ML20209C497

Forwards 14 Day Rept on Potential TS Violation Re Failure to Measure Adequately Secondary Flow Limiting Safety Sys Setting,Which Was Reported by 990608 & 10 Telcon

ML20209C497
Person / Time
Site:	05000083
Issue date:	06/29/1999
From:	Vernetson W FLORIDA, UNIV. OF, GAINESVILLE, FL
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
NUDOCS 9907120021
Download: ML20209C497 (9)

Text

-

y- 1; l .

l f

UNIVERSITY OF FLORIDA Nuclear Reactor Facility 202 Nuclew sciences Cantar Department of Neclear and Radiological Engirteering g, ,,[0gl,18*

i Tel: (352)392-1429 Far (352)392-3380 Fanail: vernet @ server 1.nuceng.ufl.edu June 29,1999 14 Day Report:

Attn: Document Control Desk Potential Tech Spec Violation-U.S. Nuclear Regulatory Commission Failure to Measure Adequately Washington, DC 10555 Secondary Flow Limiting Safety System Setting RE: UNIVERSITY OF FLORIDA TRAINING REACTOR FACILITY LICENSE R-56, DOCKET NO. 50-83 Pursuant to the reporting requirements of paragraph 6.6.2(3)(g) of the UFTR Technical Specifications, a I' description of a potential violation of the technical specifications was reported by telephone on June 8 and June 10,1999 with a following fax (telecopy) on June 10,1999 (Attachment I) and the so-called 14 day written rerort is submitted with this letter including occurrence scenario, NRC notification, evaluation of consequences, corrective action and current status. 'Ihe potentially promptly reportable occurrence involved failure to measure adequately the reactor trip on low secondary flow. This 14 day report is delayed one week per com'ersation with Senior Project Manager Ted Michaels on June 22,1999.

I SCENARIO j On May 24,1999 a Maintenance Log Page (MLP #99-20) was opened to address a perceived failure of the reed switches on the secondary flowmeter. After extensive analysis of the scram circuitry, it was determined that the switches were functioning cmrectly. After the scram check procedure (Q-1 Surveillance) was thoroughly scrutinized and related to system functions, it was found further that the scram function had been operating correctly, but the procedure used to verify it w3s flawed. , l l

The applicable Step 9 of the surveillance data sheets use 1 to perform the quarterly scram checks (Q-1 Surveillance) secondary coolant flow loss had read as fealows:

')

Step 9. SECONDARY COOLANT FLOW loss;  !

l l a. Shift secondary coolant water scram logic to city water mode. ,_-V '

(

l Ensure seconda 7 coolant flow is in the well pump mode.

Simulate reactor power bove 1 kW.

{P

b. Ti.rn off power to the W=g ; dant *vell pump (check for scram). .

I

c. Restore =~=dary coolant scram logic to the well pump mode, if appropriate.

I 9907120021 990629 PDR ADOCK 0500 EqualOpportunity/ Affirmative ActionInstitution

NRC Document Control Desk Page 2 14 Day Report June 29,1999

'Ihis procedure was found to predate the installation of the city w .ter flowmeter in November 1993. The old system (see Figure 1) prior to November 1993 used a pressure switch in the city water line to provide a go/no-go indicatma of flow. The procedure callad for shutting the city water supply .sive and supplying -

cooling water with the well pump since the two lines fkm together. Logic was switched to city water so the well pump power loss scram would be overridden Additionally, city water logic shiAs the scram signal switch from the combined flowmeter to the city water pressure switch (i.e., the scram circuitry monitored at the pressure switch and not at the combined flowmeter). When the well pump was shut off, the flow (and j pressure) in the line would fall off until the pressure switch would actuate, yielding the scram. This method would verify the operation of the scram circuitry in the control panel and had been considered adequate, but this method would not verify operation of the 60 gpm switch in the combined flowmeter, nor would it allow for verifying the 8 gpm and 60 gpm setpoints in the respective city water and well water = mad =y cooling modes In November 1993, the pressure switch was removed from the city water line and replaced with a flowmeter under MLP #93-43 and 10 CFR 50.59 Evaluaten and Determmation Number 93-09 as shown in Figure 2.

At the time of the scram chacks (Q-1 Surveillance), Step 9 was not changed to reflect removal of the pressure switch, though it probably should have been. It has been noted that the walk-through requirement for implementing procedure changes instituted in March 1997 as the result of a self-identified violation, would have caught this problem had it been in effect in 1993.

It should be noted that the city water flowmeter will read zero flow when the city w3ter supply valve is shut, regardless of whether the well pump is runnmg or not. When the scram check is conducted (in city water logic), the scram signal derives from the city water flowmeter; as a result the scram signal is always "in,"

giving the w4or the false impression that the scram has occurred as ~W For about the first year aAer the city water flowmoter was installed, it sW.4mi correctly giving operators 1 scram signals as --*d Later, the city water flowmeter began to fail and would sometunes stick in the 30 gpm position (normal city water flow rate). If the scram check was performed with the city water l flowmeter stuck in the 30 gpm position, the scram would not iniGate regardless of whether or not the well i

pump was running If the city water meter was stuck low, the scram would appear to function normally, If the city water meter was stuck high (30 gpm), the scram would appear to have failed. In these cases the j scram was marked as unsatisfactory. However, since city water mode operation was not allowed, continued  ;

reactor operation was pernutted This was an oversight since without this step completed, their was no ,

verification of a scram from loss of secondary coolant flow in the well water mode i%t ofloss of Pump power.

It should be noted, however, that ever since the replacement of the city water pressure switch with the city water flowmeter, there has been no true verification of a low flow scram (in well water mode)-it only ,

appeared to be functioning properly. 'Ihe switch that controls the low flow scram signal in w.:ll water logic (switch on the combined flowmeter) had not been being tested under low flow conditions. Nor was the 8 gpm scram setpoint in city water logic quantitatively tested until the installation of the city water flowmeter in November 1993. handary low flow scrams am verified as part of the daily checkout but are performed

. with the well pump dmysed, making it impossible to determine whether the scram was initiated as a result oflow flow or from a loss of pump power.

l I

NRC Document Control Desk Page 3 1

14 Day Report June 29,1999 On June 8,1999, the low flow scram was tested by placing scram logic in the well pump cooling mode and supplying cooling with the well pump. Under MLP #99-21, the well water valve was then throttled to reduce flow. The scram was verified to initiate at the required 60 gpm as indicated on the combined flowmeter and the well warning light was verified to actuate at the required 140 gpm. Since at no time during normal reactor operation has the well warning light initiated, it is reasonable to assume that the well pump logic 60 gpm trip point was never approached and therefore, that the UFTR never failed to trip for a condition reaching or exceedmg the Limiting Safety System Setting on the enMary flow in the well pump mode.

Nevertheless, it was clear that a potential violation of Tech Specs had occurred over some relatively lengthy l l time period.

'Ihe violation scenario was reviewed by the Reactor Safety Review Subcommittee (RSRS) Executive Committee on June 10,1999 after extensive discussions among reactor staffincluding the Actmg Reactor Manager and the Facility Director. The RSRS Executive Committee agreed that the applicable scram check procedure did not verify a low aacnadary cooling water flow trip indanandent of the loss of pump power trip and that a Tech Spec violation probably had occurred though with little or no safety implications. Changes i in procedures to remedy the shortcoming were discussed. All concluded that the best way to check the low flow scram on the quarterly checks would be to throttle secondary coolant flow to create an actual low flow condition and quantitatively venfy the low flow trip setpoint. The Executive Committee members agreed and tw+ = dad hattthis event should be treated as promptly reportable as it has been.

CORRECTIVE ACTION In response to the Tech Spec violation identified on June 8,1999 regarding deficiencies in checking the secondary cooling water low flow trip, two system hardware changes have been implemented which required that two minor procedure changes to be implemented in the UFTR SOP-A.1 (Daily Preoperational Checks) and in UFTR SOP-0.5 (Quality Assurance Program) in the surveillance data sheets for the Quarterly Scram Checks (Q-1 Surveillance). In conversations with Senior Project Manager Ted Michaels, it was agreed that assuring the mnadary flow trip worked (go/no go) on the daily checkout and then obtaining a measured value of the flow trip on the quarterly checks would meet the intent of the Tech Specs.

System Hardware Channes:

The first part of the system modification implemented under 10 CFR 50.59 Evaluation and Determmation Number 99-06 (Modifications for Improved Implementation of Secondary Flow Scram Checks) and MLP

1. 99-21 was to install a well pump power trip bypass switch. To implement the necessary changes to the daily checkout procedure, the operator must be able to bypass the loss of well pump power trip. This bypass was accomplished by installMg a bypass switch on the reactor console that temporarily (on demand only) shorts the well pump power relay contactors, cffectively bypassing the trip. There was a spare switch in the motor control panel that was unused so it has been dedicated as the bypass switch. The switch is a momentasy type switch; as such its function is only in effect when an operator is depressing the switch.

This feature elimmates the possibility that the switch can be inadvertently left in the bypass position during normal reactor operation following completion of the daily checkout. During the daily checkout procedure, the operator depresses the switch to bypass the trip and holds the switch in place while the well pump is de-energized. The switch must be continuously depressed until the scram actuates, after which the operator releases the switch, automatically retuming the well pump loss of power trip to permissive, and thereby verifying a low flow trip.

1

l NRC Dm- Control Desk Page 4 14 Day Report June 29,1999 The sacand part of the system modificatan implemented under 10 CFR 50.59 Evaluation and Determination Number 99 06 and MLP #99-21 was to install a well water mode low flow trip bypass switch. Installation af this switch allows for bypassing the aernadary low flow trip in the well water mode. Since this switch is to be utilized on a quarterly checknut, this switch is physically installed inside the console to prevent inadvertent actuation.

Procedure Changes As indentad ab 5,, two mmor procedure changes have been implemented in order to address better the intent of the Tech Specs. Procedures have been modified to allow for chkia: the low flow trip indanandaatly from the loss of well pump power trip per the first part of the modification. The trip is now to be checked qualitatively (go/no go) on the daily checkout (SOP-A.1) and quantitatively (setpoint verified) during the quarterly scram checks (SOPO-0.5 under the Q-1 Surveillance).

The following is an abbreviated summary of the procedums as they existed and the changs as implemented in response to this violation.

e UFTR SOP-A.1 (Preoperationd Checks)- Daily Chachaut Changes Anolicable Previous Stens:

1. De-energize well pump.

2. Simulate reactor power > 1 kW.

3. Verify trip, '

4. Start well pump and reestablish fkm.

5. Verify trip can be reset (scram cleared).

6. Return all controls to normal @g positions.

Corrected Steps  !

1. Bypass loss of well pump power trip. (See system modifications for details).

2. Simulate reactor power > 1 kW.

3. Verify trip.

4. Retum all controls to normal operating positions.

o UPTR SOP-0.5 (Quality Assurance Program)- Quarterly Scram CLeck Changes (Q-1 Surveillance)

'Ihe corrected Q-1 Surveillance procedure for cwklag the secondary coolant loss of flow trip addresses two cases-loss of city water and loss of well water-so that Step 9 is evaandadt o Steps 9 and 10.

Previous O-1 Survainaa~ Sten 9 (SECONDARY COOLANT FLOW lossh

1. Place secondary cooling water scram logic in city water rnode, supply cooling from

' well pump.

2. Simulate reactor power > 1 kW.

rw -

L . . .

NRC Document Control Desk Page 5 14 Day Report June 29,1999 l

1

3. De-energize well pump. j 4 Verify trip.  !

)

j l Sten 9: Corrected Sta= for SECONDARY COOLANT FLOW loss (well water):

\

1. Supply cooling water from well pump with scram logic in well pump mode.

2. 'Ihrottle secnndary cooling flow.

3. Verify trip initiates at 60 gpm.

4. Restore normal secondary flow.

Sten 10: Corrected Steos for SECONDARY COOLANT FLOW loss (city ==*ar):

l. Dm izes well pump ifrunmng

2. ShiR scram logic to city water mode.

3. Supply cooling from city water.

4. Throttle city water flow. Verify scram initiates at 8 spm.

5. Restore normal seenadary coolmg line-up.  !

Since the scram logic is now in se well water mode, a different contactor is to be shorted across to perform I

the loss of pump power trip check per the old Step 8 of the scram checks (Q-1 Surveillance). The second part of the system hardware modificatum, mstalling a switch interior to the console, implenents this change.

Since the trip on loss of =~andary enalant pump power is now assured to be chectred separately, some changes have also been made to Step 8 in the quarterly scram checks (Q-1 Surveillance) as detailed below.

Anphcable Previous Sten 8 for SECONDARY COOLANT PUMP Power T ==:

a. ShiA scenadary cooling water to city water mode. Do not shiA scram logic from well water mode. Ensure there is sufficient flow to clear scram light on motor control panel.

If there is insufficient flow, go to section 8.d of this procedure,

b. With both city water and well pump supplying flow, raise any blade to 40 units.

c. Simulate reactor power greater than 1 kW. Shut off well pump and verify scram.

Proceed to section 9.

d. If there is insufficient flow to clear scram light, shiA to well water only. Jumper TB l-9 to TB 2-9 to bypass the secondary flow scram Raise any blade to'40 units. Simulate power >l kW and turn off well pump. Verify scram and remove the jumper.

Corrected Stgp 8 for SECONDARY COOLING PUMP Power Loss-  ;

a. ' Ensure scenadary cooling is in well water mode. Place hen ~lary Cooling Flow Trip Bypass switch in the bypass position.

b. Raise any contml blade to 40 units and shut off the secondary pump. Verify trip.

Place Bypass switch (internal to console) back to permissive position.

1

1; 3

, NRC Document Control Desk Page 6 14 Day Report June 29,1999 1 MANAGEMENT REVIEW

~ The Executive Comnuttee of the Reactor Safety Review Subcommittee (RSRS) reviewed this event on i

i June 10,1999 followag internal review by Reactor Management and agreed that this event should bc j promptly reported to NRC as a potential Tech Spec violation based upon the failure to adequately measure the secondary flow trip point and that the trip occurred from lack of flow versus pump de-actuation as ,

required in the Tech Specs.1hc Executive Comnuttee also noted that the event had been reported by i "laakaaa briefly on June 8,1999 and in a more da*=M conversation on June 10,1999 when the violation of a Technical Specification was agreed upon with Somor Project Manager Ted Michaels. Subsequently, 1 a one-day notification letter was faxed on June 10,1999. This letter is Mek'=aat I. At this time the Committee also myiewed UFTR Management recommendations on the system hardware and procedure changes previously described. The Executive Committee members agreed there had been negligible impact

- on reactor safety or the health and safety of the public. Indeed, if the maranday trip did not exist, the high t+4ere, flow and other trips on the primary <=Imat system would assure no safety limit was ever approached j l

Subsequently, the full RSRS reviewed this event again at a regular meeting on June 16,1999, agam notmg that there was neghgible impact on reactor safety or the henkh and safety of the pubhc. The full RSRS ,

also reviewed and approved 10 CFR 50.59 Evaluation and Deternunation Number 99-06 to implement the changes previously described. Under this modification, the charges outlined in the Corrective Action section of this letter were implemented and tested to be satisfactory on June 17,1999 under Maintenance Log Page

1. 99-21 with requisite training naaA-*ad for licensed operators on June 17,1999.

NRC NOTIFICATION

' After determming that a possible Tech spec violation had occurred and followmg discussions between the Actire Reactor Manager and the Facility Director, NRC Headquarters was informed of this event per a brief telephone conversation between the Acting Reactor Manager and NRC Senior Project Manager Ted Michaels on June 8,1999 relative to a potential vialatian of techmcal specire=*iaas. Another telephone call was placed on June 9; however, the project manager was not available, though a message was left.

Subseqecntly, on June 10,1999, the event was discussed in detail with Mr. Michaels whose evaluation was that a vialat= had occurred. The Tech Spec required one<iny notification was submitted by fax on that date as well, since a decision was not reached on June 9 due to Mr. Michael's unavailability.

The event was again discussed with Mr. Michael's on June 22 when the Tech Spec required 14 day report was due to be maded explaining that all corrective actions had been fully implemented and receiving penniamma to delay this 14 day report one week in the interest of accuracy and completeness, especially

. since the violation itself had been completely addressed CURRENT STATUS / CONSEQUENCES As 66 ut the Reactor Safety Review SuW -W met on June 16,1999 and reviewed this occurrence and approved close out of this occurrence subject to mw~=41 checks following implementation of the 10 CFR 50.59 Evaluation and Determmation changes under Mamtenance Log Page #99-21. 'Ihe RSRS agreed with actions taken and with the initial staff evaluation that the occurrence did represent a potential violation of the UFTR Technical Specifications and should be treated as promptly reportable which was accomplished. The RSRS also agreed with the earlier evalaation that the UFTR never failed to trip for a

e

• l

. I 1

NRC De-' Control Desk Page 7 14 Day Report June 29,1999 condition reachmg or exceeding the i inie Safety System Settmg on the secondary flow in the well pump mode which is the mode in use for essentially all operations requiring secondary flow. Reactor Management and the RSRS also agree that this occurrence is not considered to have involved any reduction in reactor safety margins and it is not considered to have involved any significant effect, potential or real, on the health and safety of the public. With successful impW-*% of the changes to assure effective measurement of

==tary coolant flow trip, this occurrence is now considered closed.

If further information is needed, please advise.

Sincerely, tJ '

- William G. Vernetson Director ofNuclear Facilities WGV/dms At*=A - +

cc: T. S. Michaels, NRC Sr. Project Manager J. Wolf, UFIR Acting Reactor Manager RSRS Memben Sworn and subscribed this 79 day ofJune 1999.

Pg DANIEL J. SANETZ is COMMis5CN # CC 682050 EKMRE5 SEP 21,2001 Of ATLANTIC O D CO., INC.

NNotaryMblic I

i i

. . 4 M

I A l M, 1 V

L F L E E I W L

E /

R

- \

E V )

L D A L V 'l O

(

K

[-

E .

C m e

N t N .

O I

s y

2 T

~

7 1

EC LE S

- ~ - ,4- PN g

- MN 2

+

- AO SC n _

i

~ l g o W

~

o O C L

"2 F r R

/ e

[,lv ,l 1

TS E

V O

N O

t a

W HS I y GA T r _

IL EC SG LE a PN d MN n b ~ T ~~ _ AO SC R

o -

OEE c -

TS e j " S

• 2 E Nf1 4

/

1 h/ 1 EV L

A IR AO RT R

T V DS F

' 6 f U M

A IgV .

V, I 1

E [

' E ylX f l T ,

R 9

~

o N

. ll

)R E

G d/ UE UK R

F U

G I

• N CA 3 A AE H vR "3 C B X -

E

~ - _

T .

S A R E E N O X H H O T O T I l I S

IN WA B C -

_ NE OC UP LI L t (

M O FD E -

P N U O R

IY I F F

.. LL SL A

9 _

R 1 E . M .

T - R A O W

Y T

am N

I C . .

'e/

/

i

4 d ~

I" v p l

I a m I' V I I

8 " u P

W

• k g /

l a o " n t 2 o wO i

olF e / et T ic 1

l p k_ pe m

a T an - mn an y s s so C

)

W E

kc e N e lv. w o (

h a '

l f

Y r e e C 2 .

1

/ v i n m O L e t

s y

s n g i

o n i

et r y _

l pe c e i.o

- - ~ - - mn e n o e w C S o T S r C e t s h s ev e nm t a

4 i r a

ig la r o w tl r sC aa t e CV D S y

• r te-r a

• a V a

• r t

S U > - n.

e e

s "2

/ a 1

r _

_ r r

. 0 k e u e y 1 a t g e a n # r .

Cy C B 2

/ m E T A R u

u U c G Y ~ V a I F

V

^ v (o r l

o, e cil cCV

'v ta e

t t

s n

e #

g h

e e

n a

t n C r r

ew N, t I /

T B

t eoh l l et w o

t a

sFc t l e l

w edw Fa ns i

o re F x

s e

H T

r ov o s e tl B C t oa n iV t

l e d m e u o v F r e F r o P T w.

l o

f

- ke a

B r

e t

a F

y t

i C

. . - 4

)

ATTACHMENT I

@ UNIVERSITY OFFLORIDA Nuclear Reactor Facility 202 Nucleu sciences Center Department of Nuclear and Radiological Engineering g,. , y g",118300 ,,

Tel: 0 52)392-1429 l Fax: O 52)392-3380 Fanail: vernet @serywl.nuceng.ufl.edu June 8,1999 l

l A'lTN: Document Control Desk Potential Tech Sp . Violation U.S. Nuclear Regulatory Commission - Section 4.1 Washington, DC 10555

Dear Sir.or Madam. ,

Re: University of Florida Training Reactor Facility License R-56, Docket No. 50-83 As per te'ephone conversation on June 8,1999 between NRC Senior Project Manager Ted I Michaels and UFTR Acting Reactor Manager Jim Wolf, relative to potential inadequacies in checking secondary cooling low flow trip, this letter is to document the conversation and that there may be a potentially reportable occurrence per UFTR Technical Specifications Section 4.1 and Section 6.6.2 delineating requirements for special reports.

Sincerely, 4

James W. Wolf Acting Reactor Manager JWW/dms cc: T. S. Michaels, NRC RSRS W

Swom and subscribed this Ib day ofJune 1999.

DANIEL J. SANETZ COMM155CN # CC 682050

}I

?

8 EXPIRES SEP 21,2001

'4 nt=SDIN O., INC.

No@PubEc

-