Energy Efficiency Program for Commercial and Industrial Equipment: Efficiency Certification, Compliance, and Enforcement Requirements for Commercial Heating, Air Conditioning and Water Heating Equipment, 25103-25117 [06-3319]

Agencies

• DEPARTMENT OF ENERGY
• Office of Energy Efficiency and Renewable Energy
• Energy Efficiency and Renewable Energy Office

[Federal Register Volume 71, Number 82 (Friday, April 28, 2006)]
[Proposed Rules]
[Pages 25103-25117]
From the Federal Register Online via the Government Printing Office [www.gpo.gov]
[FR Doc No: 06-3319]



[[Page 25103]]

=======================================================================
-----------------------------------------------------------------------

DEPARTMENT OF ENERGY

Office of Energy Efficiency and Renewable Energy

10 CFR Part 431

[Docket No. EE-RM/TP-99-450]
RIN No. 1904-AB64


Energy Efficiency Program for Commercial and Industrial 
Equipment: Efficiency Certification, Compliance, and Enforcement 
Requirements for Commercial Heating, Air Conditioning and Water Heating 
Equipment

AGENCY: Office of Energy Efficiency and Renewable Energy, Department of 
Energy.

ACTION: Supplemental notice of proposed rulemaking.

-----------------------------------------------------------------------

SUMMARY: In a notice of proposed rulemaking published December 13, 
1999, (NOPR) the Department of Energy (DOE or the Department) proposed 
to adopt (1) energy conservation requirements that the Energy Policy 
and Conservation Act, as amended, (EPCA or the Act) specifically 
mandated for commercial warm air furnaces, and (2) provisions applying 
generally to covered commercial heating, air conditioning and water 
heating equipment, including furnaces, (collectively referred to as 
``commercial HVAC & WH equipment'') to assure their compliance with 
EPCA requirements. On October 21, 2004, DOE adopted a final rule 
incorporating the requirements for furnaces but only certain of the 
general provisions proposed for commercial HVAC & WH equipment. As to 
the latter, the Department did not adopt the NOPR's proposals for 
manufacturers to use to determine and certify compliance, and or most 
of its enforcement proposals, which remain under consideration. These 
include proposals about manufacturers' use of testing and calculation 
methods to rate the efficiency of their equipment, the role of 
voluntary independent certification programs in assuring the accuracy 
of the ratings, and the testing regimen and criteria that DOE would use 
in enforcement proceedings, which are the subjects of today's notice. 
The Department is now soliciting comments on several additional 
proposed options that DOE is now considering for the rule.
    In addition, the Energy Policy Act of 2005, Public Law 109-58, 
(EPACT 2005) created a new category of covered equipment and set forth 
definitions, test procedures, and energy conservation standards for 
very large commercial package air conditioning and heating equipment. 
The Department has codified the definitions and energy conservation 
standards in Title 10, Code of Federal Regulations, Part 431. 70 FR 
60407 (October 18, 2005). The Department is applying to that equipment 
the proposed compliance and enforcement requirements that are the 
subject of this supplemental notice. (The Department notes that the 
recent amendments to EPCA set forth in EPACT 2005 do not otherwise 
affect the issues raised in today's notice.)

DATES: The Department will accept comments regarding today's proposals 
until June 12, 2006.

ADDRESSES: You may submit comments, identified by docket number EE-RM/
TP-99-450 and/or RIN number 1904-AB64, by any of the following methods:
     Federal eRulemaking Portal: https://www.regulations.gov. 
Follow the instructions for submitting comments.
     E-mail: commercial-- HVACandWH_rule@ee.doe.gov. Include 
EE-RM/TP-99-450 and/or RIN number 1904-AB64 in the subject line of the 
message.
     Mail: Ms. Brenda Edwards-Jones, U.S. Department of Energy, 
Building Technologies Program, Mailstop EE-2J, Reopening Notice for 
Efficiency Certification and Enforcement of Air Conditioning and Water 
Heating Products, EE-RM/TP-99-450 and/or RIN 1904-AB64, 1000 
Independence Avenue, SW., Washington, DC 20585-0121. Telephone: (202) 
586-2945. Please submit one signed paper original.
     Hand Delivery/Courier: Ms. Brenda Edwards-Jones, U.S. 
Department of Energy, Building Technologies Program, Room 1J-018, 1000 
Independence Avenue, SW., Washington, DC 20585.
    Instructions: All submissions received must include the agency name 
and docket number or Regulatory Information Number (RIN) for this 
rulemaking. For detailed instructions on submitting comments and 
additional information on the rulemaking process, see section IV of 
this document (Submission of Comments).
    Docket: For access to the docket to read background documents or 
comments received, go to the U.S. Department of Energy, Forrestal 
Building, Room 1J-018 (Resource Room of the Building Technologies 
Program), 1000 Independence Avenue, SW., Washington, DC, (202) 586-
9127, between 9 a.m. and 4 p.m., Monday through Friday, except Federal 
holidays. Please call Ms. Brenda Edwards-Jones at the above telephone 
number for additional information regarding visiting the Resource Room. 
Please note: The Department's Freedom of Information Reading Room 
(formerly Room 1E-190 at the Forrestal Building) is no longer housing 
rulemaking materials. The docket will also be posted to the Federal 
Docket Management System through the Federal eRulemaking Portal (http:/
/www.regulations.gov) after the comment period closes. You can also 
electronically obtain a copy of this notice and related background 
documents from DOE's Building Technologies Program's Web site at the 
following URL address: https://www.eere.energy.gov/buildings/
appliance_standards/notices_rules.html.

FOR FURTHER INFORMATION CONTACT: James Raba, U.S. Department of Energy, 
Office of Energy Efficiency and Renewable Energy, Mail Station, EE-2J, 
1000 Independence Avenue, SW., Washington, DC 20585-0121, (202) 586-
8654. E-mail: jim.raba@ee.doe.gov. Thomas DePriest, U.S. Department of 
Energy, Office of the General Counsel, GC-72, 1000 Independence Avenue, 
SW., Washington, DC 20585, (202) 586-9507, E-mail: 
Thomas.DePriest@hq.doe.gov.

SUPPLEMENTARY INFORMATION:

I. Background
II. Discussion
    A. Methods for Manufacturers To Follow To Determine Energy 
Efficiency Ratings of Their Equipment
    1. Background
    2. General Standards for Testing by Manufacturers
    3. Test Sampling by a VICP Participant
    4. Criteria for AEDM Validation and Use of AEDMs
    B. Voluntary Industry Certification Programs (VICPs)
    1. Background
    2. General Standards for Testing by a VICP
    3. Determining the Validity of Manufacturers' Efficiency Ratings
    4. Manufacturer Challenges of Equipment Ratings
    5. VICP Reporting to the Department
    C. Enforcement by the Department
    1. Enforcement Testing--General
    2. Enforcement Testing--Defective Units and Retention of Sample 
Units
    3. Enforcement of Design Standards
    D. Conclusion
III. Procedural Requirements
    A. Review Under Executive Order 12866
    B. Review Under the Regulatory Flexibility Act
    C. Review Under the Paperwork Reduction Act
    D. Review Under the National Environmental Policy Act
    E. Review Under Executive Order 13132
    F. Review Under Executive Order 12988
    G. Review Under the Unfunded Mandates Reform Act of 1995.
    H. Review Under the Treasury and General Government 
Appropriations Act, 1999

[[Page 25104]]

    I. Review Under Executive Order 12630
    J. Review Under the Treasury and General Government 
Appropriations Act, 2001
    K. Review Under Executive Order 13211
    L. Review Under Section 32 of the Federal Energy Administration 
Act of 1974
IV. Submission of Comments
V. Approval of the Office of the Secretary

I. Background

    The Energy Policy and Conservation Act (EPCA or the Act) (42 U.S.C. 
6311-6316) establishes energy conservation requirements for certain 
commercial and industrial equipment. For commercial heating, 
ventilating, air conditioning and water heating (HVAC & WH) equipment, 
EPCA provides energy conservation standards and authorizes the 
Department of Energy (DOE or Department) to amend these standards. (42 
U.S.C. 6313(a)) The Act also provides test procedures for this 
equipment, and authorizes the Department to amend these test 
procedures. (42 U.S.C. 6314(a)) Finally, EPCA authorizes the Secretary 
to implement these energy conservation requirements by issuing the 
necessary rules requiring manufacturers of covered commercial and 
industrial equipment to submit information and reports, and taking 
enforcement action. (42 U.S.C. 6316(b))
    As indicated in the SUMMARY above, the notice of proposed 
rulemaking (NOPR) included proposed rules covering manufacturers' 
compliance with energy conservation requirements for all commercial 
HVAC and WH equipment and DOE enforcement of these requirements. 64 FR 
69598 (December 13, 1999). Specifically, the Department proposed 
methods for manufacturers to use to implement the DOE test procedures 
to determine the efficiency or energy use ratings of this equipment, 64 
FR at 69602-06 and 69612-14, procedures for certifying such ratings to 
the Department, 64 FR at 69604, 69614-16, and criteria and procedures 
for enforcement actions by the Department for alleged violations of 
energy conservation standards, 64 FR at 69605, 69616-18.
    On January 27, 2000, DOE convened a public hearing to receive oral 
comments on the proposed rule. The Department also received written 
statements in advance of the hearing and written comments after the 
hearing. These oral comments and written submissions, as well as the 
Department's further review of the proposed rule, raised the issues 
addressed in today's supplemental notice of proposed rulemaking 
(SNOPR). While still considering adoption of the proposals contained in 
the NOPR, the Department seeks comment on the alternative language and 
options that it is proposing in this SNOPR. The DOE wishes to emphasize 
that it will continue to consider for adoption all of the proposals set 
forth in the NOPR and the SNOPR.
    The Department also notes that the proposed rule language in 
today's SNOPR, which would be incorporated into Title 10 Code of 
Federal Regulations (10 CFR Part 431), uses subpart designations and 
section numbers that correspond to those used in the NOPR. However, 
since the issuance of the NOPR, the Department has reorganized and 
renumbered the rules in part 431. It did so, first in the final rule 
for furnaces and commercial HVAC and WH equipment, referred to above, 
69 FR 61916 (October 21, 2004), and more recently in a final rule to 
incorporate certain requirements contained in EPACT 2005. 70 FR 60407 
(October 18, 2005).
    The Department has retained the subpart designation and numbering 
approach it used in the NOPR to facilitate stakeholder comparison of 
the NOPR proposals with today's proposals. When the Department adopts a 
final rule that addresses the issues raised by the NOPR and this SNOPR, 
it will base the structure and numbering of the provisions in that rule 
on part 431 as it exists at that time. Given the current structure of 
part 431, DOE anticipates that it would include provisions as to 
compliance determination for commercial HVAC and WH equipment in 
subpart J, and for enforcement in subpart U. See 10 CFR Part 431 
subparts J and K (2005) and 70 FR at 60416. Today's proposals would not 
affect the recent amendments to part 431 that incorporated requirements 
contained in EPACT 2005. 70 FR 60407. Rather these proposals would add 
to, but not replace or alter, provisions currently in part 431.
    Finally, sections 136(a)(3), 136(b)(5), and 136(f)(1) of EPACT 2005 
amend sections 340(8), 342(a), and 343(a)(4) respectively, of EPCA, 42 
U.S.C. 6311(8), 6313(a), and 6314(a)(4) to add definitions, energy 
conservation standards, and test procedures, respectively, for very 
large commercial package air-conditioning and heating equipment rated 
at or above 240,000 and below 760,000 British thermal units per hour 
(Btu/h) cooling capacity. The Department has incorporated the new EPCA 
energy conservation standards and definitions under subpart F of 10 CFR 
part 431. 70 FR 60415. In particular, the Department inserted a 
definition of ``very large commercial package air-conditioning and 
heating equipment'' into Sec.  431.92 of 10 CFR part 431. Thus, that 
equipment is now included in the equipment covered by this rulemaking.

II. Discussion

A. Methods for Manufacturers To Follow To Determine Energy Efficiency 
Ratings of Their Equipment

1. Background
    In the NOPR, the Department proposed to require manufacturers to 
determine initially the efficiency of each of their types of commercial 
HVAC and WH equipment either by testing the equipment \1\ using the 
applicable DOE test procedure, or by calculating the efficiency of the 
equipment through use of an alternative efficiency determination method 
(AEDM). To use an AEDM, a manufacturer would have to establish the 
AEDM's validity through the following process: (1) Apply the AEDM to a 
limited number of basic models to calculate their efficiency, (2) 
measure the efficiency of these same basic models by testing them, and 
(3) compare the test results with the calculations. The proposed rule 
would allow manufacturers to participate in Voluntary Industry 
Certification Programs (VICPs) to help establish the accuracy of 
manufacturer efficiency ratings and their compliance with Federal 
efficiency standards. Firms participating in VICPs would be subject to 
less stringent requirements for test sampling of equipment and for 
determining the validity of AEDMs than firms that did not participate 
in VICPs.
---------------------------------------------------------------------------

    \1\ The Department commonly refers to such testing as 
``certification testing.'' Under DOE's regulations for consumer 
appliances in 10 CFR Part 430, each manufacturer must certify to DOE 
the efficiency rating of each of its basic models, and manufacturer 
generally derives that rating from testing it performs to determine 
initially the model's rating. The Department contemplates adoption 
of this same scheme for commercial HVAC and WH equipment.
---------------------------------------------------------------------------

2. General Standards for Testing by Manufacturers
    Section 431.481(b) of the proposed rule contains general 
requirements for certification testing and for testing to validate 
AEDMs for commercial HVAC and WH equipment. Paragraph (3) of that 
section states that such testing must ``[m]eet industry standards for 
the accuracy of testing and of rating results for the equipment being 
tested * * *.'' 64 FR at 69612. In its comments, the Gas Appliance 
Manufacturers Association (GAMA) asserts that the meaning of the term 
``industry standards'' is unclear. (GAMA, No. 3 at 4) \2\
---------------------------------------------------------------------------

    \2\ A notation in the form ``GAMA, No. 3 at 4'' identifies a 
written comment DOE received in this rulemaking after issuance of 
the NOPR. This notation refers to a comment (1) by GAMA, (2) in 
document number 3 in the docket in this matter, and (3) appearing at 
page 4 of document number 3.

---------------------------------------------------------------------------

[[Page 25105]]

    This provision is designed to require that measurements performed 
during testing meet the industry standards for accuracy that exist at 
the time a test is performed. Although the term ``industry standards'' 
may appear vague, DOE believes specific numerical criteria would be 
inappropriate in the rule because industry measurement standards can 
vary for different test procedures and types of equipment, and over 
time. The Department's intent is that ``industry standards'' as it uses 
that term in the proposed regulation would be evidenced by sources such 
as accuracy requirements in applicable test procedures and in ratings 
of measurement equipment, and would require, for example, that 
measurements conducted under DOE test procedures be performed using the 
laboratory-grade equipment, calibration standards and methods that 
represent the ``best practices'' used in the industry. In sum, the 
Department would require each manufacturer to perform the testing so as 
to minimize measurement uncertainty, in accordance with currently 
accepted industry measurement practices.
    The Department is proposing a revision to proposed Sec.  
431.481(b)(3) that would incorporate these concepts, and that would 
make clear that the rule is referring to measurement accuracy. The 
revised language DOE is considering would eliminate the reference to 
``rating results'' and add the term ``measurement accuracy.'' The DOE 
solicits public comment on the alternative proposal that if a 
manufacturer tests a basic model to determine its efficiency or to 
validate an AEDM, it must meet industry standards for the measurement 
accuracy of testing for the equipment being tested including accuracy 
requirements in applicable test procedures, accuracy achieved by 
laboratory-grade equipment, and the accuracy of calibration standards.
3. Test Sampling by a VICP Participant
    In the NOPR the Department proposed in Sec.  431.483 that when a 
manufacturer not participating in a VICP tests equipment under the 
regulations, it would have to use a test sampling procedure similar to 
what DOE requires in 10 CFR Part 430 for consumer appliances. 64 FR at 
69613. By contrast, DOE proposed no specific sampling procedure for 
testing by VICP participants, and instead proposed that when a 
participant tests a basic model it ``must use statistically valid and 
accurate methods to arrive at the efficiency rating of such basic 
model.'' 64 FR at 69613 (proposed Sec.  431.482(b)). The Department 
proposed less stringent requirements for initially establishing the 
efficiency of equipment from VICP participants because, unlike the 
equipment of non-participants, the efficiency ratings of their 
equipment would be subject to verification and other oversight by the 
VICP.
    The Department continues to believe that VICP participants should 
be subject to less stringent test sampling requirements than non-
participants and that they should have substantial discretion to choose 
a sampling plan. Nevertheless, upon further consideration DOE believes 
the ``statistically valid and accurate methods'' standard for testing 
by VICP participants may be too vague. Furthermore, the goal of any 
testing to determine a basic model's rating is to give reasonable 
assurance that the rating accurately reflects on average the efficiency 
of all units sold, and the regulations should require that 
manufacturers' testing programs meet this standard. Therefore, the 
Department is proposing to revise proposed Sec.  431.482(b) as follows:

    A VICP participant that tests a basic model pursuant to this 
subpart must use statistically valid and accurate methods to arrive 
at the efficiency rating of the tested basic model. Such methods 
must give reasonable assurance that the manufacturer's efficiency 
rating for a basic model does not exceed the mean energy efficiency 
of the population for that basic model.
4. Criteria for AEDM Validation and Use of AEDMs
    An AEDM is a method for determining the efficiency of equipment by 
means of a calculation, rather than by testing the equipment. In the 
NOPR, the Department proposed in Sec.  431.481(a) to allow each 
manufacturer to determine the efficiency of each of its commercial HVAC 
and WH basic models either by testing the model or by using an 
appropriate AEDM. 64 FR at 69612. A manufacturer could use an AEDM that 
met certain general criteria and had been validated (i.e., the 
manufacturer had established its accuracy). 64 FR at 69612-13. 
Validation of an AEDM by a manufacturer not participating in a VICP 
would be based on comparing the efficiency ratings derived from testing 
three or more basic models with the efficiency ratings derived from 
applying the AEDM to those same basic models. A VICP participant would 
have to make such a comparison for one or more basic models. When a 
manufacturer made the comparison for two or more basic models, the 
proposed rule would permit use of the AEDM only if the average 
efficiency rating, derived from applying the AEDM to these basic 
models, is within one percent of the average rating derived from 
testing them, and if the AEDM and testing results are within five 
percent of each other for each of the basic models. (See proposed 
Sec. Sec.  431.482(c) and 431.483(b), 64 FR at 69613.) For VICP 
participants who made the comparison for only one basic model, the 
Department proposed that the difference between the AEDM and test 
results must be within one percent for the AEDM to be valid. (See 
proposed Sec.  431.482(c), 64 FR at 69613.)
    In its comments, the California Energy Commission (CEC) objects to 
the five-percent provision. It appears to assert that DOE should not 
permit use of an AEDM unless the AEDM produces the same results as 
testing. The CEC also claims that the proposed AEDM provisions would 
allow use of an AEDM to rate each basic model at a level up to five 
percent higher than test results for that model would warrant, and that 
this would unfairly penalize manufacturers who base their ratings on 
physical testing, which CEC asserts is the preferred method. (CEC, No. 
7 at 8)
    The Department believes that some of CEC's concerns may have merit, 
and, upon further consideration, also has other concerns about the 
proposed provisions for validating AEDMs. First, as stated above, the 
proposed rule would permit VICP participants to validate an AEDM by 
comparing AEDM and test results for only one basic model. The 
Department now questions whether such a limited comparison provides a 
sufficient basis for concluding that an AEDM is accurate.
    Second, the Department is concerned about the possibility that use 
of AEDMs under the proposed rule could result in overrating equipment. 
The five-percent criterion provides that when a manufacturer validates 
an AEDM by applying it to more than one basic model, it must predict an 
efficiency for each that is within plus or minus 5 percent of the test 
results for that model. This means that the proposal would allow an 
AEDM to have a range of uncertainty of 10 percent, and a built-in 
potential for overrating and under-rating of five percent each. This 
may allow too great a potential for overrating, and may also raise 
questions about the accuracy of ratings. The proposed tolerances for 
validating AEDMs, coupled with the lack of limitations on the basic 
models that manufacturers can use for such validation, also may create 
potential for abuses in using AEDMs. A manufacturer

[[Page 25106]]

could, for example, validate an AEDM based on comparison of AEDM 
results and test results for a group of basic models that consists of a 
high-selling model for which the AEDM produces a rating five percent 
above results from testing, and low-selling basic models, 
unrepresentative of those generally sold by the manufacturer, that the 
AEDM under-rates by off-setting amounts. As the CEC indicates, in such 
a situation the proposed rule would not preclude the manufacturer from 
using the AEDM result to rate the high-selling basic model at a level 
five percent above the level of the test results for that basic model. 
In addition, the manufacturer's use of the AEDM to calculate the 
efficiency of other relatively high-selling basic models could result 
in their being overrated as well. Such overrating could cause 
substantial sales in violation of Federal energy conservation 
standards, and result in substantially more energy use than the 
standards contemplate.
    No evidence presented thus far in this proceeding contradicts the 
Department's reason for proposing to allow AEDMs, namely that the 
potentially large number of basic models for commercial equipment 
warrants use of AEDMs to mitigate the test burden on manufacturers. 64 
FR at 69604. Thus, the Department is not inclined to require, as CEC 
suggested, that AEDMs always produce the same results as testing. This 
would virtually eliminate their use, since it is extremely difficult to 
develop an analytical model which has that degree of accuracy.
    The DOE is considering, however, adoption of alternatives to some 
of the proposed provisions concerning AEDMs in order to address the 
other issues that CEC raised and the concerns discussed above that the 
Department now has about these provisions. Several of these 
alternatives concern the requirements for validating AEDMs and are 
designed to address concerns about accuracy in the initial ratings of 
covered equipment. The use of an AEDM to determine the energy 
efficiency of a basic model of covered equipment is already one step 
removed from an actual measurement of that equipment, and it is 
essential that the AEDM produce a reliable result.
    First, the Department is considering a requirement that VICP 
participants validate their AEDMs by comparing test results and AEDM 
results for three or more basic models, as the NOPR proposed for non-
participants. This is an alternative to the proposal that VICP 
participants validate their AEDMs by comparing results for one or more 
basic models. Mathematical or computer-based simulations, such as 
AEDMs, are most reliable when validated over a range of conditions, 
rather than for one condition. When a manufacturer validates an AEDM 
for only one basic model, applying the AEDM to other models is an 
extrapolation of that single basic model, with an uncertain 
reliability. By contrast, validation of an AEDM by reference to three 
basic models would encompass a range of conditions, and establish its 
accuracy over a wider range of variables. This would help ensure that 
each AEDM accurately reflects variations among the basic models it 
covers. Three validation points is also the minimum number needed to 
establish or verify a simulation that reflects a non-linear correlation 
among variables. This is the most common correlation among variables, 
including those that affect the efficiency of equipment. In sum, 
requiring VICP participants to validate AEDMs using three basic models 
rather than one should permit more accurate verification of their 
AEDMs, should improve the accuracy of their AEDM results, and would 
still limit the testing burden because DOE would not be requiring 
testing for many basic models. Although verification testing would 
provide an incentive to VICP participants to use accurate AEDMs, this 
incentive might not offset the risk that use of AEDMs validated by 
reference to a single point would result in inaccurate initial 
equipment ratings. Finally, given the greater risk of inaccurate 
ratings from use of a single validation point, the Department believes 
it may be unreasonable to allow VICP participants to use only one 
validation point while requiring non-participants to use at least 
three.
    Second, the Department is considering a requirement that, for any 
basic model used to validate an AEDM, the predicted efficiency 
calculated from applying the AEDM must be within two percent of the 
test results for that basic model, instead of five percent as proposed 
in the NOPR. Adoption of today's proposal would mean that an AEDM could 
have a range of error of no more than four percent, and a potential for 
overrating of two percent. For ratings derived from testing, the 
Department is proposing that the rating must either have approximately 
a 95-percent degree of confidence (for non-VICP participants) \3\ or be 
generated by methods that give reasonable assurance that it does not 
exceed the mean for the population of the equipment (for VICP 
participants). Given these requirements, the NOPR proposal to allow an 
AEDM to have an error of five percent for the validation points could 
provide too much potential for an AEDM to produce erroneous results. To 
reduce this possibility, the AEDM should be as accurate as practicable 
for the validation points. A tolerance band of 2 percent 
appears sufficient to allow for a reasonable amount of measurement 
uncertainty and modeling error.
---------------------------------------------------------------------------

    \3\ This confidence limit requirement would not permit a 
manufacturer to rate any equipment at a higher efficiency or lower 
energy use than the mean of test measurements for that equipment. 
The requirement would not, for example, provide a five-percent 
``tolerance'' that would allow a model to be rated five percent 
above test results. Rather the requirement that a rating be at or 
above the 95-percent confidence limit is a statistical test as to 
the accuracy of a rating, and would sometimes require a manufacturer 
to rate equipment below the level of the mean of the test sample.
---------------------------------------------------------------------------

    Third, DOE is considering a requirement that the basic models a 
manufacturer uses to validate an AEDM must be the manufacturer's 
highest-selling basic models to which the AEDM could apply. Such a 
requirement would reduce the likelihood that a manufacturer could 
validate an AEDM using low-sales-volume equipment and then apply it to 
high-sales-volume equipment, and would prevent a manufacturer from 
meeting the validation requirements for average accuracy by overrating 
a high-selling basic model and under-rating of one or more low-selling 
models. It would also give greater assurance that each manufacturer's 
AEDM(s) would represent the characteristics of equipment it commonly 
sells.[FEDREG][VOL]*[/VOL][NO]*[/NO][DATE]*[/
DATE][PRORULES][PRORULE][PREAMB][AGENCY]*[/AGENCY][SUBJECT]*[/
SUBJECT][/PREAMB][SUPLINF][HED]*[/HED]
    Fourth, DOE is considering the option of requiring that a 
manufacturer, for any basic model it tests in order to validate an 
AEDM, rate the efficiency of that basic model using the test results 
(not AEDM results). This would preclude a manufacturer from using an 
AEDM to rate equipment at a higher level than the validation test 
results permit. The proposed rule was not intended to give a 
manufacturer a choice between using existing AEDM and test results. 
Rather, the purpose of allowing use of an AEDM to calculate efficiency 
is to relieve the undue burdens DOE understood would result from a 
requirement that manufacturers do efficiency testing on every basic 
model of commercial HVAC and WH equipment. Thus, there is no 
justification for permitting a manufacturer to use an AEDM to rate a 
basic model for which it has already determined the efficiency rating 
through testing.
    This requirement, in combination with the requirements the 
Department is considering that all manufacturers use at least three 
basic models to validate each of their AEDMs, and use the highest-
selling basic models to which

[[Page 25107]]

the AEDM could apply, would have the effect of requiring that a 
manufacturer rate its three highest-selling basic models based on 
testing rather than use of AEDMs. This would help ensure more accurate 
ratings for the high-selling models. Requiring a manufacturer to rate 
only the highest-selling basic models based on testing would still 
allow the intended benefit from the use of AEDMs because lower-selling 
basic models are relatively numerous, and therefore represent a 
substantial testing burden.
    Fifth, because the Department is also concerned about the general 
potential for manipulating AEDMs to overrate equipment, DOE is 
considering the addition of general language to its regulations to 
prohibit a manufacturer from knowingly using an AEDM to overrate the 
efficiency of a basic model. For example, this provision would preclude 
a manufacturer from using an AEDM, after a basic model has been tested, 
to create a higher rating than is warranted by the test results.
    The Department is proposing several changes to the regulation 
language in the NOPR, to implement the foregoing five proposals. As 
presented in this SNOPR, DOE proposes to include a new Sec.  431.481(c) 
and deletion of proposed Sec. Sec.  431.482(c) and 431.483(b)(1). The 
new paragraph would require a manufacturer that uses an AEDM under this 
subpart to validate it as follows: (i) Using the AEDM, the manufacturer 
must calculate the efficiency of three or more of its basic models, 
which must be the manufacturer's highest-selling basic models to which 
the AEDM apply; (ii) the manufacturer must test each of these basic 
models in accordance with Sec.  431.481(b) of this subpart, and either 
Sec.  431.482(b) or 431.483(a), whichever is applicable; and (iii) the 
predicted efficiency calculated for each such basic model from 
application of the AEDM must be within two percent of the efficiency 
determined from testing that basic model, and the average of the 
predicted efficiencies calculated for the tested basic models must be 
within one percent of the average of the efficiencies determined from 
testing these basic models.
    The DOE also proposes to add language to proposed Sec.  431.481(a) 
to provide that a manufacturer must determine and rate the efficiency 
of a basic model from test results if it has tested that basic model to 
validate an AEDM. In addition, DOE would add a new paragraph (4) to 
Sec.  431.481(c) that would prohibit a manufacturer from knowingly 
using an AEDM to overrate the efficiency of a basic model.
    The Department is also considering, and requests comment on, a 
number of other alternatives to the NOPR's proposals on AEDMs. With 
regard to validation of an AEDM, the Department is concerned about 
whether the permissible deviations it is considering between test 
results and AEDM results are at the proper levels. In addition to 
considering the allowance of a two-percent deviation for any single 
basic model used to validate an AEDM, as set forth above, and five 
percent as proposed in the NOPR, the Department is also considering 
whether some level between those figures is more appropriate. The DOE 
also is concerned that these levels and the one-percent average 
deviation for all basic models used to validate an AEDM, may be too 
generous and may underestimate the levels of accuracy an AEDM can 
achieve. Therefore, DOE is also considering adoption of an average 
permissible deviation between test and AEDM results of 0.5 percent, 
instead of the one percent proposed in the NOPR, with a maximum 
permissible deviation of one percent for any given basic model.
    With regard to the proposal to prohibit a manufacturer from 
knowingly using an AEDM to overrate equipment, the Department is 
concerned that other ways may exist in which a manufacturer seeking to 
evade energy conservation requirements under EPCA could misuse an AEDM. 
For example, a manufacturer might use an AEDM that provides accurate 
ratings for the models used for validation, but overrates other models. 
Thus, as an alternative to the proposed general language to prohibit 
use of an AEDM to overrate equipment, the Department is considering 
broader language that would prohibit ``using an AEDM to circumvent 
applicable requirements.''
    As previously stated, the effect of certain alternative options 
described in this notice would be to require each manufacturer to 
determine from testing the efficiency ratings of at least its three 
highest-selling basic models. The Department is concerned that such a 
requirement might be viewed as arbitrary, since it would apply to each 
manufacturer regardless of its size and the number of basic models it 
produces. The Department's reason for proposing to allow use of AEDMs--
to reduce the testing burden on manufacturers that produce numerous 
basic models of commercial HVAC and WH equipment--cuts two ways in this 
respect. First, it could support requiring each manufacturer to perform 
a uniform, minimum amount of testing, and as a result allowing 
manufacturers of large numbers of basic models to use AEDMs to rate a 
larger proportion and number of their models. But second, it could also 
support requiring each manufacturer to test the same proportion of its 
basic models, with manufacturers of large numbers of basic models 
testing more models than manufacturers of fewer basic models. This 
would still reduce the test burden of manufacturers of larger numbers 
of models far below what it would be if DOE prohibited use of AEDMs. 
Moreover, it might be unreasonable for the Department to require in 
effect that the three highest-selling basic models be tested, for 
example, by both a firm for which those basic models constitute forty 
percent of production and a firm for which they are ten percent of 
production. For these reasons, DOE is also considering adoption of one 
or more of the following approaches for a manufacturer to follow in 
testing its highest selling basic models: (1) A manufacturer would 
determine from testing the ratings for some minimum proportion of its 
total number of basic models, (2) a manufacturer would determine from 
testing the ratings of basic models that account for some minimum 
proportion of its sales, or (3) a manufacturer would determine from 
testing the rating of each basic model that exceeds a certain 
percentage of its overall sales. For any of these approaches it adopts, 
the Department would specify the applicable proportion or percentage in 
the final rule. The Department is undecided as to what these figures 
would be, but is considering a proportion in the range of one-third to 
two-thirds and 15 to 40 percent for the first and second approaches, 
respectively, and three to ten percent for the third. The Department 
specifically requests comment on this issue.

B. Voluntary Industry Certification Programs (VICPs)

1. Background
    As discussed in more detail in the NOPR, the VICP is a voluntary 
program (usually run by a trade association) that collects, 
disseminates and verifies information as to the performance of one or 
more types of equipment. 64 FR at 69603. The Department proposed that 
manufacturers could participate in DOE-approved VICPs to help assure 
that the manufacturers' efficiency ratings are accurate and comply with 
applicable requirements. The DOE also proposed the features that a VICP 
would need to have in order to receive DOE approval. The program would 
have to include, for example, collection and dissemination

[[Page 25108]]

of efficiency ratings for each basic model of equipment, periodic 
testing of each basic model to determine the accuracy of the 
manufacturer's efficiency rating for the model, action when a 
manufacturer's rating was inconsistent with the test results, and 
reporting of certain information to DOE. The NOPR also addressed how 
the organization operating a VICP could obtain DOE approval of the VICP 
and the duration of that approval.
    Sections B.2. through B.5., which follow, concern elements that the 
organization operating the VICP would have to include in the VICP in 
order to receive approval for the VICP from DOE. Section B.5. also 
addresses the proposed requirement that the organization operating an 
approved VICP must report changes in its program to the Department.
2. General Standards for Testing by a VICP
    The NOPR proposed that verification testing under the VICP meet 
``industry standards for the accuracy * * * of rating results.'' 64 FR 
at 69613. A similar provision applicable to manufacturer testing, is 
discussed in section II.A.2. above. The GAMA indicated that DOE should 
explain what is meant by ``industry standards'' in this context. (GAMA, 
No. 3 at 6) For the reasons discussed in section II.A.2, the Department 
is proposing adoption in the final rule of language on VICP observance 
of industry standards in verification testing that is virtually 
identical to the revised language it is considering for manufacturer 
testing. That language, which would replace proposed section 
431.484(a)(8), is as follows:
    The program's verification testing meets industry standards for the 
measurement accuracy of testing for the equipment being tested. This 
includes accuracy requirements in applicable test procedures, accuracy 
achieved by laboratory-grade equipment, and the accuracy of calibration 
standards.
3. Determining the Validity of Manufacturers' Efficiency Ratings
    Section 431.484 of the proposed rule would require a VICP to have 
``an appropriate standard'' for determining whether a manufacturer's 
claimed efficiency rating for a product is valid. 64 FR at 69613. This 
provision concerns two facets of verification of manufacturers' ratings 
under a VICP. First, it applies to the method (such as a sampling plan) 
by which the organization operating the VICP determines a basic model's 
efficiency from the verification testing it has conducted. Second, it 
applies to the criteria (such as tolerances) that the organization 
operating the VICP uses when it compares the manufacturer's rating for 
a basic model to the efficiency that the organization has determined 
under the VICP, to decide whether the manufacturer's rating is valid. 
The provision requires the use of methods and criteria that are 
sufficiently rigorous so as to give reasonable assurance that any 
rating the organization finds valid under the VICP would, on average, 
apply to all units of the model. The Department is concerned that an 
``appropriate standard'' test for determining the validity of 
manufacturers' ratings may be overly vague, and that organizations 
seeking approval from DOE of VICPs under the regulations might not 
understand that these concepts are implicit in the rule and might 
submit inadequate programs to DOE.
    The Department also expressed concern in the NOPR that 
manufacturers, knowing the criteria used under the VICP to verify the 
accuracy of their efficiency ratings, might systematically overrate 
their equipment. 64 FR at 69605-06. Typically, the organizations 
operating the VICPs currently test one or at most two units when doing 
verification testing of a basic model under a VICP. If the efficiency 
measured from the single unit, or from the average of the two units, is 
within a set percent (such as five percent) of the manufacturer's 
rating for the basic model, the organization operating the VICP accepts 
the manufacturer's rating as valid. To address the possibility that 
manufacturers participating in a VICP might systematically overrate 
equipment by five percent or slightly less, so as to be able to pass 
verification testing while claiming a higher rating than is warranted, 
the Department proposed to require the organizations operating the 
VICPs to submit to the Department annually summary data on verification 
test results under the VICP and the ratings of tested models. The 
Department could then take action with respect to a particular VICP if 
it appeared that systematic overrating of equipment covered by that 
VICP had occurred. The Department is concerned that this approach might 
address any overrating only prospectively and might be insufficient to 
deter VICP participants from overrating their equipment.
    To address these concerns, the Department is considering two 
additions to the proposed rule. First, it is considering additional 
language to clarify what would constitute an ``appropriate standard'' 
under a VICP for determining the validity of manufacturers' efficiency 
ratings. Second, DOE is considering the option of adding criteria for 
DOE approval of any VICP that would find a manufacturer's rating for a 
basic model valid when the verification test results are within a given 
percentage of the rating. These criteria would require that the VICP 
include the specific percentage(s) used, that the size of each 
percentage relate to the equipment to which it applies, and that the 
organization operating the VICP revise its program if, during any 
calendar year, it finds valid manufacturer ratings that average more 
than one percent above the verification test results under the VICP.
    Therefore, the Department is proposing substitute language for 
proposed Sec.  431.484(a)(9) of the NOPR. The DOE solicits public 
comment on this alternative proposed language.
    The Department is also considering, and seeks comment on, other 
options to assure that VICPs operate under appropriate standards for 
determining whether manufacturers' efficiency ratings are valid. For 
the efficiency figure from verification testing of a basic model under 
the VICP, DOE is considering a requirement that such figure must be 
valid at the 95-percent confidence limit, or at some other fixed 
confidence limit based on the inherent manufacturing variability or 
measurement uncertainty for the equipment in question. If the 
manufacturer's rating were higher than that, the organization operating 
the VICP would have to find the rating invalid. (This is the same 
approach that would apply to testing by non-VICP participants.) For 
comparison under the VICP of the performance from verification testing 
with the manufacturer's rating of a basic model, the Department is also 
considering a requirement that, where the measurement under the VICP is 
below the manufacturer's rating (or above for an energy use rating), 
the organization operating the VICP must require the manufacturer to 
justify its rating. Absent a satisfactory justification, the 
manufacturer's rating would be invalid under the VICP. A satisfactory 
justification would have to be based on other measurements of the 
model's efficiency, to show either or both of the following: (1) The 
manufacturer's rating is valid at the 95-percent confidence limit, or 
at some other fixed confidence limit based on the inherent 
manufacturing variability or measurement uncertainty for the equipment 
in question (this would be

[[Page 25109]]

the same approach applicable to testing by non-VICP participants); (2) 
the verification test results fall within the lesser of two standard 
deviations or 95 percent of the manufacturer's rating.
    The Department is considering the types of verification 
requirements described in the previous paragraph for several reasons. 
First, they might provide greater assurance than is provided by the 
proposals in the NOPR, or above in this notice, that organizations 
operating VICPs would use rigorous standards to verify manufacturer 
ratings. Second, although certification testing requirements for VICP 
participants would still be less stringent than for non-participants, 
such requirements might ensure that participants and non-participants 
would be subjected to the same type of standard. And finally, these 
proposals would provide clearer criteria for DOE to use in its 
determination of whether to approve a VICP.
4. Manufacturer Challenges of Equipment Ratings
    The CEC suggested that the Department add as a condition of its 
approval that each VICP include a provision allowing a manufacturer to 
challenge ratings by other manufacturers. (CEC, No. 7 at 6). It is 
DOE's understanding that, as stated by CEC, the existing program of the 
Air-Conditioning & Refrigeration Institute (ARI) has long allowed for 
such challenges. The possibility of such challenges may deter 
overstatement of efficiency ratings, and therefore the Department is 
proposing to add to the final rule the following conditions set forth 
in proposed Sec.  431.484(a) for DOE approval of a VICP:

    The program contains provisions under which each participating 
manufacturer can challenge ratings submitted by other manufacturers, 
which it believes to be in error.
5. VICP Reporting to the Department
    As indicated above, in the NOPR the Department proposed that each 
organization operating a VICP would have to report to DOE annually on 
verification testing results under the VICP. Another proposed condition 
of DOE approval of a VICP is that each basic model covered by a VICP be 
tested under the program at least once every five years. To enable the 
DOE to monitor compliance with this latter requirement, the Department 
is considering, and seeks comment on, a requirement that each 
organization operating a VICP report to DOE annually the model numbers, 
organized by type of equipment and manufacturer, covered by the basic 
models it has tested during the previous twelve months.
    Addressing the duration of DOE's approval of VICPs, proposed Sec.  
431.484(b) provides as follows:

    Approval will remain in force for five years, unless material 
changes occur in the program. In the event of changes, the VICP must 
promptly notify the Department, which may then rescind or continue 
the approval.

    The Department designed the second of these sentences to require 
the organization operating any DOE-approved VICP to ``notify the 
Department'' immediately whenever the organization made any changes in 
its program, so as to allow the Department to evaluate the changes and 
to rescind approval of the program if such changes were material. 
Because the word ``promptly'' might be considered vague, and given the 
obvious importance to DOE of immediate receipt of information as to any 
changes in an approved VICP, the Department is proposing inclusion of 
the following sentence in the final rule, in place of the second 
sentence just quoted:

    If the organization operating an approved VICP makes any changes 
in its program, the organization must notify the Department of such 
changes within 30 days of their occurrence, and the Department may 
then rescind or continue its approval.

C. Enforcement by the Department

1. Enforcement Testing--General
    Although most of the NOPR's proposed enforcement provisions are 
very similar to those currently in 10 CFR parts 430 and 431 (for 
consumer appliances and electric motors, respectively), the proposals 
for enforcement testing of commercial HVAC and WH equipment deviate in 
a few significant respects from the enforcement testing provisions now 
in those parts. The Department proposed in the NOPR to test initially 
two units of a basic model to determine its compliance with the 
applicable energy conservation standard, except that under certain 
circumstances DOE would test one unit. 64 FR at 69616. The proposed 
rule also provides that DOE would find the model to be in compliance if 
the average result for the two tested units (or the result from testing 
a single unit) is 95 percent or more of the applicable efficiency 
standard, or 105 percent or less of an energy use standard. 64 FR at 
69617. If the test results are outside the five-percent tolerance, and 
would thereby result in a determination of non-compliance, a 
manufacturer could elect to have DOE test one or two more units. The 
Department would then determine whether the model was in compliance by 
averaging the results from both rounds of testing, and then applying 
the five-percent criterion. By contrast, parts 430 and 431 contemplate 
an initial round of enforcement testing of a minimum of four or five 
units, and a maximum of 20, as well as application of sophisticated 
statistical tests to determine whether the test results establish that 
the basic model is out of compliance.
    In their comments, CEC and the Oregon Office of Energy (OOE) assert 
that the proposed five-percent criterion provides insufficient 
assurance of compliance, stating that it would allow a model to be 
found in compliance even if each sample unit tested at a level below 
the minimum standard. (CEC, No. 7 at 6-7 and 8-9, Tr.\4\ 139, 140-41; 
OOE, Tr. 138, 141, 144) Upon further review of the proposed provisions 
for enforcementp testing, DOE believes this concern has substantial 
merit. In addition, by allowing a basic model to pass so long as the 
test results were no more than five percent below the standard, this 
provision appears to be considerably more lenient than part 430, 
particularly in instances where the spread in test results is small. 
The proposed methodology and much smaller sample sizes might also 
provide much less accurate results and a greater possibility of errors 
than the methodology in part 430.
---------------------------------------------------------------------------

    \4\ ``Tr.'' followed by a number or numbers, refers to a page or 
pages in the transcript of the January 2000 hearing.
---------------------------------------------------------------------------

    The CEC and OOE seem to be advocating that the Department revise 
the enforcement testing proposal to provide that a basic model would be 
found in compliance only if the mean of the model's enforcement testing 
results meets or exceeds the applicable standard. The Department is not 
inclined to adopt this approach because it could create too great a 
risk of erroneously finding a manufacturer out of compliance. As long 
as the mean of all units of a basic model (the ``population'') met or 
exceeded the minimum standard, the basic model would be in compliance 
with the regulations. From a statistical standpoint, for any given 
basic model with a normal distribution of performance, half of the 
units produced will perform better than the mean for the population of 
all units and half will perform worse. Thus, if the mean performance of 
the population were at the standard level, the basic model would be in 
compliance but half of its units would be expected to perform above the 
standard and half below, and

[[Page 25110]]

there would be a 50-percent chance that the mean of a test sample would 
be below the standard. If the DOE's enforcement rules were to provide 
that a basic model would be found in compliance only if the mean 
performance of the test sample was at or above the applicable standard, 
the Department would have a 50-percent chance of finding equipment out 
of compliance even if the mean of its entire population meets the 
standard. The Department is reluctant to adopt rules that would entail 
such a large risk of an incorrect decision of noncompliance, since such 
a decision would require a manufacturer to discontinue distribution of 
the equipment and subject the manufacturer to other remedial actions 
and penalties.
    The Department did not incorporate part 430's enforcement testing 
provisions into the proposed rule because of the significant 
differences between consumer products and commercial equipment. Each 
manufacturer of a consumer appliance tends to produce a relatively 
small number of basic models, each in a relatively large quantity. The 
size of the product, as well as the cost of each unit, tend to be lower 
than commercial equipment. At any time, a sufficient number of units of 
any residential equipment model will likely be available to allow 
sample sizes to be large. Thus, part 430 uses a statistical method that 
is more rigorous than would be possible with smaller sample sizes. 
Specifically, the method of part 430 is based on a double sample, with 
a maximum sample size of 20 units. The size of the combined sample 
provides a 95-percent confidence level in the accuracy of the sample 
mean. Under this method, the Department computes an efficiency level 
that constitutes a lower control limit. This level is based on the 
applicable standard, the test sample measurements, and the variance 
among these measurements, but can be no lower than five percent below 
the standard. As long as the sample mean is at least equal to the lower 
control limit, DOE considers the basic model to be in compliance.
    This approach helps to avoid false negative determinations (i.e. 
erroneously finding a basic model out of compliance). By allowing a 
finding of compliance in some instances where the sample mean of a 
basic model is slightly lower than the standard, it takes into account 
situations where the sample mean may be below the standard even though 
the population of the product is not. On the other hand, the rigorous 
statistical basis for the enforcement determination promotes accurate 
ratings by manufacturers, and provides some control of overrating. This 
is because the enforcement methodology creates a substantial risk for a 
manufacturer of a finding of non-compliance where it produces a basic 
model that clearly fails to meet the applicable standard.
    On the other hand, it is the Department's understanding that each 
manufacturer of commercial HVAC and WH equipment tends to produce a 
large range of models, many of which it produces in small quantities. 
Purchasers often select a model from a catalog to suit a specific 
application, and some models are manufactured only on order. Commercial 
equipment is more costly in general, and may also be quite large in 
size. Although not all of these factors apply to every model of 
commercial HVAC and WH equipment, the enforcement regulations need to 
take these market characteristics into account. Thus, sample sizes of 
up to 20 units, as provided in part 430, would generally be prohibitive 
for commercial HVAC and WH equipment, and enforcement testing 
provisions for this equipment must accommodate a sample size as small 
as one. The NOPR proposals to test initially two units and to find a 
basic model of equipment in compliance if test results were within five 
percent of the applicable standard, were a response to these concerns. 
But for the reasons stated above, the Department is now reconsidering 
whether these proposals are the best approach for addressing the 
characteristics of commercial equipment.
    As an alternative to these proposals, the Department is now 
considering for commercial HVAC and WH equipment an enforcement testing 
approach resembling that in part 430. This approach would approximate 
the statistical method used there, using smaller sample sizes. Compared 
to the NOPR proposal, the sample sizes would generally be larger, DOE 
would do more tests, and the pass/fail criterion would be more 
stringent. The Department believes this approach would provide more 
accurate results than the proposed method, and reduce the possibility 
that DOE might erroneously find a basic model to be in or out of 
compliance. It would serve the goals of providing a fair and accurate 
determination of the energy efficiency (or use) of the model being 
tested, and of fairly balancing the manufacturer's risk of being 
falsely found to be non-compliant with the risk to the consumer of a 
false finding of compliance. As with the NOPR's proposal, the sample 
sizes would be consistent with the constraints imposed by the volume 
and nature of commercial HVAC and WH equipment. Thus, the Department's 
new approach would serve the goals of being neither unduly burdensome 
nor excessively time-consuming or expensive to conduct.
    The specifics of the approach the Department is now proposing are 
as follows. First, DOE would generally test four units of a basic 
model, but would test fewer if only a lesser number were available or 
if testing of such lesser number were otherwise warranted. (The 
circumstances under which DOE would test fewer than four units are 
discussed below.) If DOE were to test three or four units, it would 
test each unit once; if it tested two units it would test each twice; 
and if it tested one unit it would test that unit four times. Second, 
DOE would compute the mean of the test results, as provided in the 
NOPR, but would also calculate a lower control limit. The lower control 
limit would be the greater of either: (1) 97.5 percent of the 
applicable energy efficiency standard, or (2) the applicable energy 
efficiency standard minus the product of the sample standard error and 
the t-value for a 97.5-percent, one-sided confidence limit. The sample 
standard error would be the same as in part 430 (Appendix A to subpart 
F, steps 3 and 4). (For an energy use standard, DOE would calculate an 
upper control limit, which would be the lesser of either 102.5 percent 
of the applicable standard, or the standard plus the product of the 
sample standard error and the t-value for a 102.5-percent, one-sided 
confidence limit.) Third, a basic model would be in compliance only if 
the mean measurement for the sample meets or exceeds the lower control 
limit in the case of an efficiency standard or is less than or equal to 
the upper control limit in the case of an energy use standard.
    From the standpoint of statistical accuracy, testing more units of 
a basic model and conducting multiple tests on each model would provide 
greater accuracy and less chance of making an error in a compliance 
determination. Concerns over the testing burden and availability of 
test units, however, limit the number of tests that DOE can reasonably 
require for commercial equipment. Thus, some compromise must be 
reached. A test sample size of four units would at least allow the 
statistical calculations to provide the basis for evaluating confidence 
limits, and would equal the minimum sample size in part 430. In cases 
where four units are not available, testing three would still allow 
confidence limits to be determined, as would making multiple 
measurements of one or two units. Multiple measurements of a single 
unit

[[Page 25111]]

would not incorporate the effects of equipment variability, but would 
help account for the effects of measurement uncertainty. The 
determination of a control limit based on confidence limits would allow 
for some tolerance to avoid falsely finding a basic model to be out of 
compliance, but still encourage manufacturers to accurately rate their 
equipment.
    The Department believes that using 97.5- and 102.5-percent, one-
sided confidence limits, and allowing the mean of the enforcement test 
sample to be a maximum of 2.5 percent below the applicable standard, 
would provide sufficient tolerances to reflect the normal manufacturing 
and measurement variability that might affect sample units for the 
equipment involved here. The ARI and GAMA operate VICPs to verify 
manufacturer efficiency ratings of residential and commercial air 
conditioning equipment and water heaters, respectively. The ARI finds a 
rating valid if it is no more than five percent above the results of a 
single verification test ARI performs, or above the average of two 
tests if the first test result is more than five percent below the 
rating. The GAMA uses the same approach, but with an allowed deviation 
of two percent for commercial equipment and 3.5 percent for residential 
products. In addition, under today's proposal, the initial round of DOE 
enforcement testing would typically involve four units, or three or 
four tests, and, as discussed below, several more tests could result 
from manufacturer option testing. Because this approach involves more 
than the one or two tests performed by ARI and GAMA, it would involve 
much less risk that the sample test results will be below the mean of 
the population. For these reasons, DOE believes that although the five-
percent figure proposed in the NOPR for enforcement tolerances is 
appropriate in the context of part 430's methodology for consumer 
products, for the equipment here and for the methodology DOE is now 
considering a 2.5-percent tolerance seems reasonable. Moreover, use of 
the 2.5-percent figure rather than five percent would create less of an 
incentive for manufacturers to produce equipment with high variability 
in order to obtain a greater tolerance during enforcement testing. 
Nevertheless, DOE encourages interested parties to provide to the 
Department, in response to this notice, any data they have that 
indicates a tolerance other than 2.5 percent might be warranted for any 
or all of the equipment involved in this proceeding.
    As indicated, the above-described approach for enforcement testing 
would allow the number of units tested to vary depending on the 
circumstances. The same is true to some extent of the proposal in the 
NOPR, which provides that DOE would initially test two units of a basic 
model to determine its compliance, except in two situations. First, the 
Department proposed to test only one unit, and base the compliance 
determination on that test, if that is the only unit available for 
testing. Second, if a basic model is very large or has unusual testing 
requirements, DOE proposed to allow itself the discretion to test only 
one unit upon a manufacturer's request supported by sufficient 
justification. 64 FR at 69616. The GAMA advocated expansion of the 
second exception to include situations where a manufacturer 
demonstrates limited availability of a basic model because it has a low 
sales volume or is produced only for special orders. (GAMA, No. 3 at 8, 
Tr. 120)
    The GAMA's concern would seem to be covered by the first exception, 
which would address any situation, including low sales volume or 
limited production of a basic model, that resu