[Federal Register Volume 61, Number 223 (Monday, November 18, 1996)]
[Notices]
[Pages 58711-58715]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 96-29459]


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


NUCLEAR REGULATORY COMMISSION

All Nuclear Power Plants; Issuance of Director's Decision Under 
10 CFR 2.206

    Notice is hereby given that the Director, Office of Nuclear Reactor 
Regulation, has taken action with regard to a Petition dated March 5, 
1996, by Mr. C. Morris. The Petition pertains to all operating nuclear 
power plants.
    In the Petition, the Petitioner requested that the operating 
licenses of all nuclear power plants be suspended within 90 days and 
remain suspended until such time as the licensees of those plants 
discovered the reason for what the Petitioner asserts are repeated 
errors in the undervoltage relay (UVR) setpoints (SPs) and electrical 
distribution system (EDS) designs and provided convincing evidence that 
these deficiencies had finally been corrected. Since the Petitioner had 
requested action within 90 days, the request was treated as a request 
for immediate relief. The Petitioner also requested that the 
aforementioned evidence be reviewed by a competent third party, in 
addition to the staff of the U.S. Nuclear Regulatory Commission (NRC), 
and that if the NRC concludes that plants may safely operate with UVRs 
that cannot be properly set for long periods, the NRC should reach 
these conclusions by way of a public meeting.
    The Director of the Office of Nuclear Reactor Regulation has denied 
the Petition. The reasons for this denial are explained in the 
``Director's Decision Under 10 CFR 2.206'' (DD-96-12), the complete 
text of which follows this notice and is available for public 
inspection at the Commission's Public Document Room, the Gelman 
Building, 2120 L Street, NW., Washington, DC.
    A copy of the decision will be filed with the Secretary of the 
Commission for the Commission's review in accordance with 10 CFR 
2.206(c) of the Commission's regulations. As provided by this 
regulation, the decision will constitute the final action of the 
Commission 25 days after issuance unless the Commission, on its own 
motion, institutes review of the decision in that time.

    Dated at Rockville, Maryland, this 26th day of September 1996.


[[Page 58712]]


For the Nuclear Regulatory Commission.
William T. Russell,
Director, Office of Nuclear Reactor Regulation.

Director's Decision Under 10 CFR 2.206

I. Introduction

    On March 5, 1996, Mr. Charles Morris (Petitioner) filed a Petition 
with the Executive Director for Operations pursuant to Section 2.206 of 
Title 10 of the Code of Federal Regulations (10 CFR 2.206). The 
Petitioner requested that the operating licenses of all nuclear power 
plants be suspended within 90 days and remain suspended until such time 
as those plants have (1) discovered the reason for what the Petitioner 
asserts are repeated errors in the undervoltage relay (UVR) setpoints 
(SPs) and electrical distribution system (EDS) designs and (2) provided 
convincing evidence that these deficiencies have finally been 
corrected. Since the Petitioner had requested action within 90 days, 
the request was treated as a request for immediate relief. The 
Petitioner also requested that the aforementioned evidence by reviewed 
by a competent third party, in addition to the Nuclear Regulatory 
Commission (NRC) staff, and that if the NRC concludes that plants may 
safely operate with UVRs that cannot be properly set for long periods 
of time, the NRC should reach these conclusions by way of a public 
meeting.
    On April 17, 1996, the Petitioner was informed that the request for 
the suspension of all nuclear power plant licenses within 90 days for 
the purposes of remedying repeated errors in UVR SPs and EDS designs 
was denied because licensees have, to a large degree, already addressed 
the issues which the Petitioner had raised. Also the Petitioner was 
informed that the request was being evaluated pursuant to 10 CFR 2.206 
of the NRC's regulations and that a decision, as provided by 10 CFR 
2.206, would be made on the request within a reasonable time.
    On the basis of my review of the issues raised by the Petitioner as 
discussed below, I have conclude that no substantial health and safety 
issues have been raised that would require the initiation of the action 
requested by the Petitioner.

II. Discussion

    In his Petition, the Petitioner stated his concern that the 
``enduring and widespread nature of the electrical distribution system 
(EDS) an undervoltage rely (UVR) setpoint (SP) errors (e.g., incorrect 
UVR and thermal overload setpoints) was recognized by neither the 
licensees nor the NRC staff,'' and was not included in NRC Information 
Notice (IN) 93-99, ``Undervoltage Relay and Thermal Overload Setpoint 
Problems.''
    IN 93-99 did, in fact, inform all holders of operating licenses or 
contruction permits of the widespread nature of the setpoint errors by 
listing approximately 40 licensees with incorrectly set UVRs or thermal 
overload (TOL) protective devices. The identification of these problems 
was not inadvertent, but was the result of concerted NRC staff 
attention to these issues. As was indicated to the Petitioner in an 
April 17, 1996, letter acknowledging receipt of his March 5, 1996, 10 
CFR 2.206 Petition, the Petitioner himself recognized that Electrical 
Distribution System Functional Inspections (EDSFIs) were highlighting 
these issues and that licensees were conducting self-initiated design 
basis reviews (possibly in anticipation of pending EDSFIs) to identify 
problems and were undertaking corrective actions.
    In his March 5, 1996, Petition, the Petitioner listed seven 
specific reasons that he believed caused repeated EDS and UVR 
deficiencies. The following is a description of each concern 
accompanied by the NRC staff's response:
    1. The Petitioner stated that NRC Branch Technical Position PSB-1, 
``Adequacy of Station Electric Distribution System Voltages,'' 
contained in NUREG-0800, ``Standard Review Plan for the Review of 
Safety Analysis Reports for Nuclear Power Plants,'' which requires a 
degraded voltage relay with a long delay and a loss of power relay with 
a short delay, is inadequate because it does not recognize the 
complexity of the matter. Except for the arbitrary time delays 
associated with the UVRs, no recognition has been made of voltage 
dynamics and time dependence. Signal bandwidths, responses of tap 
changing transformers, and UVR time delays have been overlooked and 
should be considered.

Response

    NRC Branch Technical Position PSB-1 does not recommend that 
licensees arbitrarily select time delays for UVRs. On the contrary, 
PSB-1 states that ``the selection of undervoltage and time delay 
setpoints shall be determined from an analysis of the voltage 
requirements of the Class 1E loads at all onsite system distributions 
levels.'' Further, it states that ``Tap settings selected should be 
based on an analysis of the voltage at the terminals of the Class 1E 
loads. The analyses performed to determine minimum operating voltages 
should typically consider maximum unit steady state and transient loads 
* * *'' Additionally, ``the first time delay should be of a duration 
that established the existence of a sustained degraded voltage 
condition (i.e., something longer than a motor starting transient)'' 
and ``the second time delay should be of a limited duration such that 
the permanently connected Class 1E loads will not be damaged.''
    Therefore, the staff concludes the NRC Branch Technical Position 
PSB-1 is adequate as it addresses those topics which the Petitioner 
believes are neglected by the Branch Technical Position.
    2. The Petitioner asserted that UVR tolerances are statistical in 
nature and not, as the staff and design engineers often regard 
them,limits to the errors in the relay setpoints. This is a significant 
problem which may not be solved if previous approaches are utilized and 
decision analysis is not applied to study the consequences of 
attempting to prevent the occasional loss of the most vulnerable safety 
load at the expense of transferring a complete division to another 
power source with attendant problems.

Response

    Regulatory Guide 1.105, ``Instrument Setpoints for Safety-Related 
Systems,'' states that ISA-S67.04-1982, ``Setpoints for Nuclear Safety-
Related Instrumentation Used in Nuclear Power Plants,'' establishes NRC 
staff guidance for ensuring that instrument setpoints in safety-related 
systems are initially within and remain within the technical 
specification limits. Section 4.3.1 of ISA-S67.04 states that 
instrument accuracies (uncertainties, errors or tolerances) may be 
combined in one of five ways: algebraically, square root of the sum of 
the squares, statistically, probabilistically, or combinations of the 
first four. Justification is to be provided for the method used.
    Regulatory Guide 1.105 expands upon this point:

    Paragraph 4.3 of the standard specifies the methods for 
combining uncertainties is determining a trip setpoint and its 
allowable values. Typically, the NRC staff has accepted 95% as a 
probability limit for errors. That is, of the observed distribution 
of values for a particular error component in the empirical data 
base, 95% of the data points will be bounded by the value selected. 
If the data base follows a normal distribution, this corresponds to 
an error distribution approximately equal to a ``two sigma'' value.

    Although the use of ``two sigma'' values (value equal to twice the 
standard deviation of the errors) does

[[Page 58713]]

not completely ensure that the measured parameter will not exceed the 
safety analysis limit without accompanying protective action, the 
probability of all the individual error occurring simultaneously at 
this extreme, non-conservative, random values is very low. Therefore, 
the regulatory guide and the industry standard together support a 
credible, statistical approach for establishing setpoints that 
considers such things as sample size of error values, random versus 
non-random errors, and independence of errors.
    The preparatory training for EDSFI team members also did not 
overlook the statistical nature of the UVR tolerances. In Section 4.8.2 
of the EDSFI training textbook, a discussion of instrumentation 
setpoint problems was provided with a sample application of ISA-S67.04 
to degraded voltage relays. This methodology was also discussed in the 
course itself. Using this knowledge EDFSIs were conducted and findings 
were written covering improper degraded voltage relay setpoints. As a 
result, licensees then followed this action with event notification and 
other activities as described in Information Notice 93-99.
    Additionally, in response to a request from Region III pertaining 
to an unanalyzed degraded voltage concern at Perry Nuclear Power Plant, 
the Electrical Engineering Branch (EELB) of NRR in an April 13, 1992, 
memo provided inspectors in NRC Regional Offices with guidance for 
establishing an adequate setpoint for the degrade voltage relays by way 
of reference to Section 4.8.2 of the EDSFI training course manual and 
Regulatory Guide 1.105. Furthermore, the staff informed all holders of 
operating licenses about a statistical approach for establishment of 
UVR setpoints when 91-29, ``Deficiencies Identified during Electrical 
Distribution Functional Inspections,'' made reference to ISA-S67-04-
1982 for useful guidance in determination of setpoints.
    The staff therefore has regarded the UVR setpoint determinations as 
statistical in nature.
    3. The Petitioner stated that although General Design Criterion 
(GDC) 17, ``Electric power systems,'' requires all EDS to be testable, 
only parts are tested because plants cannot conveniently be placed in a 
condition where actual loads can be placed on the EDS and measured.

Response

    The staff has already been aware that in certain situations it is 
not practical nor safe to test each and every component in the exact 
way it is used. General Design Criterion 18, ``Inspection and testing 
of electrical power system,'' states that ``systems shall be designed 
with a capability to test periodically * * * the operability of the 
systems as a whole and, under conditions as close to design as 
practical * * *.'' Regulatory Guide 1.118, ``Periodic Testing of 
Electric Power and Protection Systems,'' Revision 2, endorses, IEEE Std 
338-1977, ``Criteria for the Periodic Testing of Nuclear Power 
Generating Station Safety Systems,'' which states that ``the test 
program of each system shall be designed to provide for interference 
with related operational channels, systems, or equipment.'' It further 
states that `'wherever possible, tests shall be accomplished under 
actual or simulated operating conditions, including sequence of 
operations, for example, diesel load sequencing,'' but also

    Where it is not practicable to initiate the protective action, 
the system shall be designed such that * * * Designs * * * shall be 
justified on the basis that there is no practical system design that 
would permit operation of the actuated equipment without adversely 
affecting the safety or operability of the plant, and that the 
probability of failure of actuated equipment not tested during plant 
operation is acceptably low, and that the actuated equipment can be 
routinely tested when the plant is shut down.

    It is the staff's goal to have all components of the EDS 
periodically tested in a manner that is both reasonable and practical. 
Various practical test methods such as the use of miniflow paths, 
overlap testing, simulated loads, etc. have been found acceptable by 
the staff.
    NRC Temporary Instruction 2515/107 (which provided guidance for 
performing EDSFIs) required the EDSFI teams to ``verify that the 
surveillance and test procedures are adequate to demonstrate the 
functionality of the equipment or system being tested or the design 
assumptions being verified.''
    Therefore, as shown above, testing of the EDS is evaluated in terms 
of satisfying NRC requirements (GDC-17 and GDC-18) utilizing the 
guidance provided by Regulatory Guide 1.118 for a reasonable and 
practical approach (in lieu of testing each system as a whole), and 
tests are properly implemented in the manner described above.
    4. The Petitioner pointed out that load nameplate ratings are used 
in voltage analyses even when common knowledge shows that most loads 
are operated at a fraction of their ratings. Furthermore, worst-case 
ambient temperatures are used to select motor protection time delays 
even though few loads, if any, see those conditions except during a 
loss-of-coolant accident when the motor protection is bypassed. 
Additionally, UVR output delays are treated as known quantities, when 
the protection of loads by time delays and inverse time over current 
relays is a crude mitigating approach. As a related matter, the 
Petitioner objects to the inconsistent use of significant figures to 
represent EDS and UVR SP parameters.

Response

    The aforementioned temporary institution (TI) for the EDSFIs stated 
that the inspectors should verify that values for mechanical loads used 
for electrical calculations are based on actual system operating points 
during both normal and accident conditions. The staff expects licenses 
to perform accurate, conservative, and bounding calculations involving 
worst-case estimates for parameters such as ambient temperatures and 
loads. The licensees' analyses are reviewed by the staff utilizing 
engineering judgment and applicable industry guidance to ensure that 
reasonable, yet adequately safe solutions are provided.
    It is true that, occasionally, designs proposed by licensees do 
involve basic approaches (such as inverse time delay relays) and that 
some calculations performed by licensees involve the use of ultra-
precise numerical values. What the staff does require is that the 
designs utilized by licensees meet applicable NRC regulations and that 
adequate protection of public health and safety is ensured.
    The staff, therefore, concludes that component characteristics are 
treated and utilized properly in calculations that support EDS and UVR 
designs.
    5. The Petitioner believed that when licensees have discovered that 
UVR SPs are set too low, the typical response has been to raise the 
setpoints. This, in turn, reduces the safety advantage of providing 
UVRs for the EDS due to more frequent and unnecessary UVR actuations 
accompanied by possible undesirable power systems transfers.

Response

    In a letter dated August 8, 1979, addressed to all power reactor 
licensees regarding the adequacy of station electric distribution 
systems voltages, the staff stated that:

    Protection of safety loads from undervoltage conditions must be 
deigned to provide the required protection without causing voltages 
in excess of maximum voltage ratings of safety loads and without 
causing spurious separations of safety buses from offsite power.
Moreover,


[[Page 58714]]


    Voltage-time settings for undervoltage relays shall be selected 
so as to avoid spurious separation of safety buses from offsite 
power during plant startup, normal operation and shutdown due to 
startup and/or operation of electric loads.

NRC Branch Technical Position PSB-1 states that:

    * * *imporper (sic) voltage protection logic can itself cause 
adverse effects on the Class 1E systems and equipment such as * * * 
spurious separation of Class 1E systems from offsite power due to 
normal motor starting transients.

    Additionally, in IN 95-37, ``Inadequate Offsite Power System 
Voltages during Design-Basis Events,`` the staff informed power reactor 
licensees that although raising UVR setpoints ensures that adequate 
voltages exist at equipment input terminals, the higher setpoints also 
increase the potential for separation from the offsite power system 
during design-basis events over the range of normally anticipated 
offsite grid voltages.
    In a more specific example, a February 23, 1995, staff safety 
evaluation of the degraded voltage design for the Edwin I. Hatch 
Nuclear Plant, determined that combination of automatic and manual 
actions was an acceptable alternative approach to meet the branch 
technical position in lieu of raising the degraded voltage setpoints 
which could lead to unwanted plant trips. That safety evaluation and 
the above staff guidance provide evidence that the staff has considered 
avoidance of spurious bus trips as one objective to be considered when 
selecting an adequate setpoint of UVRs.
    The staff, therefore, has repeatedly and in detail both considered 
the determental effects of raising the UVR setpoints and communicated 
its concerns to licensees.
    6. The Petitioner stated that in IN 95-05, ``Undervoltage 
Protection Relay Settings Out of Tolerance Due to Test Equipment 
Harmonics,'' the staff discovered the peak reading voltmeters 
calibrated for root-mean-square (RMS) are affected by the proportions 
of harmonics in the AC bus voltages and in the calibrators used to set 
the UVRs. Additionally, the harmonics affect the UVR responses by 
changing their setpoints when the harmonic content of the bus voltage 
changes.

Response

    IN 95-05 discusses three occurrences, reported by licensees, where 
harmonics in the output voltage of the power supplies used during 
testing and calibration of UVRs resulted in the relay setpoints being 
out of tolerance. The setpoint errors were also affected by the use of 
digital voltmeters which do not respond to the harmonic content of the 
test input voltage as do the UVRs. The purpose of the IN was to inform 
all operating power plant licensees that harmonics in the voltage 
inputs (test source voltage or normal bus voltage) to the UVRs impact 
the actual operating points of those relays, as the Petitioner 
believes, and to instruct the licensees to take appropriate action 
(i.e., install filters, adjust setpoints, select proper test equipment, 
etc.) to ensure that UVR setpoints are adequate.
    The staff, therefore, has addressed this concern and brought it to 
the attention of licensees who are taking appropriate action as 
discussed above.
    7. The Petitioner concluded that impedances and inrush currents to 
motors and other loads are not known to the precision with which the 
staff and the licensees' engineers have been trying to set UVRs. Both 
groups must recognize that their task may be impossible and that their 
attempts to do so have increased the risk of a nuclear accident.

Response

    Branch Technical Position PSB-1 states that voltage analyses 
(including effects of impedances and inrush currents) should be 
performed with analytical techniques and assumptions verified by actual 
measurement. It also states that, in general, test results should not 
be more than 3% lower than the analytical results. This level of 
precision has been determined to be acceptable based on engineering 
judgment.
    Furthermore, as stated in the response to the Petitioner's fourth 
concern, even though licensee propose solutions involving different 
equipment and unique, precise calculations (which should be supported 
by actual test data as mentioned above), staff reviews are conducted 
utilizing both guidance from Branch Technical Position PSB-1 and 
engineering judgment to ensure that all applicable regulations are met 
and that adequate protection of public health and safety is ensured. 
This approach provides reasonable assurance that the level of risk of a 
nuclear accident is not increased and remains acceptable.
    Choosing a setpoint above an analytical limit based on minimum 
voltage requirements and below nominal votage ranges while accounting 
for instrumentation errors and analytical inaccuracies is often a 
challenge which leads licensees to use more precise equipment and more 
precise calculations. It is concerns such as these that have led the 
staff to consider alternative approaches to its position on degraded 
voltage protection on a plant-specific basis as noted above in the 
staff's response to the Petitioner's fifth concern.
    Therefore, although the staff has concluded that the task is not 
impossible, it has recognized alternative approaches that address 
degraded voltage concerns without increasing the risk of an accident.
    To continue the discussion, identification of problems with UVRs 
and EDSs was not inadvertent. The NRC staff had undertaken more global 
measures to ensure that concerns such as those raised by the Petitioner 
were addressed satisfactorily. Because previous NRC inspection teams 
had observed that the required functional capabilities of certain 
safety-related systems (including EDSs) were compromised due to a lack 
of proper engineering support and the introduction of various design 
deficiencies, EDFSIs were scheduled to be conducted for all operating 
plants beginning with pilot inspections in 1989. NRC Temporary 
Instruction (TI) 2515/107 was issued on October 19, 1990, to be made 
part of the NRC Inspection Manual. That TI stated that calculations to 
establish protective relay setpoints had not been initially performed 
or were not updated to reflect setpoint changes and plant 
modifications. These failures constituted some of the deficiencies that 
had been encountered by previous inspection teams. The TI stated, with 
regard to those concerns voiced by the Petitioner, that the forthcoming 
inspections should verify:
     That ratings and setpoints have been correctly chosen and 
controlled for protective and control relays and circuit breakers to 
assure proper coordination, protection, required automatic action, and 
annunciation.
     The adequacy of the load study, voltage profiles, voltage 
drop calculations, motor starting study, load shedding, engineered 
safety features (ESF) bus load sequencing and overload trip settings 
for ESF loads including consideration of steady-state and accident-
transient loads and consideration of acceleration of the loads during 
degraded voltage conditions that may occur during various modes of 
plant operation and accident mitigation scenarios.
     The adequacy of short circuit calculations, design of 
protective relay logic and relay setting calculations, grounding 
calculations and schemes, and protective device coordination studies.
     That setpoints for overcurrent protective relays are 
correctly chosen (1) to assure proper breaker coordination

[[Page 58715]]

between different voltage levels; (2) to prevent exceeding the vendor-
specified thermal limits on motors, containment electrical penetrations 
and cable insulation systems; (3) to allow starting of electrical 
equipment under degraded voltage conditions; and (4) to provide 
adequate pre-trip alarms, when applicable.
     The adequacy of setpoints and time delays for other 
protective relays for attributes such as undervoltage, underfrequency, 
reverse power, ground faults, differential current, thermal overload 
and phase synchronization to assure functionality of the EDS.
     That mechanical loads, such as pump horsepower, correspond 
to actual system operating points during normal and accident conditions 
and have been correctly translated to electrical loads and incorporated 
in the electrical load list as appropriate.
     That surveillance and test procedures are adequate to 
demonstrate the functionality of the equipment or system being tested 
or the design assumptions being verified.
    NRC inspectors (including NRC contractors) assigned to the EDSFI 
teams attended a week-long course (held in September and December 1990) 
to enhance their knowledge of EDSs, the TI and related requirements. 
Using the guidance provided by the TI and the EDSFI training course, 
the EDSFI teams then conducted inspections of the EDSs through early 
1994 at most operating nuclear power plants. As a result, numerous 
deficiencies were identified and documented in plant-specific EDSFI 
inspection reports, and corrective actions were taken. Those corrective 
actions were subsequently evaluated, found acceptable by the staff and 
documented in follow-up inspection reports. Many of these deficiencies 
and corrective actions were listed in IN 93-99 and include incorrect 
UVR relay and thermal overload setpoints caused by design errors, as 
well as other points raised by the Petitioner.
    In summary, as stated in my April 17, 1996, letter, I believe the 
NRC staff recognized the existence of repeated errors and widespread 
EDS design deficiencies, including those associated with UVR SPs, took 
appropriate actions (conducted EDSFIs, identified deficiencies, 
required corrective actions) based on those observations, and made all 
licenses aware of typical design deficiencies encountered during EDSFIs 
and licensees' self-initiated efforts by issuing INs such as IN 91-29, 
``Deficiencies Identified During Electrical Distribution System 
Functional Inspections,'' its supplements, and IN 93-99. Additionally, 
the staff has continued to inform power reactor licensees of other 
design deficiencies when they are encountered (e.g., IN 95-37 which 
discusses UVR setpoints in relationship to inadequate offsite power 
system voltages during design-basis events) and will continue to do so 
in the future when necessary. Such action by the staff is appropriate 
to address repeated errors in UVR setpoints and EDS designs and to 
provide reasonable assurance of adequate protection of public health 
and safety.

III. Conclusion

    The institution of proceedings pursuant to 10 CFR 2.206 is 
appropriate only if substantial health and safety issues have been 
raised. See Consolidated Edison Co. of New York (Indian Point Units 1, 
2, and 3) CLI-75-8, 2 NRC 173, 175 (1975) and Washington Public Power 
Supply System (WPPSS Nuclear Project No. 2) DD-84-7, 19 NRC 899, 924 
(1984). This is the standard that has been applied to the concerns 
raised by the Petitioner to determine whether the action requested by 
the Petitioner, or enforcement action, is warranted.
    On the basis of the preceding assessment, I have concluded that no 
substantial health and safety issues have been raised by the Petitioner 
that would warrant the action requested by the Petitioner. I further 
conclude that the Petitioner's concerns have been adequately addressed 
by the staff and that there is no need for a third party review. 
Additionally, with regard to plants with UVRs that cannot be properly 
set, the staff has shown in plant-specific evaluations, such as 
described above, that other alternative designs are acceptable.
    The Petitioner's request for action pursuant to 10 CFR 2.206 is 
denied. As provided for in 10 CFR 2.206(c), a copy of the decision will 
be filed with the Secretary of the Commission for the Commission's 
review. The decision will constitute the final action of the Commission 
25 days after issuance unless the Commission, on its own motion, 
institutes review of the decision in that time.

    Dated at Rockville, Maryland, this 26 day of September, 1996.

For the Nuclear Regulatory Commission.
William T. Russell,
Director, Office of Nuclear Reactor Regulation.
[FR Doc. 96-29459 Filed 11-15-96; 8:45 am]
BILLING CODE 7590-01-M