance and
that may be removed once installed.
                                                                                           25

-------
     Date of installation: The date on which the con-
     servation measure begins normal operation.

     Levels of service: Conditions describing the
     quality of energy-related services provided by
     an electric utility at the point of end use.   , •

     Lifetime: The median length of time (in years)
     that a measure produces energy savings.

     Lighting circuit: The electrical service to a group
     of lights serviced by its own circuit breaker or
     panel.  In large buildings, lighting circuits are
     often 277 Volts; in smaller buildings 120 Volts is
     more common. In complex buildings, the light-
     ing circuit is the combination of circuits servic-
     ing the lights of a distinct area, such as one or
     more floors.

     Net-to-gross conversion factor: Anumber that,
     when multiplied by the gross energy savings,
     yields net energy savings. The factor is typically
     (but not always) between 0 and 1 and is deter-
     mined by analysis conducted by an electric util-
     ity company.
Passive conservation measure: A conservation
measure that does not require regular mainte-
nance and is unlikely to be removed once in-
stalled.

Reference case: The energy use and set of des-
ignated levels of service from which the savings
are estimated.

Service-adjusted energy use: The estimated
energy use of a given group after adjusting to
different levels of service.

Start date: The date on which the program be-
gins normal operation.
26


-------
                                 Appendix  B
               Service-Adjusted Savings and Hypothesis Test
 General Approach for
 Service-Adjusted Savings

 The goal of the service-adjusted savings (SAS) is
 to estimate the energy savings resulting exclu-
 sively from the conservation measures and not
 from any changes in levels of service. It begins
 with the assumption that the estimate is based
 on the difference in energy use between an un-
 treated and a treated group. The approach seeks
 to adjust energy use in the untreated and treated
' groups to similar levels of services before esti-
 mating the difference. It is sometimes convenient
 to think of the untreated case as the "pre-retro-
 fit" and the treated case as the "post-retrofit" case.
 This analogy applies for very simple situations
 but requires modifications when the levels of
 service change or the verification involves cross-
 sectional analysis (such as in situations involv-
 ing the estimation of energy savings from new
 construction programs).

 Service-Adjusted
 Energy Use (SAEU)

 The energy use for a given installation can gen-
 erally be described as a function of the level of
 services supplied. Some of the common service
 levels determining the variation in energy use
 are:
 • Severity of heating or cooling season.
 • -Indoor temperature.
 • Operating hours.
 • Conditioned floor area.
 • Production rate (or mix).
    Other levels of service may be more appro-
 priate for special end uses or applications. The
 key aspect is that changes in these levels of ser-
 vice are responsible for differences in energy use.
 The relationship between service levels and en-
 ergy use can be expressed as a function:
 For lighting energy, use, the function might be of
 the form, E. = a + JJH., where a and p are con-
 stants and H. is the number of hours per year the
 site operates. (These coefficients might have been
 derived from a regression of submetered data.)
 The coefficients, a and p, reflect a level of effi-
 ciency of delivery of that service which is unique
 for the hardware and operating procedures in
 place. This relationship must be derived through
 monitoring and (where appropriate) market re-
 search data for both the treated and untreated
 conditions:
 The coefficients in the functional relationship will
 be different for the two situations because they
 reflect the efficiency at which the service is be-
 ing delivered. In the lighting case, for example,
 the coefficients a and P will be smaller in the
 treated case because less energy will be required
 to obtain the same level of service.
    The "reference case" is the energy use' and
 set of designated levels of services from which
 the savings are measured. The energy use asso-
 ciated with the reference service levels is called
 the service-adjusted energy use, SAEU^ By defi-
 nition, the reference case consists of the untreated
 group's energy use adjusted to service levels from
 the treated group (with exceptions described
 below):
 This differs from many definitions of reference
 cases, where the untreated (pre-retrofit) energy
 use and service levels establish the reference case.
 In this protocol, service levels from the treated
, case were chosen as the reference because ser-
 vice levels often change as a result of the retrofit
 and treated case service levels are more likely to
 prevail for the subsequent years. For example, a
 building exit light retrofit will probably include
 retrofitting lights that were not operating (i.e.f
                                                                                           27

-------
burned out) in the first place. In die course of
HVAC retrofits, ventilation rates are often raised
to meet code requirements.  In both cases, the
post-retrofit conditions are a more appropriate
baseline from which to measure the savings.

Calculation of
Service-Adjusted Savings

The "service-adjusted savings" (SAS) is the dif-
ference between the SAEUs:
           SAS * SAEUnl-SAEUtatt
The SAS must satisfy the 75% confidence hypoth-
esis test (discussed later) so that SAS will always
be less than or equal to the difference in SAEUs:
           SAS & SAEUat-
In most cases, the service levels will change in
the untreated and treated groups. For the light-
ing retrofit example, the lights may have oper-
ated 3000 hours before the retrofit and 4000 after
the retrofit. The reference level of service is thus
4000 hours. TheSAEl/ .is therefore an estimate
                    BCi
of the energy use of the original inefficient lights
operating at the higher number of hours occur-
ring after the retrofit:

             SAEl/w, =/„„, (4000)
             \
The SAS for this situation is:
          SAS = SAEU^-
           =/«r<«W) -f^ (4000)

Note that this assumes no other changes of ser-
vice level occurred in the treated case. If the light-
ing retrofits also included adjusting illumination
levels, this factor should be included in/^ and
/neat- A possible relationship would be:
where L is the average level of illumination, in
lumens, as determined by a series of representa-
tive spot measurements.  The SAEUref and
        , are calculated similarly.
Weather Adjustments

The severity of the heating and cooling seasons
fluctuates considerably from year to year. Simi-
larly, other meteorological factors, such as wind
speed and insolation, have long-term, average
conditions that may differ from the period of
measurement. Using the measurement periods
to project subsequent-year savings can lead to
obvious errors. Fortunately, long-run average .
data for these services are available and should
be used for the reference case. The appropriate
reference service levels for space heating would
be the long-term average number of degree-days
rather than those for the treated year.
   The difference between the conditions dur-
ing measurement and the average means that the
reference case must contain some levels of ser-
vice that differ from the treated period. As a re-
sult, both the untreated and treated groups will
need service adjustments. A new set of services
must be defined such that:
where Sivg becomes the S^
    Energy use often depends on a combination
of service levels determined by the user and the
weather. For example, a building air condition-
ing system's energy use might be determined by
the building's operating hours and the cooling
degree-days experienced:
In this case, SAELJref will contain service levels
from both the treated and average conditions:
                                                The SAEUtiat requires a similar adjustment:

-------
As a result, the service-adjusted energy use will
reflect a combination of service levels derived
from the measurement periods and long-term
average values.

Aggregation

The above procedure for calculating the service-
adjusted savings can also be applied to a group
of sites by aggregating the individual results or
pooling data and performing a single analysis
on the aggregate data. The pooled approach may
be necessary when level-of-service information
(such as that derived from surveys) is not avail-
able for individual sites.

Sampling and Hypothesis Test

In all cases, the verified savings must satisfy a
75% confidence hypothesis, that is, the utility
must have 75% confidence that the actual sav-
ings equal or exceed the estimated amount. As
indicated above:
         . SAS * SAEUmf-
The SAS must be lowered until the 75% confi-
dence in the difference SAEU . ~ SAEIJ., actu-
                           Irl       DNi
ally occurs.  The precise technique to evaluate
the 75% confidence hypothesis will depend on
the sampling procedure and experiment design.
    Ideally, all participants in the treated and un-
treated groups should be included in the verifi-
cation. Here, simple tests for confidence in the
savings can be applied. (Note that uncertainty
in results exists even when 100% of the popula-
tion is evaluated.)
    the high cost of monitoring, surveys, and
data management will often prevent evaluation
of 100% of the participants.  If comprehensive
evaluation is not possible, then a statistically rep-
resentative sample of participants is acceptable.
    If a sample of n installations is monitored,
out of a population of N, then
where SASu  .  is the average SAS for the
sampled installations. In this case, a hypothesis
test must be used to determine the population
mean, SAS  ^^  that can be extrapolated to
have occurred within a certainty of 75%.
  '  A variety of hypothesis tests may be used.11
A calculation using a one-tailed t-test is presented
here as an example. The SAS     is calculated
as follows:
       SAS.
           popaktioa
iSAS
                                  SAS
                             ••021
where tOK is the sample standard deviation of
SAS, and is the critical value of the t-statistic for
a one-tailed hypothesis test at the 75% confidence
level for n - 1  degrees of freedom (determined
from a table).  A table of critical values of the t-
statistic is included as Appendix C. The total
service adjusted savings that may be claimed for
the population is then N x SAS   ubtioo.


When the Sample is the Population

In certain cases, the evaluation covers every pro-
gram participant. This occurs when, for example,
the program targets only a few large customers,
or where the evaluation has access to billing data
for every participant. These situations eliminate
any uncertainty in the savings due to sampling
because the availability of data for the total popu-
lation eliminates the need to sample. The only
remaining uncertainty results from the proce-
dures for estimating savings due to changes in
services (that is, adjustments for differences in
weather, operating hours, etc.)
   When the utility verifies energy  savings
through analysis of the entire population of af-
fected customers, the adjustment to achieve 75%
confidence is not required, ^Instead, the utility
may calculate  the service adjusted savings as-.
suming 100% confidence. In a utility weather-
ization program, for example, the most likely
service adjustment is for differences in degree-
days between the two retrofit years and the long-
term average. After this adjustment is made, the
11 P. W. John, Statistical Methods m Engineering and Quality Assurance, John Wiley & Sons, New York, 1990:
                                                                                       •I  29


-------
first-year savings result, which can be used in,.
subsequent calculations for net-to-gross and sub-
sequent-year savings.
    A verification plan designed to include all
the affected customers may in the end obtain
usable data for less than the entire population.
Sample attrition is due to unavoidable events
such as terminated customers, data collection
errors, participants opting out, etc. This verifi-
cation path can still be used as long as 90% of
the original energy use is captured  in the evalu-
ation. (This corresponds to 90% of the custom-
ers except where a few, large participants drop
out of the evaluation.) An additional 10% attri-
tion per year is permitted if the utility chooses to
continue monitoring energy savings.
Step-by-Step Procedure

The following steps summarize the procedure for
calculating the service-adjusted savings:
1. Identify reference case and levels of service de-
  termining energy use. This usually happens
  after the pre-retrofit phase of metering is com-
  pleted!

2. Identify levels of service requiring average
  conditions (usually weather-related).
          ,'
3. Quantify relationship between service levels
  and energy use based on energy data, manu-
  facturing data, surveys, etc.
                                               4. Calculate SAEU
                                                                n{
                                                5. Calculate SAS at 75% confidence level.

-------
       Appendix C

 T-Statistics for 75% Confidence
 One-Tailed Hypothesis Testing
n-l
T~
"2"
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
 i:000
~816~
 .765
 .741
 .727
 .718
 .711
 .706
 .703
 .700
 .697
 .695
 .694
 .692
 .691
 .690
, .689
n-l
18
20
21
22
23
24
25
27
26
28
29
30
40
60
120
.688
.688
.687
.686
.686,
.685
.685
.684
.684
.684
.683
.683
.683
.681
.679
.677
.674
                                                       31

-------

-------
                                 Appendix D
                 Stipulated Savings, Procedures, Lifetimes,
                              and Conversion Factors
           Engineering Estimates of Energy Sayings
Technology Description

In certain circumstances, an engineering estimate
(including a building simulation) of energy sav-
ings will be acceptable in lieu of measurements.
An engineering estimate typically requires use
of recognized calculation procedures and tech-
nical specifications of materials and equipment.12
Engineering estimates or building simulations of
savings often differ from monitored savings be-
cause of incorrect assumptions, especially regard-
ing initial energy use and operating schedule."
Therefore, when measurement is  not,used,
greater emphasis must be placed on verification
of actual measure installation and correct opera-
tion.
    Engineering calculations may be used in any
one of the following situations:
• Monitoring costs would exceed 10% of pro-
  gram cost (including cost of the conservation
  measures).
• No single customer accounts for more man 20%
  of total savings.
• Energy savings are expected to be less than 5%
  of electricity use of the smallest isolatable cir-
  cuit. For example, energy savings from resi-
  dential lighting conversions will appear on
  many different circuits. In this case, the small-
  est isolatable circuit is the whole house. Ifen-
'  ergy savings are expected to be less than 5% of
  total household electricity consumption, then
  engineering estimates may be used.

 Energy Savings Calculations

 Even within engineering calculations, there ex-
 ists a range of methods and associated confidence
 in the estimates.14 Calculations performed with
 few verified assumptions will be less reliable than
 enhanced engineering estimates, .which rely
 partly on site-specific information, such as oper-
 ating hours, initial energy use, etc.  Thus, engi-
 neering calculations drawing upon site-specific
 information will receive greater credit than those
 relying on national averages or "typical" values
 derived from handbooks or other references.
    A simple engineering calculation that does
 not draw  upon site-specific information to ad-
 just assumptions will receive only a 50% realiza-
 tion rate of predicted savings. • The formula is
 given below:

  Verified savings = (0.5) x (Predicted Savings)

    The inclusion of site-specific adjustments
 will improve the confidence in the enhanced en-
 gineering estimate or building simulation. It is
 impossible to predict which site-specific adjust-
 ments are available and will be used. Therefore,
" Architectural Energy Corporation and RCG/Hagler, Bailly, Inc., Engineering Methods for Estimating the Impacts of Demand-Side
  Management Programs, 3 vols., EPRITR-100984, Electric Power Research Institute, Palo Alto, CA, 1992-1994; American Society
  of Heating, Refrigeration, and Air-Conditioning Engineers, "Energy Estimating Methods," in 1993 Fundamentals, ASHRAE,
  Atlanta, GA, 1993.
a M Pels and K. M. Keating, "Measurement of Energy Savings from Demand-Side Programs in US Electric Utilities," Annual
  Renew of Energy and the Environment, Vol. 18,1993, p. 57; S. M. Nadel and K. M. Keating, "Engineering Estimates vs. Impact
  Evaluation Results: How Do They Compare and Why?" in N. E. Collins, ed., Energy Program Evaluation: Uses, Methods, and
  Results, Argonne National Laboratory, 1991, pp. 24-33.   • .
14 M. Pels and K. M. Keating, "Measurement of Energy Savings from Demand-Side Programs in US Electric Utilities," Annual
  Review of Energy and the Environment, Vol. 18,1993, p. 57; RCG/Hagler, Bailly, Inc., Impact Evaluation of Demand-Side Manage-
  ment Programs, EPRI CU-7179, Electric Power Research Institute, Palo Alto, CA, February 1991.
                                                                                            33

-------
    credits will be available for different adjustments.
    Combinations of these adjustments can be used
    to obtain greater .proportions of the predicted
    savings (up to a 100% realization rate). The for-
    mula is given below:

    Verified savings =
                    (0.5 + Credit #2 + Credit #2 +
                    Credit #3) x (Predicted Savings)

        Up to four adjustments may be selected as
    long as the sum of the credits does not exceed
    0.5. The amount of credit for each adjustment is
    given in the. table below:
        Adjustment
Credit
        Verification of installation.

        Calibration of savings based on
        actual site's utility bills.

        Calibration of savings based on
        another program's results.

        Calibration of savings based on
        site-specific operating schedules.

        Short term measurements.

        One-year later verification of
        installation.
 0.15


 025


 0.10


 0.15

 0.15


 0.10
        For example, an engineering estimate pre-
    dicted 400,000 kWh savings in the first year of a
    commercial lighting program. This prediction
    included adjustments for verified installation of
    the fixtures and lights and a further downward
             adjustment to reflect the lower savings observed
             in another, similar program. The verified sav-
             ings calculation would then include 0.15 for
             verification of installation and 0.1 for calibration
             based on another program's results. Thus,
               Verified
               Savings
                            (0.75)  x  400,000 kWhlyear

                            300,000 kWhlyear
Verification Requirements
                                  /
The following requirements must be satisfied in
order for the engineering estimates to qualify as
verified:'
• Each submittal must include brief documen-
  tation explaining why the verification is eligible
  for engineering estimates of savings.
• Each submittal must include a brief descrip-
  tion of the technologies employed in the con-
  servation program and the assumptions and
  algorithms or procedures used to estimate en-
  ergy savings.
• Each submittal must indicate the credits se-
  lected.
• If the third credit is used, an SAS calculation
  must be made to normalize for any differences
  in energy services between buildings in the ap-
  plicant utility's program and the reference pro-
  gram (see Appendix B).
34

-------
     Reduced Transmission  and Distribution Losses
      Due to Demand-Side Efficiency Improvements
Technology Description
                                  v   *
Electric Utilities in developed countries typically
experience 2 - 8% annual losses in transmission
and distribution, although peak losses can be sig-
nificantly above that.15 The exact percentage de-
pends on the network design and other transient
factors, such as ambient temperature and system
loading.16 In addition, if the utility has more ef-
ficient transformers on all or part of its system,
the avoided transmission and distribution losses
could be as much as 25% lower.
   Reduced energy demand at the end use leads
to a reduction in transmission losses. These sav-
ings will not normally be captured in the verifi-
cation schemes required by the CVPs, even
though some savings are to be expected.  This
measure gives credit to utilities for the savings.
Energy Savings Calculations

The savings are proportional to the total savings
due to the verified conservation programs (with
the exception of amorphous core transformers).17
Transmission and
Distribution Savings
for Residential and
Commercial Programs   =0.07 x

Transmission and
Distribution Savings
for Industrial Programs = 0.035 x
Verification Requirements

The initial Verified Savings must not be greater
than the submittals for savings from the other
stipulated and monitored savings.  The utility
should also indicate whether its distribution
transformers are made of conventional silicon-
steel or high-efficiency amorphous-core metal.
15 N.-J. Bergsjo and L. Gertmar, "Superconductivity and the Efficient Use of Electricity," in T. Johansson, B. Bodlund, and
  R. Williams, eds., Electricity, Lund University Press, Lund, 1989; R. Loftness, Energy Handbook, Van Nostrand Reinhold, NY,
  1984.
16 Barakat & Chamberlin, Inc., End-Use Technical Assessment Guide (End-Use TAG) Volume 4: Fundamentals and Methods, EPRI
  CU-7222, Electric Power Research Institute, Palo Alto, CA, 1991; R. Dorf, ed.. The Electrical Engineering Handbook, CRC Press,
  Boca Raton, FL.1993.           ,                                    ,      '
17 Y. Bae, Portland General Electric Company, "1988 Line Loss Study/' Portland General Electric Co., Portland, OR, 1988.
                                                                                        35


-------
                          Energy-Efficient Refrigerators
     Technology Description
       t
     Refrigerators consume more electricity than any
     other end use in American homes. As a result of
     federal efficiency standards and utility incen-
     tives, new units consume less than half that of
     the 20-year old models now being retired. A va-
     riety of utility programs have been created to re-
     duce refrigerator energy use, including rebates
     to purchase especially efficient units and to pick
     up older, inefficient units. The rebates are typi-
     cally awarded for refrigerators that use 10 - 20%
     less than the current standard. These programs
     are often linked so that the utilities retrieve old
     refrigerators from those homes that received re-
     bates on the purchase of efficient models.
        The federal efficiency standards are causing
     significant savings to occur without any inter-
     vention by the utilities. For mis reason, the stipu-
     lated savings are meant to reward  savings be-
     yond those resulting from federal standards. This
     may take the form of replacing existing refrig-
     erators with new, efficient models and/or retriev-
     ing old units.  (Some utilities will just rebate new
     refrigerators, others will just pick up second
     units, and still others will do both.)

     Energy Savings Calculations

     Many field studies have demonstrated energy
     savings from installation of new, efficient refrig-
     erators and from removal of second units.18 The
     following table summarizes the stipulated gross
     savings from the different actions by the con-
     sumer and utility:                   '
   Consumer/
   Utility Action
Gross Savings Per
Unit (kWh/year)
    Pick-up alone (true abandoning
    of second unit).

    Pick-up of old unit and purchase
    of rebated replacement unit.

    Pick-up alone (where consumer
    recently purchased a
    non-rebated refrigerator).

    Efficient purchase alone.
        1200


         600



         450

         300
Note that these estimates are consistent with the
average new refrigerator consuming about 750
kWh/year.

Verification Requirements

The following requirements must be satisfied in
order for the stipulated savings to qualify as veri-
fied:
• The name and address of each incentive recipi-
  ent must be collected (but not submitted).
• The utility must submit a certified statement
  that the old refrigerators were operating.
• For second refrigerators, the utility must verify
  that a primary refrigerator exists and,that the
  consumer does not intend to purchase another
  refrigerator in the next year.
     " W. Bos, "SMUD's Graveyard—Conditions of the Deceased," Home Energy, January/February 1993, pp. 18-19; A. Meier, "Field
      Performance of Residential Refrigerators," ASHRAE Journal, August 1993, pp. 36-40; D. S. Parker and T. C Stedman, "Mea-
      sured Electricity Savings of Refrigerator Replacement: Case Study and Analysis," in American Council for an Energy Efficient
     . Economy, 1932 ACEEE Summer Study on Energy Efficiency in Buildings in Pacific Grave, CA, 1992, pp. 199-211; ]. Proctor and
      T. Downey, "Pacific Gas and Electric Refrigerator Rebate Evaluation: 1992 Field Monitoring Report," Pacific Gas and Electric
      Company; San Francisco, CA, 1992; E. A. Rogers, "Evaluation of a Residential Appliance Rebate Program Using Billing Record
      Analysis," in N. E. Collins, edv Energy Program Evaluation: Uses, Methods, and Results, Argonne National Laboratory, 1989, pp.
      263-269; P. M. Witte and M. G. Kushler, "An Evaluation of the Michigan Second Refrigerator Bounty Pilot Demonstration/'
      Public Service Commission of the Michigan Department of Commerce, Lansing, MI, 1990.
36


-------
    Residential Water Heating Conservation Measures
Technology Description

This measure involves reducing the energy con-
sumed by residential water heaters by address-
ing standby losses, hot water consumption, and
heating efficiency. .The measures covered here
are: water heater blankets,  anti-convection
valves, pipe insulation, low-flow showerheads,
and heat pump water heaters.

Energy Savings Calculations

Energy savings from insulation blankets depend
primarily on the temperature of the hot water
being stored, the ambient air temperature, and
the amount of insulation already around the
tank.19 Actual savings for individual units will
vary somewhat, but average savings from retro-
fits are reliable and more consistent. Most of the
savings measurements were undertaken 5-10
years ago.20 Since then  tank insulation has in-
creased substantially, so savings from the addi-
tion of the blanket have fallen somewhat.21
   Laboratory measurements have demon-
strated savings from both anti-convection valves
and pipe insulation, especially when insulating
the pipes closest to the water heater (both hot
and cold).32 Stipulated savings are given in the
table on the following page.
    Low-flow showerheads reduce the flow in
residential showers. Lower consumption of hot
water translates into reduced electricity con-
sumed for water heating. Beginning in 1994, all
showerheads sold must be rated at less than 2.5
gal/min.23. Since many old, high-flow shower-
heads are still in place, however, savings can be
obtained by replacing existing showerheads with
new, low-flow units. Showerheads with even
lower flows—some less than 2 gal/min—are
available and often preferred by consumers.24 For
this conservation measure, it is assumed that the
utility makes no special attempt to target high-
flow units.
    Recent monitored savings demonstrated that
replacement of the old showerheads with a low-
flow device by utility contractors saved about 500
kWh/year per household.25 This is somewhat
less than engineering estimates but reflects the
presence  of some low-flow showerheads and
lower flows than expected for older units. In ad-
dition, occupants had slightly shorter showers
than typically assumed.
    Water-faucet aerators have been recently
found by Puget Sound Power & Light through
19 Architectural Energy Corporation and RCG/Hagler, Bailly, Inc., Engineering Methods for Estimating the Impacts of Demand-Side
  Management Programs, 3 vols., EPRI TR-100984, Electric Power Research Institute, Palo Alto, CA, 1992-1994.
20 M. A. Brown, D. L. White, and S. L. Purucker, "Impact of the Hood River Conservation Project on Electricity Use for Residential
  Water Heating," Oak Ridge National Laboratory, Oak Ridge, TN, 1987; D. H. Sumi and B. Coates, "Persistence of Energy  .
  Savings in Seattle City Light's Residential Weatherization Program," in N. E. Collins, ed., Energy Program Evaluation Conference:
  Uses, Methods, and Results, 1989, pp. 311-316; A. Usibelli, "Monitored Energy Use of Residential Water Heaters," in Proceedings
  oftheACEEE 1984 Summer Study on Energy Efficiency in Buildings, American Council for an Energy-Efficient Economy, Santa
  Cruz, CA, 1984.,                      -
21 Architectural Energy Corporation and RCG/Hagler, Bailly, Inc., Engineering Methods for Estimating the Impacts of Demand-Side
  Management Programs, EPRI TR-100984, Electric Power Research Institute, Palo Alto, CA, 1992-1994; R. G. Pratt, B. A. Ross, and
  W. F. Sandusky, "Analysis of Water Heater Standby Energy Consumption from ELCAP Homes," Energy end Buildings, Vol. 19,
  No. 3,1993, pp. 221-234.                                                                        '
22 A. Usibelli, "Monitored Energy Use of Residential Water Heaters," in Proceedings oftheACEEE 1984 Summer Study on Energy
  Efficiency in Buildings, American Council for an Energy-Efficient Economy, Santa Cruz, CA, 1984.
23 US. Department of Energy, "National Appliance Energy Standards," Washington, D.C., 1990.                 ,
24 B. Mandark, "Low-Flow Showers Save Water—Who Cares?" Home Energy, July /August 1991, pp. 27-29.
KC. Hickman and W. M. Warwick, "Cleaning Up: An Efficient Approach for Estimating Showerhead Savings," in Proceedings of
  the ACEEE1994 Summer Study on Energy Efficiency in Buildings, American Council for an Energy-Efficient Economy, Pacific
  Grove, CA, 1994, pp. 8.89-8.95; M. Warwick and C. Hickman, "Everything 1 know about Energy-Efficient Showerheads I
  Learned in the Field," Home Energy, Jan/Feb 1994,3943.
                                                                                                37

-------
     metering studies to provide reliable energy sav-
     ings. The aerators were found to generate the
     greatest savings in kitchen faucets, with some-
     what lower savings occurring in bathrooms.
     Accounting for the rather high first-year removal
     rate for these measures (25%), the studies con-
     cluded that on average each aerator would save
     25 kWh/year, or 58 kWh/year for a single-fam-
     ily home and 68 kWh/year for a multi-family
     home.26
        This conservation measure consists of plac-
     ing aerators in the water faucets in the kitchen
     and/or the bathroom(s).
        Heat pump water heaters have been avail-
     able for about a decade but have not been popu-
     lar owing to their high initial cost.  Recent mod-
     els are much cheaper and just as efficient, how-
     ever. Recent studies have found that a heat pump
     water heater cuts heating energy use by about
     50%.      .    ,         '         .     -
        Stipulated savings for the residential water
     heating conservation measures are given in the
     table below.

       Measure                Stipulated Electricity
     	Savings (kWh/year)
       Insulation blanket around tank.            300

       Anti-convection values.                 100

       Pipe insulation.                        150

       Low-flow showertieads (savings per home.
       for 1 or 2 per home):              -      .

          Installed by utility representative.        500
          Installed by customer.                250

       Water-faucet aerators (savings per home,
      . for kitchens and bathrooms):               50

       Heat pump water heater.                1500
Verification Requirements

The following requirements must be satisfied in
order for the stipulated savings to qualify as veri-
fied:    ,
• Each installation must be inspected by a util-
  ity representative, except for customer-in-
  stalled low-flow showerheads.  .
• The name and address of each participating
  customer and date of inspection must be re-
  corded (but not submitted).
• Insulation blankets must have a thermal resis-
  tance of at least R-7.
• Anti-convection valves must be installed on
  both the incoming cold water and outgoing hot
  water pipes.
• Insulation must cover the first three feet from
  the water heater for both hot and cold pipes.
• These measures apply only to homes with elec-
  tric water heaters.
    26 M. Schuldt and R- Mazzucchi, "Energy Efficient Showerheads and Faucet Aerator Metering Studies: Single and Multifamily
      Residences/' SBW Consulting, Inc., Presented to the Northwest Regional Evaluation Group, Portland, OR, November 10,1994
38

-------
            Ground Source Heat Pumps For Homes
                              (New and Retrofit)
Technology Description

Ground source, or "geothennal," heat pumps use
the earth (or solar energy absorbed by the
ground) as a heat source and sink. The unit uses
a buried  network of pipes to transfer the heat
between the earth and a heat transfer fluid (usu-
ally water). The ground source heat pump can
operate more efficiently than a traditional air-
source heat pump because the heat reservoir is
wanner in the winter and cooler in the summer.
However, ground source units are more costly
to install and some open-loop systems use a sig-
nificant amount of energy to pump the working
fluid.  On the other hand, life-cycle  costs will
usually be lower for most systems.
    Ground source heat pumps are increasingly
popular due to improvements in their installa-
tion, efficiency, and reliability. These units, are
being installed primarily in rural areas, where
installation of a pipe network is particularly
easy.27 Improved drilling techniques, however,
are making this technology available even on
small lots.
Energy Savings Calculations

The typical alternatives to ground source heat
pumps are electric baseboard heating, air-source
heat pumps, or possibly Liquid Petroleum Gas
(LPG). The exact system also depends on the cli-
mate, that is, the relative size of the heating and
cooling requirements. Thus, energy savings must
be compared to these alternatives.  Many utili-
ties and government agencies have monitored
the performance of ground source heat pumps
in a wide range of climates and conditions.2*
   The annual stipulated savings for each
ground source heat pump installation is
2.0 kWh/square foot of gross floor area of the
served building.29
                       •

Verification Requirements

The address and floor area of each house built
with a ground source heat pump .must be re-
corded (but not submitted). The utility must in-
dicate the system design that would have been
used if ground source heat pumps had not been
selected.
v Architectural Energy Corporation and RCG/Hagler, Bailly, Inc., Engineering Methods far Estimating the Impacts of Demand-Side
  Management Programs, EPRITR-100984, Electric Power Research Institute, Palo Alto, CA, 1993; National Rural Electric
  Cooperative Association, The Heat Pump Manual, EM-4110-SR, Electric Power Research Institute, Palo Alto, CA, 1986.
* R. Freifelder, "Ground-Source Heat Pumps: Earth as Heat Source and Heat Sink," Home Energy, November/December 1990,
  pp. 32-38; F. Lenarduzzi, "Monitored Results from Commercial and Residential Heat Pump Projects," in S. Sami, ed., Heat
  Pumps in Cold Climates in Moncton, NB (Canada), Caneta Research, Inc., 1993.
* M. LTicuyer, C. Zoi and J. S. Hoffman; Space Conditioning: The Next Frontier, US. Environmental Protection Agency, EPA430-R-
  93-004, Washington, D.C., April 1993.     ,                                             .          .
                                                                                           39

-------
              Higher Efficiency Lights in  Office  Buildings
     Technology Description

     This measure involves replacing lights in office
     buildings with higher efficiency units. It applies
     only to office buildings and usually consists of
     replacing incandescent or old fluorescent fixtures
     with high-efficiency fluorescent lamps, improved
     fixtures, and electronic ballasts. A retrofit con-
     sists of identical lights replaced on a single cir-
     cuit. A lighting circuit is the electrical service to
     a group  of lights serviced by its own  circuit
     breaker or panel. In large buildings, lighting cir-
     cuits are often 277 Volts; in smaller buildings 120
     Volts is more common. In more complex build-
     ings, the lighting circuit is the combination of cir-
     cuits servicing the lights for a distinct area, such
     as one or more floors. No credit is given in this
     measure for improved switches, occupancy sen-
     sors, or daylight controls (although qualified sav-
     ings from these measures can be earned when
     verified through monitoring). Acreditfor air con-
     ditioning savings is available for buildings with
     central crullers (instead of window or rooftop
     package units).30

     Energy Savings Calculations

     Energy savings from commercial lighting retro-
     fits will depend on the number  of operating
     hours and the energy intensities of the original
 and new lighting'systems.31  In addition, light-
 ing levels are often adjusted during the retrofit.32
 The exact values for these variables are difficult
 to determine without monitoring.  Monitored
 savings have varied depending on the conditions
 outlined above.33 The stipulated savings listed
 below put caps on the variables that are most
 likely to be  overestimated.  As a result, use of
 these values will in some cases lead to underes-
 timates of actual energy savings.
    The energy savings are determined by cal-
 culating the difference in power consumption be-
 tween the old and new fixtures and multiplying
 by the number of hours of operation and num-
•ber of fixtures replaced.   Annual savings (in
 kWh/year) for a single retrofit are calculated as
 follows:

   Annual Energy
   Savings         = H x L x (PM - Paew) x A
   where:
     Annual Energy Savings = stipulated
     energy  sayings (in kWh/year) for the
     measure.
     H = number of operating hours in the
     year following the retrofit.
     L = number of identical fixtures in
     building  on same circuit operating
     prior to the retrofit.
     30 R. A. Rundquist, K. F. Johnson, and D. J. Aumann, "Calculating Lighting and HVAC Interactions/' ASHRAE Journal Vol. 35,
      No. 11,1993, pp. 28-37.
     31}. Eto, £. Vine, L Shown, R. Sonnenblick, and C. Payne, "The Cost and Performance of Utility Commercial Lighting Programs,"
      LBL-34%7, Lawrence Berkeley Laboratory, Berkeley, CA, December 1993; A. Lovins and R. Sardinsky, The State of the Art:
      Lighting (Contpetitek), Rocky Mountain Institute, Snowmass, CO, March 1988; S. M. Nadel and K. M. Keating, "Engineering
      Estimates vs. Impact Evaluation Results: How Do They Compare and Why?" in N. E. Collins, ed.. Energy Program Evaluation:
      Uses, Methods, and Results, Argonne National Laboratory, 1991, pp. 24-33; M. A. Piette, F. Krause, and R. Verderber, 'Technol-
      ogy Assessment: Energy-Efficient Commercial Lighting," LBL-27032, Lawrence Berkeley Laboratory, Berkeley, CA, 1989;
      Rochester Gas & Electric Company, "Basis for Compensation," Document Reference # ED.H.8 (RG&E DSM #1503), Rochester
      Gas & Electric Company, Rochester, NY, March 11,1991; Xenergy, "Demand-Side Management Program Evaluation Scoping
      Study," ESSERCO Project EP90-34, Empire State Electric Energy Research Corporation,  November 1990.
     92 Architectural Energy Corporation and RCG/Hagler, Bailly, Inc., Engineering Methods for Estimating the Impacts of Demand-Side
      Management Programs, 3 vols., EPR1TR-100984, Electric Power Research Institute, Palo Alto, CA, 1992-1994.
     33 B. A. Atkinson, et al., "Analysis of Federal Policy Options for Improving U.S. Lighting Energy Efficiency:  Commercial and
      Residential Buildings," LBL-31469, Lawrence Berkeley Laboratory, Berkeley, CA, 1992.
40

-------
     POM = power consumption of original
     light fixture (in KW) that was operat- -
     ing prior to the retrofit.
     P^ = power consumption of replace-
     ment light fixture {in kW).

     A=air conditioning savings credit.
The number of hours, H, must be less than or
equal to 3300.34 The air conditioning savings
credit, A, may not exceed 1.10. The difference
(PM - P^) for fluorescent lights may not exceed
the values given in the following table:
   Original Fluorescent
   Lamp Type
Maximum Allowable
.(•**,-p,J inkW
   2-lamp fixture

   3-lamp fixture

   4-lamp fixture
         0.025

         0.037

         0.050
Verification Requirements

The following requirements must be satisfied in
order for the stipulated savings to qualify as veri-
fied:
• The name and address of the building in which
  the retrofit occurred must be recorded (but not
  submitted).
• The number, type, and power rating of re-
  placed lights must be recorded for each build-
  ing (but not submitted).
• The number of non-functioning (removed or
  broken) lights must be recorded for each retro-
  fit (but not submitted). Retrofits replacing bro-
  ken or missing lights cannot be counted for
  credit with this stipulated measure.
* Rated power for the replacement must include
  the entire system, that is, the fluorescent light
  and ballast.
34 Pacific Gas and Electric Co., Commercial, Industrial and Agricultural Direct Rebate Programs Hours of Operation Study, PG&E,
  Measurement and Evaluation Planning Section,.San Fransisco, CA, August 199Z
                                                                                              41

-------
              De-Lamping  In  Commercial Buildings
Technology Description

This measure involves de-commissioning oper-
ating fluorescent lights and fixtures in commer-
cial buildings. De-lamping generally occurs
where removal- of lights does not significantly
affect safety or performance. De-lamping is an
effective means of reducing electricity used for
lighting. Ideally, the ballasts are removed or dis-
connected so as to insure mat the lights will not
be replaced during the next maintenance/light-
replacement cycle.
    Studies of actual savings from de-lamping
are rare because de-lamping is often undertaken
in conjunction with other lighting retrofits.35 Nev-
ertheless, energy savings are reasonably certain
as long as tabulations of de-activated units and
assumptions regarding operating hours are ac-
curate.36 Estimates of operating hours are par-
ticularly susceptible to overestimation.37

Energy Savings Calculations

The general procedure is to calculate the num-
ber of operating lamps that are removed and
tabulate their total power consumption. This
power consumption is multiplied by the num-
ber of operating hours. The formula is given
below:
  Annual Energy Savings = H x P(

  where:
xA
    Annual Energy Savings = stipulated
    energy savings (in kWh/year).for the
    measure.
                  H = number of operating hours in the
                  year following the retrofit.

                  ^removed = P°wer consumption of oper-
                  ating lamps (in kW) that were removed
                  or deactivated.

                  A * air conditioning savings credit.

             The number of hours, H, must be less than or
             equal to 3300. The  air conditioning savings
             credit, A, may not exceed 1.10.  The total reduc-
             tion in power demanded for any lamp cannot
             exceed the values given below:
                Lamp type
                          Maximum Credit (tcW)
                4 foot straight

                4(ootU-lamp

                8 fool straight
                                   0.030

                                   0.030

                                   0.050
Note that this measure does not apply to incan-
descent lights.

Verification Requirements

The following requirements must be satisfied in
order for the stipulated savings to qualify as veri-
fied:
• The name and address of the building in which
  the retrofit occurred must be recorded (but not
  submitted).
• The number, type, and power rating of re-
  placed lights must be recorded for each build-
  ing (but not submitted).
• The number of non-functioning (removed or
  broken) lights must be recorded for each retro-
  fit (but not submitted).
35 J. Eto, E Vine, L. Shown, R. Sonnenblick, and C. Payne, "The Cost and Performance of Utility Commercial Lighting Programs,"
  LBL-34%7, Lawrence Berkeley Laboratory, Berkeley, CA, December 1993; S. M Nadel and K. M. Keating, "Engineering
  Estimates vs. Impact Evaluation Results: How Do They Compare and Why?" in N. E. Collins, ed., Energy Program Evaluation:
  Uses, Methods, and Results, Argonne National Laboratory,1991, pp. 24-33.
* F. Bryant, "Office Building's Delamping, Lighting Retrofit Yields Payback in 11 Months," Energy User News, December 1993, p.
  10.                             , •               •
37 Architectural Energy Corporation and RCG/Hagler, Bailly, Inc., Engineering Methods far Estimating the Impacts of Demand-Side
  Management Programs, 3 vols., EPRITR-100984, Electric Power Research Institute, Palo Alto, CA, 1992-1994.

-------
                    Exit Sign Light Replacements
Technology Description

niurrunated exit signs are required by law in al-
most all non-residential and many multifamily
residential buildings. They are required to oper-
ate 24 hours/day in most situations.38 This mea-
sure involves replacing existing lights in exit
signs (which are mostly incandescent) with fluo-
rescent lights or light-emitting diodes.39

Energy Savings Calculations

The energy savings for each fixture replacement
are determined by calculating the difference in
power consumption between the old and new
fixtures and multiplying by the number of hours
of operation. Annual savings (in kWh/year) for
a single fixture retrofit are calculated as follows:

 Annual Energy Savings = 8760 x (PoW - PMW)

 where:
     Annual Energy Savings = stipulated
     energy savings (in kWh/year) for the
     measure.
    8760 = number of hours per year.

    Poid = power consumption of original
    exit light (in kW).
    P^ = power consumption of replace-
    ment exit light (in kW).
The power consumption of the original exit light
may be assumed to be 0.03 kW. A larger value
for P^ may be used if verified by inspection or
measurement of each model replaced.

Verification Requirements

The following requirements must be satisfied in
order for the stipulated savings to qualify as
verified:
• The number of replaced signs must be recorded
  (but not submitted) for each address.
• The original exit sign must be functioning and
  illuminated at the time of replacement. If only
  one lamp is illuminated, the default is 0.015 kW.
• Rated power for the replacement must include
  the entire system, that is, fluorescent light and
  ballast.
M International Conference of Building Officials, Uniform Fire Code, Whittier, CA, May 1991.
" A. Lovins and R Saidinsky, The State of the Art: Lighting (Competitek), Rocky Mountain Institute, Snowmass, CO, March 1988.


                                                                                        43

-------
                       Higher  Efficiency Street Lights
    Technology Description

    This measure involves replacing existing street
    lights with higher efficiency units.  .

    Energy Savings Calculations

    The energy savings are determined by calculat-
    ing the difference in power consumption between
    the old and new fixtures and multiplying by the
    number of hours- of operation. Annual savings
    (in kWh/year) for a single retrofit are calculated
    as follows:
      Annual Energy Savings = 4000 x (PM -
      where:

        Annual Energy Savings - stipulated
        energy savings (in kWh/year) for the
        measure.

        4000 = number of operating hours per
        year.         '

        ^oid ~ power consumption of original
        street light (in kW).

        Pnfw - power consumption of replace-
        ment street light (in kW).
Verification Requirements

The following requirements must be satisfied in
order for the stipulated savings to qualify as veri-
fied:
• The number, type, and power rating of re-
  placed street lights must be recorded (but not
  submitted) for each community.
• The street light must be functioning and in use
  prior to the retrofit; otherwise each such retro-
  fit cannot be counted for credit with this stipu-
  lated measure.            :
• Rated power for the replacement must include
  the entire system, that is, high-intensity dis-
  charge (HID) light and ballast, etc.
• The street light must be controlled by a func-
  tioning photocell.
44

-------
                     Higher Efficiency Motors For
                     Constant Load  Applications
Technology Description       r~

This measure consists of replacing or upgrading
inefficient motors being used to power a constant
load. The motor must be operating continuously
or controlled by a timedock. The retrofit will
.typically be undertaken in motors operating fans
or pumps in factories, buildings, and agricultural
sites*
    Note that this measure does hot apply to
variable speed motors.

Energy Savings Calculations

The energy savings calculations are determined
by calculating the difference in power consump-
tion between the old and new motors and multi-
plying by the number of annual operating hours.

Annual Energy Savings =
         (No. of operating hours) x (Pold - PB(W)
where:
    Annual Energy Savings = stipulated
    energy savings (in kWh/year) for the
    measure.
    No. of operating hours = number of op-
    erating hours per year, as controlled by
    timeclock. A maximum of 8500 hours
    is permitted.
  •  '^ow = P<>wer consumption of old mo-
    tor (inkW).
    P^ = power consumption of replace-
    ment motor (in kW).
True power measurements of P^ and Pm must
be performed during typical operation.

Verification Requirements

The following requirements must be satisfied in
order for the stipulated savings to qualify as veri-
fied:                      ,
• The name and address of the building in which
  the retrofit occurred must be recorded (but not
  submitted).
• The motor application (e-gy pumping, air han-
  dling, etc.) must be recorded (but hot submit-
  ted).
• The measured power of the original and re-
  placement (or upgraded) motors must be re-
  corded (but not submitted).
• The motor must be operating continuously or
  be controlled by a timeclock.
• If a timeclock is used, the timeclock must be
  operating and actually controlling the motor.
* Architectural Energy Corporation and RCG/Hagier, Bailly, Inc., Engineering Methods for Estimating the Impacts of Demand-Side
  Management Programs, 3 vols., EPR1TR-100984, Electric Power Research Institute, Palo Alto, CA, 1992-1994; S, M. Nadel, M.
  Shepard, S. Greenberg, G. Katz, and A. T. de Almeida, Energy-Efficient Motor Systems; A Handbook on Technology, Programs, and
  Policy Opportunities (Series on Energy Conservation and Energy Policy, Carl Bhunstein, ed.}, American Council for an Energy-
  Efficient Economy, Washington, DC, 1991; RCG/Hagler, Bailly, Inc., Impact Evaluation of Demand-Side Management Programs,
  EPRICU-7179, Electric Power Research Institute, Palo Alto, CA, February 1991.
                                                                                          45

-------
         Amorphous Metal Distribution Transformers
Technology Description

Amorphous metal-core transformers reduce no-
load losses by 70% or more over those in con-
ventional silicon-steel transformers.41  No-load
losses are the power required to energize the
transformer and are constant regardless of the
load. These are dominated by core losses. There
are also load  losses, which vary depending on
the load and are dominated .by winding losses.
For some transformers replaced energy savings
from decreased no-load losses may be partly off-
set by increased load losses, while in other cases
load losses will also decrease. On average, such
off sets will not be significant. No credit is given
for decreases  in load losses (although qualified
savings can be earned when verified through
monitoring).  Energy savings accruing from the
reduction of no-load losses through this measure
will occur 8760 hours each year.

Energy Savings Calculations

The energy savings are determined by multiply-
ing the 3/4 power of the replaced transformer's
rated capacity by a factor representing the de-
crease in no-load losses per unit of capacity (to
the 3/4 power) and multiplying by 8760 hours.
Annual energy savings (in kWh/year) for a single
transformer are calculated as follows:
    Annual Energy
       Savings
= C9-nx 3.1 x 10* x 8760
    where:
    Annual Energy Savings = stipulated
    energy savings (in kWh/year) for the
    measure.
    C = rated capacity of replaced trans-
    former (kVA).
                                 3.1 x 10* = decrease in no-load losses
                                 per unit capacity3'4 (kW/kVA3/4),
                                 8760 = number of operating hours per
                                 year.

                             Verification Requirements

                             The following requirements must be satisfied in
                             order for the stipulated savings to qualify as veri-
                             fied:                                 .   '
                             • Amorphous metal-core transformers must be
                               used instead of silicon steel or silicon iron trans-
                               formers:
                             • The rated capacity and the location must be
                               collected (but not be submitted).

                             To obtain greater man the stipulated savings,
                             optional testing of no-load and load losses (ANSI
                             C57.12.90-1987; IEEE C57.120 Draft 16-91) may
                             be done for the new and old transformers.
41 P. R. Barnes, J. W. Van Dyke, B. W. McConnell, S. M. Cohn and S. L' Purucker, The Feasibility of Replacing or Upgrading Utility
  Transformers During Routine Maintenance, ORML-6804, Oak Ridge National Laboratory, Oak Ridge, TN, October 1994.

-------
                           Stipulated  Lifetimes for
                  Common Conservation Measures
Technology Description

The effective lifetime of conservation measures
is crucial in determining the cumulative impact
and cost-effectiveness of conservation measures
and programs. There remains considerable un-
certainty regarding lifetimes, and research con-
tinues to improve estimates.42

Energy Savings Calculations

The stipulated median lifetimes are given in the
following two tables. They are adapted from the
consensus reached by the California utilities.
regulatory commissions and other interested par-
ties but are based on work by many groups across
the country.43  These lifetimes may be used for
calculations of savings as described in the CVPs.

Verification Requirements

   The following requirement must be satisfied
in order for the stipulated lifetimes to qualify as
verified:
• The conservation measure  in the submittal
  must closely resemble mat described in the list.
42 L. A. Skumatz, et al. "Bonneville Measure Life Study: Effect of Commercial Building Changes on Energy Using Equipment,"
  SRC- No. 7619-R2, Synergic Resources Corporation, Seattle, WA, December 1991; Synergic Resources Corporation, "Effective
  Measure Life and Other Persistence Issues in DSM Programs," Interim Report #1.2, R77291R1, SRC, Oakland, CA, February
  1992; Northeast Utilities, A Survey of Other Utilities for Measure-life and Persistence-of-Savings Issues, Conservation and Load
  Management Department, Northeast Utilities, Hartford, CT, March 1992.
u Measurement Subcommittee, Measurement Protocols far DSM Programs Eligible for Shareholder Incentives (An Energy Efficiency
  Blueprint for California: Report of the Statewide Collaborative Process), State of California, January 1990, Appendix A; California
  Public Utilities Commission, Protocols and Procedures for the Verification of Costs, Benefits, and Shareholder Earnings from Demand-
  Side Management Programs, San Fransisco, CA, July. 1994; M. Gardner, Memorandum on Showerhead Lifetime, Northwest
  Power Planning Council, Portland, OR, November 1994.
                                                                                                47

-------
                           Useful Lives of Residential
                       Energy Conservation Measures
      Measure
Lifetime (Years)
Measure
Lifetime (Years)
      Caulking
      WeathefStripping
      Ceiling.insulation
      Wall insulation
      Low-flow showerhaads
      Water faucet aerators
      Duct wrap/insulation
      Pipe Insulation
      Water heater blanket
      Fluorescent bulbs
      Window shade awnings
      High-efficiency A/C
      Central heat pump
     10
     10
     25
     25
      8
      5
     15
     22
     10
     10
     10
     16
     18.
Evaporative coolers
Clock thermostat
High-efficiency clothes dryer
High-efficiency refrigerator
High-efficiency central furnace
Whole house fan
Double glazing
Storm windows
Furnace repair
Efficient electric water heater
                -X
Heat pump water heater
Point-of-use water heater .
Solar water heater
     15
     15
     18
     20
     20
     15
     25
     10
     15
     15
     13
     12
     1S
48

-------
         Useful Lives of Commercial and Industrial
                   Energy Conservation Measures
Equipment Type & Descriptor    Lifetime (Years)
LIGHTING
Energy-efficient fluorescent lamp          5
Same as above with built-in ballast         2
.Energy-efficient ballast                 11
Electronic ballast                      10
Metal hallde lamp        -            15
Low-pressure sodium lamp               5
High-pressure sodium lamp               5
Parabolic fixture            ,          20
Dimming systems                     20
On-off switching                       7
Motion sensor  •               ,       10
HVAC
Economizer                          12
Chiller strainer cycle system             15
Air-to-air packaged heat pump            10
Water-to-air packaged heat pump         15
Ice thermal energy storage              19
Water thermal energy storage            20
Plate type heat pipe recovery system    .   14
Rotary type heat recovery system         11
Heat recovery from refrigerator condenser  11
Low leakage damper  .                  9
Variable air volume, inlet vane controls '    11
Variable pitch fan for cooling power        13
Make-up air unit for exhaust hood         10
Air de-stratification fan—paddle type       10
Air de-stratification fan—high inlet/
   low discharge                      15
Air curtain                           10
Deadband thermostat       •          13
Spot radiant heat                      10
Equipment Type & Descriptor   Lifetime (Years)
CONTROLS
Computer logic energy-management
    systems      .               "13
Electronic controls                  11
Time clocks                       10
MOTORS, DRIVES,
   &TRANSFORMERS
Standard electric motor              15
High-efficiency electric motor         15
Variable-speed DC motor            18
Variable-speed drive—solid type       15
Variable-speed drive—belt type        10
Efficient AC electric transformer       30
REFRIGERATION
Unequal parallel refrigeration  .        14
Condenser float head
    pressure control                10
Auto cleaning system for
    condenser tubes               15
Hot gas bypass defrost              10
Polyethylene strip curtain ...           3
Refrigeration case cover             11
BUILDING ENVELOPE
Double glazing  .                 20
Heat mirror     .                   18
Low-emissivtty coating  ,            14
Solar shade film (retrofit)             7
Tinted & reflective coating            14
                                                                                                49

-------
                                                                      i.-nc : £85
                 Stipulated Net-to-Gross  Factors for
                "     "       Stipulated Measures
 Description
Conservation
  Measure
  Net-to-Gross
Conversion Factor
Conversion factors in*the following.list can be
used to convert the gross annual savings calcu-
lated in the stipulated measures to net, first-year
savings. These values should be used when utili-
ties cannot provide estimates based on their own
programs or market research. Factors such as free
riders that affect the net impacts of conservation
programs vary widely by the type of conserva-
tion program and market conditions.44 Utilities
may conduct their own analyses and conclude
that their conservation programs have larger net
energy savings, rather than rely on the default
factors given below.  Several recent studies have
found, for example, free-rider rates of 10% or less.

    Net savings are calculated with the follow-
ing formula:

Net Savings = Annual Energy Savings x NGCF
when:
     NGCF = "Net-to-Gross-Conversion
     Factor" listed in the table opposite:
Engineering estimates of
energy savings -   -.      -            0.70

.Exit sign light replacements              0.60

Higher efficiency motors                 0.60
                   ../
Higher efficiency lights in office buildings    0.60

De-lamping                           0.80

Higher efficiency street lights             0.90

Refrigerators
  pick-up                            0.70
  high-efficiency replacement             0.90

Residential water heating
  Insulation blankets                   0.60
  Anti-convection valves                 0.90
  Pipe insulation                      0.60
   Low-flow showerheads and faucet
    aerators (utility-installed)             0.70
   Low-flow showerheads and faucet
    aerators (customer-installed)          0.50 *
  Heat pump water heaters              0.95

Ground source heat pumps               0.95
44 W. Saxonis, "Free Riders and Other Factors that Affect Net Program Impacts/' in E. Hirst and ]. Reed, eds., Handbook of
  Evaluation of Utility DSM Programs, ORNL/CON-336, Oak Ridge National Laboratory, Oak Ridge, TN, December 1991, pp.
  119-34; B. M. Tolkin and P. R. Rathbum, "Quantifying Free-Ridership in Four Different Customer Segments," Proceedings of the
  ACEEE1992 Summer Study on Energy Efficiency in Buildings, American Council for an Energy-Efficient Economy, Pacific Grove,
  CA, 1992, pp. 7.243-7.249; M. A. Brown and P E. Mihlmester, "Summary of California DSM Impact Evaluation Studies,"
  ORNL/CON-403, Oak Ridge National laboratory. Oak Ridge, TN, October 1994.

-------
51

-------