ML20005G907

Summary of 891218 Meeting W/Rosemount Re Loss of Fill Oil in Rosemount Transmitters.Vendor Indicated That Failures of Both Models 1151 & 1152 Transmitters Due to Loss of Fill Oil.List of Attendees & Rosemount Technical Bulletin 4 Encl

ML20005G907
Person / Time
Issue date:	01/11/1990
From:	Berlinger C Office of Nuclear Reactor Regulation
To:	Rossi C Office of Nuclear Reactor Regulation
References
REF-PT21-90 NUDOCS 9001230272
Download: ML20005G907 (96)

Text

_ _ _......

. " " N'o UNITED STATES j,,% m// gg NUCLEAR REGULATORY COMMISSION s g%-W,e wAssiwoTow. o. c. rost4 g}

M+.,,W,,/

Januny 11, 1990 I;EMORANDUt! FOR: Charles E. Rossi, Director -

Division of Operational Events Assessment Office of Nuclear Reactor Regulation FR0ll:

Carl H. Berlinger, Chief Generic Communications Branch Division of Operational Events Assessment Office of Nuclear Reactor Regulation

SUBJECT:

DECEMBER 18, 1989 MEETING WITH ROSEMOUNT REGARDING LOS$ Of FILL-OlL IN ROSEMOUNT TRANSMITTERS

/

On December 18, 1989 theGenericCommunicationsBranch(OGCB),theInstrumenta-tion and Control Systems Branch (SICB), the Vendor Inspection Branch (RV1B),

and other NRC representatives met with Rosemount, Incorporated to further discuss the loss of fill-oil in Rosemount transmitters issue. Specific areas that the staff had recuested Rosemount to address during this meeting included their recorraended surveillance and testing procedures to detect Model 1153 and 1154 transmitters that may be leaking fill-oil and the susceptibility of Model 1151 and 1152 transmitters to failure due to loss of fill-oil. A listing of the meeting attendees is provided by Enclosure 1.

!!r. Steve Wanek, Vice President Operations and other Rosemount representatives presented a general overview of tneir proposed diagnostic guidelines. These guidelines primarily encompass trending calibration data, trending operational cata, sluggish response, and process noise analysis.

Rosemount indicated that they believe these proposed diagnostic guidelines can be utilized for all types of loss of fill-oil failures, in addition, Rosemount indicated that additional detailed information concerning these guidelines would be provided to industry by their Technical Bulletin No. 4 (a copy of which is provided by Enclosure 2).

Rosemount's presentation indicated that failures of both Model 1151 and 1152 transmitters from loss of fill-oil have been confirmed.

Rosemount also indi-cated that they intend to issue an informational letter to all U.S. nuclear facilities to inform them that Rosemount Model 1151 1152, 1153, and 1154 transmitters and Model 1199 remote stals utilize oil filled systems and, thus, could be susceptible to failure due to loss of fill-oil.

In addition, Rosemount indicated that they manufacture transmitters that are resold under private label by other entities and that they are aware that a secondary market exists for new or salvaged transmitters, r

CONTACT: Jack Ramsey, NRR I

'y, 492-1167

,(m d

N E gg CENTS C M g'O' s

i.

E8 M 8E5 M A L n

eon J

t o

M Januny 11, 1990 Charles E. Rossi A copy of the slides presented by Rosemount during the meeting is provided by.

Mr. Paul Blanch of Northeast Utilities made a brief presentation utilizing data obtained from the transmitter monitoring system that has been implemented at Millstone Unit 3.

Mr. Blanch's presentation highlighted how the transmitter monitoring system at Millstone Unit 3 can be utilized to detect and character-ize transmitter drift.

i The NRC staff informed the meeting attendees that regulatory action, such as l

issuance of an NRC bulletin, to address loss of fill-oil was being considered.

In addition, the staff noted that by letter dated December 1,1989, Rosemount was requested to provide additional information concerning transmitter failures due to loss of fill-oil.

(A copy of this letter is provided by Enclosure 4.)

Rosemount's response to this request was prosided to the staff imediately prior to the December 10, 1989 meeting.

Rosemount has indicated that they consider the information contained in their response to be proprietary.

The staff is currently reviewing the information provided by Rosemount.

Oticinal Signed Dy Carl H. Derlinar Carl H. Berlinger, Chief Generic Communications Cranch 3

Division of Operational Events Assessment Office of Nuclear Reactor Regulation

Enclosures:

As stated DISTRIBUTION w/ enclosures FJMiraglia, NRR CERossi, NRR CHCerlinger, NRR JERamsey, NRR JAcalvo, NRR CHaughney NRR

,CentralfFhesi DCB PDR DOEA R/F OGCB R/F JRamsey R/F Document Name: DECEMBER 18 MEETIN MINUTES C/0b[5A:NRR 0

aNRR JER 01 9[y CHBerlinger 0

01/f0/90 t'

LG3bUUux) U 1

t 6

j MEETING G4 30SDOGC 'IRANSMITIERS l

)OffaY IECEMBER 18, 1989 NUC12:AR REI301A'ICRY COMISSIQi Q4E WHI'IE TLDC NORn!

NAME ORGANI2ATIQi HICt1E Htky ktenru fcorm e ver-cau wr-4M-432-s rv /

A Sn as Auik k O SC A4A n k $ 4't.

(blA VO h-WS $

Swe weL

%,,,c a.,1 I n e un 1 sa stay

?

0 0 llbcr,,o 4' O 3

(~>< w.

(Ti 4 ) 7W4%V Ad94

)

/

/

l'th L

$J dt (*s/

A (trdf/3'7' l'7///7/A'

,,r3 -//f -3yry bbeb 0Nle Nkdr;e $.cer kosmeA1~db)e Nic5 755-2S 39 lavrv (te./ier Te A d,.

'd i a...,fl5tw t'awa gh yig-3pi-7799 NIn1 SR /00ll 90(NC!) C/)l-tf f 21)

(7M)31,2-9/3B

'br nne A Cen

~hu m Ed e So s-N 7 - 8 m 9, Doned e\\ Ric6-J Boys Edison 6 n - eg.q-8g37 C ( C......' ~

M. mm t

.L,

6 ( t 9 Q 11 c (_

lGrk Yan N ovn kn umwd Re_

4D -St&%5Wl b% b 6J0MY4C 2d2 - 872 -/7E0 denneA U~. CditJc.B Due by C,%

(704)373,7 esp Rsw,el\\ P. Tur n tr D4s L tr (',rr(%,

C70 4) 3 7 F 1334-e l

3' Alv CRRH PSEiC

/M - 737-'/83Y R m m b At m r FL. L N,-

9m-Nc sm E u h / O, @ i'.sc.1]

0 2 ~2.,

3o( -yc2-I' W i

hA.d Adowna Psc>u,s io,e i Gs,e<e,cic.

,nt - t s o - p>m bE/EY $ AO</l V Ai/M Gd/A helwK ps, let/.

Ji> ~ 3 Y f-7//.$

b ri. P. E' tie,N GPt)Aluelea<-

(het) 3/4 - 7sby hn Reci<toeu.

klE (st) Dc-Ib9 l kaw.u xiore

/v y 7tsvoe m nery Da-Liv 9.- em Gaw 3.Tewn 4,,mAu,'af.,<sbW6, 6e-6 54r-sue.

j)p w b 8, B m n kRC-RIiI (708)776-5 776

1

,o ME1 TING CN WEDt1NT 'IRANSKITIERS M]NDAY IIX2MBER 18,1989 NUCIIAR REGULA'DDRY 03 MISSION CNE WITTE M2NT NORM NAMC CECANIETION HDE Ronald E. Rogers m n.'k nay Aut M kis) 3ss-ssst.

y

~

r tu <,

/Qcuri t rz z ( i;t ccm llu g. (1. s x ; ). y. m., ;,,

i 1

.s.. :

8o62Lv l0tSTEb he SEAw udr7'.ZAb

[h*N S'LS'3 M k se apt L C M y L IDW o$$MoaWTT'D C..

(o i 2 829 M 8 03 d o ^ <. band 6%

Tc& swyd 'T 6 L

(, l2 - % 2. I

~ 2 E f

T~m n honun kne. {ul hent. AEEMLDS bD b S*1l-511r0 DJ R Moutwo TuEticme (in) est-sc79 SAVA W ~7 A6+ue

/Yna bitev $ tera &sq h/C] YS/-W//

_ Jcc td., is s G P R "I.

(4 t50 855 - 215 6 o.N 3. 5% v's s o n soleikka Elede;e C.

Ned wi'Isoo L L1&w w-A G &HJt//an (s-Mw-xeY Acle

%r rta,na %tuubr I44 ass.ts4 Aavrrno neug,aaee uunouiwer co iris 977-assi w wb S?bem L waam n &

Fr$d sm x,R.

G a t e l. S i l e,-

11M

(&id 7s7-<n16

.,lfic H Alzo t. %mM

'T\\/ A (ces; Gu - Rn3

/*h 0sf hb Y- $

S CJ /2/n - f0 2;<

6 ~a.: a h

%% !. cm,,

,J vs.3-yoo T h.Td A m ts Tu6 bMt ad &Q K9/

G L.n.x,D))?A ddark is 20s -cccswi l

A /3 d o,e 7am um. w m ee w z3o2cm i

L

(

3 i

s;

,a i

MTI'ING ON ROSDOUNr TRANSKI'ITeRS i

PONIRY DOCDEER 18, 1989 NUCLEAR REDUIRIORY O2 MISSION me wIrre rz2xr Ncam ll

[

ume onawrzmm noe NGsta l6 mis Anne gas!Ezscnce 8/C 973-079c dEFFM1L.PcRTNG4 P4coPlc Ghs t ElsLTelc

(+S ) 9 72-2775' aem v < =. a. o,

w ar, %,v. e ca.m c sit)3ms+n MATr CHIAfMkL Wb l ~' A-e -

0*l) MA - + s-Yf d erts kl%

NA c /As.gr

/sei)vn-3.24y 0

V?neanT' % emes A/A c / A/A'k

[3oi) Vf % -0784 i

V

_EH / es, A Abf'c/AWA//N'

.5'c ) - WE ' /Eid /

Tuk W me.1 A/ac)/4'c4 /o/sc A M i -4 91-Il O n

//. 2 Alh/DU N28/NU /VfB 30/ - $92-mfd i aA bbr), nse r mt cl pan)oGc B 201 - 4 92 -tis > 9 v

I W

mem l

. Enclosure 2 1

. cet

. O ir hgy D,e Atcytical -

Icen Pr;itt M<nrebote 66M4 U $ A Vaws istpve 161219416ff0 feen 43t0012 o< 431D024 F AX 1012) 606 300B

[

t December 22,1989 i

Reference:

Rosemount Technical Bulletin Number 4 i

i

~

The attached Rosemount Technical Bulletin is the fourth and final in a series that addresses the loss of fill fluid issue on the Model 1153 end 1154 transmitters. This

/

bulletin provides the results of Rosemount's five stea program on this issue and provides the diagnostic guidelines for monitoring at ris < transmitters.

Also attached is Rosemount Report D8900115 " Leak Rate Model and Risk

[

Assessment for Model 1153 and 1154 Transmitters". This report describes the basis

- for the guidelines published in Technical Bulletin 4.

A complete list of Rosemount's publications regarding loss of fill fluid is found in Technical Bulletin 4. These are available on request from Rosemount.

In response to numerous requests for a complete list of all suspect transmitters

[

identified in our February 7,1989 mailing, we have attached that list. This list also includes new lots that have opened up since February.

if you have seen symptoms of loss of fill fluid as described in our technical bulletins, we ask that you contact us directly and provide us the following information:

Transmitter serial number Model number Exact symptoms and trouble shooting performed The time in service e.nd the application parameters (such as calibration and static pressure) 1 L

f

+1i iL

• • hteeno6 pal int.

t=

=.r,:::,;, b s Zoen Pr:,rc, MN $5344 Pagetwo Please call Neil Lien (612) 828 3100, Tim Layer (612) 828 3540, or Jane Sandstrom (012) 828 3286 if you have any information or questions on this situation.

)

Rosemount remains committed to providing q'uality products and services to the nuclear power industry. We appreciate your cooperation on this matter and look j

forward to meeting your future needs.

y Sincerely,

{

ROSEMOUNT INC,

• x%A Mark Van Sloun Nuclear Business Unit Manager e

attachments p'

-, ~ _

Rosemount*

i Technical Bulletin No.4 December 22,1989 (Supercedes December 15 version) i

1. INTRDDUCTION This bu'!stin is the fourth and final bulletin in a series that has addressed the loss of fill fluid in the Model 1153 and 1154 transml.tters. The result of fill fluid loss may be slow transmitter response and/or an inability of the transmitter to meas 0re input accurately. The focus of this fourth bulletin will be to supply the nuclear

)

utilities sufficient information to:

1

)

a) Identify which transmitters are suspect b) Perform diagnostics on these suspect -

transmitters to reduce the risks p

This is the final step in the program to understand and quantify the loss of fill fluid failure on the model 1153 & 1154. The program included:

o Identifying all symptoms of oilloss e Ouantifying oilloss symptoms in the laboratory e Defining diagnostic guidelines for Industry review e Field Testing the proposed diagnostic guidelines e Publishing Ros6 mount recommended oil loss diagnostic guidelines These diagnostics guidelines can be used to detect all forms of fill fluid loss.

l l

Page 1

The format of this bulletin will be as follows:

Section Subject 1

introduction 2

Executive Summary 3

Appendices A. Drift Analysis Guideline B. Sluggish Response Guldeline C. Process Noise Guideline D. Additional Range 9 Data i

i References

1. Rosemount Report D8900023 Revision A.

1

• Description of Process improvement on Model 1153 and 1154 transmitters."

/

2. Rosemount Report D8900115 Revision A 1

' teak Rate Model and Risk Assessment

~

for Model 1153 and 1154 transmitters."

i

3. Rosemount Technical Bulletin No.1 May 10,1989

4. Rosemount Technical Bulletin No. 2 July 20,1989

5. Rosemount Technical Bulletin No. 3 October 23,1989 b

t Page 2

2. Executive Summary A small, but statistically significant number of model 1153 and 1154 transmitters with sluggish response were returned to Rosemount. Andysis indicated that the root cause for these failures was a glass to metal seal failure internal to the sensor. This glass to metal seal failure allowed fill fluid to leak out of the sensor at a very slow rate. The major cause of tf4 glass to metal seal failure was the use of metal process o-rings which more than doubled the amount of force on the sensor compared to other Rosemount transmitter designs..

Manufacturing improvements were implemented to eliminate the potential for leaking modules. At the same time diagnostic guidelines were developed to be used to identify suspect transmitters in the field.

Rosemount has tried to quantify the degree of risk associated with using 1153's and 1154's. This is a complex issue and involves a number of interacting factors..

Because of this Rosemount has developed a number of methods to quantify the risk.

r Rosemount first evaluated the raw failure versus ahipment numbers.

We currently have confirmed 91 glass to metal ssal failures with 16 potential units in failure analysis. This is compared to approximately 14,145 Model 1153 and 1154 units shipped. Excluding 1,158 Model 1153A transmitters which did not use metal o rings, this results in an overall failure fraction of 0.82%. We have also estimated the actual in service time on the installed base to be in the order of 110,000,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />. This indicates the overall failure rate is approximately 1 x 10 6 failures /hr.

A mathematical model of leaking transmitters has been developed. The basis for this modelis standard viscous flow equations and detailed application data from the failed units. This model shows a strong relationship between the time at pressure before the failure occurred and the actual pressure applied to the transmitter. This relationship indicates that a transmitter at low pressure (<250 psi) has an acceptably low failure rate through out its life.

Rosemount understands that 5 x 10'7 failures /hr is acceptably low" in the Nuclear Industry.

At higher pressures the transmitters have a higher failure rate for some finite oeriod of time, then drop to the acceptably low failure rates.

Page 3

1 i

For example, transmitters operating at 1000 ps! gage pressure or static line pressure will have a peak failure rate in the 10 6 failures /hr range declining to 5 x 10'7 failures /hr at 60 months. A transmitter operating at 2500 psi gage pressure or static line pressure will have a peak failure rate in the 10 6 failures /hr range declining to 5 x 10*7 failure /hr at 24 months. Therefore, the product of operating pressure multiplied by the time sensing pressure (or psi-months) can be established as a criteria to indicate.when acceptable failure rates are achieved.

]

The volume of oil in each sensor and the amount required to be lost before j

performance is affected varies by sensor range code.

Rosemount has I

established that all sensors except differential units of range codes 6 through 8 that have 60,000 psi months without experiencing oil loss symptoms will have j

failure rates in the acceptable range of less than 5 x 10'7 failures /hr. The differential units of range codes 6 through 8 will obtain this acceptable failure rate J

in 130,000 psi-months.

j Other risk factors were evaluated and indicate that:

a) Transmitters normally not pressurized (stand by service) are at acceptable low failure rates due to limited time in service b) The low range codes (3 through 7) absolute and gage transmitters are at acce stable low nsk due to low pressures appied.

The last risk factor is one Rosemount has previously discussed and is the susceptibility of select manufacturing lots to experience oil loss. Evaluating the failure data indicates failures are often clustered into manufacturing lots.

Previously identified, suspect manufacturing lots have an overall reported failure fraction of over 6% of shipments while the remaining non suspect lots have a 30 times lower failure fraction of 0.2%. However, Rosemount must caution the industry we cannot be positive we have identified all the suspect manufacturing lots.

Rosemount combined all these risk assessment factors and has developed a decision tree for utilities to follow. This is included here in figure 1 for the utilities to use as guidance in determining which transmitters to put into a surveillance program.

Page 4

Figure 1 Utility Decision Tree Determine all 1153 &

1154 transmitters received *

..r u

YES Do you have any 1153 No safety concern Series A units?

4 --

NO p

No current safety YES Are transmitters concern. Return to spares or not Rosemount for in service?

l module replacement U

l I

technician operators les th n psi?

on low oil symptoms and continue operation g

n U

YES Are transmitters in

)

stand by service?

l U

YES is the pressure x time

> 60k psi-mo(all R3 5)

=

>60k psi-mo(AP/GP R610)

>130k psi mo(DP/HP R6-8) ?

NO y

ll Continue to operate using diagnostic guideline (s) until critical pressure x time in service value is met or t

until transmitter is replaced.

l l

l.

• Model 1153 and 1154 with serial numbers less than 500,000 l

containing the sensor module manufactured before July 1989 i

Plant specific analysis of application and transmitter diagnostics I

should be considered L

Page5 l:

i

Rostsmount has developed diagnostic guidelines that can be used to monitor and reduce the risk of operating a model 1153 or 1154 that may experience a loss of oil failure.

Three guidelines were developed and can be used independently or in combination to identify transmitters suspect of oil loss.

These guidelines are in the general categories of:

a) Output Drift Analysis b) Sluggish Response c) Process Noise Analysis The Output Drift Analysis can be divided into two optional analyses. First, normal calibration data (as found data versus as left data) can be evaluated to determine any cumulative positive or negative drift trends.

Second, trending and comparison of actual operatirig data on processes with redundant transmitters, can identify suspect oil loss transmitters. Both techniques have been field tested and have shown the ability to detect leaking sensors.

Sluggish Response of transmitters can be detected with two optional methods.

The first method is a qualitative test where experienced tech'niclans can detect slow response during normal calibration. Second, a bench test of the transmitter could be performed on suspect units to quantify the slow response and confirm oilloss.

Rosemount has investigated using the Process Pressure Noise signal normally present in the transmitter output as a diagnostic tool for loss of oil. We have studied the amplitude variation in the frequency domain. Variation of the output signal from normal can identify a suspect transmitter. Other process noise symptoms have been reported but have not been verified.

Process noise analysis is a vlable diagnostic tool, however it appears very application dependent and typically the data is difficult to analyze. Due to interpretation difficulties and lack of universal applicability, Rosemount is no longer evaluating additional noise diagnostic guidelines.

These 3 diagnostics are further detailed in appendices A, B and C.

Rosemount stands ready to assist the industry in regards to this issue. We also ask your assistance by reporting directly to us any transmitters with the Page 6

n.

symptoms identified in our correspondence 13 y:u. Ple:se contact Neil Lien (612) 828 3100 Tim Layer (612) 828 3540, or Jane Sandstrom (612) 828 3286.

The selection of which diagnostic test (s) are to be used will depend on specific application data that each utility will have to consider,

~.

4 a

0 m

Page 7

l

?

- APPENDlX A.

Drift Analysis Guideline i

1. Introduction h has been shown that a transmitter leaking oil will have a zero drift and a small amount of span drift as soon as oil is lost from the sensor. The amount of the shift will depend on the transmitter's calibrated range. The zero shift before response degradation varies from 4% Upper Range Limit (URL) to.15% URL, depending on range and whether the high or low side of the sensor is leaking. Thus by monitoring sensor drift, most leaking transmitters can be identifieci before they actually fail to perform correctly, t

The individual drift graphs were presented in Technical Bu!!stins 1,2 and 3 for your reference.

2. Test Guidelines Utilities supplied transmitter data from 15 failed and 75 normal operating transmitters. The data was studied to determine if zero drift and/or operation data are viable techniques for identifying leaking transmitters.

Two techniques were used in this study; a.) Trending of CalibratioriData b.) Trending of Operating Transmitter Data i

The study verified that these techniques can detect falling units, and identify leaking transmitters before significant degradation in response time occurs.

l Page A1

1 j

.o i

3. Trendina Calibration Data I

3.1 Introduction Trendina Calibration Data Calibration data is taken on most pressure transmitters on a regular basis.

)

Calibration intervals range from quarterly to calibration at refueling l

outages only. If a transmitter is leaking oil,it will show a consistent trend i

in the cumulative zero pressure calibration point. Range Codes 3 through 7 usually have large enough zero pressure shifts to be an early warning indicator. Range codes 8 through 0 will also exhibit the drift symptom, but it is very small and difficult to detect. This technique gives very limited early waming for range codes 8 through 0. The span will begin to shift as l

the transmitter is falling. Please note range code 4 may also give limited early warning if not ranged down.

3.2 Guidelines for Trendino Calibration Data Calibration data must have sufficient accuracy to resolve sustained drift rates. On some range codes and turndown conditions, zero pressure calibration drift would be large enough to be read off of 1% to 3% panel meters.

All of Rosemount's zero drift test data was taken with the transmitter calibrated from zero to URL (upper range limit), if the calibrated pressure span of the transmitter is ranged down, then either allowable % URL zero drift erroro must be converted to % of calibrated span by multiplying by the range down factor or else the utilities calibration zero pressure error must be expressed as a % of the transmitters upper range limit. To do this the range down factor (RDF) is used. The RDF is defined as URL divided by calibrated pressure span, in summary, an error in % of calibrated span can be converted to % URL by dividing by RDF, similarly, a % URL error can be converted to % of calibrated span by multiplying it by the RDF.

Page A2

i Graphs for zero drift vs. oil loss were presented in previous Technical Bulletins.

Drift limits before response time begins to degrade are summarized below in Table A1.

t TABLE A1 j

KAXIMUM ALI4WABLt CUFJff!ATIVE DRITT Toft 1153/1154 OIL 1 CSS TFANSMITTERS M10H SIDE CIL 1458 14W SIDE CIL 2458 RANOC TRANSMITTER ttR0 STAN 2ERO SPAN C00E URL

$HITT SHIPT SHIFT SHITT (t URL) p READW3)

(4 URL) m READN3) 3 30 IN H2O

+4.00

+0.20

+3 00

+0.4 4

150 IN H2O

=0 50

+0 30

• • 1.5

'+0 2

5 750 IN H2O

-0.80

+0.20

• 1.45

+0.95 6

100 PSI

-2.00

+0 60

+3.40

+1.45

/

7 300 PS!

-0.90

+0.44

+1.00

+ 0. 4 5 8

1000 PS!

-0.40

+0.15

• C.5

+0.15-9 3000 PSI

-0.15

+0.10 NA NA 0*

6000 PSI

-0.07

+0.07 NA NA

• HOT TESTED, EXTRAP0 LATED TROM EANCE 9 DATA p

e if a transmitter has a zero based calibration, calculating the zero shift as a

% of calibrated span is very simple. The zero pressure shift is simply the difference between the as found and as left zero readings divided by the nominal calibrated span of the transmitter expressed as a percentage.

The zero shift in % URL is then determined by dividing this number by RDF.

l Page A3

l Note that the percent drift data in Table A1 has asymmetrical drift limits for

)

low and high side oil loss. This is due to the fact that the table is based on a zero to + URL calibration, which is an asymmetrical range for the sensor. When a sensor is calibrated from URL to zero (fully elevated) the 1

asymmetry is reversed, and the magnitudes (not the signs) of the high side and low side values.In table Ai must be interchanged.

For a symmetrical 4 to 20mA calibration, such,as 1/2 URL to + 1/2 URL, the oil volumes and diaphragm deflections sie symmetrical.

Therefore the l

percent drift data will also be symmetrical and have a magnitude equal to the average of the high and low values shown in the table. Limits for other calibrations can be interpolated from these values or else the lower of the high or low side magnitudea can be used as a conservative choice for all cases.

/

I W

f s

L t

Page A4

t Example 1 demonstrates calibration trend charting of a zero based flow 3

transmitter calibrated from 0 to 400 in. H O at 1000 psi static pressure.

2 The URL for the range code 5 ic 750 in H 0, so the RDF is 1.88 (750/400 2

= 1.88). The published zero drift limits for a range 5 from Table A1 are -

.8% URL and + 1.45% URL Since this application has an RDF of 1.88, the applicable true zero drift limits become 1.88 times larger or 1.50% and +2.73% of calibrated span before response time starts to r

increase.

i Ngute Al flange Codo $ C411bratich Data Trendthg f

tiSXe5 Ced+oied 0 to 800 W *i20 2*

1.5 -

1*

lese bitt cf.

.225 toi&oted won g

/

".% _g

\\

g,,, pai

-0.t" (5 8tendag)

-i-1.52 bitt Lmt

- 1.5-0 too e50 650 050 scho tthe few (Dori)

After the first 12 month calibration a zero drift of (.22% of calibrated span) was measured. To establish a trend after such a small drift, the sensor calibration monitoring should be continued. To determine if, at this drift rate, failure could occur before the next calibration the oil loss lifetime should be calculated.

This is done by dividing the zero drift limit calculated from Table A1, in this case either -1.50% or +2.73% of calibrated span by the observed drift rate. The zero drift limit in this case would be 1.50% because the data is trending in the negative direction, and the calibration is not elevated. This is E!so the most conservative case because -1.50% is the smaller of both limits. The oil loss lifetime at the zero drift rate of.22% of calibrated span for 12 months is 1.5%/.22% per 12 month = 6.8 years. This means that if the transmitter is leaking oil it l

Page A5 ee

---

m.

---..-,y

__y

.--.2.

E i

\\

f

. would tak? 6.8 years at 1000 pal before the performance would start to degrade. The transmitter can be left until the next regular calibration to get further trend data. At the next calibration no additional zero drift was detected. This indicatec a very low risk of oil loss. Since both the first and j

. second calibrations showed very low cumulative calibration shitto i

compared to the 1.5% limits after almost 3 years at 1000 psi, h is highly j

unlikely that this sensor is leaking oil.

i

)

1 1

i 1

t f

I l.

L Page A6

t 4

4 y

If a transmitter is elevated or suppressed the drift at zero differential will not be known, as no measurement is made with the sensor at zero i

differential pressure. In this case the shift at zero differential must be extrapolated.

i The extrapolation is done by imagining that the transmitter output range exceeds 4 to 20 mA and extends as fat as necessary in the positive or negative direction to cross through zero' differential pressure. Thus for each successive elevated or suppressed calibration, the slope of the line 1

must be calculated, and extrapolated to determine the output current that would exist at zero d;fferential pressure. An example is shown in Figure A2 below.

n, r. u

c. a.i.u i.,. enits r.,. n...sw e.t ar.si tl 7

)

/

A i.ii s

46 -

too ' [t f,-

f-

-..Y.n

/,0. ~,'r.

. j#/

c-tt St =

(ku ten f

f

/

s.

,r y-e,x

.,,si

>.s..

15 -

'O '

utm

[

j o-

-72 too

-tM 0

mi e

. m. m>

The transmitter used in this example is a range code 5 calibrated 750 in H O to -500 in H 0.

RDF=URL divioed by Calibrated Span = 750 2

2

/250=3.0 Page A7

...=

The as found and as left data for this transmitter is as follows:

Differential As Left As Found 1

Pressure (Previous (Current input calibration) calibration)

(in H2O)

(mA)

(mA) l 750 4.000 3.920 500 20.000 19.940 The following steps need to be done to determine the transmitter zero shift.

)

i 1.) Ti.e slope of the as left data is 16 mA divided by (750 500 in H2O)=0.064 mA/in H O 2

2.) The extrapolated zero output reading for the as left data is 750 in H O 2

times 0.064 plus 4ma=52.000 mA.

3.) The slope of the as found data is (19.940 minus 3.920mA) divided by (750 minus 500 in H O)=0.06408 mAlin H O i

2 2

4.) The extrapolated zero output reading for the as found data is then 750 in H O times.06408 plus 3.920 mA=51.980 mA 2

5.) The zero shift is then the difference between the 2 wxtrapolated zero readings divided by the output span:

(52.000 51.980 mA) divided by 16 mA= 00125 Expressed as a percentage it is.125% calibrated span l

6.) To get the shift in % URL, divide by range down factor:

Zero error in % URL=.125 divided by 3.0=.04 % URL.

l 7.) This cumulative zero shift would then be trend charted by plotting it l-versus cumulative time at pressure, if it is plotted as calendar time instead of time at pressure, linear extrapolations of lifetime, as shown on page AS, may not be valid. Large outage periods between calibrations could result in forecasting a longer lifetime if they were not subtracted from the time at pressure used in the extrapolation.

Page A8

i s

The following are two case history examples of calibration trending with an j

elevated zero.

Example 2 is shewn in Figure A3. This data is from a Range Code 5 level transmittar which is calibrated with a 200 in. H O span from 389 to 2

189 in. H O and operated at 1000 psi static pressure. The URL is 750 in 2

H O for this transmitter, so RDF is 3.75 (7, 50/200). From Table A1 the 2

5 Range 5 zero drift limits are.8 to + 1.45% URL or in this case 3.75 times larger or 3 to + 5.44 of calibrated span. The cumulative zero drift is + 2%

URL or +.75% of calibrated span (.2% URL x 3.75 RDF) at the 12 month point. This drift may indicate a possible oilloss so transmitter calibration monitoring should be continued until t, significant trend is established. To determine if, at this drift rate, fature could occur before the next calibration the oilloss lifetime should be calculated. This is done by dividing the zero drift limit calculated from Table A1, in this case either 3 or +5.44 of calibrated span by the observed drift rate.

To be conservative the absolute value of the smaller of the two or 3% of calibrated span was used. The oil loss lifetime at this drift rate is 3/.75%

per 12 months or 4 years. This means that if the tra'nsmitter is leaking oil it would take 4 years at pressure (1000 psi) before the performance would start to degrade. The transmitter can be safely left in service until the next regular calibratiori.

N gure A3 mange case 5 Cattbrotten tote trending 1153005 CoEpowd

• 29 to *109 tot H70 I*

'A*

Ice &dt (, Wested spea)

+

g.

\\,,e.u,,-

e4-g

,n se

=0$'

.t.

= 13 '

7b ab 6b

.bb 1000 1700 0

im Page A9

The next three calibrations actually showed that the transmitter had a very stable zero output, with cumulative drift of.28% URL or 1.05% of calibrated span (.28% x 3.75 RDF).

This consistent trend confirmed that the transmitter was not leaking o11.

n,.r. s.

m.,,,. c 4 c.uw.u

w. w. et,,,

m e..... m i...is =

• o i-e.-

-"., 7 i

% wins. a

. I.t ;

.r -

{. t.5 -

3 3s.

... win s....i.e - >

\\.

u-6 C

200 800 600 000 1000 1200 Tene (Dort)

Example 3 is shown in Figure A4 This data is from a Range Code 4 level transmitter calibrated every 3 months with a 70 inch span from.111.9 in H O to -41.9 in H O and operated at 1000 psi The URL is 150 in H 0, so 2

2 2

the RDF for this transmitter is 2.14 (150/70). From Table A1 the Range Code 4 Zero Drift ilmits are. 5 to + 1.5% URL.

This unit exceeded the.5% URL limit after 6 months in service but continued to drift an additional 18 months before sluggish response was observed.. Clearly trend chart:ng gave early warning of an oilleak in this installation.

Page A10 1

l 3.3 Coutions for Trendina Transmitter Calibration Data Zero calibration drift can be caused by other factors, including temperature effects, sensor repeatability, line pressure effects and overpressure effects. Therefore if a sensor does not have stable zero calibration data it cannot be concluded that it is definitely leaking oil.

However field data has shown that the sustained drift caused by oil leakage is generally greater than the sum'of the other drift errors, i

Elevated and suppressed calibrations require extrapolation to determine the actual zero drift. The technique to be used has been explained in this bulletin.

When analyzing elevated or suppressed calibration data, any electronic or pressure source errors will cause extrapolation errors at zero pressure.

The magnitude and importance of the error will depend on the RDF, the amount of elevation or suppression, and the amount and source of other errors, if the extrapolation errors are too large to establishs trend another analysis technique should be used.

I For range codes 8,9 and 0, the zero drift is very small and would only provide a detectable drift rate under precision calibration conditions or in a highly ranged down condition. The span shifts would be evident at calibration but, for the required amount of oil loss for these span shifts to show, the transmitter would be readily identified by sluggishness during c8tlibration.

l L

l Page A11

3.4 Conclusions for Trendina Transmhter Calibration j

D.318 i

Calibration trending of historical zero pressure calibration data can be a 1

relatively easy method to determine whether or not a Range Code 3 through 7 sensor is leaking oil. Data shows that for Range Code 8,9 & O zero drift prior to transmitter failure is very small and calibration trending is not as effective.

1 Analysis of numerous actual field failure calibration trend charts, such as example 3, clearly show an obylous upward or downward trend over time.

in cases where a large initial zero drift or sustained low level zero drift trend is noted, the unit must be considered suspect and evaluated in a different manner or replaced to assure continued proper operation.

/

i Calibration drift data substantially confirms data and theory that field f

failures were leaking at a constant rate since they were initially pressurized. Thus, if a transmitter shows stable zero readings over a given period of time, it can be concluded that the transmitter is not leaking and no further analysis need be done, t

Page A12

4

,..n..

=

Az-&->

4m a w

i y.

l

4. Trendina Ooeratina Data 4.1 latroduction Trendino Ooerstino Data l

The purpose of this section is to provide the guidelines for diagnosing oil loss by trend charting transmitter operating data. This technique can be more accurately used with redundant pressure trtinsmitter installations.

The basic idea behind the analysis it'that redundant measurements should very precisely track each other.

Using this technique in a non redundant application is possible if the process pressure being measured is known by some other measurement, 4

or is known to be very stable, For range codes 3 through 7, trending operating point data will show that a transmitter is leaking wellin advance of any increase in response time.

Fledundant range codes 8 through 0 will show a very small operating point trend apart prior to actual failure, but as the transmitter bees response time the data should begin to trend apart very quickly. Because there is little advanced warning of response time loss, more frequent data sampling may be needed to assure a low mean time to failure on range codes 8 through 0.

It will be shown from data on page D2, that range codes 8 through 0 transmitters drift low very rapidly after they begin to lose response time to increasing pressures above their operating points.

- Sensitive drift measurements may be able to detect this drift before their response time dtgrades. Several utilities have demonstrated the ability to resolve 3% ej calibrated span over many months using computerized data acquish'an systems.

Page A13 l

C i

4.2 Trendino Oooratino Data Guidelines Periodic operating point data must have sufficient accuracy to resolve sustained drift rates. On some range codes and for sufficiently ranged down calibrations, drift would be large enough to read off of 1 to 3% panel j

meters. Actualloop measurements have adequate resolution if taken with j

quality digital ammeters or voltmeters.

Surveillance data from utilities on failing transmitters has shown that j

calibration data taken off of 1 to 3% control room meters is adequate only in certain applications where expected shifts due to oilloss woulo be over several percent of calibrated span.

j Comparing the drift in operating point data between redundant 1

transmitters eliminates all of the true zero pressure extrapolation uncertainties inherent in trending elevated or suppressed calibration data.

/

Also, the more frequent periodic data points will significantly reduce the mean time to detect a falling transmitter. As transmitters,1,ncrease time in i

service, the time interval between trend chart points can typleally be increased, j

Actual trending can be done in several different ways. One technique that has been successfully used is to call the $rst available data point for each I

transmitter baseline zero. Then periodically plot the change from initial reading in % of calibrated span. - Transmitters that are tracking each other will change like amounts over time. A transmitter that is drifting will trend away from the other transmitters. However, when the piocess pressure changes, all of the redundant transmitters will usually track right off the page on high resolution plots, p.

l l

l-1 i

1 1

Page A14

~

f E

Another technique that has been successfully used for this analysis is deviation from average. The average of all redundant transmitters is calculated, then deviatior. from everage is calculated. The drawbsek to this technique is that both leaking and non leaking transmitters will move away from the average, thus it is not always clear which transmhter is-actually drifting. However, all transmitters will remain on scale when the f

process pressure changes so drifting is e.asier to resolve.

The simplest tochnique is to plot the periodic outputs of redundant transmitters and look for long term drift trends.

The frequency for data gathering should be often enough to establish statistically valid trends.

3 -

Allowable drift limits can be U.:>tnined from Table A1. Generally it is simply the zero drift in % of URL However, when trend charting the operating

. points, the operating drift limits can be calculated from table A1 by adding the zero shift (%URL) to the span shift times operating point /URL.

4.3 Examoles of Trendino Ooeratina Data Example 4 (shown in figure A5a) shows how redundant transmitter data

+

might be plotted for 2 groups of 3 redundant transmitters. The trend charts are plotted as change from the first reading in % of calibrated span, nereu.

Trending Transsitter Data s.

2.3 -

1.5 -

i-D.*$

I

'I N

g v

I~ _

?

g

...e-W

?

-I i.e-w-j-WE====%===r

-22 n so n

ioo in iso m = m m m m w mm)

Page A15

n.

p an' y

4. -

As the process changes, all three trcnsmitters change in unison, cithough.'

- there is some drifting apart over the first 25 days of trending. However, there is no sustained drift trend over the first year of operation. This would.

L h

indicate normal operation of the transmitters.

[

Figure Alb e

- tre,.dtng transettter Data 3-g a

j

/g -x x

u.

l-.

g 5-c.b '

o 4

e f

1

-o.s-.

<,-,-x-y-x l.,.

. u..

-t" o

n wn in m iso in := rrs no en

>x iww (Ooys)

Feure A5b shows the other group of 3 transmitters.

' One of the units is drifting away from the other two which are tracking each other quite closely. The drift trend is sustained and would indicate.

the possibility of loss of oil, in this example, further investigation or analysis would need to be done to determine the cause of drift.

l 1

f.-

l l

l l.

~i 1:

!~

Page A16 I

1

~

=; y

..L

' 72;i

. _.04'

,,4j

);

Example 50 shown in Figure A6a.- This data was mOnuCIly taken from the trip unit panel meters and is for 2 groups of-4 redundant range code 9 gauge pressure transmitters. A deviation rate of about.4% URL per day

(.8% calibration span) started about day 528 days after installation. This abrupt change in drift rate, as detailed on page D2, could likely have been resolved if precision readings had been trended instead of just 3% panel meter readings.

Figure A6a mange code 9 Surveillanco Date trending 1153089 CoI&cted 0 to 1500 ti g i

i 1500 <

1400 '

1300 '

i 1200 <

1100 '

j relied one' S?6 Weys et,<essee 800 W,1 _- - M-

• • pe -w=

. =rarm=g.c3 -

t..,.-,.... ~,e

=

~ ex w100, t s

2

..es cowetee sona/ser erin rote

$10 - 570

$30 640 550 St0 S?0 550

$90 400*

61'.

0 875 The range code 9 gauge pressure transmitters in figure A6b were operating normally with no sign of drift, then one of the units began to trend down from the other 3 units. This unit was also subsequently replaced and found to be sluggish.

Figure A4D Bange codo 9 Surveillance Data trendity 1IS3C89 Coi&cted 0 to t$00,s:q 1900 '

1400 '

1300 '

1200 '

1100 '

f 1000 '

=- -_

• • a=

'* ** =_'ae a Storting to Ord'

- 3.x *

=.

,. w one.

.o,..

e.,,e g00 -

-6 67, C.5-m-

400 -

m-m-

100 0

636 648 856 e6't 678 eso e9e 9e 918 97-9M Om Page A17 i

l

f a

h

~

4.4 Cautions for Trendino Ooeratina Transmitter Data Operating transmitter drift can be caused by other factors, including

- temperature effects, sensor repeatabilhy, line pressure effects and overpressure: offects.

Therefore if a sensor does not have stable operating data it cannot be implied that it is definitely leaking ol!, However field data has shown that the drift caused by oil leakage is significantly greater than the sum of the other errors, if flow data is being trended in engineering units, care must be taken to extract actual transmitter pressure output from the flow data or else express the pressure drift. limits from taole A1 on page A3 in the _ correct flow units using the square root relationship. For example, due to the

~

square root of pressure flow function, a 2% flow error could be a very large or a very small pressure transmitter error depending on where the e

transmitter operating point is and the calibrated range of the transmitter.

However, trending engineering flow data ls valid if drift limit,s are converted into equivalent flow units.

For range codes 8 through 0, the drift is very small prior to transmitter failure and very rapid as failure is occurring.

Transmitter data which has significant process noise on the signal must be properly averaged to reduce the effect of the noise signal, Page A18

o

' 4.5 Conclusions for Trendina Ooeratina Transmitter Datt Trending of transmitter. operating data is an accurate method of determining whether a redundant transmitter is drifting and could indicate

' loss of fill fluid, For range codes 3 through 7, the data will indicate oil.

leakage before any transmitter malfunction due to oil loss occurs, in the case of range codes 8 through 0, the drift rates prior to failure will be very low unless the transmitter is calibrated in a highly ranged down condition.

For those range codes, very high resolution trend charting can detect the failure as soon as they affect the operating point and greatly reduce the mean time to detect a failure.

t i

Page A19

3..

L.'

_' Appendix B

-l Test Guidelines for Detection of Sluggish Transient Response 1

- Rosemount Model 1153 and 1154 Pressure Transmitters j

1. Introduction This section contains two guidelines. The first guideline covers detection of

. sluggish transmitters during a normal calibration. The second guideline can be used as a bench test to accurately confirm if a suspect transmitter has lost enough oil to affect operation of the transmitter.

. When a step or ramp pressure signal is applied to a pressure transmitter, there is a lag between application of the input pressure and the indicated pressure from the 4 20 ma transmitter output.

Normal sensors have a very fast response to a step or ramp pressure change, thus.the leg time is very short for a transmitter that is operating -correctly.

Beyond a threshold fill fluid loss, the response time of the transmitter will get

~

longer with continued fill fluid loss. This knowledge can be.used by plant

_i personnel to identify transmitters that are low on fill fluid during normal plant L

maintenance.

l The individual sluggish' response curves were presented for each range in Technical Bulletins 1,2 and 3 for your reference.

l'

2. Guideline for Detection of Sluoaishness durina a l

L Standard Transmitter Calibration l+

. The purpose of this guideline is to provide a very simple technique for detecting sluggishness during a standard calibration.

.in a standard calibration of a pressure transmitter, a precision pressure is applied to the transmitter so that the output of the transmitter can be verified to be correct. Typically, precision pressure sources do not apply a step pressure change, rather they ramp the pressure up over a short period of time.

Page B1

e>

e

~ 4

_ When a transmitter is being calibrated, the technician learns from experience how I

.much time the transmitter requires to reach a stable output. If the output of the transmitter is observed when pressure is applied, the technician can generally tell that a transmitter is sluggish based solely on past experience _When pressure is reduced, the transmitter output should also be observed to verify correct response.

t For range codes 5 through 9, the transmitter output current should always follow the input pressure to within 1 second, Range codes 3 and 4 require slightly longer to respond, due to the low differential pressure applied. Range 4 should respond within 6 seconds, and range 3 should respond within 35 seconds.

These are the expected times for response to 99.5% of the step or ramp input.

Typical field failures have been returned with response times ranging from several minutes to over 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />, thus a sluggish transmitter at calibration is relatively easy to detect.

3. Guideline Bench test for Confirmation of Oil Loss i

if a transmitter is suspected of loss of fill fluid, a bench test can be done to verify a

loss of fill fluid. A transmitter that has lost fill fluid on either the high pressure side or the low pressure side of the transmitter can be detected with this test. For this testing the transmitter must be recalibrated twice. First it must be calibrated 0 to URL to check for high side oilloss. Then it must be calibrated 0 to -URL to check for low side oil loss. These 2 calibrations are the worst case calibrations and will

. be most sensitive to oilloss.

In order to calibrate the transmitter O to -URL, the amplifier board must be removed and the jumper or switch set for elevated zero calibration.

First the transmitter should be calibrated zero to URL (upper range nmit) pressure. Pressu e should then be applied to the high pressure port and full scale pressure should be achieved as quickly as practical. A suggested test set-up is shown in Figure B1.

The output response should meet the criteria described in section 2 of this guideline.

Page B2

f f

3 If the output does not meet the criteria, this would generally indicate high side oil loss. For gage and absolute transmittsrs, this is all that needs to be done. This test will determine whether or not the transmitter has failed due to low oil in the high pressure side. The low side does not have to be checked, because it not pressurized. For differential transmitters, the next step is to check for a low side leak.

.The transmitter should be calibrated 0 to URL. Pressure should then be applied

- to the low pressure port and full scale pressure should be achievcd as quickly as practical. The response should again meet the criteria in section 2 above.

Failure to meet this criteria would generally indicate low side oli loss. The response time should be substantially equal during both of these tests or else an unequal amount of oil on each side would be indicated.

/

The O to URL and 0 to -URL calibrations are the worst case for oilloss. Most transmitters that have lost significant amounts of oil will be identified immediately when calibration of the transmitter is done, as the transmitter will be particularly slow in responding to calibration pressures.

This test could also be done using a ramp pressure input. A fast ramp will have greater sensitivity to oil loss than a slow ramp, e.g. a ramp rate of zero to full

. scale in 1' minute would be less sensitive than a ramp of zero to full scale in 5 seconds. The fastest possible ramp should be used for oilloss detection.

4. Summarv L

This guideline provides information to aid in detecting a sluggish transmitter during a standard calibration. With some relatively simple techniques, low oil sensors that should have been detected by other monitoring techniques, will be detected by sluggish response at standard calibration intervals.

if a transmitter is then removed from service for suspected oil loss, a thorough bench top test to check the transmitter for oil loss can be done. In order to verify loss of oil, both the high and low side must be checked for a differential l

transmitter, whereas for a gage or absolute transmitter, only the high side need Page B3 LL -

y be checked.: The transmitter will be most sensitive to oilloss when it is calibrated J o the upper range limit. Thus recommendations on how to verify both high and -

t low side oilloss were madec Measurements made using greater than iURL -

steps can provide a high9 range dependant additional warning of oliloss.

Caution, that will be of little value on the higher range codes.

9 P

9

/

I f -

1 Page B4

Figure.B1

'~

Example of Sluggish Response Test

't

-Basic Equipment 12 Transmik to k CaliW Reference Pressure Indication 11 1

4-20 mA Calibration Value

1. . 10 2 ;

2

^

9 3

f i

e 4 -

7 6

l

- Boudon Tube Gage I

l

-Reference Transmitter q/ (@/

g g

- Pressure Transducer a

y

_1 3

?!..

l l 3 Way Valve

)

Voltage or Current Meter Ci2 Pressure Source

- Dead Weight Tester

- Manometer l

-Hand Pump,(Pneumatic / Hydraulic)

- Mensor i

j

- Ruska

. - Etc -

i i

l 1

- Note delay time for XMTR under calibration to reach same output as reference indication.

h

. ~,, -

~

1

.1 Appendix C Process Noise Analysis Rosemount has done a great deal of testing and analysi.s to determine if analysis of the process noise signalis a viable diagnostic ~.est to determine -

loss of oil. The testing was centered on studying the amplitude of the 1

transmitter. output signal in the frequency ~ domain.

Rosemount's approach was to characterize variable ~. volume sensors for frequency response with various amounts of oil loss, then characterize transient response for the same amounts of oliloss.

1 The results were encouraging for transmitters that operate near their trio y

p_ghts and that are on a orocess with sufficient orocess noise if either of these conditions are 'not m'et, amplitude versus frequency data may not'

~

detect a failure until aflgr the unit has lost ability to respond.

/

Rosemount limited the analysis to tracking the amplitude of the various spectral components. There are other analysis techniques that do more detailed signal analysis. The Rosemount program was based on creating and analyzing the' Bode plot versus oil volume lost, then applying this information to actual spectral analysis plots.

p

(

There are many applications in a nuclear power plants that do not have a great deal of process noise present In these applications the amplitude L

versus frequency gives very little information. Comparing consecutive p

Power Spectral Density plots may or may not show trends of failing l'

transmitters.

a L

Noise analysis yields information about the frequency response of the transmitter in the region that the diaphragm is operating, if trips points are i

near the normal operating point of the transmitter, noise analysis can indicate that a transmitter is falling and perhaps even provide some advance knowledge of an incipient failure.

If trip points are in a significantly different operating region, it is Rosemount's judgement that frequency response can only tell you that the transmitter is operating correctly in the immediate vicinity of the operating Page C1

9t 4,

s 7

point not near the trip point. For example, consider a transmitter that is

~'

_ calibrated 0 to 750 in H O and is normally operated at 700 in H 0. If the-

-t 2

2 transmitter is set to trip at 750 in H 0, then noise analysis may be helpful 2

. in detecting a failure, perhaps even an incipient failure. Now consider the case where the transmitter trips at 200 in' H 0. If the transmitter is losing 2

oil on the low side, ability to respond to a decreasing pressure differential will be affected.. The first effect that will occur will be inability to respond near 0 DP. As more oil is progressively lost, eventually the transmitter will lose ability to respond near 200 in H O and then 700 in H 0. Thus in the 2

2 period of time it took to lose that amount of oil, an undetected malfunc+Jon would have occurred.

v Noise analysis is valuable as a rough cut, if a normally nolsy transmitter

{

output becomes quiet, it is possible that the transmitter has lost oil.' it can be very accurate for transmitters that operate near their trip points, as frequency response will be affected prior to the failure. The issue in this case is how close does the set point have to be to the operating point for process noise analysis to be effective. The answer to this has not been researched, but is likely to be related to sensor rarige and noise

' amplitude.

The end result.of all of this is that noise data is difficult to analyze and F

interpret.' For this reason Rosemount will not publish general guidelines.

Some customers have reported the presence of negative or positive L

spikes on the output signal prior to a transmitter failing due to loss of oil.

l' While Rosemount has not been able to reproduce this phenomenon in the l

laboratory, it is believed that this is an earlier symptom of oil loss than full attenuation of the noise signal. The theory is that spikes are caused by modulated process noise and the low oil sensors ability to respond to a signal in one direction, but not in the other.

ll.

Page C2

h p

7-This' diode like response to pressure changes demodulates the process.

i jp noise causing the transmitter output to track the upper or lower process noise envelope," depending on.which side loses oil.' As process noise:

)

changes with time, the transmitter output will then track the noise envelope up or down instead of Just averaging out the noise and if observed in the-(

, proper time frame will look like increased noise spikes in the output. A reduction in all noise transmission will follow as more oli is lost.

~

);,

I 1

h

'si..

e i

9 I

-,\\

i Page C3

. E '. w k

4

.. +

-.-r,

c.;

, ' =

• APPENDIX D Additional Range 9 Data Additional Data and Corrections to Range 9 Data As oil is lost from the range code 9 transfnitter, the zero output drift prior to malfunction of the transmitter is very small compared to other ranges. However, once enough oil has been lost to cause malfunction, the operating point begins to drift downward quite rapidly. The purpose of the-additional data is to show precisely when the transmitter begins to malfunction versus the amount of operating point shift. The operating point shift enunciates the failure, therefore this data provides information that will allow an estimate of how long it takes after failure for enunciation to occur.

/

In Technical Bulletin No. 3 a range code 9 calibration shift versus oil loss curve was presented (Figure 13) along with a family of responseturves for a ramp pressure input (Figure 14). The zero and operating point shift versus volume lost have been retaken with additional data points between _65 and 70 % oil loss to get a more precise shape to the curve. The new curve is shown below in Figure D1. The transmitter was calibrated 0 to 3000 psi for this test.

Page D1

j' 0654 hesense to C) sei k!):

\\

63 Febe to respond y..

l ~-

Figu,e D3 m.' ex

.o,.......... u...

0.,723 4 C; g.

0 -1.87%

4.

_j $:j

• 2" E1 im

/

C f 000 0 0

~.

'%R.s w

I.im J, C.m :

J..> cc - {

l 4

g. doc,I

+

.$ o0 io n

u w

w

,o

.o s.

Od % dome LosI(%)

G.1,.. wn a om.i, e. ij

/

The span shift is very small and positive up to 64% oil loss. At 67% oilloss the l

operating ' point shift turns negative and accelerates as the isolator diaphragm starts to bottom out.

This graph is slightly different than the graph published in Technical Bulletin No.

. 3.;The span shift for 66% oil volume lost was shown as -1.4% and is now known j

to be.05.1 This' difference is primarily due to the interpolation used in Technical Bulletin No'3 see (Figure 13).

Since our leak size does not change with time, the % of oil lost per day is.

1 constant.- If we lost 1% of the oil per day, after 67 days the output would abruptly a

. start to decrease at the rate of about -1% per day.

This rate is inversely proportional to leak rate, so if it lost.1% of the oil per day it would last 670 days before decreasing at the rate of about.1% per day.

.Next, response to a 120 psi peak to peak square wave vs. oil volume loss was evaluated. The period of the wave is 1 second. The waveform starts at 2940 psi, E

increases to 3000 psl, then decreases to 2880 psi and is designed to simulate typical trip points. The results are shown graphically in Figure D2.

j Page D2 j

6

V' k

s d

t g

p

- p rtw r. os r.. s. 3,.r.

c

.., s..e see ou.it -

- mo.

.. O m es.n

, / a' f

.m.

l um -

rm -

- m --

[

- rito -

,,.n (re~so - ]r

\\

8 L,,...

L 2930 -

g 2:20 -

2910 2900 -

. 2090 '

\\

ese0

.267C 0

0.1 0.2 -

0.,3 0.4 0.5 0.6 0.7 0.6 09 i

<i.a

?

/l Note that response to the decreasing portion of the wave form is normal. With up to 66.5% oil volume lost the response to the increasing portion of this curve is l

normal. At 67% oil volume lost, the transminer is no longer able to reach the trip-point in 500 msec. Response continues to be reduced as more oil is removed from the sensor, o-s.

?

- Calibration data at 2950 psi was also taken. The drift at 2950 psi was within 10%

4 of the drift shown in Figure D1 versus oil volume lost. It is clear from this data

-j that asithe transmitter begins to malfunction due to sluggish response, the.

' transmitter simultaneously ' begins an operating point shift.

Assuming the transmitter is considered failed at 67.0% oil volume loss, it would need to lose another 1% of the oilin order to have a shift of 0.72% (68% oil loss) ~

y t

i.

iL 1-I l

Page D3

' a1

• L

]-

• r At the NRC meeting on August 23,1989 in Washington D.C. a preliminary Bode r

t plot for the range code 9 transmitter for various percentages of oil loss was

'shown and later distributed ' y the NRC with the meeting minutes.- In subsequent b

testing a problem was found with the transmitter which affected the results from this test. The problem was corrected, and the correct Bode plot is presented in Figure D3.

+

l The transmitter was calibrated 0 to 3000 psi The operating pressure was 2950 psi and the pressure signallnput was sinusoidal at 50 psi peak to peak.

'G

}

Figure 03 f

Frequency, Response todo Plot Aenge 9. vs.etde 01 volume.

i Eh

= m

- E !"

$ o.i N

N m

1 o.i i

io ioo Feveue.wy (ha)

The response is slightly attenuated at 67% oilloss where the sensor begins to malfunction. At 68% oil volume loss the sensor is attenuating over 50% of the signal above 10 bz and at 70 % oil volume loss, the unit attenuates over 90% of.

a the signal above.5 hz, thus it is essentially not responding to the noise signal.

l

. Since response time begins to decrease at 67% oil loss, no advance warning is provided as could be concluded from a preliminary range code 9 Bode plot f

1 shown during the August NRC rneeting.

ll l

l Page D4

.j

+

.. l Conclusions As a range code 9 transmitter loses oil, there are virtually no symptoms with the exception of a verv small zero output shift of approximately.16% URL and some slight attenuation of any noise input signal. The sensor abruptly begins to i

malfunction between 66.5% and 67% oil volume lost. At 67 % oli volume lost, the zero shift is.15% (URL), the operating pol'nt shift is.55% and response to a j

small sinusold is slightly attenuated. With a precise monitoring program against I

i, a redundant transmitter, the transmitter could be detected at 67.5% oilloss. After another 1.5% oilloss (69%), the operating point decreases by another 1.32% and the frequency response is greatly diminished.

Trending transmitter output is effective for detecting failure of a range code 9-transmitter. Amplitude versus frequency analysis of the noise signal is also effective for detecting a failed unit. Neither technique provides advance warning j

to the failure, although the data indicates that monitoring output against a redundant transmitter would provide the earliest detection of fallure.

Page D5

t t-j

.i.

I

^

s. ROSEMOUNT ="*"'

1;=h""%..

t, Analyteco -

Erlen Preme VN 3344 tt 1, A j

'= -

voivos -

(6 SF; 9414%C j

g.

Twi 4310017 or4310024

- 1 Fas te th 6#4.4F-

]

I

.)

i i

NUCLEAR OPERATIONS GROUP LEAK RATE MODEL AND RISK ASSESSMENT FOR MODEL' r

1153 AND 1154 TRANSMITTERS.

l i

l Rosemount Report D8900115 Revision A Originated By

_ /A _

Date IA-A1 79 Stan Rud - Sr. Principal Engineer

. Approved By asul A

Date az l22lM er Anderson - Quality Assurance Supervisor Approved By bM OA b-Date 27 b M Mark Van Sloun - Business Unit Manager i

l' n

l.f ? -

REV1013N OTATUD CHEET

+.,

s.

> DOCUMENT NUMBER D8900115--

1 DOCUMENT-TITLE i.av Data Mnrtal anti Risk Attestment for Model 1153 and 1154 Transmitters.

' REVISION CHANGE / DESCRIPTION PAGE/ PARAGRAPH APPROVED BY DATE

[

t Aj Original-release A11/A1.1 -

g p.g.g m.-.

t

/

p' L

I u

y' 1,

i i.

1 L

L i

I

. Form No. 60122 Rev. A I.

L-

---

1

r Leak Rate Model and Risk Assessment for Model 1153 and 1154 Transmitters RMT Report D8900115 Rev. A e

1.-

TRANSMITTER RISK ASSESSM,1HI 1.1 Risk Assessment Comments The risk of an undetected oil loss failure in a.model 1153 or 1154 during any given time period is dependent on several factors. It is not possible to make a general blanket statement without considering a number of application specific factors.

The factors that must be considered in determining loss of oil risk include:

- Observed failure rates

- Time in service at operating pressure g

- Applied process pressure i

j

- Ambient temperature L

- Volume of oillost from the module

-Type of transmitter (DP or GP)

- Desired response under event conditions.

These factors have been evaluated and a number of risk assessment methods have been used to provide utilities with information to identify suspect transmitters at their specific location ' These methods will be discussed here and y

include:

+

- Unadjusted failure rate statistics

- Mathematical modeling of leak phenomenon

- Failure rate vs. operating pressure and time in i'

service l

Page 1 of 10

l~

l 1.2 Unadjusted Failure Rates There are many different ways to interpret complex failure statistics. We will explain several ways of looking at the raw failure data to give an estimate of risk based on time in service and manufacturing date, First we can use some gross measures to establish failure rates. Rosemount has shipped 14,145 Model 1153 and 1154 transmitters. As of December 15,1989 there are 91 confirmed failures with 16 additionalin evaluation.

Excluding 1,158 model 1153A transmitters which did not use metal o rings, this results in a failure fraction of 0.82% of shipments.

Another gross measure of failure rate is to estimate the hours of service and to divide this by the number of failures. Rosemount has estimated service hours by using data collected from our Mall Survey of the failed units and during our field test visits. The assumptions, used in the service hour calculation included:

e Average lag time from shipment to pressurization is 12 months e 50% of our shipments are installed e 60% of installed units are actually under static pressure e 70% time at pressure l

l With these assumptions we estimated that Rosemount Model 1153 and 1154 L

transmitters have accumulated 110 million hours of service. Dividing 107 failures.

and potential failures by 110 million hours equates to 1 x 10 6 failures /hr or 116

. years MTBF. This is an average failure rate over the last 9 years, h

l Page 2 of 10 l

l

~

[,

y j

5 Time in service for the installed base of transmitters was estimated using the l

stated assumptions and failure rates were calculated. Overall failure rates for j

each year in service are shown in Figure 1.

Hgure5 Overall falluto Rato i

FWn/w et Preswe 3-s 7

12-th' 2.4 -

2.7 -

g4 i,:

1.5 -

l 3

la on -

04 -

{,

0-

.m.

+

0-12 13 24 ?$-36. 37-46 49 60 51-72 73 64 $$-96-97-400 109-120 m,..I r nn, w, fm Rosemount estimates that over 30% of the transmitter service hours are from transmitters that have been in service for over 48 months. However, only 1 of the failures is from transmitters with over 48 months in service. This data clearly fits the definition of early life failure.

Early life failures are actually strongly related to operating static pressure. The data presented in Figure 1 was not adjusted for static pressure. A more precise analysis was done to account for the actual pressure applied to each transmitter.

1 The remainder of this report documents this analysis and presents estimated

- failure rates as a function of pressure and time in service (or psi months).

Page 3 of 10

g

-i i,

W:

m'

• 1.3 Mathematical Modeling of Leaking Transmitters Rosemount performed extensive analysis on failed sensors and was able to build a mathematical model to simulate a leaking module. The key variables we were able to determine were time in service, pressure, and to a lesser extent, ambient -

i temperature.

~.

l The equation used is based on normal viscous flow. That is, the leak rate (O) is defined as the volume lost (V) in time period (t) and is directly proportional to the dimensional leak coefficient (k) and static pressure (P) and inversely proportional-to oil viscosity (u) which is a function of temperature. We have called this a leak j

rate model.

I Therefore, we have Q = V / t = (k *P) / u

/

Each sensor range has a slightly different oil volume. The critical volume each sensor range can lose before degrading beyond its response time specification f

was determined and used in this calculation, 1

Information from the failure surveys, such as time at pressure, static pressure

]

and average temperature was then used in the mathematical model to calculate.

the dimensional leak coefficient (k) for each failed sensor. ' Ali this information j

was-used to compare'the failed units under similar conditions. Range code 5 l

- was chosen in this analysis because it is representative in critical oil loss volume, of range codes 3 through 5 all types and range codes 6 through 10 gage and absolute units. Range code 6 was also chosen because it is representative of l

range codes 6 through 8 differential only.

Transmitters with static pressure exceeding 2000 psi will fail the fastest, and therefore will be a leading indicator of the general populations failure fraction. For this reason, this analysis was developed from only those failed transmitters that were used at 2000 psi or above. This limited the study to transmitters used at L

Pressurized Water Reactors (PWRs). There have been a total of 21 failures in l'

applications at or above 2000 psl.

Page 4 of 10 i

._ ~ _

.. -s-i ;; : {

,4 p

5, l.'

- The frequency distribution of leak coefficients for transmitters used at pressure-1-

exceeding 2000 poiis shown in Figure 2.

l ii l- -

l'igure 2 Imek Caetticient frequency Distribution'

~

so-1 s-9 i

kj p.

s.

t E

j n',

3 5-0 I

i

[

4

.3-p q

f 2-h, N

4 c

Jp g.

i O

100 200 300 800 900 teoh Coef4ernt in order to evaluate how these leaking _ sensors would have performed under sim!!ar. operating conditions, their lifetimes were calculated using the computed leak coefficient for each sensor, and standard operating conditions.

Figure 3 shows the failure rate we would expect for range code 5 versus time in

. service for 3 typical operating pressures. Figure 4 is the same plot for range code 6.

Both of these plots were developed using these additional assumptions to determine hours at pressure:

w o Only transmitters installed in PWRs were considered o 20% of these installed units are operated above 2000 psi Failure rates were developed for every 6 months at pressure based on the 21 returns and the smoothened data is plotted in figure 3.

Page 5 of 10 4

1 4

n,

+w 2

,,,,m sr m.

.-~ *.

-n.. _,. -

-., - e.,-

n p.

[

'se 3

g; i.

s:

Figure 3-

.1 Failure itetes for Range S et Stended Cone!tions j

g V-s.

3

-1 3.5 -

.p 2$00 ps-3-

)

i e

f 2.5 -

k il E t-

-i ;

6 e

1.5 '

,e 1000 pil l

1' 60.000 90-mo 60.000 tW-me

.l

/

../

/

03 E

vi 12 24 36 48 to 72 84 96 803 170 Time et eresawe

/.

^i o

Figure 4 Failure mates for Itange 6 DP at stenderd conditione.

it 33-4 a

3-th-y t

f1b00 W E

13-1000 psi

== s ~, /

$ a' imo= p,%

i-C3 E

l N.

0 i '

O 12 24 36 48 80 72 44 96 106 120 w. me.

l p.

l l

[f L

Page 6 of 10 i

.15

. ' g Failure rates are proportional to the failure fraction and increase rapidly to a peak rate that is directly proportional to static pressure. The time in service to reach that peak varies inversely with static pressure. Then the failure rate decreases exponentially with time in service until it falls below an acceptable level after a constant number of psi months.

At 2500 psi the failures are compressed into a shorter time. At a line pressure of -

250 psi, the failures are spread out over a 10 times longer period of time and thus have a 10 times lower peak failure rate. The number of failures does not change, just the rate at which they happen. Our model is based on a 3.0%

failure fraction.

At 2500 psi the failure rate for the range code 5 is reduced to 5.0x10 7 failures-per hour after the units have been in service for 24 months. At 1000 psi, the peak failure rate is 2.5 times less and reduced to 5.0x10'7 failures per hour after 60 months. Note that for any pressure a constant pressure time, product of 60,000

- psi months results in a failure rate of less than 5.0x10'7 failures per hour. At 250 psi, the peak failure rate never exceeds 5.0x10'7 failures per hour. These numbers apply to range codes 3 through 5 all types and range codes 6 through -

10 gage and absolute due to similarity.

I At'2500 psi the failure rate for the range code 6 (which is 2.2 times less than a range code 5)'is reduced to 5.0x10'7 failures per hour after the units have been in service for 53 months. At 1000 psi, the failure rate is reduced to 5.0x10'7 failures per hour after 130 months, although it never exceeds 1.0x10 6 failures per hour throughout it's life.. Note that a constant pressure time product of-approximately 130,000 psi-months results in a failure rate of 5.0x10'7 failures per hour. At 250 psi, the peak failure rate never exceeds 5.0x10'7 failures per hour.

These numbers apply to range codes 6 through 8 differential only due to similarity.

Page 7 of 10 1

.I h is interesting to note that if there were 750 transmitters simultaneously installed at 2500 psi, we would expect the following number of fallures:

P(redicted) (Actual Failures Model

>2000 psi GLm,!g)

(Rate x 10 6/hr)

Reoorted) 0 - 6 mo

.2

.65 4

7 12 mo -

2.8 9.2-8 13 - 18 mo 3.8 12.5 '

3 19 24 mo 1.7 5.6 4

25 30 mo

.3

.99 1

31 - 36 mo

<.2

&5 1

29.6 21 The model closely predicts the observed distribution shape.

However, the magnitude was based on an estimate of the installed at pressure transmitter population and is a function of the resulting failure fraction, it appears to be a very conservative bounding function. We have had no failures over 36 months at greater than 2000 psi as predicted by this leak rate model.

~

1.4 Review of Lot Specific Failure Rates Part of the. Rosemount failure analysis was a detailed investigation of manufacturing lots, it was immediately obvious that failures were clustered. Our February 7,198910CFR part 21 notice included approximately 1000 transmitters which had 62 failures or a failure fraction of 6.2%.

The remaining manufacturing lots had 24 confirmed or suspected failures for a failure fraction of.20%.

This is a 31 times lower failure fraction and for this reason all lots were not included in the first 10CFR part 21 notification. These are still valid comparisons but since suspect groups require time at pressure to initiate failures, we can anticipate some additional lots to enunciate themselves after more operational time. Therefore, a second notice was issued to very conservatively cover all groups, since they are still at some risk.

l i

l Page 8 of 10

o j

t A consuking firm was retained to perform a probablistic risk assessment (PRA) for both BWR and PWR plants. The work was done by Pickard, Lowe and Garrick inc. (PLG). The PLG analysis was done using failure data supplied by Rosemount and considered the effect of suspect lots. The assumptions in this analysis are slightly different, but the failure rates that were developed were subsequently applied to a simplified model of a BWR and a PWR to provide bounding estimates of the risk significance of the failures. The PLG analysis does not find any dgnificant risk for BWRs. For PWRs, PLG did make note of an identified plant specific scenario involvinp pressurizer pressure and level measurements which should be evaluated. However, the PLG scenario did not consider operator intervention and assumed no redundancy of measurement, it also did not consider the fact that gage pressure transmitters will respond properly to a decrease in pressure even if loss of fill fluid occurs.

The PLG arialysis also noted that under high pressure transmitters susceptible to a glass to metal seal induced olt loss failure reveal themselves relatively early in life.

1 Page 9 of 10

i'C g

i i

1.5 Other Risk Factors Two other risk factors have been studied. The following is a discussion of these factors.

)

i i

Transmitters normally not pressurized (in stend by service)

Transmitters that are operating in stand by service generally see static pressure for only several hours each quarter during functional testing.

Their risk is generally very low, even if they are leaking, because their time at pressure is very j

low. That is, if either the pressure x time in service product is less than 5,000 psi-I months for range codes 3 through 5 or 12,000 psi months for range codes 6

)

through 10, then they haven't had time to lose a critical volume of oil. However, if they have accumulated more than this amount of a pressure x time in service product, they must become at risk units and be treated as such.

Low Range Absolute and Gage Transmitters Rosemount does not have any confirmed glass to metal seal oil loss failures of j

range codes 3 through 7 absolute or gage transmitters. We believe this is due to two factors. First, only the high side of the transmitter is pressurized, thus the odds of a failure are reduced by 50%. Second, the pressure is always less than 300 psi on these units. Therefore, the failure rates are acceptably low.

Page 10 of 10

=-

33 u

.n.w moeemount inc.

aL

,8 Sg C4 Pit'01 173D1 TecttM040py DWe i

ce.ww Eo n Fair 2, ww nua u s A vei.

citi 9ano j

Tws 43tD0tt os a)100N i

i ter air, ers noe I

l December 22, 1989 i

ROSEMOUNT MODEL 1153 AND 1154 PRESSURE TRANSMITTER SUSPECT LIST 1

i The attached list represents Model 1153 and 1154 Pressure Transmitters deemed suspect under the February 7, 1989 10CTR21 Notification.

The listing has been updated to include additional transmitter groups identified through

/

December 22, 1989.

Transmitter serial numbers labeled with an asterisk (*)

represent new identified transmitters since th's February Notification.

If there are any questions concerning this list, please contact the Rosemount Nuclear Group at (612) 828-3540.

I 7

1 I

i 1

.n,

p S/N MODULE MODEL S/N MODULE MODEL J-362483 984469 1153GB4 405996 751804 1153DB5 397819 954192 1153HD5 406003 751786 1153DB5 l

-402578 1329633 1153DD4 406005 751771 1153DB5 405542 754305 1153GB5 406006 751793 1153DB5 i

i '

405545 754249 1153DB5 406007 751785 1153DB5 405566 754245 1153DB5 406009 751878 1153DB5 405569 754264 1153DB5 406010 754294 1153DB5 405570 754266 1153DB5 406011 751795 1153DB5 405571 754263 1153DB5 406013 754329 1153DB5 405573 754278 1153DB5 s

406014 754302 1153DB5 405574 754252 1153Da5 406015 754288 1153DB5 i

405587 754247 1153DB5 406016 751877 1153DB5 405588 754283 1153DB5 406017 754327 1153DB5 405589 754267 1153DB5 406019 751775 1153DB5 405596 754253 1153DB5 406020 754321 1153DB5

-405597 754277 1153DBS 406021 754309 1153DB5 405599 754256 1153DB5 406022 754318 1153DB5 405614 754330 1153DD5 406024 751784 1153DB5 405616 754251 1153DD5 406025 751790 1153DB5 405618 754260 1153DD5 406026 751794 1153DB5 405621 754248 1153DD5 406027 751805 1153DB5 405622 754332 1153DD5 406029 754326 1153DB5 405638 754242 1153HD5 406030 754312 1153DB5 405643 754290 1153DB5 406031 751792 1153DB5 405646 754304 1153DB5 406032 734308 1153DB5 405649 754331 1153DB5 406033 754340 1153DBS

'405652 754265 1153DB5 406034 751789 1153DB5 405653 754301 1153DBS 406035 751776 1153DB5 405654 754297 1153GB5 406036 754317 1153DB5 405684 754295 1153DB5 406044 751777 1153DDS i

405685 754291 1153DB5 406046 754339 1153DB5 405709 754344 1153HB5 406053 754325 1153DB5 405716 754334 1153HB5 406054 751781 1153DB5 405721 754262 1153DB5 406056 754319 1153DB5 l

405797 754269 1153HD5 406057 751773 1153DB5 405798 754257 1153HD5 406058 751788 1153DBS 405799 754270 1153HD5 406059 751796 1153DB5 405802 754274 1153HD5 406061 751778 1153DB5 405803 754342 1153HD5 406062 751787 1153DB5 I

405824 754243 1153DB5 406063 751799 1153DB5 405827 754280 1153DD5 406064 754338 1153DB5 405828 754271 1153DD5 406066 751780 1153DB5 405829 754276 1153DD5 406067 751797 1153DB5 405831 754324 1153DDS 406069 751779 1153DB5 405832 754255 1153DD5 406070 751798 1153GB5 405833 754281 1153DDS 406071 751803 1153GB5 40587J 754293 1153GB5 406163 751870 1153DB4 405928

-754337 1153DB5 406167 751774 1153DB5 405929 754341 1153DB5 406218 751838 1153DB5 405930 754306 1153DB5 406219 751834 1153DB5 405931 754307 1153DD5 406220 751826 1153DB5 405932 754343 1153DD5 406221 751820 1153DB5 405933 754285 1153DD5 406236 751807 1153DB5 405989-754273 1153 ABS 406238 751815 1153DB5 405993 754275 1153AB5 406239 751837 1153DBS

.-._.,_....-,~,..~..,.e,-

,--.v.

6'-

.e e

406240 751830 1153DB5 407719 751875 1153DB5 406242 751822 1153DB5 407721 '

751873 1153DB5 406244 751818 1153DB5 407724 751860 1153DB5 406245 751809 1153DB5 407785 751850 1153DB5

)

406246 751833 1153DB5 407795 751853 1153DB5 406247 751808 1153DB5 407804 751861 1153DBS 2

406255 751825 1153DD5 407979 954141 1153DB5 i

406256-751835 1153DD5 407980 954121 1153DB5 i

406283 751840 1153DB5 407981 954126 1153DB5 406284 751836 1153DB5 407982 954124 1153D85

~406287 751810 1153DB5 407983 954113 1153DBS

+

406289 751812 1153DB5 407984 954137 1153DB5 406291 751823 1153HD5 407985 954136 1153DB5 406292 754322 1153HD5 407986 954118 1153DB5 406293 751819 1153HD5 407988 954116 1153DB5 406294 754323 1153HD5 407993 954078 1153DB5

.406295 751813 1153HD5 407994 954133 1153DB5 406296 751827 1153HD5 407995 954153 1153DB5 l

f406297 751832 1153HD5 407996 954179 1153DB5 406302 751824 1153DDS 407997 954203 1153DB5 406303 754300

~1153DD5 407998 954130 1153DB5

/

i 406304 754286 1153DD5 407999 954125 1153DB5 406305 754299 1153DDS 408000 954114 1153DB5 406318 751811 1153DB5 408001 954159 1153DB5 4

406319 751821 1153DB5 408002 954119 1153DB5 406387 751806 1153DBS 408003 954131-1153DB5

>406655 1019739 1153DB5 408004 954115 1153DB5 406658 754315 1153AB5 408038 954145 1153DB5 406660 754346 1153AB5 408039 954129 1153HB5

-406661 754311 1153 ABS 408040 954084 1153DD5 406662 754328 1153AB5 408050 954170 1353DB5 406663 754314 1153 ABS 408051 954147 1153DB5 406665 754310 1153AB5 408052 954146 1153DB5 406667 754347 1153AB5 408053 954139 1153DB5 406669 754316 1153 ABS 408054 954150 1153DB5 407478 754244 1153GB5 408055 354109 1153DB5 407638 751856 1153DB5 408056 954172 1153DB5 407641 751866 1153DB5 408057 954142 1153DB5 407647 751848 1153DB5 408059 954160 1153DB5

~407652 751846 1153DB5 408060 954111 1153DB5 407674 751864 1153DB5 408061 954151 1153DB5 407684 751851 1153DB5 408062 954122 1153DB5 407686 751865 1153DB5 408073 954187 1153DB5 407688 751845 1153DB5 408074 954082 1153DB5 407696 751868 1153DB5 408076 954104 1153DB5 407698 754241 1153DBS 408078 954100 1153DB5 407704 751867 1153DB5 408079 954181 1153DB5

'407707 751852 1153DB5 408088 954188 1153DB5 1407710 754303 1153DB5 408103 954205 1153DB5 407711 751862 1153DB5 408153 954198 1153DB5 407713 751874 1153DB5 408154 954067 1153DBS 407715 751855 liS3DBS 408157 954086 1153DBS 407717 751841 1153DB5 406158 954101 1153DB5 407718 751857 1153DB5 408159 954072 1153DBS

,.. ~,

__.____.________..____,_________.m_

b j

i

'408162 554201

1153DB5, 408741 984493 1153DB5 408171 954184 1153GB5 408742 984543 1153DB5

)

408175.

954196 1153DB5 408743 984556 1153DB5 1

408176 954077 1153DB5 408744 984479 1153DB5

)

408178

'954071 1153DB5 408747 984500 1153DB5

'408181 954094 1153DBS 408748 984504 1153DB5 i

408182 954096 1153DB5 408749 984484 1153DB5 1

408184 954091 1153HD5 408750 984521 1153DB5

)

.408187 954098 1153HD5 408751 984562 1153DB5 1

408188 954095 1153HD5 408752 984485 1153DB5 408189 954089 1153HD5 408753 984472 1153DB5

~408190 954070 1153HD5 408754 984475 1153DB5 i

408191 954144 1153HD5 408755 984467

'1153D85 i

~ 4 08193,

954128 1153HD5 408756 984519 1153DB5 408194 954140 1153HD5 408757 984502 1153DB5 1

408195 954183 1153HD5 408758 984473 1153DB5

)

408197' =954097 1153HD5 408759 984490 1153DB5 I

408198 954186 1153HD5 408760 984492 1153DB5

'408249 954092 1153DB5 40P761 984487 1153DB5 I

.'408250 954090 1153DB5 408762 984480 1153DB5 i

~ 4 082 51 -

954171 1153DB5 408763 984496 1153DB5 408252 954190 1153DB5 408764 984555 1153HB5

/

408253 954093 1153DBS' 408766 984561 1153HB5 L

408254

.954182 1153DB5 408767 984495 1153HB5 L

408255 954185 1153DB5 408768 984549 1153DB5 l

408284 954105 1153DB5 408769 984508 1153DB5

~

408293 954079 1153DB5 408770 984471 1153DB5 L

408294 954194 1153DBS 408771 984557 1153DB5

.408295 954189 1153DB5 408772 984466 1153DB5 i

-408297~

954103 1153DB5 408773 984501 1153DB5 408298

954200 1153DB5 408774 984498 1153DB5 408299 954202 1153DB5 408775 984534 1153DB5 408300 954088 1153DB5 408776 984482 1153DBS 408302 954080 1153DB5 408777 984481 1153DB5 408303 954106 1153DB5 408778 984483 1153DB5 408306 954107 1153DB5 408781 984503 1153DB5 408307 954180 1153DB5 408796 984468 1153DD5 408308 954087 1153DB5 408797 984509 1153DB5 408309 954204 1153DB5 408798 984461 1153DB5 408310 984514 1153DB5 408799 984491 1153HB5 408311.

954199 1153DB5 408806 984499 1153DB5

'408312 954117 1153DB5 408877 984497 1153DD5 408313 954191 1153DB5 408891 984510 1153DB5 408314 954073 1153DB5 40S894 984522 1153DB5 l

408346 954195 1153GB5 408897 984470 1153DB5 L

408476-954158 1153DB5 408898 984527 1153DB5 L

408645 754279 1153HB5 408932 984515 1153HD5 408646 754345 1153HB5 408960 984528 1153DB5 403654 751817 1153HB5 409102 984518 1153HD5 408736 984535 1153DB5 409107 984545 1153HD5 408737 984486 1153DB5 409154 984533 1153DB5 408738 984529 1153DB5 409155 984547 1153HD5 408739 984474 1153DBS 409157 984526 1153DB5 408740 984556 1153DB5 409189 984507 1153 ABS Y

-.---..-..-,,,n.--~.

~ - -,---..

~c-.

- - -, -.- - - - -, ~

n l

t

\\

?409370 984546 1153DD5 410286 1027094

• 1353GD9 409386 984540' 1153HB5 410301 1060793 1153HD5 t

409497 984512-1153DB5 410302 1060796 1153HD5 409508 984538 1153DB5 410303 1060781 1153HD5 409511 984531 1153AB5 410304 1060770 1153HD5 409524 1019757 1153DD5 410305 1052991 1153HD5 409526 1019716 1153DD5 410306 1053013 1153HD5 4

409647 1019747 1153DB5 410307 1052992 1153HD5 469649 1019742 1153DB5 410308 1060799 1153HD5 409650 1019759 1153DB5 410309 1060792 1153HD5

)

409735 1053048 1153DD5 410310 1060771

1153HD5 409736 1053056 1153DD5 410311 1060775 1153HD5 409739 1052990 1153DB5 410312 1053020 1153HD5 409740-1053041 1153DB5 410320 1027102 1153GD9 1

409741 1053031' 1153HB5 410329 1060745 1153HD5 409743 1053039

-1153HB5 410330 1060767 1153HD5 i

409755 1053030 1153DD5 410331 1053028 1153HD5 409756 1053042 1153DD5 410335 1060746 1153HD5 f

409757 1053059 1153DD5 410337 1060809 1153HD5

. 409758 1053045 1153DD5 410338 1060813 1153HD5 409759 1053035 1153DD5 410339 1060766 1153HD5 409760 1053051 1153DD5 410340 1060756 1153HD5

/

J409761 1052983 1153DDS 410488 1060776 1153DB5 i

409780-1053033 1153GB5 410493 1039053 1153DB5 409876 984476 1153GB5 410497 1039015 1153DB5 l

409880 984477 1153GB5 410499

~1039047 1153DB5 i

410031 1053054 1153DB5 410503 1039007 1153DB5

'410032 1052993 1153DD5 410510 1039042 1153HB5 5

410033 1053052 1153DD5 410511 1039043 1153HBS 410034 1052999 1153DD5 410532 1075045 1153GB9 4

~ 410035 1053046 1153DD5 410536 1039048 1153DB5 1

410050-1053036' 1153DBS 410539 1075025 1153GB9 410053 1053044 1153DB5 410541 1075041 1153GB9 410054 1053047 1153DB5 410542 1075052 1153GB9 410086 1027133 1153GB9 410543 1075053 1153GB9 410089 1019749 1153DB5 410546 1075027 1153GB9 410112 1053006 1153DD5 410547 1027095 1153GB9 j

410113 1053040 1153DD5 410548 1075050 1153GP9 410114 1053008 1153DD5 410549 1075054 1153GB9

)

410115 1053061 1153DD5 410550 1075057-1153GB9 410120 1060801 1153HB5 410551 1075048 1153GB9 410157 1053001 1153DD5

• 410552

• 1060633 1153DB4 410170 1060794 1153HD5 410572 1060747 1153DB5

>410171 1053007 1153HD5 410573 1060824 1153DB5 j

410172 1053021 1153HD5 410574 1039063 1153DB5 0410203

• 1060646 1153DD4 410578 1060808 1153DB5 410271 1075037 1153GD9 410581 1039049 1153DB5 410272 1027101 1153GD9 410583 1039046 1153DB5 410273 102?l17 1153GD9 410584 1060788 1153AB5 410274-1027107 1153GD9 410585 1060755 1153GB5 410275 1027123 1153GD9 410586 1060778 1153GB5 410276 1027111 1153GD9 410588 1039066 1153GB5 410277 1027098 1153GD9 410590 1039062 1153GB5 410282 1027132 1153GD9 410594 1039060 1153DB5

i I4 410600

'103'9052 1153DB5 410848 1109356 1153GD9 i

l410641 1075043 1153GB9 410852 1075035 1153GD9 410643 1075024 1153GB9 410880 1109310 1153GD9 410644 1075036 1153GB9 410902 1039065 1153DB5 410658 1075046 1153GB9 410960 1109344 1153GB9 I

410659 1075022 1153GB9 410961 1109329 1154GP9 410674 1075039 1153GB9 410962 1109314 1154GP9

'410675 1075034 1153GB9 410963 1109315 1154GP9 410676 1075031 1153GB9 410964 1109319 1154GP9 410677 1075049 1153GB9 410965 1109317 1154GP9 410678 1075032 1153GB9 410966 1109343 1154GP9 4106.79 1075056 1153GB9 410967 1109328 1154GP9

.f 410680 1075028 1153GB9 410972 1060744 1154DP5 410682 1060768 1153DB5

• 410973

• 1101949 1154DP5 410686 1060742 1153DB5

• 410975

• 1101915 1154DP5 410f93 1060819 1153GB5

• 410976

• 1101916 1154DP5 i

i-410694 1060764 1153GB5

• 410978

• 1101952 11b4DP5 413695 1060789 1153GB5

• 410979

• 1101944 1154DP5 410696 1060823 1153HD5 410980 1109321 1154GP9 l

410697 1060760 1153HD5 410981 1109324 1154GP9 l

.410698 1060814 1153HD5

• 410990

• 1101928 1154HP5-L

.410699 1060786 1153DD5

• 410992

• 1101909 1154HP5

/

410701-1060757 1153DD5

• 411047

• 1101874 1153HBS

-410702 1060751 1153DD5

• 411050

• 1101903 1153DB5 410703 1039056 1153DD5

• 411051

• 1101947 1153DB5 410704 1039061 1153DD5

• 411055

• 1101837 1153HB5 410705 1039059 1153DD5-

• 411056

• 1101879 1153HB5 410707 1060825 1153DD5

• 411059

• 1101850 1153DB5 1410709- - 1060821 1153DD5

• 411061

• 1101951 1154DP5

'410710 1060820 1153DD5

• 411062

• 1101927 1154DP5 410712 1060748 1153DB5

• 411064

• 1101935 1154HP5 410713 1060817 1153DB5

• 411089

• 1101858 1153DB5 0410734

• 1060647 1153DD4

• 411090

• 1101956 1153DB5 410737 1060750 1153DD5

• 411091

• 1101835 1153DB5 410740 1060815 1153HB5

• 411092

• 1101866 1153DB5 410744 1060749-1153DBS

• 411093

• 1101871 1153DB5

-410746 1060826 1153DB5

• 411094

• 1101880 1153DB5 410747 1060762 1153DB5

• 411095

• 1101873 1153DB5 410753 1060782 1153DDS

• 411096

• 1101844 1153DB5 L

0410759

• 1060632 1153HD4

• 411097

• 1101832 1153DB5 0410761

• 1060640 1153DB4

• 411098

• 1101902 1153DB5 l

410792 1019713 1153DB5

• 411099

• 1101919 1153DB5 2

410795, 984478 1153HB5

• 411100

• 1101890 1153DB5 410830 1052984 1153DB5

• 411101

• 1101882 1153DB5 410831 1053049 1153DB5

• 411102

• 1101881 1153DB5 410834 1053015 1153DB5

• 411117

• 1101950 1154 GPS 410835 1052982 1153DB5 411118 1109345 1154GP9 410836' 1053058 1153DB5 411119 1109318 1154GP9 410837 1053000 1153DBS 411120 1109289 1154GP9

~410838 1053055 1153DB5 411121 1109313 1154GP9 410839 1053038 1153DB5 411122 1109301 1154GP9 0410843

• 1101940 1153DB5

• 411128

• 1101870 1153DB5 410844 1039058 1153DB5

• 411129

• 1101899 1153DB5 411130 1116214 1153DB5

• 411135

• 1101831 1153DD5

• 411136

• 1101875 1153DD5

... -. e...................

L

V I

411143 3309335-1154GP9 411270 1116210 1153DB5 i '

411144 1109331 1154GP9 411271 1116243 1153DB5

c411148,-*1101906 1153DB5

• 4il272

• 1101954 1153DB5 c411149 '*1101860 1153DB5 411276 984560 1153GB5 c411373' *1101888 1153AD5

• 411279

• 1101918 1154GP5 l

i c411174

• 1101836 1153AD5

• 411280

• 1101934 1154 GPS c411175. *1101945 1153AD5 411287' 1116259 1153DD5 I

0411176

• 1101845 1153AD5 411288 1116251-1153DD5 411195 1019752 1154HP5 411289 1116228 1153DD5 c411212- *1101931 1154DP5 411290 1116264 1153DD5 i

'411213

'1019745 1154DP5 411291 1116253 1153DD5 c411214

• 1101929 1154DP5 411292 1116256 1153DD5 0411215

• 1101908' 1154GP5 411293 1116300 1153DD5 1411216 1109297 1154GP9 411294 1116242 1153DD5 411217 1109302-1153GD9 411295 1116212 1153DD5 411218 1109291 1153GD9 411296 1116220 1153DD5 411219 1109312 1153GD9 411297 1116306 1153DD5 411220 1109360 1153GB9 411298 1126299 1153DD5 411221 1109308 1154GP9 411299 1116279 1153DD5 0411222

• 1060652 1153DB4 411300 1116217 1153DD5

• 0411224

• 1101862 1154DP5 411303' 1019737 1154HP5 0411225

• 1101943 1154DP5

• 411309

• 1060624 1153GB4 0411226

• 1060625 1153DD4 411447 1109295 1153GD9-411228 1039029

-1154DP5 411448 1109311 1153GD9 c411229

• 1101891 1154DP5 411449 1109288 1153GD9 0411230

• 1060620 1153DB4

• 411511

• 1060621 1153DB4 0411231

• 1060622 1153DB4

• 411518

• 1060629 1153DB4 4

411233 1109320-1153GD9 411537 1116230 1153DB5 411234 1109298 1153GD9 411538 1116227 1153DB5 6411236

• 1060638 1153DB4 411544 1116224 1153DB5 0411237

• 1101889 1153HB5 411545 1116237 1153DB5 i

L 411239 1109353 1153GD9 411553 1116254 1153HB5 411240 1116263 1153DB5

• 411569

• 1060651 1153DB4 411241 1116257

-1153DB5 411632 1109316 1153GB9 L

411242 1116268-1153DB5 411633 1116185 1153HB5 E

e411243

• 1101863 1153DB5 411654 1109354 1153GB9 411244 1116269 1153DB5 411664 1116233 1153DB5 411245 1116249 1153DB5 411665 1116282 1153DB5 l

411246 1116223 1153DBS 411670 1116215 1153GB5 l

411247 1116248' 1153DB5 411671 1116283 1153GB5 l

411248 1116265 1153DB5 411674 1109282 1153GB9 411249 1116246 1153DB5 411714 1109269 1153GD9

'411250 1116233 1153DB5 411718 1109350 1153GD9 411251 1116241 1153DB5 411724 1109365 1154GP9 411252 1116285 1153DB5 411726 1109307 1154GP9 0411253

• 1101855 1153AB5 411728 1109325 1154GP9

-*411254

• 1101932 1154DP5 411729 1109337 1154GP9 0411255

• 1101923 1154DP5 411730 1109366 1154GP9

'411256 1060743 1154DP5

• 411472

• 1101930 1154DP5 411257 1116255 1153DB5

• 411773

• 1101867 1153GB5 6411261- *1060628 1153GB4 411859 1109268 1153GD9 0411262

• 1060634 1153GE4 411863 1109281 1153GD9 c411263

• 1060643 1153GB4 411865 1116222 1153 ADS e411265

• 1060627 1153GB4 411866 1116231 1153AD5

'0411267

• 1060636 1153GB4 411872 1116297 1153GB5 411269 1116186 1153DB5 411892 1129229 1153GB5

• i l

[

D

' 411893 1129106 1153GB5 412007 1129111 1153AD5 411894 1129179 1153GB5 412008 1129112 1153AD5 411895 1129130 1153GB5 412009 1129196 1153AD5

411896 1129128 1153GB5 412010 1129120 1153AD5 411897 1129142 1153GB5 412011 1129190 1153AD5 411898 1129176 1153GB5 412012 1129138 1153AD5

~411899 1129164 1153GB5 412013 1129225 1153AD5 411900 1116291 1153GB5 412014 1129127 1153AD5 411901 1129194 1153GB5 412015 1129136 1153DD5

.411911 1109274 1153GB9 412016 1129187 1153DD5 411912 1109271 1153GB9 412017 1129228 1153DD5 411923 1129154 1153HD5 412018 1129244 1153DD5 411924 1129174 1153HD5 412019 1129182 1153DD5

411925 1129163 1153HD5 412020 1129248 1153DD5 411926 1129231 1153HD5 412021 1129172 1153DD5 411927 1129230 1153HD5 412022 1129151 1153DD5

'411935 1129117 1153DD5 412023 1129251 1153DD5 411942 1129220 1153HD5 412024 1116284 1153DD5 411943 1129207 1153HD5 412025 1129235 1153DD5 411944 1129242 1153HD5 412026 1129226 1153DD5

'* 411945 1129144 1153DB5 412027 1129178 1153DD5 411946 1129134 1153DD5 412028 1129234 1153DD5 411947 1129188 1153DB5 412029 1129205 1153DD5 411948 1129181 1153DB5 412030 1129240 1153DD5 411949 1129141 1153DB5 412031 1129153 1153DD5 411950 1129203 1153DB5 412032 1129168 1153DD5 411951 1129125 1153DB5 412033 1129155 1153DD5 411952 1129183 1153DB5

• 412035

• 1060630 1153DD4 411953 1129169 1153DB5 412045 1129115 1153HD5 411954 1129232 1153DB5 412046 1129140 1153HD5 411955 1129165 1153DB5 412047 1125132 1153HD5 411956 1129250 1153HB5 412048 1129243 1153DB5 411957 1129227 1153HB5 412049 1129237 1153HD5 411958 1129135 1153DD5 412168 1129292 1154DPS 411959 1129252 1153DD5 412175 1129216 1153AD5 I

411960 1129247 1153DD5 412176 1129137 1153 ADS 411961-1109287 1153GD9 412183 1129180 1153DB5 i

411965 1129297 1153DB5 412184 1129162 1153DB5 411966 1129239 1153DB5 412197 1129166 1153GB5 411967 1129158 1153DBS 412198 1129160 1153GB5 l411968~

1129201 1153DD5 412199 1129184 1153GB5 411969 1129238 1153DD5 412200 1129121 1153GB5 l

-411970 1129290 1153DD5 412201 1129161 1153GB5

+

411971 1129124 1153DD5 412202 1129193 1153GB5 411973 1116304 1153DB5 412216 1129175 1153DB5 411975 1129197 1153DB5 412217 1129145 1153DB5 411976 1129208 1153DB5 412218 1129133 1153DB5 411977 1129126 1153HB5 412219 1129146 1153DB5

-411978-1129195 1153HB5 412220 1129200 1153DB5 411982 1116276 1153DB5 412221 1129177 1153DB5

'411983' 1129219 1153DB5 412222 1129157 1153DB5 411986 1129198 1153GB5 412223 1129233 1153DB5 e411996

• 1060645 1153DB4 412224 1129236 1153DB5 412003

- 1129119 1153DB5 412225 1129246 1153DB5 412005 1129202 1153AD5 412226 1129152 1153DB5 412006 1129291 1153AD5 412227 1129110 1153DB5 4

.m

. ~~ -

I I

l r

412228 1129222 1153DB5 412618 1163592 1153DB5 412229 1129186 1153DB5 412627 1129149 1153AB5 412230 1129131 1153DB5 412636 1163608 1153AD5 412231 1129173 1153DB5 412707 1163613 1153DB5 412232 1129249 1153DB5 412708 1163555 1153HD5

.412233 1129159 1153DB5 412709 1163510 1153DDS 412234 1129113 1153DB5 412710 1163535 1153DD5 i

412235 1129118 1153D85 412717 1163539 1153DB5 L412236 1129143 1153DB5 412718 1163529 1153DB5 j

'412237 1129156 1153DB5 412719 1163581 1153DB5 412238 1129224 1153DB5 412720 1163621 1153GB5 412239 1129216 1153DB5 412722 1163523 1153GB5 5

412240.

1129147 1153DB5 412734 1116189 1153AB5 l412359>

1129254 1153DD5

• 412735' *1101842 1153AB5 412360.

1163515 1153DD5 412736 1116235 1153AB5 412364 3129221 1153DB5 412737 1116207 1153AB5 412365 1129150 1153DB5 412771 1129305 1154HP5 412366 1129223 1153DB5 412776 1163627 1153DD5 412374 1163516 1153DB5 412777 1163594 1153DD5 1

412376 1129315 1153DB5 412786 1163601 1153DB5

• 412377 1163518.

1153DB5 412875 1109338 1154GP9

'412379-

-1163584 1153DB5 412885 1163543 1153DB5 j

f 412380 1163625 1153DB5 412905 1163507 1153DB5 412381 1163578 1153DB5 412913 1163517 1153DD5

)

412410 1163536 1153DB5 412915 1163564 1153HD5 412411-1163624 1153DB5 412930 1163511 1153DB5 1

412501 1163628 1153DD5 412938 1129312 1153AB5 i

412502.,1163577 1153DD5 412941 1163524 1153DB5 412503 1163569 1153DD5 412956 1129255 1153DB5 412504

.1163632-

- 1153DDS 413036 1163559 1153DB5 412506 1163522 1153HD5 413042 1163560 1153HD5

)

412521 1163568 1153DBS 413045 1163557 1153HD5

-)

412522 1163609 1153DB5 413046 1163545 1153HD5 412523 1163589 1153DB5 413261 1163533 1153DBS

)

412524 1163587 1153DB5 413268 1163551 1153DBS 412525 1163622 1153DB5 413279 1163552 1153DB5 i

412526 1163629 1153GB5 413281 1163512 1153DB5 j

412527 1163602 1153GB5 413288 1129316 1153HB5 i

412528 1163633 1153GB5 413291 1163514 1153HB5

412531 1163580 1153GB5 413309 1129245 1154DP5 412532 1163583 1153GB5 413383

. 1163550 1153GB5 1

412533 1163570 1153GB5 413411 1109336 1154GP9 I

412534 1163597 1153GB5 413460 1163525 1153DB5

'412535 1163528 1153GB5 413670 1129309 1153DB5 412536 1163563 1153GB5 413674 1129262 1153DB5 412537 1163537 1153GB5 413744 1163544 1153HD5 1412546 1163620 1153DD5 413763 1163603 1154DP5 L

412595f 1129295 1153DB5 413766 1163573 1154HP5 412596 1163626 1153DB5 413889 1163588 1154DPS l412598 1163630 1153DB5 413890 1163586 1154DP5 412599 1163591 1153DBS 413891 1163600 1154DP5 412600 1163623 1153DB5 413892 1163565 1154DP5 412609 1163576 1153DD5 413893 1163619 1154HP5

.412615 1129306 1153DB5 413895 1163579 1154HP5 412616 1163590 1153DB5

• 414113

• 1295007' 1153DB4 412617 1163575 1153DB5 414114 1277820 1153DB4

~

n.,n

, -,,,,-e.

.--c--..n.,, - -. - - -, - - -.,,..,,,., -

,--,-,,-.-.n,.,

- - -. - - -,. - -. -.. - - -, ~, -

(

d.

~*1295001 1153DB4

• 414772

• 1240930 1153DB8 0414115 414116.

1329625 1153DB4

• 414777

• 1240931 1153DB8 i

414145 1163540' 1153GB5 414844 1265811 1153DB4 l

414147 1163519 1153DB5 414845 1265814 1153DB4 414345' 1129319 1154DP5

• 414846. *1294976 1153DB4 414346 1039035 1154DP5 414868 1277826 1153DB4 1C414347

• 1101910 1154DP5 414869 1277813 1153DB4 414388 1277827 1153DB4 414870 1277828 1153DB4 0414392

• 1295003 1153DB4 414871 1277806 1153DB4

• 414672

• 1295010 1153DB4 414393' 1277810 1153DB4 1

414455

'1129299 1154DP5 414873 1277808 1153DB4

'414550' 1277815

• 1153DB4 414877 1329720 1153DB5 414596 1307538 1153DB5 414884 1307509 1153DB5 i

414604 1277799 1153HD4 414885 1307480 1153DB5 i

0414610- *1294998 1154GP4 414897 1163567 1154GP5 414611 1265813 1154GP4 414898 1163604 1154GP5 414612 1265808.

1154GP4 414899 1277809 1154HP4 0414613

• 1294999 1154GP4 414900 1329657 1154HP4 1

0414'614

• 1294978 1154Gp4 414902 1307519 1153AB5

'414615 1265818 1154GP4

• 414913

• 1295007 1153DB4 I

0414616- *1295005 1154GP4

• 414915

• 1295001 1153DB4

'414617 1265821 1154GP4 414916 1329625 11S3DB4

/

t 414626 1307476 1153DBS 414918 1329700 1153GB5 414637 1277829 1153DB4 414919 1329716 1153GB5 0414638

• 1295002

-1153DB4 414920 1307514 1153DB5 I

-414639 1277825 1153DB4 414921 1329727 1153DB5 414640 1277807

' 1153 DB4 414924 1329725 1153DB5

-414641 1277796 1153DB4 414925 1329721 1153DBS

'414642 1277822 1153DB4 414926 1329726 1153DB5 414643 1277801 1153DB4 414927 1329723 1153DB5 0414644

• 1294995 1153DB4 414930 1329711 1153DB5 414645 1277814 1153DB4 414937 1329706 1153HB5 414647 1307501 1153DD5 414938 1129702 1153HB5 414653 1307511 1153DB5 414939 1329707 1153DB5 E414659 1277805 1153DB4 414940 1329712 1153DB5 L414662 1307504 1153AB5

• 414992

• 1295011 1154DP4 414663 1307500 1153AB5 414993 1329646 1154DP4

-414664 1277803 1153DD4

• 414994

• 1294978 1153DD4 414665 1329651 1153DD4

• 414995

• 1295009 1153DD4 7

'414666 1307527 1153GB5 414997 1277795 1153DB4 1

414667' 1307487 1153GB5 415015 1307477 1154HP5 414668 1307523 1153GB5 415026 1265807 1153DB4 414669

~1307507 1153GB5 415028 1265806 1153DB4 414671 1307479.

1153GB5

• 415029

• 1295008 1153DB4 0414695

• 1294982 1153DB4

• 415030

• 1294981 1153DB4

'414698 1307483 1153HD5

• 415031

• 1294993 1153DB4 414699_

1307515 1153HD5 415032 1307540 1153DB5 414700 1307485 1153HD5 41503S 1329673 1153DB5 0414703

• 1294988 liS3uB4 415036 1329734 1153DB5 0414704

• 1294989 1153DB4

• 415037

• 1294991 1153DB4 414741' 1307517 1153DB5

• 415038

• 1294986 1153DB4 414745 1265797 1153DB4 415066 1265819 1153DB4 1

414746 1265795 1153DB4 415067 1277836 1153DB4 l

414750 1307534 1153GB5 415068 1277834 1153DB4

-414764 1265802 1153DB4

• 415069

• 1294983 1153GB4

)

.414765.

1265817 1153DB4 415074 1265804 1153DB4 l

b415075

• 1294979 1153DB4 417258 1329619 1153GB4 415078 1329687..

1153HD5 417259 1329650 1153GB4 c415097

• 1294992 1153DB4 417262 1493164 1153AB$

415098 1329656 1153HB4 417266 1163531 1153GB5 415099 1329644 1153HB4 417300 1329629 1153DB4 415100-1329639 1153HB4 417307 1329647 1153GB4 0415101- *1295004 13,53HB4 417308 1329636 1153GB4 415102 1265805 1153HB4

• 417473

• 1240931 1153GD8 415103 1265816 1153HB4

• 417474

• 2240930 1153GD8 415106 1329627 1153DB4 418155 1696286 1153DB3 415110 1329717 1153DB5 418156 1696264 1153DB3 415112 1329688 1153DB5 418161 1696295 1153DB3' 415118 1307482 1154DP5 418164 1696271 1153DD3

" 415195 1329630 1153DB4 414165 1696299 1153DB3 415200 1329652 1153GB4 418169 1696293 1153DD3 415210 1329696 1153DB5 418170 1696277 1153DD3

.415211

.984542 1153DB5 418210 1696285 1153DB3 415212 1329620 1153GB4 418339 1696297 1153DB3 415213 1329674 1153DB5 418340 1696298 1153DB3 415214 984513 1153DB5 418341 1696276 1153DB3 415215 984517 1153DB5 418342 1696288 1153DB3 415216 1329730 1153DB5 418356 1696289 1153GB3

/

415217 1329710 1153DB5 418404 1696291 1153DB3 415319 1329686 1153DB5 418405 1696269 1153DB3 415226 1329637 1153DB4 418410 1696296 1153DB3 415227 1277820 1153DB4 418411 1696284 1153DB3 i

415389 1329658 1154HP4 418822 1696266 1153DD3 1

415390-

'1329645 1154HP4 418907 1696268 1153DB3 415392 1329659 1154HP4 418908 1696267 1153DB3 415393 1329655 1154HP4 419006 1792289 1154DP4 415394 1329631 1154HP4 415630 1329729 1153DB5 l

415631 1265794 1154DP4 415643 1329649 1153DB4 415644 1329648 1153DB4 i

415686 1329623 1154HP4 415687 1329628 1154HP4 415689 1265803 1154HP4 415690 1265815 1154HP4 410691 1329635 1154HP4-415692 1329621 1154HP4 415693

-1329643 1154HP4 415694 1265796 1154HP4 415896 1163572 1154HP5 410898 1163618 1154DP5

~ 416219 1493140 1153AB5 416232 1493163

-1153DB5 l

416263 1493149 1153DD5 416283 1493131 1153DBS l'

416818' 1493136 1153 ABS 417008 1493145 1153AB5 417109 1493148 1153AD5 L

417110 1493137 1153AD5 l

417111 1493138 1153AD5 417132 1493134

-1153AB5 417191 1329640 1153HD4 i

1 v.~

"8~"e md a

i J

l e

o, O

I w

en z

.5 R

yW

• Cn s

. Y 00 8

?

a i

oe I

'EZ.g e

5 hO D

E 5#

O v-g m

e

~

~

Q k

E p

c w

+

G e

o i

-m u

m m

.o i

i O

5 5

lc, E

g a

t 3

e m

25 m

g

.N x

E De s

m O

g h

E E

s s

t m

s o

E y

2

?

8 e

s a

w 5

m V

O O

O O

C h

M yy rw-e is-p arvi-w<vw=*ea'meR-*is h~

-,wa.

e ma m

-**--a-'

wa -

2-

---a-~--a-

Rosemount/NRC Meeting-18 Dec 1989 1

Introduction Technical Bulletin 4 Summary.

o Technical Bulletin #4 provides sufficient information for the industry to impieirimd appropriate action on the loss of fill fluid condition i

o Summary of data being provided in Technical Bulletin #4 i

o The information will be presented in two sections

- Risk Assessment

- Diagnostic Guidelines i

s w..,

e4

.m,,.._

o..

m

.m

. ~,

- n Rosemount/NRC Meeting 18 Dec 1989 2

5 Steo investiaation Proaram on Model 1153 and 1154 Glass to Metal Seal Failures 1

o identify all symptoms of oil loss 8

o Quantify oil loss symptoms in the laboratory o Define diagnostic guidelines for iridustry review 4

4 o Field test proposed diagnostic guidelines o Publish Rosemount recommended oil loss diagnostic guidelines 4

9&*

Qm*

• b i

L

=

u l'

Rosemount/NRC Meeting 18 Dec 1989 Risk Assessment o 1153/1154 Failure Data (glass to metal seal failure) 91 Confinned Failures i

16 Units in Failure Analysis l

107 TOTAL l

Approximately 14,145 Units Shipped (Exclude 1158 Model 1153 Series A units which use eissic,iTieric o-ring)

Failure fraction:

0.82%

4 o Estimate Failure Rate

- Rosemount has estimated that the 1153 and 1154's have accumulated appsviiiistely 110,000,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of service

- Estimated overall failure rate of 1 x 104 failures /hr s

j i

t

,,.y.,

~,, -

?-,~.-

~ * *, >

- ~. - - - - - - - ~ ~

'-s-

y

~

Rosemount/NRC Meeting 18 Dec 1989 i

Rosemount Mathemittical Leak Rate Model i

o Rosemount developed a mathematical leak rate model based on viscous flow equations o Actual field failure data supports this model i

o A strong pressure, time-in-service relationship has been established i

- Transmitters at low pressure have an acceptably low failure rate in the order of magnitude of 10-7 failures /hr during their service life l

l

- Transmitters at higher pressure fail sooner j

- Transmitters at high pressure which surpass a critical l

time-in-service achieve an acceptable failure rate l

l i

i

w s

v Rosemount/NRC Meeting 18 Dec 1989 1

l i

i Rosemount Mathematical Leak Rate Model o Pressure / time-in-service factor i

- Criteria to establish when a transmitter is in the acceptable failure rate range o The amount of oil that can be lost before performance is effected is range dependent i

3 l

o Two pressure / time-in-service factors have been established to account for the oil loss range dependence

- Ranges 3 thru 5 (AIITypes) 60,000 psi-mo 1

- Ranges 6 thru 10 (Gage / Absolute) 60,000 psi-mo l

- Ranges 6 thru 8 (Differential) 130,000 psi-mo i

i

}

l

~

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

c, Rosemount/NRC Meeting i

18 Dec 1989

]

1 Other Risk Assessment Considerations o Transmitters in stand-by service have an acceptably low failure rate due to the limited time at pressure 1

o Low Range (Ranges 3 thru 7)' absolute and gage liaisnitters have an, acceptably j

low failure rate due to the low pressure applied 1

i l

l f;

i

-n

...,-e,

--m en r,-e w.+-

w.,.w.-

-..,.., -+-... _,

_,._.n..._n.

-.m

t

..p, Rosemount/NRC Meeting 18 Dec 1989 Susoect Manufacturina Lots i

o Rosemount clearly sees the failures of 1153's and 1154's as clustered i

o The lots identified with failures have.had a failure fraction of over 6%

o The remaining manufacturing lots have a failure Tigrui of less than 0.2%

i o The 30 times difference is significant 1

o Rosemount cannot confirm that all suspect lots have been identified I

i.,

s Utility Decision Tree Determine all 1153 &

1154 transmitters l

received

• J

\\

1 I YES Do you have any 1153 No safety concern Series A units?

No g

No current safety YES Are transmitters concern. Return to spares or not

{

Rosemount for in service?

module replacement U

l technicians operators I',0Pera n g,,

n psl?

on low oil symptoms and continue operation l

h U

I YES i.

Are transmitters in stand by service?

f U"

YES is the pressure x time

-=

> 60k psi mo(all R3 5)

>60k psi mo(AP/GP R610)

L

>130k psi-mo(DP/HP R6 8) ?

NO p

Continue to operate using diagnostic guideline (s) until critical pressure x time in service value is met or until transmitteris replaced.

l.

i l

'Model 1153 and 1154 with serial numbers less than 500,000 l

containing the original sensor module l

Plant specific analysis of application and transmitter diagnostics should be considered 0

4 m --

Rosemount/NRC Meeting

~

18 Dec 1980 Conclusions from PRA Conducted by PLG l

o There does not appear to be any clear risk-significant scenarios arising from l

the use of 1153's and 1154's in BWR's.

l l

j o For PWR's only one identified scenario was observed. This was a reactor trip i

i followed by failure of sensors monitoring pressurizer pressure and level which assist in cool down

- No credit for weior intervention

- Assumes no redundancy of measurement o Under high pressure, the transmitters susceptible to a glass to metal seal induced oil loss failure will reveal themselves relatively early in life l

l r

i

~.

~

-w

~-

.--.s

.-.~..

m

-.v--

.. _ _ -_._.. *. -. +.-.. ~,

Rosemount/NRC Meeting 18 Dec 1989 l-Diaanostic Guidelines i

i o Rosemount has developed 3 diagnostic guidelines which can be used individually or in combination

- Drift Analysis

- Sluggish Response

- Process Noise Analysis o A number of factors (application data, transmitter. range, age of unit,etc.) must be considered by each utility as it determines which diagnostic guideline to use-s i

s tn s

i d

r s

u e

e r

t e

c4 t

t i

t c

m d

t i

e s

m p

n d

s s

a r

n u

r o

a s

t c

r t

t y

c e

t f

e b

c i

t e

n p

g p

e s

t s

d u

s s

d u

n u

i t

s e

m i

n y

r e

i y

a f

u d

i e9 t

f c

n q

n i

k t

M8 n

s e

e a

r 9

e d

d i

t C1 d

n a

e f

i i

e t

n r

r r

a a

a d

R c o

t d

c t

N e s

t tf r

u y

a e

p

/D d

ir r

t t

t t

e d

o a

t n

d u

i 8

t u

n e

a d

m o

r s

o1 e

v o

a n

e l

r i

t t

m a

b n

a s

e g

a o

tr u

l a

e b

e t

i s

n m

a t

c r

n o

n o

e a

n r

a r

p d

a o

f R

c o

n c

a e

d u

f t

v e

o d

s t

f e

n i

a i

i i

r o

d s

t o

n g

it n

p e

n s

i e

d i

o d

d s

n m

i i

t n

a o

a e

e r

v s

e c

c t

r i

ib i

i t

n ta m

s e

t r

l l

e e

a u

v o

h l

i l

d c

m m

t i

t ta O

i f

u l

i G

a u

r l

m C

D u

n I

m e

s r

t o

u o

is N

C N

y lan o

o o

A tf irD

~~

Rosemount/NRC Meeting 18 Dec 1989 A

Stuaaish Response Guideline o A qualitative assessment has been used effectively during normal calibration Technicians knowledgeable with Rosemount transmitters can detect suspect units

- A number of confirmed failures have beenideiinfred by i

this method 4

i o An optional quantitative bench test has been developed to confirm a transmitter with low oil j

)

s

=<.

,w aw.

+ +

ee.

-,.m. -.

x

,.,r.--,

~

, -, - - <~.,. = -. +.n

)

l i;

I;k :

I I

t i i

~

~

n o

t i

i t

n u

u a

c t

c e

i.

p ni s

u s

e a

pe y

d l

id n

a o

i d

ta g

i n

c n

o n

i l

a p

i i

t t

e a

c p

r a

e e

r p

e y

M9 o

tt r

i e

8 m

v l

C9 a

s i

n e

R1 n

a b

i r

Nc o

t n

o

/e e

t tnD g

h s

t n

ra u8 ir f

u o

e o1 d

p e

p m

d d

a e

e u

t s

r i

s o

p s

l i

o t

y in m

l R

o a

na m

a l

a e

n e

b g

s i

i s

o e

n n

n a

e i

c h

s le t

s d d

e e

e f

iu s

o c

s i

G o

o u

r n

g p e e

n b

s s

t f

i s

e o

io e

t iu s

N c

e u

o q

s s

P A

U m

s r

ec o

o o

o rP 4

i i!:

l:

i '

1 1l i!;)1

. ii*

~.

Rosemount/NRC Meeting 18 Dec 1989 i

i Conclusion f

I o Sufficient information for the industry to implement appropriate action I

on a loss of fill fluid condition i

o The diagnostics discussed here can be used for all types of loss of oil failures o Rosemount is now prepared to support the industry as it implements the l

required actions b

b 1

I f

i i

k i

k s

-'_E.

f_

t

[

Rosemount/NRC Meeting-l'

.18 Dec 1989 4

Susceptibility of Oil Loss on Rosemount Transmitters o The NRC-has. requested information from Rosemount on the loss of oil with respect to Model.1151,1152i 1153,1154

~

o Rosemount has responded to this request i

j o This is an overview of the public information covered i

7

.k 1

i q.

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

y s

l

.u Rosemount/NRC Meeting-

'18 Dec 1989 u

l l

l Susceptibility of Loss of Oil on Filled or Evacuated Measurina Devices

+

i o Any pressure measuring device that uses a sealed, processisolated system (liquid orgas) to transmit and sense measured pressure is susceptible to fill fluid loss i

This equipmentincludes:

i

- Pressure transmitters f

- Remote seal systems

- Leveltransmitters i

i

- Other similar devices q

i o Loss of fill fluid in these systems can result in transmission of an inaccurate measurement I

s

~

Rosemount/NRC Meeting 18 Dec 1989 Suoolv of Rosemount Manufactured Sensor Modules o Rosemount ' supplied equipment under its own name and trade mark o Rosemount manufactured transmitters that are resold under private label by

.l another entity 1

i o Rosemount manufactured equipment supplied to OEM's which are resoiqi and carry the Rosemountidentification i

o A secondary market exists'for new or salvaged equipment l

o Unauthorized remanufacturers / refurbishers exist for Model 1151 transmitters 1

f N

,7...

,._.m.......y.,

+

~

l Rosemount/NRC Meeting.

E 18 Dec 1989 l

l.

Susceptibility of Model1151 to Loss'of Oil o Review of Model 1151 commercial records indicates that approximately 0.2% of the c

sensing components have been returned for all reasons.

A portion of these l

returns are due to'ioss of fill-fluid from all causes including damage and corrosion i

j o Rosemount considers oil loss on Model 1151 to be random i

4 h

i l

i s

n.

_ ;, ;.3

' N ~, j. S Rosemount/NRC! Meeting

.18 Dec 1989 Susceptibility of Model1152 to Loss of Oil f

o The Model 1152 is similar-to the model 1151, especially with respect to the _

process-o-ring o 1 Loss of' oil-failure (alltypes) has been confirmed by Rosemount out of 8800 U.S. units o Rosemount considers this to be random 6

s em J".,

4 y

+p pq v#h-r-

r y-g

• q g'-

N

,m,

=r' app,.

"i'&r w

+

m-s-

"-_____y

n

~

~

~

m Rosemount/NRC Meeting 7

.18 Dec 1989

~

Conclusions and Actions Related to Oil Loss ' Susceptibility o Any pressure transmitter which uses a sealed,. process isolated system (liquid or gas) can lose fillfluid. This event can result in the transmission of an inaccurate i

measurement.

I Rosemount will issue an informational letter to all U.S. nuclear faciliti$ to inform o

them of this fact and indicate that Rosemount Model 1151, 1152, 1153, 1154 pressure transmitters and Model 1199 remote seals use oil filled-systems _ and' therefore can lose oil and transmit an-inaccurate measurement.

This letter also states that the diagnostic tests presented in Technical Bulletin #4 can detect l

incipient' oil loss.in any Rosemount pressure products.

It further requests that

'l any information regarding loss of fil! fluid on Rosemount transmitters be brought to our attention.

i

- ?

I w

e 1O, i

Rosemount/NRC Meeting

,l

- 18 Dec 1989 s

l Conclusions and Actions Related to Oil Loss Susceptibility-0 i

t o There is an identified problem -with Rosemount Model 1153 and 1154 Transmitters manufactured before July 1989.

i Engineering and returned-product failure analysis confirms this problem as a o

failure of the sensor glass to metal seal which is caused by forces from a metal o-ring used exclusively in these models.

Rosemount has taken action in the form of design and manufacturing changes o

to substantially eliminate the problem on product shipped after July 1989

}

- - c

"^e'"

WL' Yh d e

rd'-#

4

-W;V d'

V'

=---

a 4:

Rosemount/NRC Meeting:

18 Dec 1989 Conclusions and Actions Related to Oil Loss Susceptibility" o Rosemount has reviewed failure data relating to Model 1151 transmitters used in k

commercial applications.

This includes data relating to suspect.and identified loss of oil failures.

Based on evaluation of the data and consideration of relevant product application conditions, Rosemount concludes that the failure of i

Model 1151 transmitters 'due to loss of oil is random.

a i

i

[

-i

.w'

.-,.,.___....y,

, ~..

... ~

-m-.

s

(;

x Rosemount/NRC Meeting w

18 Dec 1989 Summary o A loss of fill fluid problem was identified on Model 1153 and 1154 transmitters' l

o Rosemount undertook a comprehensive engineering program to fully undeishsrni this problem Rosemount implemented ~ design and manufacturing changes to deal with the.

o identified problem in July 1989 Rosemount has made a comprehensive review of the loss of oil susceptibility o

~

issue as it applies to allproducts. Pertinent facts and recommendations related to l

this issue are being transmitted to all U.S. nuclear facilities l

2 o These actions provide nuclear facilities with the. necessary information, tools and j

guidelines for continued safe operation i

1 3

J-

---u

. J

4.

' 4 n,

- Rosemount/NRC Meeting i

18 Dec 1989

[

Summary l

l o Diagnostics are available to detect incipient failure of Rosemount transmitters due to loss of oil i

i o Guidelines are in place for diagnostic use on Model 1153 and 1154 tr.;nsmitters-manufactured prior to July 1989 i

o Rosemount will issue a letter to all U.S. nuclear facilities to inform 1 Hem of-

~

1 i

l Loss of oil susceptibility l

Availability of diagnostics

-Sources of supply of Rosemount components The necessity to report symp. toms / failures to Rosemount o Rosemount will continue to monitor and evaluate the entire situation and take appropriate action as required g

E e

I

,.m_

, _., ~... _

---x

Y W.' Stephen J. Wanek Vic) President Operations Rosemount, Inc.

12001 Technology Drive Eden Prairie, NN 55344

Dear Mr. Wanek:

SUBJECT:

LOSS OF FILL-Olt IN TRANSMITTERS l'iANUFACTURED BY ROSEHOUNT During our discussions of October 24, November 20, and November 21, 1989, we requested that Rosemount provide information op the susceptibility of Model 1151 and 1152 transmitters to loss of fill-cil.- The primary purpose of this letter is to reiterate that request for information and to request additional infortnation on loss of fill-oil in transmitters manufactured by Rosemount.

It is our understanding that failures of Model 1151 transmitters due to loss of fill-oil have been confirmed by Rosemount. Model 1151 transnitters, although supplied by Rosemount as commercial grade equipment, are utilized in safety-related systeras in nuclear power plants.

A public meeting between Rosemount and the NRC has been scheduled for December 18, 1989 to discuss Rosemount's recommended surveillance and testing procedures to detect, prior to failure, Model 1153 and 1154 transmitters that are leaking fill-oil and to discuss the susceptibility of Model 1151 and 1152 transmitters to loss of fill-oil. We request that, prior to this meeting, Rosemount provide responses to the enclosed questions for Model 1151, 1152, 1153, and 1154 transritters.

l In accordance with NRC practice, a copy of this letter will be placed in the NRC's Public Docunent Room.

l Sincerely, Original Signed By Carl H. Berlinger Carl H. Ber11nger, Chief Generic Communications Branch Division of Operational Events Assessment Office of Nuclear Reactor Regulation

Enclosure:

Questions on Transmitters DISTRIBUTION w/ enclosure FJMiraglia, NRR CERossi, NRR CHBerlinger, NRR JERamsey, NRR CHaughney, NRR JAcalvo, NRR SNewberry, NRR AThadani, NRR WBrach, NRR BGrimes, NRR PDR Central Files l

DCB DOEA R/F OGCB R/F JRamsey R/F

'~

$5EE A2EL'lcub couvaJ24re EA:NRR

$GCB:D0EA:hRRk/SICB: DST:NRR$/ DST:NRR C

i :DRIS:NRR I

RR JERamsey SNewberry AThadani WBha~ch

[nBK3 rimes f1(JJ/89 11/a>/89 11/D/89 11/$/89 11/g/89

Dscarbar 1, 1989 Enclosure n.

1)

Provide the model, range, ar.d lot (weld) numbers, the date of manufacture, the date of return, and the date and amount of time in service for all Model 1151, 1152, 1153, and 1154 transmitters returned to Rosemount that may have exhibited symptoms indicative of loss of fill-oil. This should include transmitters returned to Rosemount in Eden Prairie, Minnesota, as well as Rosemount's regional service centers (Los Angeles, California; o

Charlotte, North Carolina; Houston, Texas; Cleveland, Ohio; and Baton Rouge, Louisiana).

l 2)

Provide the model, range, and lot (weld) numbers and the date of confirma-

_i tion for all Model 1151, 1152,~1153, and 1154 transmitters that Rosemount has confirmed have experienced a loss of fill-oil.

3)

Providethestatusoftransmittersidentifiedin1)abovethathavenot 1

been subjected to Rosemount's failure analysis.

4)

Provide the model, range, and lot (weld) numbers, the date of manufacture, and the date and amount of time in service for all Model 1151, 1152, 1150, and 1154 transmitters that Rosemount is aware of in which symptoms indica-tive of loss of fill-oil are believed to have been seen; however, the transmitter cannot or has not been returned to Rosemount for failure analysis.

In addition, describe Rosemount's program to identify addition-al transmitters that may have shown symptoms indicative of loss of fill-oil.

c 5). Provide a listing of all customers, both domestic and international, that i

have purchased Model 1152,1153 or 1154 transmitters.

6)

Provide information regarding the total number of Model 1151, 1152, 1153, E

and 1154 transmitters manufactured by Rosemount.

In addition, identify transmitters Rosemount considers not to be covered by Rosemount's 10 CFR Part 21 notification on-loss of fill-oil.

7)

Provide a listing of all companies that Rosemount believes may have refurbished transmitters and then sold the refurbished transnitters, as well as a listing of customers that may have purchased these refurbished transmitters.

8)

Provide a listing of all manufacturers that have purchased Resemount

~ transmitters and components, including sensing modules, and also indicate whether these manufacturers have been informed of the loss of fill-oil L

issue.

9)-

Describe actions taken by Rosemount and other nuclear industry representa-tives, prior to development of Model 1152, 1153 and 1154 transmitters, to l

qualify Model 1151 transmitters for service in nuclear power plants.

10) Provide confirmation that the sensing module utilized in Model 1151, 1152, 1153, and 1154 transmitters is not utilized in other Rosemount transmitter models.

{

- - -... -..