ML18283A535

Affidavit of Jack R. Calhoun

ML18283A535
Person / Time
Site:	Browns Ferry
Issue date:	06/21/1976
From:	Calhoun J Tennessee Valley Authority
To:	Atomic Safety and Licensing Board Panel
References
Download: ML18283A535 (45)

Text

UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION

~4 3/ J94' Before the Atomic Safety and,Licensing, Board J

In the Matter of )

)

TENNESSEE VALLEY AUTHORITY ) Docket Nos

) 50-260 (Browns Ferry Nuclear Plant )

units 1 and 2) )

AFFIDAVIT OF JACK -R. CALHOUN Jack R. Calhoun, being duly sworn, deposes and says:

business address is Tennessee Valley Authority, 702 Edney Building, Chattanooga, Tennessee. I am employed by the Tennessee Valley Authority as the Chief of the Nuclear Generation Branch, Division of Power Production. I am familiar with these proceedings snd have personal knowledge of the matters contained herein.

alifications I have been continuous employed by the Tennessee Valley Authority since 1949. Prior'o Chat time I,served for eight years in the United States Navy. Part of this time I was an Electrica1 Officer on a light cruiser and an aircraft carrier'and was qualified as Engineering Officer-of-the Watch at sea on 'both 'ships.

I have received the Bachelor of Science degree in electrical engineering from Tennessee Technological University in.1949. During this period. I was the Executive Officer and ELectronics Officer of the U.S. Naval Reserve ELectronics Warfare Company located at Cookeyille, Tennessee.

I began my employment with TVA in 1949 as a student in the steam generating plant operator training program at the Watts Bar Steam Plant and later became an instructor in that program. I was transferred. to the, Johnsonville Steam Plant in 1952 as a unit operator and later assumed the position of an electrical engineer. In 1954 I was- placed. in charge of all electrical maintenance at the Johnsonville plant.

In 1958 I became assistant plant superintendent at the 1,500-MW Shawnee Steam Plant at Paducah, Kentucky.

In 1960 I became superintendent of the Experimental Gas-Cooled Reactor (EGCR) at Oak Ridge, Tennessee. During this period. I attended. the Oak Ridge School of Reactor Technology. In 1961 I spent five months at the Berkeley Nuclear Power Station in Bristol, England., assisting in the startup of that reactor. While at Berkeley I completed the reactor operator training course on a nuclear plant simulator used, to train all reactor operators for the Central Electricity Generating Board..

In 1963 I was appointed assistant Prospect Manager of the Experi-mental Gas-Cooled Reactor and. was responsible for assisting the prospect manager in all phases of technical and operational work.

From 1963 to 1966 I was a member '(for reactor operation) of a panel created. by an agreement between the United Kingdom Atomic Energy Authority and the United States Atomic Energy Commission to exchange information on gas-cooled reactors. As a memberof this panel, I twice traveled to England to investigate and. to,observe the operation oi'he British Advanced Gas-Cooled Reactor in preparation for the startup of EGCR.

From February 1966 to February 1968, I held the position of Assistant to the Chief', Power Plant Maintenance Branch, Division of Power Production in TVA. I assisted in the engineering and, coordination of the electrical and mechanical maintenance of all TVA steam and hydro plants.

I was also responsible for the operation and maintenance planning relating to future TVA nuclear po~er plants.

From February 1968 to July 1971, I held the position of Plant Superintendent of the Browns Ferry Nuclear Plant in Athens, Alabama.

From. July 1973: to April 1974, I was nuclear operations coordinator; and in April 1974 my title was changed. to Chief, Nuclear Generation Branch.

In this position I am responsible for and. in charge of staffing, startup testing, and. operations of a11 TVA nuclear power plants, including the Browns Ferry Nuclear Plant, units 1 and 2. I am also responsible for the coordination of the restoration and modifications activities, including fire protection improvements, of the Browns Ferry Nuclear Plant, units 1 and 2, following the March 22, 1975, fire.

I sm presently a member of the Advisory Council at Pennsylvania State University (advisor to the Nuclear Engineering Department) and. serve as Vice Chairman, Reactor Operations Division, American Nuclear Society.

I am familiar with this proceeding and have personal knowledge of the matters stated herein.

Statement The modification and restoration of units 1 and. 2 in accordance with TVA's "Plan for Evaluation, Repair and. Return to Service of Browns Ferry, Units 1 and, 2, (March 22, 1975, Fire)" have been substantially completed. The Atomic Safety and Licensing Board. has authorized the

Director of Nuclear Reactor Regulation to make appropriate findings and authorize fuel loading.

The control rods are fully inserted and disarmed throughout refueling. After refueling is completed, TVA proposes to conduct the following subcritical testing, which is a part of the startup retest program: Control Rod. Drive System tests (Startup Test No. 5) scheduled at zero reactor pressure and the Full Core Shutdown Margin test (Startup Test No. 4). The control rod, drive tests proposed are position indication, insert/withdraw times, coupling, friction, and scram testing at zero reactor pressure. The proposed. Full Core Shutdown Margin test will be conducted, by withdrawing the analytically strongest rod. and one or more additional rods. The testing to be performed is a portion of the unit 2 program previously approved by the Nuclear Regulatory Commission, and will be conducted, as described in the Browns Ferry Final Safety Ana1ysis A

Report, Section 13.5 (pages 13.5-17 through 13.5-19) and Table 13.5-5 h

(Attachment 1) except for those changes discussed in Part XZ, Section D, of TVA's "Plan for Evaluation, Repair and. Return to Service of Browns Ferry Units 1 and. 2, (March 22, 1975, Fire)" (Attachment 2) and, on page 1 of the enclosure of a letter from J. E. Gilleland, TVA Assistant Manager of Power, to Benard C. Rusche, Director of Nuclear Reactor Regulation, dated, May 17, 1976 (Attachment 3).

The purpose of conducting the. Control Rod Drive System tests will be to determine initial operating characteristics of the Control Rod Drive System and. to ensure that no control rod interference exists in the fully loaded core. The Full Core Shutdown Margin test will demonstrate

that the reactor will be subcritical throughout. the remainder of this fuel cycle with any single control rod fully withdrawn. On successful completion of these tests TVA will install the reactor vessel head which will reduce. the startup retesting period by approximately 10 days when permission is granted to operate units 1 and 2.

During this testing, it will be necessary for a portion of the control rods to be operational. To conduct the Control Rod, Drive System test, the ability to move control rods one at a time is needed,, and in order to conduct the Pull Core Shutdown Margin test, operability of two and three rods is needed. for units 1 'and 2, respectively.

An evaluation of core shutdown margins f'r units 1 and 2 for the worst accident possible during these tests has been performed and. is presented below. This accident is one in which all operational control rods are fully withdrawn from the core. No eva1uation is presented for the Control Rod Drive System test because the operability of only one rod at a time would result in a less severe accident than that for the Full Core Shutdown Margin test, in which the use of'wo or three rods is needed.. The evaluation presented below is based on shutdown margin curves supplied by the General Electric Company which take into account data taken from previous operation of units 1 and 2. Rod. worths used are from calculations performed by TVA using methods which have been documented and verified and have been shown to accurately predict control rod effects and other reactor core performance parameters. Operational data, including that taken during various criticalities of units 1 and 2, have been used in the development and verification of these methods.

The following assumptions were used as a basis for the evaluation:

(1) all rods other than those in use will at all times be tully inserted, va1ved out, and electrically disarmed.; (2) all fuel assemblies are returned to the locations they occupied prior to the March 22, 1975, fire; and (3) an accident occurs in which all operational control rods are fully withdrawn from the core. This would, be a maximum of two rods for unit 1 and three rods for unit 2. Administrative controls used during the testing will ensure that all rods not being used. will be fully inserted., valved. out and disarmed which precludes withdrawal of these rods. Following the fire, all fuel was removed. to the spent fuel storage pool. With fuel in the spent fuel pool, a verification of the fuel serial number and pool location is performed. every 6 months. Emmediately prior to removing an assembly from the spent fuel pool for insertion into the reactor vessel, the assembly number and pool location is checked. against the most recent pool verifica-tion. After the assembly is loaded into the core, the core position is checked against the fuel loading sheets. All of these checks are documented on the refuel floor by the fuel handling supervisor. After the core loading is complete, a videotape is made of each fuel assembly in the fliLly loaded.

core, and, each assemb+ is verified to be in the same position it occupied.

prior to the fire.

Evaluation for Performin the Shutdown Mar in Test with Three Rods erational on Unit 2 and Two Rods erational on Unit 1 The technical specifications require that the reactor remain subcritical by a margin of at least .38$ Ak/k at the most reactive point in life.with the analytically strongest rod fully withdrawn., Rod 26-07 is the analytically strongest rod. The margin of'385 Ak/k is required

'6

to account for calculational uncertainties. Demonstration that this shutdown margin is available will require that rod 26-07 be fully with-drawn, and at least one other rod withdrawn to introduce .385 hk/k plus the reactivity necessary to simulate the most reactive point in core life.

Unit 1 has passed the most reactive point in life, and now has a shutdown margin of 2.10$ bk/k with rod 26-07 ful withdrawn. With rod 26-07 fully withdrawn, it is necessary to withdraw rod 22-03 an amount necessary to demonstrate the .38$ bR/k margin required by the technica1 specifications. The worst accident that could occur in this configuration is the full withdrawal of rod 22-03 from the core. The 'I worth of rod. 22-03 with rod 26-07 withdrawn is 1.2lg bk/k. Therefore, in the event of an error or malfunction in which rod 22-03 is fully withdrawn, the reactor I

will remain subcritical with an excess shutdown margin of .89% 4R/k and the shutdown margin required by the technical specifications would. not be reduced. Thus, this accident wou1d have no adverse consequences.

Unit 2 currently has a shutdown margin of 3.15$ Cs/k with rod 26-07 fliLlywithdrawn from the core. Unit 2 has not reached its most reactive point in life, and the core is expected to underg'o an increase in reactivity of 1.25$ 4k/k from its present state. Thus, an additional 1.25/a dR/k must be added to the .385 hk/k required for uncertainties. Therefore, the total shutdown margin to be demonstrated during the, test (with the strongest rod out) will be 1.63$ his/k for unit 2. The total worth of rod 22-.03 with rod. 26-07 out is 1.52$ bk/k. Therefore, rods 26-07 and 22-03 must .be fully withdrawn and a third rod (30-11) partially withdrawn to demonstrate the proper shutdown margin. The worst accident that could occur in this case would be the complete withdrawal of rod. 30-11 from the core. The total

worth of rod 30-11 with rods 26-07 and. 22-03 out is .535 hk/k. In the event that rod 22-03 is ful3Z withdrawn due to an error or ma1function, the reactor will remain subcritical with an excess shutdown'margin of 1.10$ 8t/k (3.15$ - 1.52$ .535 = 1.10$ ), and the shutdown margin required.

by the technical specifications would not be reduced. Thus, this accident would have no adverse consequences.

From these evaluations, it is shown that even in the unlikely event that an accident occurs in which a11 rods being used, (the only rods capable of being moved.) are somehow fully withdrawn from the core during the Full Core Shutdown Margin test, criticality is prevented and. the shutdown margin required by the technica1 specifications is not reduced.

for both units 1 and 2. In the case of the Control Rod Drive System test, in the event of an accident in which the operational rod is fully withdrawn from the core, proper shutdown margin will be easily maintained, and the consequences of such an accident wou1d be less severe than the Full Core Shutdown Margin test accident evaluated. Also, additional assurance that the Control Rod Drive System will function properly during the Full Core Shutdown Margin test will be gained since the Control Rod Drive System tests will be conducted. before the Full Core Shutdown Margin test.

None of the mechanica1 equipment or electrical cables that will be necessary for the conduct of these tests on unit 2 was damaged by the fire or has been modified as a result of the fire. For unit 1, no mechanical equipment was damaged. or modified, as a result of the fire. ELectrical cables damaged by the fire on unit 1 which are necessary for the'onduct of these tests have been restored, preoperationally tested, and shown to be operating

~ ~

satisfactorily. It is important to note that failure of any or all of the equipment or cables used in these tests would result in no adverse safety consequences'ased on the above evaluations, I 'conclude that the above testing can be conducted with no significant increase in the probability of occurrence or consequences of a serious accident which would endanger the health and safety of the public. In the worst case accident, for this configuration, one in which all operational control rods are ful1y with-drawn from the core, the reactor will r main subcritical and the margin

'f safety required by the technical sp ci 'cations will not be duced.

ack . Calhoun Subscribed and. sworn to me this W~

Notary Publi My commission expires

~ ~ r ~

~

~ (Attaohment 1)

BFNP.64 Criteria loading will commence with'he loailing of four fuel assemblies around the central neutron source. Fuel loading Level I Will be accomplished by loading complete control cells that Chamisal 4CtOCS daf In/id ln the TeChnlCal SpeClri sequentially complete each face of an ever Increasing square cacionsmustbacnalntalned within the limits specified. core lctading in a counterclockwise direction, The activity of gaseous and liquid el flu'ants must con Control rod drive functional tests are performed form to license Ilmlta lions.

Lwel 2 during the last week before fuel loading. Control rod Waler quality must be known at all time and should functional tests, subcriticallty checks, and shutdown margin remain within the guidelines of the Water Quality Specifica- demonstrations will be performed periodically during fuel tions. loading.

Criteria TEST NUMBER 2- RADIATIONMEASUREMENTS Level 1 Purpose The partially loaded core must be subcritical by at The purposes of this test are (a) to determine the least 0.389m hk/k with the analytically strongest rod fully background radiation levels in the plant environs prior to withdravm.

operation for base data on activity buildup and (b) to monitor radiation et selected power levels to assure the protection of personnel during plant operation. TEST NUMBER 4- FULL CORE SHUTDOWN MARGIN Description Purpose A survey of natural background radiation The purpose of this test Is to demonstcate that the throughout the plant site will be made prior to fuel reactor will be subcritical throughout the first fuel cycle loading. Subsequent to fuel loading, during reactor with any single control rod fully withdrawn.

heatup and at power levels of 25%, 50%, 75%, and 100% of rated power, gamma radiation level measure- Description ments and ~here appropriate, thermal and fast This test will be performed in the fully loaded core at neutron dose rate measurements will be made at ambient temperature In the xenon free condition. The, sign'ificant locations throughout the plant. All poten-shutdown cnargin will be measured by withdrawing the tially high radiation areas will be survey'ed. ~

analytically strongest rod or the equivalent (anothec rod .

Criteria plus an added reactivity) and one or more additional rods which have been calibrated by calculation until criticality is Level 1 reached.

The radiation doses of plant origin and the occupancy times of personnel ln radiation tones shall be controlled "

Criteria consistent with the guidelines of the standards'or protection against radiation outlined in 10CFR20 AEC General Design Criteria.

TEST NUMBER 3- FUEL LOADING Purpose (b) If (a) cannot be satisfied, then the shutdown margin .

The purpose of this test is to load fuel safely, and of the'ully loaded cora is sa'.isficd if the reactor efficiently to the full core size. remains subcritical by at least 0.28% hk/k (plus and

'additional margin for exposure to be determined Description later) during the sequential, complete withdrawal and Pcior to fuel loading, control rods'nd neutron Insertion of every control cod within the core.

sources and detectors will be installed and tested. Fuel Level 2 drrcodry should occur with/a a 1.OSS d'c/h ortho predicted rod configuration.

13.6.17

BFNP63 TEST NUMBER 5 CONTROL ROD DRIVE SYS s EM Description The CRD tests performed during Phases II through IV Purpose of the startup test program are designed es an extension of The purposes of the Control Rod Drive System test the tests performed during the preoperational CRO system are (a) to demonstrate that the Control Rod Drive (CRD) tests. Thus, after It is verified that all control rod drives System operates properly over the full range of primary operate properly when installed, they are tested periodi ~

coolant temperatures and pressures from ambient to cally during heatup to assure that there is no significant opcradng, and (b) to determine the Initial operating binding caused by, thermal expansion of the core characteristics of the entire CRD system components. A list of all coi~ol rod drive tests to be performed during startup testing is given below.

CONTROL ROD DRIVE SYSTEM TESTS Reactor Pressure with Coro Loaded Test Accumulator pslg (kg/cms )

Description Pressure Preop Tests 0 600 (422) 800 (662) Rated Position Indication all all Normal Times Inset tiWithdrawn all all 4a Coupling all ail ~ aa Friction ,all 40 Scram Normal ail all 4a 4a - all Scram Minimum 40 Scram 2ero 40 Scram (Scram Discharge Normal 4 (full core Volume Migh Level) -scram)

Scram 4a ~

a Value refers to the four sloweat CREE as determined from the normal accumulator pressure scram test et ambhmt reactor pressure.

Throughout the procedure, "the lour alaweat cRD'a'a tmpnea the four slnweat compedbla whh rod wonh mlnin4 sar and cRD sequence raqulramenu.

63l "scram times of the four slowest cRD's willbe determined at gsss, and 100gs of rated power during ptennud reactor scrama.

'"Establish lnidslly that this check ia normal operating procedure.

NOTE: Single CRD scrams should be performed with the charging valve closed (do not ride the charging pump head).

I3.5.18

BFNP.64 Criteria TEST .NUMBER 6 SRM PFRFORMANCE AND CONTROL ROD SEQUENCE Lavai 1 Each CRD must have a normal withdraw speed less Purpose then or equal to 3.6 Inches per second (9.14 cm/sec),

Tha purpose of this test is to demonstrate that the Indicated by a full 12 foot stroke in greater than or equal to operational sources,. SRM instrumentation, and rod 40 seconds.

withdrawal sequences provide adequate information to The mean scram time of all operable CRD's must not achieve criticality and increase power in a safe and efficient 54 exceed the (ollowing times: (Scram time is measured from manner. The effect of typical rod movements on reactor the time the pilot scram valve solenoids are deenergized.)

power will be determined.

Scram Timo Scram Time (Seconds) (Seconds) Descri ption Vessel Dome Vessel Dome The operational neutron sources will be installed and Pressure Pressure Percent WSO psig C950 psig source range monitor count rate data will be taKen during Inserted (6L9 kg/cms) (66.9 kg/sms) rod withdrawals to critical and colnpared vith stated 5 0.375 0.475 criteria on signal and signal count.to.noise count ratio.

20 0.90 1.100 A withdrawal sequence has been calculated which 50 2.0 2.0 completely specifies control rod withdrawals from the

64) 90 3.5 3.5 all rods.in condition to the rated power configuration.

Critical rod patterns will be recorded periodically as the The mean scram time of the three fastest CRD's reactor is heated to rated temperature.

in a two by two array must not exceed the following times: Movement of rods in a prescribed sequence is (Scram time ls measured from the time the pilot scram monitored by the Rod Worth Minimizer and the 'Rod 56 valve solenoids are deenergized.) Sequence Control System, which will prevent outwf.

sequence withdrawal and insertions.

Scram Time Scram 'Arne As the withdrawal of each rod group is completed (Seoonds) (Seconds) during the power ascension, the electrical power, steam Vessel Dome Vessel Dome Pressure Pressure flow, control valve position, and APRM response will be Percent W50 pslg <950 psig recorded.

Inserted (66.9 kg/cms ) 66 9 kg/sm Date will be obtained to verify the relationship between core power and first stage turbine pressure to 0.396 0.504 insure that the RSCS properly fulfills its intended function 20 0.954 1.166 up to the required power level,

.50 2.120 2.320

64) 90 3.600 3.800 Criteria Level 2 Each CRD must have a normal insert or withdrawn Level 1 speed of 3.0$ 0.6 inches per second (7.62 sma,b1.S2 There must be a neutron signal count.tonoise count cm/sec), indicated by a full 12.foot stroke in 40 to ratio of at least 2 to 1 on the required operable SRM's or 60 seconds. Fuel Loading Chambers.

With respect to the control rod drive friction tests, if There must be a minimum count rate of 3 56 the differential pressure variation exceeds 1S psid counts/second on the required operable SRM's or Fuel (1 kg/cms) for a continuous drive in, a settling test must be Loading Chambers.

performed, in which case, the differential settling pressure The IRM's must be on scale before the SRM's exceed should not be less than 30 psid (2,1 kg/cms) nor should it the rod block set point.

vary by more than 10 psid (0.7 kg/cms) over a full stroke. The Rod Sequence Control System shall be operable Scram times with normal accumulator charge should as specified in the Technical Specifications.

fall wIthin the time limits Indicated on Figure 5.3.1 of the Startup Test Instructions.

13.5-19

I.

5%

~i ammm~~~~~~~~~~~~~~

~mm~~~~~~~~~~~~~~~~

iH Gl

~~~~~~~~~~am~~~~~

M RA M

Kl

~~mm~~

~~53K%~~~

~~aK%%3KM~~

M ~mramrm~~~~~~~~~~~~~maaa~~

~~~~~~~~~~~~~~NPEM~~

M M ~~~~mmamm~~

~~~~~mcim~~

~~~~KSIKil~~~~~

~~~~EKBXR~~

gg s

4 Section 7

D'Page Recover FLan BFNP 11/13/7g TEST NUMBER 4 - FULL CORE SNUTDOWN MARGIN Deviation from ur ose descri tion and cx'iteria

a. Purpose << no change.

b. Description - The description will remain the same except to delete the last four vords, " . . ~ until criticality is reached." The r /P technical specifications do not require a critical rod configuration as part of shutdown margin test or x'eactivity anomaly check, since

%here will be no change in the fuel assemblies loaded into the reactor vessels. 'This obviously diffexs from the I

initial startup program where the full core load consisted of new fuel.

c. Criteria '- Delete the level 1 (b) criteria since it will~ot meet the technical specifications. Level 1 (a) is slightly more stringent and does conform'to technical specification requirements.

2. Deviation from table 13.5-5 fre uenc No change.

e. 'I p ~

Part XI Section D Page 8 TEST NE4BER g - CONTROL ROD DRXVE SYSTEM Recovery Plan BPHP

'1. Deviation from ose description and criteria 11/u/7>

a. Purpose - The FSAR calls'for demonstration of CRD system operation over the fu11 x'ange of primary coolant temperatures and pressures i

from ambient to operating. Determination of initial operating charac>>

teristics of the entire CRD system is also required. For the purposes of startup retesting it will be sufficient to determine initial operating chaxacteristics, by friction and scram testing at zexo reactor pressure aftex fuel loading (the preop teats as listed will also be performed prior to fuel loading). Scram times will also be measured at rated temperature and pressure during heatup and/or low power retesting, b, Description - According to,the FSAR the periodic CRD testing during heatup is done to assure that there is .no significant binding caused by I r

I thermal expansion of the core components. Since the thermal expansion characteristics have already been proven, they will not require periodic

.testing during heatup for the retest program.

'1 The control rod drive.

tests which will be required during startup retesting are position, indication, insert/withdraw times, coupling friction,. and scram testing at zero reactox pressure; and scram testing of all CRD's at rated temperature and pressure. Additional initial startup testing with various accumulator pressures and I I with repeated confirmatory tests for selected rods has demonstrated expected design response and expected repeatabiUty; therefox'e an extended retesting is not needed. The testing program wiU. adhere to h all technical specification xequirements.

cy Criteria - No change.

E 4

9 .

Section D Page 9 Recovery Plan

'BFNP TEST NUMBER 5 - CONTROL ROD DRIVE SYSTEM (Continued) ll/13/7g

2. Deviation from table 13.5-5 fre uenc STI 5 will only be performed during open vessel testing, heatup, and at 15-40K power as described in the purpose and description (see l.a and l.b above). The change cE the upper limit at 15-35/ to 15-40X is consistent with technical speyification requirements and rod sequence control system limitations. Further testing at test conditions 2E, 3E, and 4E is not needed for reasons given in l.b above.

(Attachment 3)

MAY .1 y )97@

}h. Dsnsrd C. Ruacho, Dire or, Off'ice of Huclear-Reactor Re'adulation Q, S. lhiclear" Regulatory Commission Wssh$ ugton, DC 20555

Dear Hr,

Ruscho!

Xn tho Matter of the Docket No, 50-296 Tannsssoe Valley Authority This letter concerns my April 6, 1976, letter to you in which TVA roapondod to several tPiC staff concerns regarding Brcr2ns Ferry Nuclear Plant unit 3 startup test program and an NRC lotter of April 30, 1976, R. C. DeYoung to Godwin Williams, Jr.g in wMch the HRC staff x'osponded to TVA's April 6y 1976'etter, lb>closed are additional chases to the staxtup test program as suggested by tho staff rerLew, explanations of staxtup test acceptance criteria incXuded in the April 6 letter, and cox'ractious to Browns Parry Final Safety Analysis Report Amendment 64 Tliesa comments should satisfactorily resolve all remaining open issues on the unit 3 startup test program. Xf you }2aVe additional questions or concerns, please let us know so that these items can be closed out in 'a timely manner ao as not to delay our staxtup schsdulem Very truly yours,

~ F., Gilleland Assistant Hsnager of Power 2>fiY}c2'26 '2 </mi Pmclosuro Subscribed and sworn tiiin ~2it.

t~ fore 1976,  ; l II mi tm

lI

~

ma dny I

lO~L} ~

No ary Public Ny Commission Expires QBf:}EW:JTL: SCR CC (Fnclosure): gC-A. W. Crevasse, 207 PRB-C H. G. Parris, 403 PRB-C R. H. Dunhsm, W11A9 C-K H. S; Sanger, 629 NSB-K H. S. Fox, 716 RB-C httn. J.R. Calhou P. A. Szczepanski, 210 PRB-C reen, rowns Ferry Godwin Williams, Jr., 830 PRB-C G. H. Kimmons, W12A9 C-K G. Rm Wolfe, 1003 CBB<<C Dr. P. Aa Krenkcl, 268 40XB-C

s ENCLOSURE BROWNS FERRX NUCLEAR PIANT UNXT 3 CHANGES TO STARTUP TEST PROGRAM h

TEST NUmER 4 FSAR Amendment 64 incorrectly deleted Level 1 criterion (a) rather than Lovel 1 criterion (b) as stated in the April 6, 1976, TVA letter to NRC.

A subsequent FSAR amendment will correct this error.

TEST KPi4BER 13 The test purpose, description, and. criteria will be changed, as follows;

~Per ose Thc purpose of this test is to verify the performance of the process computer under plant operating conditions.

Descri tion GE/PAC computer system program vexifications and. calculational program validations at static and at simulated dvnamic input conditions will be preoperationnlly tested at the computer supplier's site and following delivery to the plant site. Following fuel loading, during plant heatup and the ascension to rated power, the nuclear steam supply system and the balance-of-plant system process variables sensed by the computer as digital or analog signals will become available. Verify that the computer correct values of sensed process variables and'that the resultsis'eceiving of performance calculations of the nuclear steam supply system and the balance of plant are correct. At steady-state power conditions, the DynataLc System Tes Case will be performed.

Criteria Level l Not applicable.

Level 2 Program OD-1 and P-1 will be considered operational when:

The MCPR calculated by an independent method and the process computer either:

a. Axe in the same fuel assembly and do not differ in value by more than two percent, or,

b. If two different fuel assemblies are chosen by the two methods, the MCPR calculated by the other method in each assemb1y agrees with the MCPR calculated by the computer in that assembly within two percent

2. The LPR4 calibration factors calculated by the independent method, and. the process computer agxee to within two percent.

The remaining programs will be considered operational upon successful completion of the static anddynamic testing.

TEST NUMBER 18 Back round and Discussion The initial. purpose of the TIP reproducibility test was to verify that the TIP and XY-plotter could retrace an axial flux shape within the norm established by a base set of TIP traces taken at several sites in It was specifica11y not meant to be a check of the actual flux 1970-'971.

variation, but rather the TIP and XY-plotter m chanics and electronics.

The original xeproducibility criterion was simply + 3 percent but was not chosen on a statistical basis. Rather it was chosen on a worst-case basis of the original. data base of TIP traces. As a result, as one might expect when samling a much larger set of data, many cases resulted in failure to meet the + 3 percent criterion because of f"ux variations

~ being included in the analysis. Because of this, the criterion was increased to + 3.5 percent or + 0.15 inch, the 0.15-inch criterion being added in an attempt to handle very small values. of heat flux, especially at the upper and lower parts of the core. As the set of new data grew with more plants starting up, more failures of the + 3.5 percent criterion occuxred.. In an attempt to prevent so many failures, advantage was taken of the areas of the core where so much flux noise was not occurring (non-boiling region) and at lower power (where the flux noise'as lower).

In these cases the criteria could be regularly met. This was in line with the original purpose of testing the TIP and XY-plotter operation '(not the.

point-by-point flux variation) and was felt to be acceptable.

Relationshi of "ha e to Past Events Since the time the criterion was first applied (1972), more and. moxe emphasis has been placed. on the process computer and specifical3y its nodal power calculations. Because this switch from hand methods to computer methods has occurred, it seems more, appropriate to test the TIP reproducibility not of the XY-plotter output but the computer software and hardware interface, specifically program OD-2."

In addition, the nature of the actual comparison, of'eproducibility would seem to be more meaningful pass-fail test.

if it were statistical rather than a straight

The method would use as a criterion the random noise component of 1.2 pcxccnt, quoted in NED0-20340, "Process Computer Perfoxmance Evaluation Accuracy," J. F. Carew, June 1974, and NED0-109/8, "GETAB: Data, Correlation and Design Application," November 1973.

The concerns of this issue are deeply involved in statistica1 methods and must be carefully handled.

The referenced document NED0-20340, June 1974, by J. F. Carew discusses xandom noise effects (as discussed above) and geometric effects (loosely, differ'ences from unity in the ratio of symmetrical TIP readings 3.ess statistically the random noise effects).

The random noise one sigma uncertainty is 3..2 percent and the geometric noise one sigma uncertainty is 2.3 pexcent, based. on information in NED0-20340.

The NHC request for a criterion on symmetry in STX-18 is interpxeted as a request to confirm the 2.3 percent one sigma uncertainty in the geometric noise~ based. on TIP txaces from symmetrical locatiohs.

The geometric uncertainty of 2.3 percent was based on several sets of symmetrical TXP readings taken by GE from a startup in 3.9'70; however, this number has now been revised by GE following the ana1ysis of 41'ets of TXPS from six recent plants. This analysis showed. a lr TIP signal uncertainty of 3.9 percent which, after statistically sub-leve1 overaU.

txacting out the 1.2 percent random noise uncertainty, gives a new geometric uncertainty of,3.7ipercent (causing the effective TIP uncertainty to increase from 6.3 percent to 6.7 percent). A check with the GE group which calculates the Browns Ferry unit 3 MCPR limits shows that they are now performing analyses based. on this newer 3.7-percent geometric uncertainty, Consequently, the value to be used in cxeating an acceptability criterion is 3.7 percent rather than 2.3 percent from NED0-20340.

There are several important points to make about the xequests of issue 4.

Specifica13y,

1. Since the GEKL/GETAB correlation is statistical and the two parameters random noise uncertainty and. geometrical uncertainty are each statis-tically based on sets of data, any criterion should be based, on a statistical comparison of new sets of data, not a point-to-data base compaxison.

2. The request that TIP reproducibility have a 3..2 percent standard

~

deviation fox ~an 6-inch bundle segment is inappropriate. It is similar to collecting a set of data, calculating the average, making the average be a limit, then xequiring every data point in the set to be less than the average, and, then to make that a Level 3. criterion.

0 fi fi

3. Xt is statistically certain that a fra ti rac on of, new data 'he alone, wil1 be outside the standards of the existin data this comparison is requested to be a Le The fact that a few new data points ma be o negligible effect ec on thee over ov averages and standards of the existing

e. imp~ that these new data points (statistically certain o e outside the existing data base outweigh the entire exist9.

data base and require a reanalysis with GEXL~GETAB to generate new values of lCCPR is completely o p e ely invalnvalid and. has no engineering or s s xc justification.

5. The only reasonable, statisM. call vali stically valid approach, consistent with the tions of Level 1 and. Level 2 criteria, and the im 1 definitions e

not meeting an acceptabilit y cr erion -is to perform a statistical treatment of the sets of data collected. at the startup and corn are with the characteristics of thee ex existi s ng data d t base.

b Actual Test Outline

a. Run a complete OD-1 (complete TIP set).

b. Edit BASE values for'll TXF locations.

The analysis would then proceed as follows:

c. For each symmetrical ymm cal TXP pair, divide the nodal BASE va 1ue n the i upper left (fr a core loading map view) by its half oof thee core (from ymm t'r cal counte rp art in he lower right half.

symme t This will result in mes n calculated points for a core with n s etr i.e.', 24 values for eac f each o n TXP P pairs.

d. Calculate t average te the and. the standard deviati on o f th ese (24 times n) ratios.

e. Divide the standard deviation byy ~2 . Zxpress .as a percent.

t.

f. Statistically subtract the 1 .2 percent random noise portion from the results in e.

~ geometry = (o - 1.2 }

g. Compare this value with the 6.7 percent crite The steps of the outlined test are clear until step e. This ~2 factor

's e

needed because the geometry.c uncertainty in one TXP readinng iss thee desired parameter, but the measured, sure geometric eom uncertainty is of a ratio of two TXP r di

&5M These parameters are related by:

0 2 (Tl) = 2o (T) where cr(T1) =0(T2) =P(T)

(T2 )

r and cr (T) ~(T1/T2)/2

= 0'(Tl/T2)/v 2

'hus 0 (T)

The background of the 3.7 percent geometric uncertainty is outlined below.

The average standard. deviation of the actual TIP symmetry ratios was 5.48 percent. When divided by ~2 this yields the 3.9 percent referred to in the.

discussion on the more recent data base, above, as the value for combined geometric and xandom noise uncertainty. The scatter of 41 data points about this average showed a standard deviation of 1.47 percent (26.9 percent of the. mean va1ue of 5.48 percent).

This 26.9 percent of the mean value was then applied to the 3.7 percent geometric uncertainty (26.p~ of 3.7g - 1.+). Thus, one sigma. of data scatter is 1.0 percent relative to the mean of 3.7 percent. The. three sigma criterion is then 3.7~ + (3 x 1.+) = 6.7$ .

The data base generated. in the test to confirm the reproducibility of the computer-stored TIP system segment average values should be compared with a 34 (3 x 1.2', or 3.6g) random data .variation. This is because the sample size generated in the tes+ is significantly smaller than that used to arrive at the 1.2'$ deviation which GE determined from a much larger data base, The purpose, description, and criteria for this test will be changed, as follows.

~Pur ose The purposes of this test are:

a. To confirm the reproducibility of the computer-stored TIP system segment-averaged values.

b. To determine the core power distribution in three dimensions, and

c. To determine core power symmetry.

Descri tion Core power distribution including power symmetry will be obtained during the power ascension program. Axial power traces will be obtained at each f'r;

e l

of the TIP locations. Several TIP systems have been. provided to obtain these traces--a common location can be traversed by each TIP chamber to permit intercalibration.

A check of the nodal standard deviation of computer>>stored TIP traces is made at sevexal different power/flow conditions. The checks are made with the plant at steady-state conditions by producing several TIP traces in the same location with each TIP machine.

Criteria Leeel l Not applicable.

Level 2 In the TIP reproducibility test the overall standard deviation of segment-averaged. TIP values (BASE distribution from process computer) of'he central nodes shall be less than or equal to (3 x 1.2) or 3.6 percent.

In the analysis of data to assess core power symmetry,. the statisticaU.y calculated average standard deviation of the TIP pair x'atios shall be less than or equal to 6.7$ .

TEST NUMBER 19 The test description will be changed to read as follows.

Descri tion The core performance evaluation is intended to detexmine the principal thermal and hydraulic parameters associated with core behavior. These parameters are core flow, core thermal power, MLHGR, MCPR, and. MAPLHGR.

These core parameters will be evaluated by manual calculations, the process computer, or the off-line computer progxam BUGLE. If the process computer is used as a primary means to obtain these parameters, have been checked out in accordance with Test No. 13.

it must II TEST NUMBER 25 The Level 1 and Level 2 criteria will be changed to read as follows.

Level 1 Closure time must be greater thanQ) and less thang seconds. The initial transient rise in vessel dome pressure occurring withi'n 20 seconds of the main steamline isolation valve trip initiation and the transient rise in simulated heat flux shall not exceed the predicted values stated in NED0-21165, page 3-10 for the case of-

~ ~

closure of aU. MSIU's with a flux scram (237 psi and 33 percent heat flux). The peak reactor vessel pressure shall be less than 1250 ps+

during the transient fol1owing closure of'll MSIV's.

Level 2 The initial transient peak in vessel dome pressure occurring within 20 seconds following initiation of the MSIV closure and. the transient peak in simulated surface heat flux shall not be more limiting than the predicted transients in the Transient Analysis Design Report (100 psi and no heat flux increase).

Discussion Basis Since initial reactor dome operating pxessures often vary from the values assumed in predictive transient calculations, it is not the absolute peak values which axe to be compared but xather the oressure rises. Therefore, Level 3. criterion number 2 "'s written in that respect. However, from a xeactor ope ation point of view, it is extremely important not to open any safety valves. Consequently, a Level 1 c iterion is added which addresses he absolute peak pressure rela ive to the lowest safety valve setpoint for the MSXV transient, since the calculation with flux scram results in a peak dome pressure very close to the safety valve

'etpoint. Tto such Level 1 criterion is needed for the turbine trip transient since the highest reasonoble initial pressure plus the'predicted. pressure x'ise is less than the 3xvest safety valve setpoint.

To analyze the situation a little deeper, however, there is a smal3. several-pound pressure drop, from the steam dome to the steamlines at the point of th e safety valves, Thus, the vessel pressure could actually go above the safety valve setpoint but by an amount less than'he actual steamflow-pressure dxop and, not open. any safety valves. The selection of'the lowest safety valve setpoint is made since (1) it is conservative and (2) the exact pxessure dxop is not known to a degree that in the criterion.

it could be reliably used TEST NUMBER 26 FSAR amendment 64 incorrectly changed the test criteria for unit 2 rather than unit 3. A subsequent FSAR amendment will correct this error. In addition, the following changes will be made to the Level 1 and Level 2 criteria.

Level 1 tl The sum total of capacit)es from 11 relief valves shall 'be euual to or greater than 8.83 x 10 1b/hr of 1112 psig."

+ g corxected for an inlet pressure Level 2 Relief valve leakage shall be low enough that the. temperature .

measured by the thermocouples in the discharge side of the valves

~ ~

xeturns to within 10 F (5.6o C) of the temperature recorded before the va3.ve was opened.

Each individual relief valve shall have a minimum capacity of 720,000 coxrected for an inlet pressure of 1112 psig.

Discussion The following statement should clear up the concerns with valve capacities used in the safety analysis (NEDO 21165) versus the Level 1 criteria used in the startup test description.

N.sunderstandings have occurred in determining primary system relief valve capacities because in some cases of safety analysis, GE reports the .nominal value settings only. This was the case in the TVA 3.etter to John F. Stolz dated April 12, 1976, where a nominal value of 1080 psig was used. in the response to the first question.

The accident analysis (NEDO-21165) used. the relieving capacity of the Taxget Rock and. Crosby valves at their actual setpoints plus 1 percent.

These values are listed in the following table. Table 3-3. of NEDO-21165 as amended by the April 12, 3.976, letter from J. E. Gilleland to J. F.

Stolz listed the setpoints used in the analysis.

CAPACITIES OF THE BROWNS FERRY UNIT 3 SAFETY AND RELIEF VALVES USED IN GETAB AItALYSES Pressure Crosby Safety/ Taxget Rock Dresser Assumed in Relief Valves Safety/Relief Safety Analysis Valves Valves PSIG Flow/Valve g of Flow/Valve g of $ of Rated 4/Hr Rated FloW g/Hr Rated FloW FLow 2 va3.ves valves 2 valves 1080 + lg ~ 1093. 784,479 52-77$

1080 + 3g = 1112~ 813~000 12.15$ 800,000 53.8$

1110 + 1/o = 1121 819,000 12 25$

1250 + 3.$ = 3.262

~Rated Steamflmf - 13,380,000 j$/hr.

~lfhen tested in accordance with ASME Pressure Vessel Code Section III, 3.968, para. N-914.3.

In our April 6, 1976, letter to NRC, TVA incorrectly calculated, the required, total relief valve capacity by adding the capacity of 52.3 and 12.25 percent =

for thc Target Rock and Crosby valves, respectively, as listed in NED0-2l165.

This was incorrect because the capacities of relief valves set at different pressures are different and cannot be directly, added. The is calculated follows.

requiredtotal'elief capacity as l (Nameplate capacity per'valve)

(at 1080 + 3$ or 1112 psig )

~8 538$

a,.

13.38 x 106 'E or .5381 x 13,38 x 10 lb/hr - 7.200 x 3.0 'b/hx'.

b ~Crosb (Namep3.ate capacity per valve)

(at 1110 + 1$ or 33.21 psig .,) I (Nameplate capacity per. valve) 1112

.(corrected 1112 psig. .

), 1121 (83.9~000) =, 812~%5 lb/'

2 x 812 42 12 1 4 p of rated steamf3.o'w.

10 6'

or .1214 x 13.38 x 10 6 lb/hr.= 1.625 x 10 -lb/hr.

~

I

c. Total Ca cit Corrected'to ll12 si C

Total capacity corrected. to 1112 psig = 53;81 +'2.14 = 65.95$

of xated, steamflow 6

or 7.200 x 10 + 3..624 x 10 6 = 8.825 x 10 6 lb/hr.

In some few cases, the me'asured capacity of individual relief valves have been less than the nameplate capacity due to the inaccuracies involved, in determining valve capacity in an operating plant. Specifying a minimum capacity of an individual valve to be 720,000'lb/lm (about 99$ of nameplate capacity) corrected. to 1112 psig as Level 2 criteria is acceptable as long as the total capacity criteria is met.

TEST NUMBER 27 II The Level 1 and. Level 2 criteria is It changed. to ~ t read, as follows.

'Level 3.

I

'he initial txansient rise in vessel dome pressure occurring within 10 seconds of the turbine/generator trip initiation and the transient

0 rise in simulated heat flux shall not exceed. the predicted values stated the NEDO 21165 for the case of turbine trip or load, rejection

'n with turbine bypass valve failure (3.50 psi and ten percent heat flux).

The turbine stop valves must begin to close before the control valves for the turbine trip. The turbine contxol valves must begin to close before the stop valves during th. generator load rejection.

Following fast closure of the turbine stop and control valves, a xeactor scram shall occur greater than 205 psig.

if the turbine first sta'ge pressure is Feedwater systems must prevent. flooding of the steamline following the transients.

Level 2 The initial transient rise in vessel dome pressure occurring within-3.0 seconds of the turbine/generator trip initiation and the txansient xise in simulated. surface heat flux shall not be more limiting than the predicted transient presented in the Transient Analysis Design Report (100 psi and, no heat f3ux increase).

The pressure regulator must prevent a low-pressure reactor isolation.

The.feedwater controllex must prevent a low-level initiation of the HPCX and MSIV's as long as feedwatex remains available.

The load rejects.on within bypass capacity must not cause a scxam. The trip scaxm function for higher power level's,must meet RPS specifications.-

s Discussion Basis The discussion and, basis is the same as that for test No. 25.

We also propose that the turbine trip by stop valve closure be performed at test condition 3E (65-85$ power, 102$ flow) as indicated on attached table 13.5-6 rather. than at test condition 4E (109jf power, 10'low).

Reason Past experience and computer analysis have shown that neutron flux heat flux, and pressure rise behavior for turbine trips an'd load.

x'ejections are about the same. For example, reference (2) shows 'that a nominal 100-percent power, 100-percent flow turbine trip, and. a load rejection xesultcd. in pressure peaks of 3100 and 1090 psig, respectively, a ten psi difference. Other startup'esults from past plants confirm this close agreement,.

Transient analysis predictive calculations show that the two transients axe about the same. Reference (3.) shows an expected peak pressure of

33.00 psig for a turbine trip and 1090 psig for a load rejection from full power and, flow. The expected, difference comes from different steam shutoff times (faster for stop valves than control valves). However, the control valves may not be comp3.ete3y open, so that their length of travel divided by the closure rate (closure time) may be close to the closure time for the stop valves (stop valves are, of course~ always 100 percent open initially).

The ability of the Cransient analysis to predict the actual measured, pressure peaks is confirmed from past results at Bxowns'erry units 1 and 2, as well as all other past plants. For example, the unit 1 3.00-percent power turbine trip pressure peak 'was predicted to be 1100 psig, measured at 1080 psig; the 3.00-percent power load. rejection pressure peak predicted to be 1090 psig, measured at 1090 psig; unit 2 100-percent turbine trip pressure peak was predicted to be 1125 psig, measured at 1125 psig. These values are from references (3.) and, (2).

HRC Regulatory Guide 3.68 was first formulated using the existing test plans for a typical BWR in the days of Dresden 2 when both turbine trips and, load. rejections were specified by Genera3. Electric. The data gathered during tests from those plants and subsequent startups have given us the experience to delete an unnecessary scram from 3.00-percent power and recovery during the startup test program, based on the arguments above. NRC Regulatory Guide 158 does not take into account this experience gained since its original creation, with respect to. turbine trips and load, rejections.

The normal, conservative policy concerning power plateau increases during the startup test program is to perform an intermediate power test of a quick steam shutoff transient before reaching rated power.

This, documents that there is an intermediate, confirmatory checkpoint and prevents the unlikely embarrassment of having any unexpected surprising xesults occur at rated. power.

Xt is desirable,'though not necessary, to perform one of each type during the power test program to confirm the fast closing action of of'rip the stop valves and the control valves. Because the load reje'ction results in turbine overspeed, the fuL1 power condition is of slightly more interest for that transient,. Therefore, the logical optimum test pattern seems to design itself: a load rejection below the bypass capacity to confirm that this will not cause a scram, a turbine trip at intermediate power as a benchmark or intermediate, confirmatory point to give confidence in raising power to rated, and a load rejectidn at rated power.

Computer analysis has shown insignificant differences from turbine trip and load, rejection transients regarding reactor pressure, neutron flux, and heat flux rise, when comparing 7 x 7 with 8 x 8 fuel.

Determination of the response of plant electrical equipment to over-frequency or under-frequency conditions during, those transients from high

power is not appropriate. Contacts on the scram relays initiate 'a fast transfer (does not wait for under voltage) of the 4-kV unit boards and, the 4-kV shutdown boards to their alternate sources (common transformers for the 4-kV unit boards and an operating unit or common transformer for the 4-kV shutdown boards). In the case of turbine trip test, a scram is initiated by the stop valve position switches. In the case of load refection from high power, a scram is initiated by the. control valve fast closure relays. In actual test data, the difference in time, from the initiating event to. the time scram occurs is on the-order of a few

mil1iseconds; therefore, the response of the electrical equipment will be identical whether the turbine trip be initiated at 75$ power ox 10+

powel ~

References:

1. NED0-20747, Final Report: Startup Test Program Browns Fex'ry Nuclear Plant, Unit 1" dated January 1975.

2. 'ED0-20891, "Browns Ferry Unit 2 Startup Test Results-Final Sununary Report" dated. hhy 1975.

TEST E'ER 30 Delete both single pump trips and the two pump trip from test condition 3E.

Reason Mth the conversion from MCHFR to MCPR, the recirculation pump trip transients are no longer a concern since computer analysis shows that xecirculation pump trips are very mild from a MCPR standpoint.

had shown that'rom a MCHFR standpoint, recirculation pump Computex'nalysis trips had. been very significant. This. computer analysis on MCPR and MCHFR was confirmed by analysis of test results obtained. from BFNP units 1 and, 2 startup program. Consequent~, pump trips are rio longer required, for the purpose of thermal hydraulic transient analysis.

One pump trips have been retained. to demonstrate the level control system's ability to handle such transients. Vibration testing is now done with one two-pump trip from full power, and, steady-state measurements during slow power maneuvers along a constant rod line. Thus, the 75-percent power trips need. not be done since all the desired information is generated, by tests at other conditions or during other tests.'ESTS NUMBERS 16, 1 30 AND 34 Delete these tests from test condition 3A (37$ power, Natural Circulation).

Reason I Deletion of the two pump trip at test condition 3E precludes the perfonnance of STI's 16, 19, 3Q and 34 at test condition 3A. Sufficient

&$3 &

'I t

data will be obtained following the pump trips f'rom test conditions 2E (5P> power) and 4E (10+ power) to provide a .comp1ete spectrum of data and, thus satisfy the reguirements of STX's 16, 19, and 30. STI 34 vibration data will be taken at equally spaced points along the 75$

load, line and natural circulation is not reguired f'r this data.

TEST NUMBER 31 Level 1 and Level 2 criteria will be changed to read as f'ollows.

Level 1

'I The initial transient rise in vessel dome pressure occurring within 10 seconds of turbine/generator trip initiation and the in simulated heat f'lux shall not exceed the predicted values transient'ise stated in the FSAR for the case of turbine trip or load refection with bypass valve failure (150 psi and, 10 percent heat flux).

AU. safety systems, such as the reactor protection, system', the diesel-generator and the RCIC and HPCI, must function properly without manual assistance.

Level 2 The initial transient rise in vessel dome pressure'occurring within 10 seconds of the turbine/generator trip initiation and,.the, transient

.rise in simulated surface heat flux shall not be more limiting than the predicted transient presented in the Transient Analysis Design Report (100 psi and. no heat flux increase).

Normal reactor cooling systems should be able to maintain adequate suppression pool water temperature, maintain adequate drywell cooling, and, pxevent actuation of the auto-depressurization system.

TEST NUMBER 72 FSAR amendment 64 incorrectly changed. the Level 2 criteria for this criteria will be to read as'it did in amendment 63.

test.'he Level 2 changed PEA and. GE have analyzed the data from Browns Ferry units 1 and 2 drywell temperature tests and determined that although certain areas of the drywaLL exceed 150 F, equipment critical to temperatures higher than 150 F is not located, in these areas, and. the average ambient temperature of the dryweU. is below 150 F. Analysis has shown that equipment in the areas which experienced 'temperatures in excess 'of 150 F can operate at temperatures up to 180 F without adverse effects on design life of the equipment.

1

P Table 13.5-6 submitted. i.n FSAR amendment 64 wi11, be correct+ labeled unit 3 and the appropriate changes made to the table as listed above for test Nos. 27, 30~ 16, 19, and. 34.

Issue No. 5 of the NRC letter, R. C. DeYoung to TVA dated. Apxil 30, 1976~.

"TVA does not plan to perform a test to demonstrate that the reactor shutdown from outside the control xoom while't power."

can'e Our response to this issue is as follows:  !

We acknowledge that Regulatory Guide 1.68 recommends demonstrating the capability to shutdown and maintain the reactor in hot standby from out-side the control room, and we have care~ considered. the details of such a test. However, eich element of such a test's already included, i.n our preoperational or startup test program so that in integrated, control room abandonment test'would serve no useful purpose other than meeting a guide recommendation.

The existing control room abandonment instruction provides that the reactor be scrammed. before leaving the control room; but, if this is not possible, reactor scram is accomplished by closing the main steam isolation valves from the backup control center .thereby initiating an MSIV closure scram.

MSIV operability from the backup control center is'verified in the pre-operational test program, and the MSIV closure ..scram is demonstrated in the startup test program when at full power.

Following reactor isolation and. scram, the control room abandonment provides that water'ev'el inventory be maintained by the RCIC 'nstruction system operated, from the backup control center. Operability of. the HCIC from the backup'ontrol centex 'is demonstrate'd in the startup test program.

The instructions then provide that the reactor be depressurized by operating the main steam relief valves from the backup control center., Operability of these xeli.ef valves from this location; as well as from the control room, is demonstrated during'the heatup phase of'-.the startup test progiam.

the xeactor! i.s depxessurized,, the instructi'ons provide f'r alkgning 4'nce and. operating shutdown cooling from local control stations. This i.s the most complex portion of the control room'abandonment procedure-but is not upon the reactor being at power; and, in fact, be accomplished with the reactor shut down and the system depressurized.

it can only 'ependent The ability to perform this operation, including tests of necessary communications systems and, adequate manning to perform the operation provided, is included, as a part'f our prefuel-'loading functional test program.

The is principal objection to performing a control xoom test i.s that an unnecessary reacCor trip,'n unnecessary'-usage of reli.ef valves, it and of nobenefit since all operations are'emonstrated elsewhere. in the program. We request that this be recognized and that an extra test sol~

to fulfilla guide recommendation" not be reguired.

1/

l

!! ~ I S

~ << ~

~ Fttnss TsMe SIARTUP TEST PROCRA iLS 4

{VtttyI) 3 OVrlt Pew r i TKSSKC 50'I Flov <astrol Lint ~ I 'Firr Control Line lrrftt Flov Coxtrot tine CR lfFAT ]i ~ '0 )i ~

Tt:T Cky itic'l (C<<o Fitu. c I' CCLII TI~~T VP WI << '/0 2J wit,<< v rct iic'.. <<A vt't<<t I<<llo It<<nits a! tttcc Ir. V<<cstrtcr<<nc tr

~

I TIII'l'.

ct O'I .O ~ II I. rovn ~ ~I shtl r<<r.. "Gltrlcror tod sertr . '<<i vr 'lcutc .<<Ct

<<rl. r. 1n. 1 t1.. ~riHIn rorr ~ ~ Ix'lrr~

C.C I'

Mccccc trotccs cc cra".FI rr SCF I ) 'L.l I*l

.Ofe I'r I~~re't I Ul ~ Or I Cote ttrrotts1'<<

I . I<<xa Irr otc:. f.

~

1 t tax t csrotle ' ho ts 22

~ IS .U "I'l te1in cetan so. ' ixesl I rc; v XI<<f ,I<<C Roti ~ C .nival Ca ac:t tuat o:.s IIC ~ Ir

'ta F

t'I <<F'I rh <<

'I<<

o' ~ "r: lr<<I .-e "r. a'I 1t Ih S Ctf ~ r r tu Slee 'Fr<< t

~ t<< I~ ~

<<I r<<ciF ~ FJS~SCCn n t teo ~I ' '

csccxF I sttt 'tl1 ' 1I'.

C et O'I'\ rtta'<<OVS t " ~

rivet. ltts r.et c v"stir. - lien

~ \ ~to .Ia <<acr r ~ . stirs G ~s ~ ye ten Iperccnt of rated lover, 12r'! Wt Included only io neet Test Ic Seluircncnts L v heal Vanual Fiov Colttr)L )kris 2rerccnt of raced flov, 1c2.5 x lcS Iblhr ip ihe Generator Field Srcskcr N ~ Vastcr Vfanual First 'it.Stol Vxrde Llhlso Ob'cain dsia vith Tests 25, Full Iso a Test 27 Cbialn data vlth Test 30 5Pcrforn 'Fest 5, tining of h ~ Iotlcst control rods

'o

%cat up tests of trSITs r Relief Valves are cheek operation only 1 RSCS cleared ( bOI)

A o Autcnatte I'ICV CsntFVL "Oie X o Teat IndCPetrtent Of I lac CantrOl SP rr Scrars Possibility Stnda in con/unttion vith these strtns Slpcterninc raxicrsc povcr vithout saran SS ~ Saran Sxpccted "Ferforn ihe Itynani>> Cystcn Test Case 12Fron Test Condition 2E to 5 SC o Natural Circulatiaa tSSttot rctuircd if 50I Fover cs ing vill be done vittltn about 2 nontts ILSdl

~ I