ML20154L585

Meeting Rept 50-293/M88-312 on 880831.Discusses NRC Concerns Re Rosemount 1153 Transmitter Ringing Problem & Util Approach to Resolving Issue

ML20154L585
Person / Time
Site:	Pilgrim
Issue date:	09/06/1988
From:	Anderson C, Cheung L NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I)
To:
Shared Package
ML20154L580	List: ML20154L576 ML20154L585
References
50-293-M88-312, M88-312, NUDOCS 8809260225
Download: ML20154L585 (23)

Text

__________ _. . __. _ _ _ _ _ _ _

4 U.S. NUCLEAR REGULATORY COMMISSION iEGION !

Meeting No. 50-293/M38-312 Docket No. 50-293 License No. OPR-35 Licensee: Boston Edison Company Facility Name: Pilgr_im Nuclear Power Station Meeting Location: NRC Region I, King of Prussi_a, Pennsylvania i Meeting Oste: 3usust31,1988 u

Prepared by: C O b ~

d Cl8 ef 9/G/ff

. CIidung S . Reactor Engineer d te Approved by: O .

I- 9 6!/F  :

C. Asderson, Thief 71 ant Systems Section date  ;

l Conference Sumary: i l

A technical meeting was helt on August 31, 1988, in the Region I office to I problem  !

discussNRC'sconcernsregardngaRosemount1153 that had occured in two Region I plants, and to discuss transmitter"Ringing"sapproach Boston Edison to resolve this issue for the Pilgrim station. l t

l l I l

t 8809260225 830906 FDR ADOCK 05000293 i G PDC i

DETAILS 1.0 Conference Attendees Boston Edison Company (BECO) i d

S. Dasgupt4, Control Systems Division Manager l P. Hamilton, Compliance Olvision Manager

, D. Richard, Control Systems Principal Engineer R. Williams, Sr. System and Safety Analysis Engineer 4

United States Nuclear Regulatory Commission J. Durr, Chief. Engineering Branch C. Anderson, Chief, Plant Systems Section i N. Conicella, Operation Engineer L. Cheung, Sr. Reactor Engineer i

2.0 Background and Purpose of Meeting Rosemount 1153 transmitter "Ringing" problems which caused reactor trips and inadvertent actuation of engineering safeguard systems were identified 4 at Hope Creet and Nine Mile 2 during their startup periods. A letter was sent to the licensee from the NRC on January 29, 1988, identifying this as

! one of the six issues to be addressed for the Pilgrim Power Ascension Program. 1

The licensee responded to all six issues in their February 29, 1988 letter l 1 to the NRC. Subsequently, a management meeting was held in the Regior. I (

) office on April 8, 1988, to discuss the overall power ascension program. l 1

During that meeting, the licensee coamitted to submit a separate response l to the NRC's concerns regarding "Ringing" problems associated with L Rosemount 1153 transmitters. l

\

The licensee submitted their response (BECO Letter No.88-117) on l August 4, 1988. This letter addressed the Itcensee's approach to resolve  !

the NRC's concerns for the "Ringing" problem. Since several technical 1ssues discussed in this letter were not clear to the NRC, a meeting was r requested by the NRC to discuss these technical issues. This resulted in  !

the current meeting being held on August 31, 1988.

3.0 Meeting Summary The meeting was opened by J. Durr, Chief, Engineering Branch of NRC. f P. Hamilton, Compliance Division Msnager of BECO outlined the s: ope of i BECO's presentation. The technical elements of BECO's approach to resolve the Rosemount 1153 transmitter "Ringing" problem were presented by i j R. Williams, Senior System and Safety Analysis Engineer of BECO. j l

I

! i

1 I

! I 1

n 3

l The following summarizes the t)chnical elements of the licensee's  !

i presentation: ,

I a) The "Ringing" problems at Hope Creek and Nine Mile 2 were due to  :

i high frequency (3 to 20 Hz) harmonic type mechanical noise 1 (pressura pulses), i This mechanical noise was probably caused by:

b)

- Air in instrument lines from flexible hose or routing considerations i (e.g., sloping, low points etc.)

- Long runs of small bore tubing

- Instrument tubing support (long unsuppvrted spans) '

.t c) The Rosemount 1153 transmitters used at Hcpe Creek and Nina Mile 2 l were fast response type (60 millisec>nds), tierefore, these -

transmitters see this noise.

[

i d) The Rosemount 1153 transmitters used at Pilgrim for reactor water i

) level measurements contain filters which increase the response time  :

4 to 350 to 500 milliseconds, thus filtering out the 3 to 20 Hz harmonic !

,~

noise. l e) Pilgrims's instrument lines contain no flexible hose, are properly  :

routed and supported, and consist of short instrument tube runs.

a i

f) Other operating plants (Monticello, Hatch, Vermont Yankee, and .

i Dresden) with similar instrumtnt installation did not hav. inadvertent !

! trips of this nature. [

Other topics, such as the Plant Information Computer System (EPIC) and train L steam flow measurements, were also discussed at the meeting. The licensee  !

presentation is outlined in their handout which is provided as Attachment A to l 1 this report.

i j 4.0 Conclusion i The NRC staf f agreed with the licensee's conclusion that it is unlikely that i the problem of Rosemount 1153 transmitters will cause unnecessary trips at the  !

Pilgrim statier. F I

i i l

1 i

! I 1 1 i ,

l l i

i

^

M a4fous-Z A Rosemount 1153 Transmitter "Ringing"Issue (Monitoring at Pilarim Station)

Presentation to the NRC Auaust 31,1988

~ .

Trips at Newer BWR's Alert industry To Rosemount Transmitter Sensitivity Problems Hope Creek and Nine Mile 2 Reported inadvertent SCRAMS .

Resulting from Reactor Water Level Signal Noise Industry was Notified of the Problem by GE

- GE Issued RICSIL 12 With Supplement and S!L 463

- RICSIL 12 With Supplement - Provide 6 iraformation ,

Based on GE investigations

- SIL 463 - Provided Discussion, Recommended Actions, and Root Cause Analysis Guidelines

- Boston Edison Documented Pro 5'"m (PCAQ NED 87-142, NED 87-187) l

. / .

2- _

Hope Creek has Experienced inadvertent Trips involving Rosemount 1153 Transmitters

- Turbine Stop-valve Closure caused Pressure Transient 3

- This Resulted in inadvertent Trips From the Reactor Water i Level system I

j - Trips Were Caused by the above Amplified Pressure i

Transients Being Detected by the Fast (60 millisecond) l Response Time 1153 Transmitters 1

! - Corrective Action was to Add Filter Capacitors to Slow the

! Response Times of the 1153 Transmitters to Under 300 ms l

1 i

\ .

l 3~ ~

! Nine Mile Point 2 Has Experienced inadvertent Trips

~

i Involvina Rosemount 1153 Transmitters l

> - Inadvertent Trips Occurred During Startup Testing

- Process Perturbations Caused Pressure Transients 4

I

! - Flex Hose Instigated or Amplified Pressure Transients in j the Reactor Water Level Instrument Sensing Lines I

) - Instrument Line Slopes Contributed to Pressure Transients

- Trips Were Caused by the above Amplified Pressure i Transients Being Detected by the Fast (60 millisecond) l Response Time 1153 Transmitters

)

I 2

- Corrective Action was to Add Filter Capacitors to Slow the

) Response Time of the 1153 Transmitters to 300 ms l

y ,

l 4 -

l BECo Acts to identify and Resolve issue I

i j

BECo Surveyed 10 BWR's for Problems

! - Nine Mile Point 1&2 - Grand Gulf j

- Perry 1&2 - Limerick

! - Clinton - Monticello

) - Quad Cities - Fitzpatrick

- Dresden - Hope Creek i

i l Evaluated Pilgrim's Succeptibility i

l w _- __ _ - - -w - _______._ _ _ _ _ _

5-

)

Sensitivity Linked to Combination Of Process Noise and i Instrument Line Characteristics I

\

l Only Rosemount 1153 DB5 Transmitters with Fast i Response Times (less than 60 milliseconds) are Sensitive i to Process Noise

! - Noise (Ringing) Frequency Between 3 & 20 Hz i

I A

l Noise Effects can be Aggravated by:

l

- Air in Instrument Lines From Flexible Hose or improper j Routing (e.g., sloping, low points etc.)

- Long Runs of Small Bore Tubing l - Improper Instrument Tubing Support (unsupported spans too long)

Inadvertent Trips Have Resulted From Combination of Noise, Transmitter Sensitivity, and Instrument Line '

Anomalies.

i

l s -

Random Noise is A System Response to Process Variables l

Process Noise l - Random Noise is inherent in All Dynamic Systems j - This Noise is Independent of any Resonance in the System i

Random Process Npjise

j

- 7 .

! Ringing is a Characteristic Response to a Stimulus l Every System Has a Characteristic Resonant Frequency at Which it will Vibrate in Response to an Appropriate Stimulus The Stimulus can be Transient or Continuous in Nature Resulting in Transient or Continuous Ringing n-Transient Ringing

.OOOOO .

VVVVV Continuous Rinoing

s- .

The Rosemount Transmitter Does Not Ring The Rosemount 1153 Transmitters are inherently Overdamped and cannot Ring The Rosemount Transmitter output can Only Reflect tise System Variable it is Monitoring System Ringing will be Reproduced by the Rosemount output if the Transmitter has a Fast Enough Response Time Therefore a Fast Response Time Transmitter may give the Appearance of Ringing When it is Not

9 Pilgrim Has Two Applications Which Require Review for Transmitter Sensitivity .

Two Applications of Seven identified in SIL 463 use Rosemount 1153 Transmitters

- Reactor Water Level

- Main Steam Line Flow 1153 Transmitters not Used at Pilgrim for Many Process Variables Exhibiting Problems at Other Plants Other Pilgrim Applications use Different Sensing Instruments With Good Performance History Other Pilgrim 1153 Transmitter Applications are not Noise Problems and/or do not Cause Trips h_ __ _ _ w _ -- iw --w -- w-- ,w- - --

m ,w -- w-

10 l BECo Concludes Rosemount 1153 Transmitter -

Response Time Problems as Improbable at Pilgrim l

i I Reactor Water Level

- Transmitters are Slower Acting (350 to 500 millisec.)

l No Flexible Hose Short Instrument Tube Runs Lines Properly Routed and Supported .

" " - et * ,

11 ,

BECo Concludes Rosemount 1153 Transmitter Response Time Problems are Improbable at Pilgrim j Main Steam Line ,

J Large lJiargin Between Nori 1al Operation (100 l

psid) & Trip Set Point (121 ssid) j Industry Noise Lower Than Margin by Factor of i 10 No Flex Hose; Lines are Short & Properly

Supported and Sloped i

l Industry Has Not Experienced inadvertent

! Actuations From Main Stearn Line Ringing i

I i

l

~ 12 -

)

Pilgrim is Similar to Plants Having No inadvertent Trips l

Due to Fast 1153 Transmitter Response Time

! These Plants have no long Tubing Runs Do not use flexible Hose l

l

~

Do not use Fast Response Time Transmitters i For Water Level Applications

- Monticello

- Hatch

- Vermont Yankee

- Dresden l

, . , _ _ . . _ _ _ _ .~~ ._ . . . _ . . ._ , , _ , . .

13-Pilgrim Takes Conservative Measure to Monitor for Ringing Sensitivity  ;

Monitor All 56 Rosemount 1153 Transmitter Outputs .

Monitor With Transient Recording in EPIC Design Complete, Construction Scheduled for October ,y,yg.g 1988 Completion Tie-ins are Expected to be Complete in October, Work Can Continue During Plant Operation with Final Test and Turnover During Power Ascension Monitoring Expected to Confirm Evaluation l - Detects Sensitivity to Noise

! - Trends Noise information

- Provides Root Cause information

~

I .

14-Monitoring the Rosemount 1153 Transmitters is Not a Safety issue The Likelyhood of an inadvertent trip at Pilgrim due to Fast Response time instruments is very low.

- Proper Plant Safety Responses to Situations Requiring RPS / ECCS FunctionsWill Not Be Prevented if Monitoring is Not Available.

Events at Ollier Plants Did Not Result in Adverse Safety Consequences.

Industry Experience for Plants of PNPS Design has Shown no inadvertent Trips From Rosemount Transmitter Sensitivity i

I

,-c- r - - . . ,

- _ y--. ,. , . - - - . , , - - - . , - - . - - - - - _ - _ _ _ - _ _ _ - - - - - - - - - - - - - - - - -

15-Monitoring of Rosemount 1153 Transmitters is Not a Re-start Issue Design Review Concludes Low Likelyhood of Getting inadvertent ATS Trip / Initiation Signal i

)

! - Design Weaknesses Found at Other Plants Do not Exist at Pilgrim ,

Pilgrim Reactor Water Level Instruments Have Larger Time Constant l

Than Those Triggering Trips at Other Plants i

l i

Pilgrim Main Steam Line Flow Transmitter Setpoints are Well Above l Normal Output Providing a Large Margin

! Industry Experience for Plants of PNPS Design has Shown no l Inadvertent Trips From Rosemount Transmitter Sensitivity i

j Monitoring Capability is Scheduled to be Available Before

Completion of Power Ascension Plant Response to an inadvertent ATS Signal is within Design Basis l

1 i

9 Other Reported Rosemount 1153 Transmitter

~

Problems are not Restart issues Transmitter Lock-up

- Caused by Metallic Filings in the Dielectric Silicon Oil

- Failures are Random and Easily Detected By Daily Surveillances

- Initiation Logic Requires 2 out of 2 Failures - Improbable With Daily Surveillances Transmitter Drift

- Drift Caused by Oil Leaking From Diaphragm

- Failures are Random and Easily Detected By Daily Surveil?ances

- Initiation Logic Requires 2 out of 2 Failures -Improbable With Daily Surveil lances

., Rosemount J153 Transmitter ' Ringing' Safety Concerns Q. Why are Rosemount transmitter problems not a safety concern?

A. Problems will result in premature (conservative) safety system initiation.

The following is a discussion of the three separate problems which have been identified.

TR ANS MITTER RESPONSE TIME - When an underdamped system is exposed to a pulse or step transient, its output signal will overshoot the final steady state signal level and oscillate at the instrument's resonant frequency at a decaying amplitude.

ASSESSMENT OF RISKS:

. Fast response time could cause inadvertent trips /

initiations of RPS and ECCS systems. This would be caused by process ringing, noise or transients in the system. This is not a safety problem, as the safety systems will have performed their safety functions;  ; 1 however, challenges to these systems should be avoided because of the potential to degrade system reliability. Rosemount transmitters are inherently not susceptible to ringing as they are underdamped and will not overshoot or ring. However, if the system '

being monitored rings in response to a process transient, a monitoring transmitter with a sufficiently ,

fast response time will produce an output which mirrors the system ringing response.

The Rosemount transmitters installed at PNPS for reactor water level are typically slower than that required to respond to the noise characterized by GE in .

SII. 463. If the noise in PNPS systems have the same characteristics as has been documented at other llWR's, the installed transmitters will not be affected and no inadvertent trips / initiations will occur.

l

• A large transient in the water level system could ,

alternately initiate and trip the HPCI system. This was i observed at Hope Creek during their turbine trip event l but resulted in no loss of function to the HPCI system. ,

This is attributable to the short duration of this l transient (less than 3 seconds) which would prevent i:

the system from actuating. i

+ Noise in the main steam line may cause premature  ;

actuation of the high flow group 1 PCIS actuation and  ;

the resulting SCRAM. The main steam line system at i!

PNPS were designed to accommodate full flow in three  !

of the four steam lines. Consequently, the high flow  !

setpoints are set at approximately 25 psid higher than f the normal signal level. Therefore in order for noise to ,

trip these instruments, it must have a magnitude of '

approximately 25 psid. This magnitude of noise is ,

physically unrealistic and well beyond the range of i nose observed in the industry (Note that a high speed data acquisition system would be required to determine the nature of process noise t

at PNPS.)

TRANSMITTER LOCK-UP -Transmitters have failed due to metallic  !

filings in the dielectric silicone oil. This causes the transmitter '

output to shift and remain offscale hi or lo and not responding l to the transmitter input. ,

j ASSESSMENT OF RISKS: l This failure mode is random in nature and easily detectable by daily surveillances which would result in the immediate replacement of the faulty  !

transmitter. This is not a safety issue as the initiation  !

logic requires at least two instruments to provide j signals. The probability of two instruments failing in '

the same direction within a 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> surveillance period is vanishingly small and has never been observed in the industry. l i

f f

e TR ANSMITTER DRIFT DUE TO LE AK AGE - The transmitters have silicone oil filling their diaphragms. The oil has been observe.d to leak from the diaphragms resulting in transmitter drift and sluggish response.

ASSESSMENT OF RISKS:

The failure mode does not occur immediately but takes a length of operation before it is exhibited. It is easy to detect by performing daily surveillances which would result in the immediate replacement of the faulty transmitter. This is not a safety issue because the initiation logic requires at least two instruments to provide signals. The probability of two instruments exhibiting this failure mode within a 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> surveillance period is vanishingly small and has never been observed in the industry.

4 l

, ,-,e r --- w----t-- "P-*v.