United States
           Environmental Protection
           Agency
Air and
Radiation
(6204-J)
EPA430/8/B-92-002
March 1993
&EPA  Conservation
           Verification Protocols
           A Guidance Document for Electric Utilities
           Affected by the Acid Rain Program of the
           Clean Air Amendments of 1990
              UIIIIIII
    ' Recyclea/Hecycia«™
  7"T Xj Printed on paper that contains
  X_lC/ at least 50% recycled fiber
            ACID
                                       P R  O G R >

-------
If you are interested in obtaining more information on the Acid Rain Program, or in receiving a
    copy of the Conservation Verification Protocols Reporting Form and Instructions, call:

                                ACID RAIN HOTLINE
                                     617-674-7377
             Monday through Friday, 9:00 a.m. to 5 p.m., Eastern Standard Time

                                      or write to:

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

-------
           United States
           Environmental Protection
           Agency
Air and
Radiation
(6204-J)
EPA 430/8/B-92-002
March 1993
v>EPA   Conservation
           Verification Protocols
           A Guidance Document for Electric Utilities
           Affected by the Acid Rain Program of the
           Clean Air Amendments of 1990
                                     ACID-RAIN
                                     PROGRAM

-------
CONSERVATION VERIFICATION PROTOCOLS:

           A Guidance Document for Electric Utilities
           Affected by the Acid Rain Program of the
             Clean Air Act Amendments of 1990
                    vvEPA
                   ACID RAIN DIVISION
            U.S. ENVIRONMENTAL PROTECTION AGENCY
                     JANUARY 1993

-------
ACKNOWLEDGMENTS

The Conservation Verification Protocols (CVP) were developed and written as a team effort. Barry
Solomon was the Project Officer, under the direction of Joe Kruger, Renee Rico and Brian McLean at
the Acid Rain Division of EPA. Alan Meier was the Project Leader at Lawrence Berkeley Laboratory
(LBL), which assisted EPA in the development of the CVP. Also instrumental was the role of the
Conservation Verification Subcommittee of EPA's Acid Rain Advisory Committee, ably chaired by
Stan Hulett, which provided valuable guidance, advice, and direction for the content of the CVP. The
current Subcommittee membership is listed in Appendix E of this report. Membership of the
Subcommittee has included Tom Buckley, Cary Bullock, John Fox, Larry Frimerman, 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 CVP include Joe Eto, Chuck Goldman, Jeff Harris,
Barbara Litt, and Ed Vine of LBL, and Dan Blank, Marilyn Brown, Paul Centolella, Pat Curran, Bill
Gavelis, Phil Hanser, John Hoffman, John Hughes, Phil Hummel, Ken Keating, Anne Gumerlock Lee,
Rick Morgan, Gil Peach, Judy Tracy, Lloyd Wright, Roger Wright, and Cathy Zoi.

-------
              UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                            WASHINGTON, D.C.  20460
                                                                   OFFICE OF
                                                                 AIR AND RADIATION
Dear Reader:
    One of the goals of the Acid Rain Program is the promotion and use of energy-
efficient strategies for utility compliance.  Because the Acid Rain Program employs
a market-based system of flexible compliance, this program allows utilities to choose
least-cost strategies, including the use of energy conservation programs.  In addition
to the overall flexibility, Congress provided  EPA with two specific provisions that
provide  additional incentives for conservation,  the Conservation and  Renewable
Energy Reserve, and the Reduced Utilization planning requirements.

    One important step in promoting the use of energy conservation programs is the
development of rigorous yet flexible quantification methods for verifying that utility
conservation programs are reducing energy consumption.  Good verification will lead
to increased confidence  that  these programs produce reliable energy savings, and
increase knowledge about what kinds of programs are the most successful.

    Our Protocols, which are  designed for application to the Acid Rain  Program as
well as for review and adoption by rate-making authorities, have three goals:

    •   Conservation verification protocols should  be strongly oriented  toward
         measurement of energy savings, rather than engineering estimates.

    •   Conservation verification protocols must be flexible since energy conservation
         is a diverse activity.

    •   Verification of energy savings  should be cost effective and should require a
         level of data and analysis appropriate for specific measures and programs.

    Since the practice of verification of energy conservation savings based on program
evaluations is rapidly evolving, EPA expects to revise these protocols, perhaps several
times over the corning years. We look  forward to working with all interested parties
to increase the success of energy  efficiency programs and acid  rain control.
Eileen B. Claussen
Director, Office of Atmospheric Programs, EPA
                                                                    Printed on Recycled Paper

-------
                  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
   Forms and Data Requirements for Users	12

Section 2. Verifying Energy Savings	13
   Introduction	13
   Path I: Stipulated Savings	13
       Stipulated Savings for Specific Measures	13
       Good-Practice Engineering Estimates	14
   Path II: Monitored Energy Use	14
       Reference Case	15
       Service Adjusted Energy Use and Savings	15
       Gross and Net Savings Distinctions	16
       Confidence in Estimated Savings: Satisfying a 75% Hypothesis Test	16
       Acceptable Measurement Procedures	17
       Duration of Metering	18
   Subsequent-Year Savings	18
       Option 1: Monitoring Subsequent-Year Savings	19
       Option 2: Inspection for Presence and Operation of Subsequent-Year Savings	19
       Option 3: Default Subsequent-Year Savings	20
   Supply-Side Efficiency Improvements	20

Appendix A
   Definitions	21

Appendix B
   Service-Adjusted Savings and Hypothesis Test	23

Appendix C
   T-Statistics for 75% Confidence One-Tailed Hypothesis Testing	27

Appendix D
   Stipulated Savings, Procedures,  Lifetimes, and Conversion Factors	29

Appendix E
   U.S. Environmental Protection Agency Acid Rain Advisory Committee
   Subcommittee on Conservation Verification	43

List of Tables and Figures
   Table 1. Qualifying Criteria and Applicability for Electric Utilities Using Energy Efficiency
       Incentives Under Title IV of the CAA	10
   Figure 1. Overview of Verification Options	2
   Figure 2. Relationship Between EPA's CVP and Acid Rain Program Allowances	6
   Figures. Stipulated Savings Path	14
   Figure 4. Steps to Calculate Service-Adjusted Savings	15
   Figure 5. Subsequent-Year Verification Options	18

-------
                    Executive  Summary
The Conservation Verification Protocols (CVP)
have been prepared by EPA as part of its mission
to implement the  Acid Rain Program of the
Clean 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.
Beginning in the year 2000, most electric utilities
that burn 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
investor-owned utilities that qualify for these
incentives will probably have their energy sav-
ings verified by procedures specified by their
state Public Utility Commission (PUC). The pri-
mary  users of the CVP under the Acid Rain
Program are expected to  be public-power
utilities.
   The CVP is designed to be rigorous without
being burdensome on the utility or the regula-
tor. The CVP has the added benefit of helping to
ensure the cost-effectiveness of utility conserva-
tion programs and SO2 emission reduction mea-
sures, as well as the reliability of energy savings
from the 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, which is available to utilities
that meet electric demands with either conser-
vation 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 rates for conservation measures
(to be certified by the U.S. Department of En-
ergy). In addition,  the utility can only receive
credit for demand-side  energy conservation
measures.
   The Reduced Utilization provision applies
to the 110 Phase I plants who 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 CVP allows for two general savings paths
(Figure 1): Monitored Energy Use, or Stipulated
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 attributable to
the utility conservation  program. The Stipu-
lated Savings Path includes procedures for esti-
mating savings, as well as simple equations and
standard values for estimating stipulated en-
ergy savings from a limited number of conserva-
tion measures for which expected energy savings
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 CVP also includes
guidelines for verifying the persistence of en-
ergy savings from conservation measures.

Path I: Monitored Energy Use

The preferred verification approach is to infer
energy savings through  the measurement of

-------
      energy 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
        confidence 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
Figure 1. Overview of Verification Options
         Utility Conservation Program
                          Monitored Energy Use
                        Service-Adjusted Savings I
                            Gross
                           Savings
                             Net Savings
                   First-Year Savings
                Subsequent-Year Savings
adjustments to account for differences in meter-
ing periods, weather, indoor temperature, and
levels of service. The CVP requires these adjust-
ments for both the reference case and the cus-
tomers participating in the conservation
program.
   The CVP requires estimation of net energy
savings, i.e., 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
similar 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
normally be considered the reference case. If this
approach is not followed, the CVP allows 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-par-
ticipants.
   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 CVP is to award allowances for savings that
occurred with reasonable certainty. The CVP
requires that the savings be expressed in terms
of the utility's confidence, based on statistical
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 evalu-
ation, a 75% confidence (or 1-tailed hypothesis
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 captured 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 verify-
ing 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 for the first
year. It must measure energy use in the 1st and
3rd subsequent years. After year 3, 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 utility
can receive credit for at least 75% of the energy
savings in subsequent years for up to half of the
measure's lifetime. In the case of "passive con-
servation measures'1 (e.g., wall insulation),  a
utility may receive credit for 90% of the savings
for the full lifetime of the measure.
   The  default option provides for 50% of the
Ist-year savings to continue in subsequent years
if the utility discontinues any form of monitor-
ing or inspection,  for half the lifetime of the
energy conservation measure.

Path II: Stipulated Savings

While the CVP is oriented toward measurement
of energy savings, in some cases utilities may
use simple algorithms that have been provided
for a limited number of technologies. 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 factor, or by docu-
menting an actual "gross-to-net" ratio. Subse-
quent year savings for the this path are esti-
mated by the same procedure used in the
Monitored path (see previous discussion).
   The rationale for the Stipulated Savings ap-
proach is that the performance of some mea-
sures is well understood and may not be cost
effective to monitor. Use of stipulated savings in
these limited circumstances should prove very
attractive to small utilities and ones with new
conservation programs.  The initial list in the
CVP contains 7 measures  for which there are
stipulated savings: constant-load motors, exit
sign lights, amorphous metal transformers, com-
mercial lighting, new refrigerators, street lights,
and water heater insulation blankets.
   In circumstances in which  extensive mea-
surement and analysis are  not cost effective or
feasible, an engineering estimate of energy sav-
ings will be acceptable. These circumstances
include any of the following situations:
• Measurement costs would exceed  10% of
  program cost.
• Program-wide energy savings are expected to
  be 5000 MWh/year or less and 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
                   Behind the Approach
The U.S. Environmental Protection Agency
(EPA) has prepared the Conservation Verifica-
tion Protocols (CVP) as part of its mission to
implement the Acid Rain Program authorized
by Title IV of the Clean Air Act Amendments of
1990. Two cornerstones of the Acid Rain Pro-
gram are SO2  emission allowance trading and
energy efficiency improvements as part of com-
pliance strategies of affected electric utilities.
These cornerstones are linked through two ex-
plicit incentives for energy conservation that are
described in more detail in the Acid Rain rules:
the Conservation and Renewable Energy Re-
serve (CRER)  discussed in the Allowance Sys-
tem Rule (40  CFR Part 73), and the Reduced
Utilization (RU) provision for Phase I affected
units discussed in the Permits Rule (40 CFR Part
72). These two incentive programs allow electric
utilities to earn or save allowances through quali-
fied conservation measures.
   Each emission  allowance gives its holder
(usually a power plant) the right to emit one ton
of SO2 during or after a specified year, after
which the allowance must be retired. Alterna-
tively, a holder of allowances can bank them for
future use or  sell them to another person at a
negotiated price.  Allowances issued by EPA
under the Acid Rain Program may have consid-
erable market value. Consequently, energy sav-
ings achieved  under these incentive provisions
need to be verified by either the electric rate
regulator or the EPA.
   The main goal of the CVP is to credit electric
utility conservation programs for energy sav-
ings that they  caused, 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 CVP is intended to describe good prac-
tice for impact evaluators of electric utility en-
ergy conservation programs. It is not intended
to be prescriptive. While the formal federal veri-
fication requirements will end when the CRER
and RU programs expire, users may want to
continue to follow the CVP or to  use them in
other applications such as state-level conserva-
tion programs. A good  verification program
should be rigorous without being burdensome
on the utility or the regulator. The CVP will have
the added benefit of helping to ensure the cost-
effectiveness of utility conservation programs
and SO2 emission reduction measures, as well as
the reliability  of actual energy savings from
these programs. Eventually, energy savings
should be as reliable for providing energy ser-
vices to utility  customers as energy production
measures are currently.
   The CVP has been written to be accessible to
all electric utilities: large and small, public and
private. In addition, it can be used by utilities
well versed in evaluation of energy conserva-
tion programs, as well as those who are new to
this field.
   The CVP described  herein is intended to
provide stronger motivation for monitoring en-
ergy savings from conservation measures and
programs. The CVP embodies 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
algorithms for  calculating energy savings from
specific measures) when that approach best fits
the circumstances of the utility. The stipulated
savings approach, while generally much less
expensive and  time consuming to implement, is
designed to be conservative and will often result
in less energy savings credit than will a conven-
tional verification procedure. More typically,
representative customer groups will be analyzed
by the utility to determine the energy savings.
   The CVP is  a guidance document for electric
utilities, which may be revised  periodically.
Moreover, as explained in Section I of this re-
port, many utility applicants to the CRER will
not have to use the CVP Consequently, use of
the CVP is not a requirement, though the reader
should note that under  EPA's Allowance Sys-
tem Rule applicants to the CRER must demon-

-------
        strate to the satisfaction of the EPA Administra-
        tor that in the case of qualified energy conserva-
        tion measures, energy savings have been actually
        secured. EPA will consider the CVP as a source
        of guidance for reasonable conservation verifi-
        cation procedures. The Rule allows, but does not
        require, applicants that do not qualify to have
        verification performed by their state public utility
        commission (the criteria for which are explained in
        Section I) to do so through use of the CVP.
           The CVP has been designed to create a single
        set of protocols with sufficient flexibility to ac-
        commodate Acid Rain Program provisions as
        well as future applications by other federal or
        state  agencies.  Thus,  the CVP is self-standing
        and independent of the individual regulatory
        programs.  This necessitates language linking
        the individual allowance provisions and  the
  Figure 2. Relationship Between EPA's CVP
     and Acid Rain Program Allowances
        EPA's Acid Rain Program
    Conservation
   and Renewable
   Energy Reserve
State Energy
Conservation
 Programs
          \
       Bridging Language
                  Bridging Language^ stae BrW8ln9
                              Unsua9e -*•  Other?
                EPA's Conservation
               Verification Protocols
                       Revise
  List
  of
Stipulated
 Energy
 Savings
                    Technical Review
                      Committee
CVP. Section I also specifies the administrative
details unique to each regulatory program, such
as the timing, reporting and auditing, etc. Figure
2 illustrates the relationship between the allow-
ance schemes and the CVP.
   Energy conservation is a diverse activity. No
general protocols for verifying energy conserva-
tion savings can anticipate every kind of conserva-
tion technology, program, or activity that will be
undertaken by utilities. Procedures for verifying
the energy savings must therefore be flexible
enough to accommodate verification of the com-
mon conservation measures as well as new effi-
ciency options. Forexample,utilityprogramshave
focused on residential and commercial buildings
while the Clean Air Act will also permit utility
conservation programs in the industrial sectors.
Evaluation and verification techniques developed
for the residential and commercial sectors will not
be directly transferable to the industrial sectors.
For these reasons, the CVP will give general guide-
lines forverifyingenergy savings rather thanspeci-
fication of the verification procedure for each kind
of measure.
   The CVP is strongly oriented toward mea-
sured energy savings rather than engineering
calculations. Metering and evaluation may con-
sist of utility bill analysis, periodic inspections of
retrofitted equipment and confirmation of their
continued operation. The cost of savings verifi-
cation will depend on the kind of program and
the verification procedures used by the utility. If
the anticipated extra cost of evaluation reverses
the cost-effectiveness of the conservation pro-
gram, then a less stringent evaluation will be
permitted. For purposes of discussion, it is gen-
erally expected that 5-10% of the program cost
will be devoted to savings verification. No spe-
cific measurement technology is required; a util-
ity can use whatever approach will achieve the
verification requirements at lowest  cost.  It is
anticipated, however, that most utility verifica-
tion programs will rely on a mixture of utility
billing data, simple submetering, as well as stipu-
lated savings (to be explained below). hi limited
cases engineering  estimates of energy savings
can be used by a utility.
   A rapidly-increasing body of literature quan-
tifying the field energy savings from various
conservation measures is  emerging. In some
        6

-------
cases, the savings are sufficiently reliable that a
comprehensive verification program may not
be needed. A preliminary list of these measures,
with stipulated energy savings or a stipulated
energy savings formula, has been developed by
EPA.  The stipulated savings may come from
several sources. Where laboratory measurements
have proven to be reliable indicators of  field
electricity use, extensive metering may not be
necessary. Finally, the stipulated savings will be
given for some measures when reasonably reli-
able savings have been achieved  over a wide
range of verifications. The savings will be dis-
counted, however, to  reflect uncertainty that
occurs when field verification is not undertaken,
savings that would have occurred without the
aid of the utilities, and to encourage utilities to
verify the savings from their own programs.
    While the list of measures with stipulated
savings is not comprehensive, it will reduce the
monitoring and verification burden and allow
utilities to focus on DSM programs where im-
pacts are less predictable. The provision for stipu-
lation also offers some energy savings credit to
conservation programs that were undertaken
without adequate verification plans. The list and
the stipulated savings will be subject to periodic,
professional reviews and revisions.
   This document  uses  considerable special-
ized vocabulary, some of which has not gained
general acceptance. The ultimate reference for
meanings of terms not explicitly defined in this
document is the recent report by Hirst and Sabo.1
A list of definitions of the most important verifi-
cation terms is presented in Appendix A.
   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 own protocols.
Additionally, 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 defines integrated resource planning
to require  the verification of energy savings
achieved through conservation. The Agency
strongly believes that the CVP needs to remain
flexible in order to allow utilities ample freedom to
evaluate their energy conservation savings in the
most cost-effective manner. Thus, EPA expects to
revise the CVP based on comments from the pub-
lic and other information, in 1993 and possibly
again in later years.
1E. Hirst and C. Sabo, Electric-Utility DSM Programs: Terminology and Reporting Formats (ORNL/CON-337), Oak Ridge National
  Laboratory, Oak Ridge, TN, October 1991.

-------
   Linkage with  the  Acid  Rain  Program
                                                                                      Section
Background
The purpose of this section is to provide back-
ground and context for the CVP by linking it
with EPA's Acid Rain Permit and Allowance
System rules that were promulgated in Decem-
ber 1992 (40 CFR §§ 72.43 and 72.91, and Part 73,
Subpart F, 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
savings. For a more complete discussion of these
issues, the reader is referred 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 environmentalbenefits through
reductions in emissions of SO2 and NOx, the pri-
mary causes of acid rain. To achieve this goal at the
lowest cost to society, the program employs both
traditional and innovative, market-based ap-
proaches for controlling air pollution. In addition,
the Program encourages energy efficiency and
pollution prevention.
   Title IV of the CAAA sets as its primary goal
the reduction  of annual SO2 emissions  by 10
million tons below the 1980 level. To achieve
these reductions, the law requires a two-phase
tightening of the restrictions placed on  fossil-
fuel-fired  power plants. Phase I  begins in 1995
and affects 110 mostly coal-burning electric util-
ity plants  located in 21 eastern and midwestern
states. Phase II, which begins in the year 2000,
tightens the annual emissions limits imposed on
these large, high-emitting plants, and sets re-
strictions on smaller, cleaner plants fired by coal,
oil, and natural gas. The program affects exist-
ing plants with an output capacity of 25 mega-
watts 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
continuous emissions monitoring  of SO2 will
provide automatic verification of emissions re-
duction (40 CFR Part 75).


Conservation and
Renewable Energy
Reserve (CRER)

The CRER is a special pool of 300,000 total allow-
ances taken from the Phase 13 allocations, which is
available to utilities that meet electric demands
with either conservation or renewable energy re-
sources. Congress established this Reserve to pro-
vide an early "jump start" to energy efficiency and
renewable energy strategies for reducing SO2 emis-
sions. There are several criteria that an  electric
utility must meet in order to qualify for a share of
these bonus allowances in the CRER, which are
issued on a first-come, first-served basis (see also
Table 1):
• A utility must own or operate one or more
  affected power plant units, or be at least the
  partial owner or operator of an affected unit
  (see 40 CFR § 73.82 (a), and EPA's proposed
  rule on Acid Rain Allowance Allocations and
  Reserves at 57 FR 29965-30024 for a list of these
  affected units and EPA's proposed yearly base
  allowance 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
  savings (post-hoc) that occur from January 1,
  1992 to December 31, 1994, or until  all the
  allowances  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 II affected unit (or units) may apply to
  the CRER for allowances for verified  energy

-------
 savings (post-hoc) that occur from January 1,
 1992 to December 31, 1999, or until all the
 allowances in  the Reserve  are allocated,
 whichever occurs first (40 CFR §§ 73.82 (g)(l),
 73.80 (b)).
• Autilitymusthavea"leastcostplan"orplanning
 process for meeting future electric needs, which
 may consider  social and environmental
 externality costs (40 CFR § 73.82 (a)(4)).
• 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.
 Department of  Energy will begin certifying
 the net income neutrality of State Public Utility
 Commissions' (PUCs') electric rate-making
 procedures on January 1,1993 (40 CFR § 73.82
 (g)(2).  Bonus allowances may be awarded
 conditionally pending DOE certification (40
 CFR § 73.84 (c)).
• Utilities applying  to the  CRER must use
Table 1.
Qualifying Criteria and Applicability for Electric Utilities
Using Energy Efficiency Incentives Under Title IV of the CAA
FEATURE
Who May Apply?
Utilities Must Have a
I fjast-Cost Plan and
Net Income
Neutrality?
r H& of Efficiency
Measures That
Oualify
F minion Rate at
//hir,h Allowances
f wned/Saved
Mow Many
Allowances
A /viable?
Timing of Program
CRER BONUS
ALLOWANCES
Owner/Operator of Any
Affected Unit
Yes
Demand-Side Measures
Only
004 Ibs/kwh
(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
Average Emission
Rate of
Phase II Units
During Phase I Year
No Limit
1995-1999
(Phase I)
  qualified demand-side energy conservation
  measures (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
  appear on the list but which 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
  efficiency; must not  increase the use of any
  other fuels (other than renewables, industrial
  waste heat or 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 the 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

                                              Congress recognized in establishing the Acid
                                              Rain Program that during Phase I utilities might
                                              circumvent SO2 emissions reductions by shift-
                                              ing load to units not regulated in Phase I. To
                                              avoid this  potential problem, the rules imple-
                                              menting the CAAA require that during Phase I,
                                              a unit whose utilization is reduced below its
                                              1985-87 baseline level  must surrender  allow-
                                              ances representing a shift of emissions  to un-
                                              regulated 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 valuable to some utilities, since it
                                              allows them to account for reduced load at Phase
                                              I units (and avoid surrendering allowances) by
                                              receiving credit for system-wide energy sav-
                                              ings. Unlike  the CRER program, under RU a
                                              Phase I affected unit would effectively receive
                                              credit for conservation at the SO2 emission rate
                                              of the units to which load would otherwise have
                                              been shifted  (primarily Phase II units).  An af-
                                              fected utility could thereby potentially save al-

-------
lowances at a much greater rate than could be
earned based on the formula-based CRER.
   The criteria for a utility to use energy conser-
vation under the RU provision (40 CFR § 72.43)
have fewer restrictions than do the criteria un-
der the CRER (Table 1). This provision is appli-
cable only to electric utilities with Phase I affected
units, and that file RU plans with EPA pursuant
to § 72.43. Such utilities may use both demand-
side measures and supply-side (i.e., power gen-
eration, 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.
    Verification of energy savings is required for
each year in which energy conservation mea-
sures contribute to the RU. A Phase I unit may be
attributed the savings from any conservation
measures (other than improved unit efficiency
measures) instituted  within  the unit's  utility
system.1 A RU plan must be filed with EPA by
November 1 of the year in which RU occurs. An
initial estimate of energy conservation savings
must be filed in the annual compliance certifica-
tion report by March 1 of the following year. The
utility unit's designated representative  must
submit the verification results as part of a confir-
mation report by July 1 of each year for which an
affected unit's annual compliance certification
report is submitted to the 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 a  RU plan
by November 1995. The EPA Administrator may
grant, however, for good cause shown, an exten-
sion of the time to file the confirmation report. 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.
Finally, 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, as discussed in
the next section, can be assumed by utility appli-
cants to equal "gross" energy savings in the case
of supply-side conservation measures. Separate
application forms for RU plans will be available
at a future date.
Criteria  for  Deferring
Conservation
Verification to States
The primary users of the CVP are expected to be
public-power utilities.  Most investor-owned
utilities in states that qualify for the CRER will
probably have their energy conservation sav-
ings verified by procedures specified by  the
PUC, and EPA recognizes that it would be bur-
densome to require these utilities to use two
separate verification procedures. Therefore EPA
will defer the conservation verification require-
ment to PUCs based on criteria discussed in the
Allowance System Rule (40 CFR § 73, Subpart F)
and summarized below.
   EPA will require state PUC verification of
energy savings claims of electric utilities if the
PUC 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. The Agency believes that the verifi-
cation performed by state  regulators is likely to
be fairly rigorous when the evaluation of energy
savings will be used to determine the cost of the
conservation incentives to the ratepayers. In the
case of both "shared savings mechanisms" and
"lost revenue adjustment mechanisms", regula-
1 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. See 40 CFR § 72.43(b)(2)(iv).
                                                                                      11

-------
tors must evaluate energy savings achieved by
utility conservation programs in order to deter-
mine the proper adjustment to a utility's rates
necessary to recover lost revenues or to provide
a financial award.
   Under this approach for deferring conserva-
tion verification to a state, the use of engineering
estimates of energy savings is allowed. The re-
quirement to  periodically modify performance-
based rate adjustments through the evaluation of
net energy savings, however, will help to ensure
that a State is committed to good evaluation prac-
tices. EPA recognizes that evaluation of energy
savings is a young and rapidly evolving field and
that many States and utilities begin their evalua-
tion programs with engineering estimates. Most
evaluation experts agree that while engineering
estimates are useful, they should be supplemented
with statistically valid measurements of the de-
crease in energy use resulting from utility energy
conservation programs (i.e., net savings). These
measurements should reflect both the actual num-
ber 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 conservation
verification to states is being established, EPA is
not prescribing standard verification methods or
techniques. States have the flexibility to design
their own verification protocols, as many have.
Because of the diversity of evaluation methods
and the ongoing evaluation activities of state agen-
cies, electric utilities, and other professionals, EPA
believes that this approach will be more effective in
ensuring good practice than would mandating spe-
cific methods or requiring universal use of the CVP.


Forms and  Data

Requirements for  Users

Users of the CVP should summarize their verifica-
tions results on the CVP Reporting Forms for the
CRER and RU, to be published by EPA in early
1993. Applicants  to the  CRER should send this
Form and any attachments (or a state PUC-ap-
proved conservation verification, where appli-
cable), to EPA at the address provided at the end of
the Form Instructions. Applicants should call the
Acid RainHotline (617-674-7377) inFebruary 1993
to receive a copy of the CVP Reporting Form and
Instructions. EPA will begin its review of the veri-
fied conservation savings claims as  the CRER
application period opens on July 1,1993, on a first-
come, first-served basis. EPA will review applica-
tions  to  determine whether they meet the
requirements of Subpart F of 40 CFR§ 73. Verifica-
tion claims may be based on the CVP, a state PUC-
sanctioned method, or other methods. Use of other
methods, however, is likely to increase the time
period for verification by EPA of an application
compared to that needed when the CVP or state
PUC-sanctioned method is used, because EPA
would need to evaluate the alternative method as
well as the application claim. Deficient verification
claims will lose their place in line for bonus allow-
ances if they must be sent back to the applicant for
revision (see40 CFR§ 73.84(b)). Nonetheless, there
is no deadline for sending in  applications to the
CRER (although the program has a limited dura-
tion), nor is there a prescribed reporting period for
conservation results.
   A detailed evaluation of an energy conserva-
tion program requires a large amount of data and
statistical analyses. In an attempt to minimize the
reporting burden for verification, EPA is request-
ing applicants to submit general summary data (as
will be indicated on the CVP Reporting Forms).
Further data collection is needed for the stipulated
measures, for information that is not required tobe
submitted to the EPA, but that should be main-
tained 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 deter-
mine whether an incorrect number of allowances
was disbursed to a utility, or to collect more de-
tailed information about the performance of en-
ergy conservation programs and SO2 emissions
offsets or reductions. Adequate verification pro-
grams by utilities should minimize the number of
audits conducted by the EPA. If such an audit
determines that  an applicant's  energy savings
claims were overstated during the verification,
EPA may reduce the allocation of allowances or
require their surrender by the applicant (40 CFR §
73.82 (c)(4)).
12

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

As indicated earlier, the philosophy of the
CVP is to present general verification pro-
cedures rather than specific requirements.
This approach offers utilities the maximum
amount of flexibility in developing verifica-
tion methods (and may be the only possible
solution given the range of energy efficiency
programs and measures covered  by the
CVP).  Nevertheless, some consistency of
method is needed in order to permit quanti-
fication of the energy savings and objective
evaluation of energy savings claims.  Two
general savings verification paths are de-
scribed: stipulated and monitored. The two
paths are shown earlier in Figure  1 and are
discussed in detail below. This section also
discusses requirements for determining the
persistence  of energy savings,  and proce-
dures for quantifying the impact of supply-
side efficiency improvements.
Path  I:  Stipulated
Savings
The CVP is 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
Appendix D to earn credit for the applicable
conservation measures instead of relying on a
monitoring program. These Stipulated Savings
methods are of two  different types: (1) algo-
rithms for calculating energy savings for spe-
cific measures; and (2) a set of criteria for using
best-engineering practices.

Stipulated Savings for
Specific Measures

The rationale for the list is that 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 conserva-
tion programs (although the energy savings as-
sumptions are  often conservative).  The
Stipulated Savings List contains the following
kinds of measures:
• Common measures whose savings are well-
  documented   in  the  literature  (e.g.,
  refrigerators).

• Measures where instantaneous metering is an
  acceptable proxy for first-year savings (exit
  sign lights).

• Measures where short-term metering is an
  acceptable proxy for first-year savings
  (transformers).
In some cases the laboratory test, "label", or
other pre-specified efficiency parameter may be
a reliable indicator of energy use or savings. For
stipulated  measures, the utility may calculate
energy use and savings based on these values
rather than through metering. For example, in a
refrigerator rebate program, the utility could
calculate energy savings by calculating the dif-
ferences between the maximum energy use dic-
tated by the federal standard and the labeled
consumptions of the refrigerators purchased in
its high-efficiency program. This verification
procedure would be much less costly than me-
tering because the utility would need only to
tabulate labeled values and count installations.
   The approach for calculations of stipulated
energy savings is outlined in Figure 3. The calcu-
lations in the stipulated measures result in gross
energy savings. The utility may convert the Gross
Savings into Net Savings through simple multi-
plication by a default factor (the "gross-to-net
conversion factor") given in the list of stipulated
gross-to-net factors shown in Appendix D; alter-
natively it can use its own adjustment factor
based on its research. The stipulated gross-to-
net factors range from 0.6 - 0.9, and are based on

-------
    Figure 3. Stipulated Savings Path
        Is measure
   on stipulated savings list or
   does program meet criteria
     for use of engineering
         estimates?
        Use stipulated
     or monitored savings
   Calculate stipulated savings
    or engineering estimates
   	(Appendix D)	
       Use stipulated ai
      utility-researched
      gross-to-net factor
                         Utility-researched
                        gross-to-net  factor
                        (subject to approval)
Subsequent-year savings calculated I
       experience with utility conservation programs.
       It should be noted, however, that net energy
       savings estimates for similar conservation pro-
       grams can vary widely, and if a utility develops
       its own estimate through internal research (see
       pp.  24-26) it may be able to claim larger net
       savings than if it uses the stipulated value. The
       utility may use this factor if it can demonstrate
       the legitimacy of its analysis to the EPA. Subse-
       quent-year savings for stipulated measures are
       estimated with  the same procedure as in the
       monitored savings path (see the discussion later
       in this section).

       Good-Practice Engineering  Estimates

       In certain circumstances, measurement of en-
       ergy savings is not feasible or practical. These
circumstances include any of the  following
situations:
• Program-wide energy savings are expected to
  be 5000 MWh/year or less (earning 10 or less
  allowances).
• Extremely small anticipated energy savings
  per customer, spread over many customers.
  (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.

• High cost of monitoring and verifying energy
  savings. A utility may still want to apply for
  verified energy  savings even though
  verification through monitoring would cost
  significantly  more  than 10% of the total
  program costs.
Determination of energy savings through en-
ergy monitoring is the preferred means of veri-
fication. When this is not feasible, in limited
circumstances engineering estimates of savings
may be substituted. The precise conditions in
which engineering estimates can be used are
specified in the list of Stipulated Savings.


Path II: 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
energy consumption measurements into energy
savings. These steps include:
• 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 energy
  savings  resulting from the conservation
  program that  would not have  happened
  without the utility intervention.
• Establishing the confidence in energy savings.
This discussion concludes with an overview of
acceptable measurement procedures  and the
       14

-------
appropriate duration of metering. No particular
measurement technique, however, is prescribed.
The intent of the CVP 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
describe the adjustments  and other method-
ological 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 un-
treated group's energy use adjusted to service
levels from the treated group. This differs from
many definitions of reference cases, where the
untreated (pre-retrofit) energy use and service
levels establish  the reference case. In the CVP,
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 prevail 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 a comparison group. The reference case,
however, also includes levels of service that can
affect the level  of energy use (and hence esti-
mates of energy savings) such as weather, oper-
ating hours, and  industrial output. For an
insulation program of electric-heated homes,
the reference case  might consist of the space
hearing 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  inspections. For ex-
ample, the reference case for  a building lighting
retrofit program might consist of a similar group
of buildings unaffected by the program. In addi-
tion, the operating hours and illumination levels
should be surveyed in order to adjust for yearly
variation.
    The specification of the reference case will
also determine if the utility is estimating gross or
net energy savings. If the reference case captures
the impact of the conservation program (rather
than only the technology), then the utility 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 participate in
the program, then the utility is estimating net
energy savings.

Service Adjusted Energy
Use and Savings

Raw energy savings, that is, simple differences
in metered energy use, will nearly always re-
quire adjustments to account for differences in
metering periods, weather, and levels of service.
These adjustments, to derive the  "service-ad-
              Figure 4. Steps to Calculate
               Service-Adjusted Savings
    f Monitoring Path ...V
            Step 1.  Determine relationship between service levels and
              energy use for treated and untreated cases (through
            	monitoring)	
              Step 2. Specify reference case from which to measure
                energy savings (energy use and levels of service)
                 Step 3. Calculate service-adjusted energy use
                	for treated case
                               I.
                Step 4. Estimate Service-Adjusted Savings (SAS)
               satisfying hypothesis test at the 75% confidence level
                             Does the
                           reference case
                            capture net
                              avings'
Yes
T

I J
First-year savings
                                                                                           15

-------
justed energy use" (SAEU) and "service-adjusted
savings" (SAS), are described below while the
general approach  is shown in Figure  4. The
major steps involve calculating the SAEUs for
the reference case and for the treated case. The
difference between 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
example, a house's space heating energy  use
depends (among other factors) on the outside
and inside temperatures, a factory's energy use
will depend on the number of products manu-
factured in that period, and a restaurant's en-
ergy use will depend on the number of customers
served. Some of the most important service fac-
tors 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 ex-
ample, space heating energy is a function of the
degree-days, and lighting energy use depends
on the number of operating hours and the condi-
tioned 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 ex-
ample, EPRI1 presents one formula quantifying
the relationship between energy use and vari-
ous factors determining 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 more detailed
description, presented in steps, is given in Ap-
pendix B.
    When estimating energy savings, energy use
should be adjusted to the treated ("post-retro-
fit") levels of service in all cases except for weather
variables or where a long-term average differs
from the measurement period. This practice ac-
counts for the fact that energy-efficiency im-
provements are often made in buildings at the
same time that service levels are improved.

Gross and Net Savings Distinctions

The intent of the CVP is to estimate the net
energy savings from  conservation programs.
The utility may specify the reference case  such
that it captures the net savings. For example, the
reference  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 savings") emerge directly from the
SAS calculation.
   The net effects can also be estimated  indi-
rectly through  market research, surveys and
inspections of non-participants. Based on these
methods, a utility can determine a "gross-to-
net" conversion factor. (Multiplying the gross
savings by the "gross-to-net conversion factor"
yields net savings.)
   A special case of net savings arises when the
treated group of  customers is effectively the
entire 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.

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
savings estimates, the  interaction effect of mul-
tiple conservation measures on total energy sav-
ings, and the persistence of energy savings over
1 Electric Power Research Institute, Impact Evaluation of Demand-Side Management Programs, Volume 1: A Guide to Current Practice,
   EPRI CU-7179, Palo Alto, CA, February 1991, Prepared by RCG/Hagler, Bailly, Inc.
 16

-------
time. For purposes of the emissions allowances,
the objective is to award allowances for savings
that occurred with reasonable certainty. The
CVP requires 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 applied. In other words, in order to
qualify for 100 allowances using the CVP, a
utility must be reasonably confident that the
true savings  equaled or exceeded 100 allow-
ances. The utility must 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 presented 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, smaller sample sizes, and
the opportunity to claim some legitimate sav-
ings even when the evaluation itself was not as
successful as  planned. Hanser and Violette dis-
cuss this procedure and compare it to other
confidence determinations.2 The CVP takes 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 Ser-
vice-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 earlier
discussion on gross and net savings distinctions).

Acceptable Measurement Procedures

Energy savings must be derived from the differ-
ence between a reference case and a treated
condition. A  variety of different methods are
acceptable 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 com-
mon methods are:
• Measurement  of energy use  pre/post-
  installation of conservation measures with a
  treated group compared to pre/post- with a
  comparison group.
• Measurement  of energy use  pre/post-
  installation of conservation measures with a
  treated group only (i.e., the treated group is its
  own comparison group).
• Measurement of energy use post-treated /post-
  untreated (i.e., cross-sectional analysis).
• 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
general, however, this procedure yields gross
energy savings (rather than net) because it does
not control for what would have happened to
energy use if the retrofit did not occur. Treated
versus untreated groups might be used to esti-
mate savings from a conservation program ap-
plied  to  new building construction. This
procedure could also supplement a before/after
measurement to provide net energy savings es-
timates. A flip-flop procedure allows for alter-
nating measurements of the original equipment
and the replacement unit. One good candidate
technology for flip-flop measurements would
be where the conservation measure consists of
installing a heat pump in a  home  already
equipped with resistance heating. External con-
trols would periodically 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
Savings analysis of a retrofit program will need
to measure the energy use of participants before
2 P. Hanser and D. Violette, "DSM Program Evaluation Precision: What Can You Expect? What Do You Want?" Proceedings,
   Fourth National Conference on Integrated Resource Planning, pp. 299-313, National Association of Regulatory Utility Commis-
   sioners, Washington, D.C., September 13,1992.
                                                                                        17

-------
            and after the treatment, in addition to the energy
            use of a group of non-participants.

            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 life-
            time. 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
                          energy savings. For example, the "natural cycle"
                          for a space heating retrofit is one winter. Briefer
                          metering periods are acceptable under the CVP,
                          however, if they can be reasonably extrapolated
                          to the whole year. For conservation measures
                          applied to  constant loads (or those automati-
                          cally controlled), instantaneous power determi-
                          nations are satisfactory.
                             Once all of these procedures are completed
                          by the utility, the result is the calculation of first-
                          year energy savings.
   Figure 5. Subsequent-Year Verification Options
  Monitoring
Inspection for
presence and
 operation
   Default
  Biennial verification in
subsequent years 1 and 3
  (including inspection )

 Savings for remainder of
physical lifetime are average
 of last two measurements
X
Yes


(Active
measure)
75% of first-year savings
for units present and
operating for half of
physical lifetime
(biennial inspections)
                   50% of first-year savings for half of
                        physical lifetime
Subsequent-Year

Savings

The persistence of energy conservation savings
over time is uncertain. This uncertainty stems
from both the  deterioration in a conservation
measure's technical performance, and behav-
ioral factors such as the energy user's shifting
preferences and maintenance practices. The lim-
ited research that has been conducted on persis-
tence  of  energy  savings  indicates that
conservation programs often overestimate mea-
sure persistence, particularly in the commercial
sector and for  residential compact fluorescent
lamps.3
   Three options are available for verifying sub-
sequent-year energy savings: monitoring,  in-
spection, and a default. Figure 5 shows these
options as a flow chart, which are applicable to
both the Stipulated Savings and the Monitored
Energy Use paths. Various discounting factors
have been inserted so as to make monitoring
attractive. By monitoring, a utility can obtain
credit for a greater fraction of the savings and for
a longer period. In contrast, the default option
greatly restricts the allowable savings. This re-
flects decreased certainty that savings  persist
when they are  not monitored by the utility.
   For all three options, the gross-to-net conver-
sion factor is calculated once for the first-year
            3 See, e.g., L.A. Skumatz, et al. "Bonneville Measure Life Study: Effect of Commercial Building Changes on Energy Using
               Equipment", SRC No. 7619-R2, Synergic Resources Corp., Seattle, WA, December 1991; Northeast Utilities, A Survey of Other
               Utilities for Measure-Life and Persistence-of-Savings Issues, Conservation and Load Management Dept., NU, Hartford, CT,
               March 1992; E. Vine, "Persistence of Energy Savings: What Do We Know and How Can it be Ensured?", Energy—The
               International Journal, Vol. 17, No. 11,1992, pp. 1073-84; K.M. Keating, "Persistence of Energy Savings", in E. Hirst and J. Reed,
               eds., Handbook of Evaluation of Utility DSM Programs, ORNL/CON-336, Oak Ridge National Laboratory, Oak Ridge, TN,
               December 1991, pp. 89-99.

-------
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 gross-to-net
conversion factor of 0.85), then this factor carries
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. A measure may be physically
present for many years beyond the time it is
saving energy at its designed level. If significant
degradation in performance occurs, then the
effective lifetime is shortened.
   Utilities may shift from one verification path
to another in subsequent  years. It  is expected
that  these  shifts will be from the  monitoring
options to less extensive verification procedures.
These shifts are permissible; once a shift is un-
dertaken, however, a utility cannot reverse itself
and return to a stricter option and receive con-
tinuous credit from the latter.

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
confidence in the existence and accurate esti-
mate of the savings derived from monitoring. In
addition, the utility receives credit for savings
that occur after it terminates monitoring.
   The monitoring option requires that the veri-
fied savings requested by the utility continues to
meet the 75% hypothesis test used for the first-
year  savings calculation. As mentioned previ-
ously, conversion of Gross Savings into Net
Savings can be done through the first-year de-
fault conversion factor or by the  utility- re-
searched factor, and the first-year values can be
used in all Subsequent-Year Savings. A utility
may, if it  wishes, re-evaluate the gross to net
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 veri-
fication is performed) will be  calculated at the
inspection or default rate for no Subsequent-
Year Savings verification. When the next moni-
toring verification is performed, however, the
utility may recoup uncredited savings that oc-
curred in the off year, based on the average of the
last two measurements. More frequent verifica-
tions are also acceptable. The savings from the
fourth year onwards are calculated as the aver-
age 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
actively operating in subsequent years. A mea-
sure is considered to be inoperative if its perfor-
mance is found to be significantly impaired, that
is, about half of the first-year savings are not
occurring. For example, clogged coils in a com-
mercial refrigeration system and broken  time
clocks for a thermostat setback control are rea-
sonable 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 mea-
sure" was determined to have been removed
from 10% of the sites inspected, then the subse-
quent-year verified savings for that year must be
reduced by  10%. Following  this  adjustment,
where applicable, the utility may claim 75% of
first-year savings for half of the measure's physi-
cal lifetime.
   Certain conservation measures do not  need
regular maintenance and  are unlikely to be re-
moved once they are installed. In this sense, they
                                                                                       19

-------
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 sav-
ings, passive measures  will  receive  90%  of
first-year energy savings without  further in-
spection. Utilities 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 performing to first-year levels with-
out requiring active maintenance or operation.
Examples 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
default savings. These savings consist of 50% of
first-year savings, but are limited to half of the
measure's lifetime (as discussed previously).
Supply-Side Efficiency
Improvements

Supply-side efficiency improvements can also
be verified with these protocols. Common sup-
ply-side improvements include boiler upgrades,
transformer substitution, transmission and dis-
tribution improvements, and generator repow-
ering. In most cases the supply-side technology
will convert fuel or heat into electricity. In some
cases, such as transformers, supply-side effi-
ciency improvements will involve technologies
that modify some feature of the electricity, i.e.,
voltage, current, or power factor. In these situa-
tions, the inputs and outputs are electricity. An
efficiency improvement will translate directly
into increased electricity available for distribu-
tion with less or equal input energy. With en-
ergy as both the input and output, measurements
of energy savings from the improvement are
considerably easier to perform and the evalua-
tions more straightforward.
   Standard methods for measuring input and
output energy and efficiencies exist for the ma-
jor supply-side measures. For example, overall
heat rate measurement methods are specified by
the Federal  Energy Regulatory Commission
(FERC), and boiler rates  are covered by the
appropriate professional associations, such as
American Society of Testing Materials.
   The electricity savings from supply-side effi-
ciency improvements should be verified using
the same procedures as outlined in earlier Sec-
tions. A reference case must be specified, with
careful attention to the levels of service for that
piece of equipment. For supply-side measures,
service factors may include capacity factor, op-
erating hours, ambient temperature, and fuel
type. Service adjusted energy use for the pre-
and post-treatment cases  must be calculated,
and service-adjusted savings estimated  from
these two values.
   For example, if a Phase  I utility boiler re-
duces its heat input because of improved unit
efficiency, the reference case is the pre-retrofit,
annual, average heat rate as reported in 1992 in
the EPA's National Allowance Data Base Ver-
sion 2.1 (NADB).4 This heat rate is the applicable
boiler's Btu consumed per kilowatt-hour aver-
aged from 1985 to 1987. To calculate the energy
savings from the improved heat rate, the utility
should take the product of the NADB baseline
heat rate and the unit's baseline annual kilowatt-
hour generation, and subtract from it the prod-
uct of the new heat rate and current annual
generation. The new heat rate should be taken
from the hourly heat rates that will be included
in the quarterly monitoring reports to the Acid
Rain Division of the EPA, weighted by the unit's
generation.
4 U.S. Environmental Protection Agency, Acid Rain Division, "National Allowance Data Base", Version 2 1 EPA Washington
   D.C, July 1992.                                                                       6
20

-------
                              Appendix  A
                                    Definitions
The following definitions are used in the CVP.
Note that all energy savings described in this
document are assumed to be electricity. Several
types of energy savings are used in this docu-
ment. Since each has a different meaning and
application, the definitions are listed together
below. Definitions for other terms are listed
alphabetically following  the energy  savings
terms.

Energy Savings

Raw, metered energy savings: The simple differ-
ence in metered energy use between the treated
customers and the comparison group during the
relevant periods. Essentially no processing has
been undertaken, so  the periods need not be
similar and no normalizations have been per-
formed. (The "savings" can even be negative.)
This term is rarely used in this document but is
defined to  emphasize the distinctions in the
definitions of gross and net energy savings (be-
low).

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, floor area, inside
temperature, etc. Periods  of measurement for
the treated  and comparison groups will often
differ and need to be normalized to a common
length (typically a year). Gross energy savings
seek to quantify the savings from the technolo-
gies applied, so the comparison groups consist
entirely of similar installations that did not re-
ceive the conservation measures.

Net energy savings: The energy savings attribut-
able to the energy conservation program. A
standard method for estimating net energy sav-
ings involves comparing the gross energy sav-
ings for the treated group to the change in energy
use of a similar group not participating in the
program, some of whom may have implemented
similar conserving (or increasing) measures on
their own initiative.

First-year energy savings: A specific period of a
conservation program's energy savings. Soon
after installation, the conservation measures
begin saving energy. For purposes of verifica-
tion, however, the savings are divided into the
"first-year" and "subsequent-year". The "first-
year" begins after the conservation program is
being deployed in a routine manner and the
conservation 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 includes adjust-
ments for changes  in service  that may have
occurred between the untreated  and treated
groups.

Subsequent-year energy savings: The energy sav-
ings occurring in the annual periods following
the first-year.

Stipulated Savings: Energy savings for specified
measures (or procedures) that a utility may use
instead of savings determined through monitor-
ing.
Other Terms

Active conservation measure: Conservation mea-
sures that require regular maintenance, and may
be removed once installed.

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

Gross-to-net conversion factor: A number that,
when multiplied by the gross energy savings,
yields net energy sav ings. The factor is typically
(but not always) between 0 and 1, and deter-

-------
mined by analysis conducted by an electric util-
ity company.

Level of service: Conditions describing the qual-
ity 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.
Passive conservation measure: Conservation
measures that do not require regular mainte-
nance, and that are unlikely to be removed once
installed.

Reference case: The energy use and set of desig-
nated levels of services from which the savings
are estimated.

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

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


-------
                              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 due exclusively
to 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 untreated
and an untreated group. The approach seeks to
adjust energy use in the untreated and treated
groups to a similar level of services before esti-
mating the difference. It is sometimes  conve-
nient to think of the untreated case  as the
"pre-retrofit" and the treated case as the "post-
retrofit" case.  This  analogy applies for very
simple situations  but  requires modifications
when the level of services changes or the verifi-
cation involves cross-sectional analysis (such as
those used for estimating the energy savings
from new construction programs).

Service-Ad justed
Energy Use (SAEU)

The energy use for a given installation can gener-
ally 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 appropriate
for special end uses or applications. The key
aspect is that changes in these levels of services
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 + fifi, where a and 6 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 6, 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:

             ^treat = /treat ^treat'

               unt ~~ / unt ^ unr
The coefficients  in the functional relationship
will be different for the two situations because
they reflect the efficiency at which the service is
being delivered. In the lighting case, for ex-
ample, the coefficients a and 6 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 associ-
ated with the reference service levels is called
the service-adjusted energy use, SAEU ref . By
definition, the reference case consists of the un-
treated 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 be-
cause service levels often change as a result of
the retrofit and 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., burned out)
in the first place. In the course of HVAC retrofits,
                                                                                    23

-------
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.

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
services that differ from the treated period. As a
result, both  the untreated and treated groups
will  need service adjustments. A new set of
services must be defined such that:

             SAEUKf = f^ (Savg)

where Savg becomes the Sref.
   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:

            E; = a + fiH;

In this case,  SAEL/refwill contain service levels
from both the treated and average conditions:
          SAEUref = /un, (S,,.,,;,,,Savg)

The SAE14.^ requires a similar adjustment:

          SA£Ulreat = /treat (^treat/^avg'
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.
Calculation of
Service-Adjusted Savings

The "service-adjusted savings" (SAS) is the dif-
ference between the SAEUs:
              = SAEU«f-SAEUrtrea,

            = /unt ^ref) " /treat (Sref)
These SAS must satisfy the 75% confidence hy-
pothesis test (discussed later) so that SAS will
always be less than or equal to the difference in
SAEUs:

          SAS < SAElZref - SAEtftrea,
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. So the SAEUref is an estimate of the
energy use of the original inefficient lights oper-
ating at the higher number of hours occurring
after the retrofit:

            SAEUref = /on, (4000)

The SAS for this situation is:

          SAS = SAEUKf - SAELTtreat

          = /„,, (4000) -/t^ (4000)

Note that this assumes no other changes of ser-
vice  level occurred in  the treated case.  If the
lighting retrofits also included adjusting illumi-
nation levels, this factor should be included in
/u^and /trea,. A possible relationship would be:

               EJ = a + 81^

where L{ is the average level of illumination, in
lumens, as determined by a series of representa-
tive  spot measurements. The SAEUref and
SAEL4.eatare similarly calculated.

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
24

-------
may be necessary when level-of-service infor-
mation (such as that derived from surveys) is not
available 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 requested amount. As
indicated above:

          SAS < SAEUKf - SAEU^
The SAS must be lowered  until the 75% confi-
dence in the difference SAEUref - SAEU^^ actu-
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
untreated groups should be included in the veri-
fication. Here, simple tests for confidence in the
savings can be applied. (Note that uncertainty in
results exists even when 100% of the population
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
representative sample of participants is accept-
able.
    If a sample of n installations is monitored,
out of a population of N, then
          SAS
               sample
= 4- I  SAS.,
where SASsample  is the average  SAS  for the
sampled installations. In this case, a hypothesis
test must be used to determine the population
                            mean,  SASpopuUhon, that can be extrapolated to
                            have occurred within a certainty of 75%.
                               A variety of hypothesis tests may be used.1 A
                            calculation using a one-tailed t-test is presented
                            here as an example. The  SAS	,.u._ is calcu-
                            lated as follows:
                            population
                                                             'SAS
                                 SAS   1(.  = SAS    -*
                                      population       sample  0.25
where SSAS is the sample standard deviation of
SAS, and t0 a is the critical value of the t-statistic
for a one-tailed hypothesis test at the 75% confi-
dence level for n -1 degrees of freedom (deter-
mined 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 V0fu\atmn.

Stepwise  Procedure

The following steps summarize the procedure
for calculating the Service-Adjusted Savings:
1. Identify reference case and levels of service
  determining energy use. This usually happens
  after the pre-retrofit phase of metering is
  completed.
2. Identify levels of service requiring average
  conditions (usually weather-related).
3. Quantify relationship between service levels
  and energy  use based on energy data,
  manufacturing data, surveys, etc.
4. Calculate SAEUief and SAEU^.
5. Calculate SAS at 75% confidence level.
1 P.W. John, Statistical Methods in Engineering and Quality Assurance. New York: John Wiley & Sons, 1990.
                                                                                       25

-------
    Appendix C

T-Statistics for 75% Confidence
One-Tailed Hypothesis Testing
n-1 t025
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
1.000
.816
.765
.741
.727
.718
.711
.706
.703
.700
.697
.695
.694
.692
.691
.690
.689
n-1
18
19
20
21
22
23
24
25
26
27
28
29
30
40
60
120
oo
TJ.25
.688
.688
.687
.686
.686
.685
.685
.684
.684
.684
.683
.683
.683
.681
.679
.677
.674

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

In certain circumstances, an engineering esti-
mate of energy savings will be acceptable in lieu
of measurements. An engineering estimate typi-
cally requires use of recognized calculation pro-
cedures and technical specifications of materials
or equipment.1  Estimated savings often differ
from measured savings because of incorrect as-
sumptions, especially regarding baseline condi-
tions. Therefore, when measurement is not used,
greater care must be placed on verification of
actual measure installation and correct operation.
   Engineering calculations may be used in any
one of the following situations:

• Measurement costs would exceed 10%  of
  program cost.
• Program-wide energy savings are expected to
  5000 MWh/year or less (1-10 allowances) and
  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. For
  example, energy savings from residential
  lighting efficiency improvements will appear
  on many different circuits. In this case, the
  smallest isolatable circuit is the whole house.
  If energy savings from the improvements are
  expected to be less than 5% of total household
  electricity use, then engineering estimates are
  acceptable.


Verification Requirements

The following requirements must be satisfied in
order for the stipulated savings to qualify as
verified:

• Each submittal must include a brief description
  of  the technologies employed in the
  conservation program, and  the assumptions
  and procedures used to estimate energy
  savings.
• The utility must employ  quality control
  measures in lieu of direct verification of energy
  savings, i.e.,  short-term  measurements,
  inspections, interviews, etc.  Such inspections
  must include  a representative sample of
  installations.
• All estimates  of energy savings must be
  calibrated to  measured consumption. If
  program-specific data from that utility are not
  available, then other data from other utility
  programs are acceptable.
1 Electric Power Research Institute, Engineering Methods for Estimating the Impacts of Demand-Side Management Programs, Volume
  1: Fundamentals of Engineering Simulations for Residential and Commercial End Uses, EPRITR-100984, Palo Alto, CA, July 1992,
  Prepared by Architectural Energy Corporation.

-------
       Higher Efficiency Motors for Constant Load
            Applications  Operating Continuously
Technology Description

The measure  consists of replacing or up-
grading inefficient motors being used to
power a constant load for at least 8500 hours
per year. The retrofit will typically be un-
dertaken in factories or buildings with con-
stant operation (such as hospitals) for fans
and pumps.2

Energy Savings Calculations

The energy savings are determined by cal-
culating the difference in power consump-
tion between  the old and new motors and
multiplying by 8500 hours. Annual savings
(in kWh/year) for a single retrofit are calcu-
lated as follows:

 Annual Energy Savings = 8500 x (PM - Pnev)

where:
• Annual Energy Savings = stipulated energy
  savings (in kWh/year) for the measure.
• 8500 = number of operating hours per year
  (which assumes 3% average downtime for
  maintenance).
• Pow = power consumption of motor (in kW).
• p^^ = power consumption of replacement
  motor (in kW).
True power measurements of Pold and Pnew must
be performed during typical operation.

Verification Requirements

The following requirements must be satisfied in
order for the stipulated savings to qualify as
verified:
• The name and address of the building in which
  the retrofit occurred must be recorded (but not
  submitted).
• The motor application must be recorded (but
  not submitted).
• The measured power of the original and
  replacement (or upgraded) motors must be
  recorded (but not submitted).
2S.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 Blumstein, ed.), American Council
  for an Energy-Efficient Economy, Washington, D.C., 1991.
30

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

Illuminated exit signs are required by law in
almost all non-residential and many multifam-
ily residential buildings. They are required to
operate 24 hours/day in most situations.3 This
measure  involves replacing existing lights in
exit signs (which are mostly incandescent) with
fluorescent lights or light-emitting diodes.4

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 (Pold - Pnew)

where:
• Annual Energy Savings = stipulated energy
  savings (in kWh/year) for the measure.
• 8760 =  number of hours per year.
• P ld = power consumption of original exit light
  (in kW).
• Pnew = power consumption of replacement exit
  light (in kW).
The power consumption of the original exit light
may be assumed to be 0.03 kW. A larger value
for Pold 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.
'International Conference of Building Officials, Uniform Fire Code, International Conference of Building Officials, Whittier, CA,
  May 1991.
4 A.B. Lovins and R. Sardinisky, The State of the Art: Lighting (Competitek), Rocky Mountain Institute, Snowmass, CO, March 1988.
                                                                                    31

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

Amorphous metal-core transformers reduce
no-load losses by 60 to 70% over those in
conventional silicon-steel transformers. No-
load losses are the power required to ener-
gize 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 av-
erage, such offsets 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 mul-
tiplying  the  3/4  power of the replaced
transformer's rated capacity by a factor rep-
resenting the decrease 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 trans-
former are calculated as follows:

         Annual Energy Savings =
           O75 x 3.1 x lO-3 x 8760

where:
• Annual Energy Savings = stipulated energy
  savings (in kWh/year) for the measure.
• C = rated capacity of replaced transformer
  (kVA).
• 3.1 x 10"3 = 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
verified:
• The transformer must be used  instead of a
  silicon steel or silicon iron transformer.
• The rated capacity and the location must be
  collected.  (This information need not  be
  submitted.)
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, to
obtain greater than the stipulated savings.
32

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

This measure  involves replacing lights with
higher efficiency units in offices. It only applies
to office buildings and usually consists of re-
placing incandescent or old fluorescent fixtures
with high-efficiency fluorescent lamps, im-
proved fixtures, and electronic ballasts. A retro-
fit consists of identical lights replaced on a single
circuit.5 No credit is given for improved switches,
occupancy sensors, or daylight controls in this
measure (although qualified savings from these
measures can be earned when verified through
monitoring).

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.6-7 In addition, lighting
levels are often adjusted during the retrofit. The
exact values for these variables are difficult to
determine without monitoring. Monitored sav-
ings have varied depending on the conditions
outlined above.8-9'10-11 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 calcu-
lating 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 (inkWh/
year) for a single retrofit are calculated as fol-
lows:

 Annual Energy Savings = HxLx (Pold - Pnew)

where:
• Annual Energy Savings = stipulated energy
  savings (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 retrofit.
• Pold = power consumption  of original light
  fixture (in kW) that was operating prior to the
  retrofit.
• Pnew = power consumption of  replacement
  light fixture (in kW).
The number of hours, H, must be less  than or
equal to 3300.12 The difference (Pold  Pnew) for
fluorescent  lights may  not exceed the values
given in the following table:
Original Fluorescent
Lamp Type
2 Lamp fixture
3-lamp fixture
4-lamp fixture
Maxium Allowable
Difference*
0.025
0.037
0.050
5 A lighting circuit is 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 lighting circuit
   is the combination of circuits servicing the lights for a distinct area, such as one or more floors.
6 A.B. Lovins and R. Sardinsky, The State of the Art: Lighting (Competitek), Rocky Mountain Institute, Snowmass, CO, March 1988.
'Rochester Gas & Electric Co., "Basis for Compensation," Document Reference # III.H.8 (RG&E DSM #1503), RG&E, Rochester,
   NY, March 11,1991.
8S.M. Nadel and K.M. Keating, "Engineering Estimates vs. Impact Evaluation Results: How do They Compare and Why?",
   Proceedings of the 1991 International Energy Program Evaluation Conference, Evanston, IL, 1991, pp. 24-33.
9M.A. Piette et al., "Technology Assessment: Energy-Efficient Commercial Lighting," LBL-27032, Lawrence Berkeley Labora-
   tory, Berkeley, CA, 1989.
10 Xenergy, Inc., Impact Evaluation of Commercial Lighting Rebate Program, Vol. 1992, Integrated Least-Cost Resource Plan,
   Prepared for Potomac Electric Power Co., Washington, D.C., 1992.
11 The Results Center, Sacramento Municipal Utility District: Commercial Lighting Installation Program, IRT Environment, Inc.,
   Aspen, CO, 1992.
12 Pacific Gas and Electric Co., Commercial, Industrial and Agricultural Direct Rebate Programs Hours of Operation Study, PG&E,
   Measurement and Evaluation Planning Section, San Francisco, CA, August 1992.

-------
 It is technically possible to achieve energy sav-
 ings approximately double those given in the
 above table, in cases where state-of-the-art fluo-
 rescent lamps, ballasts, and fixtures replace very
 inefficient (but fully functioning) ones. Many
 utilities will want to monitor a sample of their
 commercial lighting retrofits, which could lead
 to these larger savings.

 Verification Requirements

 The following requirements must be satisfied in
 order for the stipulated savings to qualify as
 verified:
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 replaced
lights must be recorded for each building (but
not submitted).
The number of non-functioning (removed or
broken) lights must be recorded for each retrofit
(but not submitted). Retrofits replacing broken
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

-------
        High  Efficiency Refrigerator Replacement
Technology Description

This measure involves replacing an existing re-
frigerator-freezer with a higher efficiency unit.
It applies to all new residential refrigerator-
freezers sold with Energy Guide labels.

Energy Savings Calculations

The field energy use of a refrigerator has been
shown to correlate well with the labeled energy
use.13 Net energy savings are the difference be-
tween the labeled energy use of the purchased
unit and a comparable new unit meeting the
applicable efficiency standard. A single stock-
wide value, based on  1991  AHAM data14, is
provided to avoid determining the labeled en-
ergy use of each comparable unit.
                        all refrigerators
 Annual Energy Savings =
where:
• AnnualEnergySavings=program-wideenergy
  savings due to refrigerator  replacements
  occurring in year n (in kWh/year).
• i = each new refrigerator.
• All refrigerators = all new refrigerators
  participating in the program.
• e = labeled energy use of each purchased
  refrigerator unit (in kWh/year).
Year replacement
took place (n)
1985-1989
1990-1992
1993-1996
e,,
(kWh/year)
1200
950
750
Verification Requirements

The following requirements must be satisfied in
order for the stipulated savings to qualify as
verified:
 • The name and address of each participating
  customer, date of delivery of rebate check, and
  the manufacturer and model number of the
  associated refrigerator, must be recorded (but
  not submitted).
• An older unit must be  taken from the
  customer's premises for each new unit
  supplied.
13 A.K. Meier and R. Jansky, "Field Performance of Residential Refrigerators: A Comparison with the Laboratory Test," LBL-
  31785, Lawrence Berkeley Laboratory, Berkeley, CA, 1991.
14 Association of Home Appliance Manufacturers, Energy Efficiency Trends, AHAM, Chicago, IL, June 20,1991. An average 18
  cubic foot model refrigerator-freezer is assumed here to represent the entire stock. The 1993-1996 value accounts for the sale
  in 1993 of a small number of units carried over from the 1992 stock, and reflects consumers' growing preference toward
  purchase of larger refrigerator-freezers.

-------
                 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 be-
tween 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 (Pold - Pnew)

where:
• Annual Energy Savings = stipulated energy
  savings (in kWh/year) for the measure.
• 4000 = number of operating hours per year.
• Pold = power consumption of original street
  light (in kW).
• P   = power consumption of replacement
  street light (in kW).


Verification Requirements

The following requirements must be satisfied in
order for the stipulated savings to qualify as
verified:
• The number, type, and power rating of replaced
  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 retrofit
  cannot be counted for credit with this stipulated
  measure.
• Rated power for the replacement must include
  the entire system, that is, high-intensity
  discharge (HID) light and ballast, etc.
• The street light must be  controlled by a
  functioning photocell.

-------
                Water  Heater Insulation  Blankets
Technology Description

This measure involves reducing the standby
losses of residential water heater storage tanks
through insulation blankets and anti-convec-
tion valves.

Energy Savings Calculations

Energy savings from insulation blankets de-
pend primarily on the temperature of the hot
water being stored, the ambient air temperature,
and the amount of insulation already around the
tank. Savings from individual units will vary
somewhat but average energy savings from ret-
rofits of many units are reliable and more consis-
tent. 1S-16-17 Actual savings from individual units
may, in some cases, be greater than the  stipu-
lated savings if the savings are monitored.
   If the utility choses to  use the stipulated
value, annual savings (in kWh/year) for a single
retrofit are calculated as follows:

  Annual Energy Savings = Expected Savings

where:
• Annual Energy Savings = stipulated energy
  savings (in kWh/year) for the measure.
• Expected  Savings - typical  savings for a
  carefully installed measure, as determined
  from the table below:

        Measure       Expected Electricity Savings
	(kWh/year)	
  Insulation blanket around tank     400
  Anti-convection valves           200
  Pipe insulation                 200


Verification Requirement

The following requirements must be satisfied in
order for  the stipulated savings to qualify as
verified:
• Each installation must be inspected by a utility
  representative.
• The name and address of each participating
  customer  and date of inspection  must be
  recorded (but not submitted).
• Insulation blankets must have a thermal
  resistance  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.
15 A. Usibelli, "Monitored Energy Use of Residential Water Heaters," Proceedings of the ACEEE 1984 Summer Study on Energy
   Efficiency in Buildings, American Council for an Energy-Efficient Economy, Santa Cruz, CA, 1984.
16 D.H. Sumi and B. Coates, "Persistence of Energy Savings in Seattle City Light's Residential Weatherization Program/'
   Proceedings of the 1989 International Energy Program Evaluation Conference, Evanston, IL, 1989, pp. 311-16.
17M. A. Brown et al., Impact of the Hood River Conservation Project on Electricity Use for Residential Water Heating, Oak Ridge National
   Laboratory, Oak Ridge, TN, 1987.
                                                                                        37

-------
                        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.18'19'20

Energy Savings Calculations

The stipulated median lifetimes are given in the
following two tables. They are adapted from the
consensus21 reached by the California utilities,
regulatory  commissions and other interested
parties but are based on work by many groups
across the country. These lifetimes maybe used
for calculations of savings as described in the
Protocols.

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 that described in the
  list.
18 L.A. Skumatz, K.M. Lorberau, R.J. Moe, R.D. Bordner, and R.D. Chandler, "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, R7729 mi, SRC, Oakland, CA, February 1992.
20Northeast Utilities, A Survey of Other Utilities for Measure-Life and Persistence-of-SavingsIssues.'NortheastUtilities, Conservation
   & Load Management Department, Hartford, CT, March 1992.
21 Measurement Subcommittee, Measurement Protocols for 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.
38

-------
  Useful Lives of Residential
Energy Conservation Measures
Measure
Caulking
Weatherstripping
Ceiling Insulation
Wall insulation
Low-flow showerheads
Water faucet aerators
Duct wrap/insulation
Water heater blanket
Fluorescent bulbs
Window shade awnings

Lifetime (Years)
10
10
25
25
10
10
15
10
10
10

Measure
High-efficiency A/C
Central heat pump
Evaporative coolers
Clock thermostat
High-efficiency refrigerator
High-efficiency central furnace
Whole house fan
Double glazing
Storm windows
Furnace retrofit
Efficient gas water heater
Lifetime (Years)
18
18
15
15
20
20
15
25
10
15
13
                                    39

-------
         Useful  Lives  of Commercial  and Industrial
                  Energy  Conservation  Measures
Equipment Type & Description      Lifetime (Years)
LIGHTING
Energy-efficient fluorescent lamp          5
Same as above with build-in ballast        2
Energy-efficient ballast                 11
Electronic ballast                      10
Metal halide lamp                     10
Low-pressure sodium lamp              5
High-pressure sodium lamp             5
Parabolic fixture                      20
Dimming systems                     20
On-off switching                      7
Morion sensor                        10
HVAC
Economizer                          11
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 & Description       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            17
Variable-speed DC motor               18
Variable-speed drive—solid type         15
Variable-speed drive—belt type          10
Efficient AC electric transformer          15
DOMESTIC HOT WATER
Heat pump water heater                10
Point-of-use water heater               12
Solar water heater                     15
Change electric to gas booster            —
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-emissivity coating                 14
Solar shade film (retrofit)                7
Tinted & reflective coating              14

-------
             Stipulated  Gross-To-Net Factors for
                          Stipulated  Measures
Description

Conversion factors in the following list can
be used to convert the gross annual savings
calculated in the stipulated measures to net,
first-year savings. These values should be
used when utilities cannot provide estimates
based  on their own programs  or market
research. Factors that affect the net impacts
of conservation programs such as free rid-
ers vary widely by the type of conservation
program and market conditions.22 Utilities
may conclude  that their conservation pro-
grams have larger net energy savings when
they conduct their own analyses rather than
rely on the default factors given below. Sev-
eral 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 GNCF

where:
GNCF = "Gross-To-Net-Conversion Factor"
listed in the table below:
   Conservation Measure
                             Gross-To-Net
                            Conversion Factor
Engineering estimates of energy savings      0.6
Exit sign light replacements               0.6
Higher efficiency motors for constant load     0.6
   applications operating continuously
Higher efficiency lights in office buildings     0.6
Amorphous metal distribution transformers   1.0
High efficiency refrigerator replacement      0.9
Higher efficiency street lights              0.9
Water heater insulation blankets           0.6
22 W. Saxonis, "Free Riders and Other Factors that Affect Net Program Impacts," in E. Hirst and J. Reed, eds., Handbook of
  Evaluation of Utility DSM Programs, ORNL/CON-336, Oak Ridge National Laboratory, Oak Ridge, TN, December 1991, pp.
  119-34.
                                                                                    41

-------
                           Appendix  E
                 U.S. Environmental  Protection Agency
                     Acid Rain Advisory Committee
             Subcommittee on Conservation Verification
Stanley Hulett, Subcommittee Chair
CEO, Chairman of the Board
Intersource Technologies Inc.

Thomas A. Buckley
Director of Energy Services
Burlington, Vermont Electric Department

Cary Bullock
President
Kenetech Energy Management

John Fox
Manager, Energy Efficiency Services
Pacific Gas & Electric Company

Larry Frimerman
Federal Liaison Officer
Ohio Office of Consumers' Counsel

William C. Harding
Retired Energy Engineer
Bethlehem Steel Corporation

Jeff Harris
Conservation Analyst
Northwest Power Planning Council
Elizabeth Hicks
Director, Demand Planning
New England Power Service Company

Dr. Eric Hirst
Corporate Fellow
Oak Ridge National Laboratory

Dr. Martin Kushler
Supervisor of the Evaluation Section
Michigan Public Service Commission

Steven M. Nadel
Deputy Director
American Council for an Energy-Efficient Economy

Dr. Robert L. San Martin
Deputy Assistant Secretary for Utility Technologies
U.S. Department of Energy

Vincent Schueler
Manager, Program Research
Washington State Energy Office

Sam Swanson
Deputy Director, Office of Energy Efficiency
 and Environment
New York State Public Service Commission
                                                                             43

-------