Nuclear Engineering and Design 134 (1992) 283-294 283 North-Holland

Evaluation of check valve monitoring methods *

H.D. Haynes

Oak Ridge National Laboratory, Oak Ridge, USA

Received 4 April 1990, revised version 30 October 1990

Check valves are used extensively in nuclear plant safety systems and balance-of-plant (BOP) systems. Their failures have resulted in significant maintenance efforts and, on occasion, have resulted in water hammer, overpressurization of low-pressure systems and damage to flow system components. Consequently, in recent years check valves have received considerable attention by the Nuclear Regulatory Commission (NRC) and the nuclear power industry. Oak Ridge National Laboratory (ORNL) is carrying out a comprehensive two phase aging assessment of check valves in support of the Nuclear Plant Aging Research (NPAR) program. As part of the second phase, ORNL is evaluating several developmental and/or commercially available check valve diagnostic monitoring methods; in particular, those based on measurements of acoustic emission, ultrasonics, and magnetic flux. These three methods were found to provide different (and complementary) diagnostic information. The combination of acoustic emission with either ultrasonic or magnetic flux monitoring yields a monitoring system that succeeds in providing sensitivity to detect all major check valve operating conditions. The three check valve monitoring methods described in this paper are still under development and are presently being tested as part of a program directed by the Nuclear Industry Check Valve Group (NIC) in conjunction with the Electric Power Research Institute (EPRI). Phase 1 of this program (water testing) is being carried out at the Utah Water Research Laboratory located on the Utah State University campus.

1. Introduction

1.1. Background

Check valves are used extensively in nuclear plant safety systems and balance-of-plant (BOP) systems. Their failures have resulted in significant maintenance efforts and, on occasion, have resulted in water ham- mer, overpressurization of low-pressure systems and damage to flow system components. These failures have largely been attributed to severe degradation of internal parts (i.e., hinge pins, hinge arms, discs, and disc nut pins) resulting from instability (flutter) of check valve discs under normal plant operating condi- tions. Check valve instability may be a result of misap-

Correspondence to: H.D. Haynes, Engineering Technology Di- vision, Oak Ridge National Laboratory, P.O. Box 2009, Oak Ridge, TN 37831, USA. * Research sponsored by the Office of Nuclear Regulatory

Research, U.S. Nuclear Regulatory Commission under In- teragency Agreement DOE 1886-8082-8B with the U.S. Department of Energy under contract No. DE-AC05- 840R21400 with Martin Marietta Energy Systems, Inc.

plication (using oversized valves) and exacerbated by low flow conditions and/or upstream flow distur- bances. [1] Present surveillance requirements for check valves have been inadequate for timely detection and trending of such degradation, because neither the flut- ter nor the resulting wear can be detected prior to failure. Consequently, in recent years, check valves have received considerable attention by the U.S. Nu- clear Regulatory Commission (USNRC) and the nu- clear power industry.

Oak Ridge National Laboratory (ORNL) has car- ried out a comprehensive two phase aging assessment of check valves in support of the Nuclear Plant Aging Research (NPAR) program that was established by the USNRC Office of Nuclear Regulatory Research (RES) primarily as a means to resolve technical safety issues related to the aging of electrical and mechanical com- ponents, systems, and structures used in commercial nuclear power plants [2]. ORNL check valve research provides information applicable to the NRC inspection and regulation of nuclear power plants, the ASME Pressure Vessel Code, the ASME Operation and Maintenance (O&M) Standard, and the resolution of NRC generic issues. It also is consistent with the

0029-5493/92/$05.00 1992 - Elsevier Science Publishers B.V. All rights reserved

284 H.D. Haynes / Evaluation of check l,alce monitoring methods

objectives of the recently organized Nuclear Industry Check Valve Group (NIC) in coordinating the nuclear industry response to the Significant Operating Experi- ence Report (SOER 86-03) prepared by the Institute of Nuclear Power Operations (INPO). SOER 86-03 recommends that nuclear power plants establish a pre- ventative maintenance program to ensure check valve reliability which should include periodic testing, surveillance monitoring, and/or disassembly and in- spection.

1.2. Objective and scope

The primary objective of the check valve aging assessment program is to identify and recommend methods of inspection, surveillance, and monitoring that would provide timely detection of check valve degradation and service wear (aging) so that mainte- nance or replacement can be performed prior to loss of safety function(s). In that regard, ORNL has been carrying out an evaluation of several developmental and/or commercially available check valve diagnostic monitoring methods, in particular, those based on mea- surements of acoustic emission, ultrasonics, and mag- netic flux. The evaluations in each case have focused on the capability of each method to provide diagnostic information useful in determining check valve aging and service wear effects (degradation), check valve failures, and undesirable operating modes.

A description of each monitoring method is pro- vided in this paper including examples of test data acquired under controlled laboratory conditions. In some cases, field test data acquired in-situ are also presented. The methods are compared and suggested areas in need of further development are identified.

2. Check valve function and types

The function of a check valve is simply to open and thus permit flow in only one direction. When the flow stops or reverses direction, the check valve closes. Check valves are self actuating - that is, they require no external mechanical or electrical signal to either open or close. As a result, most check valves have no capability to be actuated other than by changing flow through the valve. Several types of check valves are commonly used, such as the swing, piston-lift, ball, and stop-check designs. The differences between types per- tain to the type of obturator used. The discussions presented in this paper refer entirely to the swing check valve, shown in fig. 1. However, all monitoring

/% o :,L,~

o F

"~-_ ~"~. 'C~, ~ I ~ ~ HINGE ARM /:~ " ~ ',,t~::: : ~ %~ O HINGE PIN PLUGS

:: "~ '~ ~ 0 q ~(~l l~ 0 ~ CAP STUDS

', ",\ ",,,,;\y O'PG- -ET " \~ J ", r,\ ~,~/~'" ::/'j e SEAT RING

': " ' '~ )~/ ; '~ ~'~ ' D,sc W,SHE. + H,.GEP, N

DISC NUT

Fig. ]. Typical swing check valve.

methods described herein have the potential for being applied to other check valve types.

3. Acoustic emission monitoring

3.1. Basic principles

Acoustic emissions (pressure waves) can be gener- ated in a variety of ways. Of particular interest are those generated either when solids contact each other or when liquids or gases flow through pipes and fit- tings. Acoustic emissions are detected by sensors which respond to pressure waves over a wide range of fre- quencies, such as piezoelectric-type accelerometers or microphones. Signal conditioning electronics can be used to amplify selected acoustic noise signals while attenuating others, e.g., unwanted environmental back- ground noise. Analyses of acoustic emission signals obtained from check valves can be used to monitor check valve disc position, movement, and mechanical condition, as well as internal flow/leakage through the valve.

3.2. Detection of valve disc movement

Acoustic emission monitoring has been shown to detect check valve disc movement. As an example,

H.D. Haynes / Evaluation of check valve monitoring methods 285

II I

SYSTEM SCHEMATIC

'~ - COLD LEG ACCUMULATOR CYCLE TEST OF A TEN INCH COLD LEG ACCUMULATOR DISCHARGE CHECK VALVE

MOTOR-OPERATED VALVE BY DUKE POWER COMPANY - (STROKE TIME = 10 SECONDS) IN MARCH 1984

jl(- MONITORED CHECK VALVE TO REACTOR VESSEL

r'--,J l",,,J ~ COLD LEG

I ACOUSTIC EMISSION SIGNATURE

TH ROI-I-LING NOISE

' , i IL : i FULLY'c)PEN i ' ' CLOSURE-~- I I ' T . : ~ I i ~ : i II ~ .L | I ~ FLOWNOISE--I..

I : : ~ , l l I I I I | ,< . . , I . . . . - " I , : L ~

i . . . . q I I i I , I , I I 1 !1 ] , I i i l l , - - TIME " l" 20(S) PJ

Fig. 2. Acoustic signal vs. time for a ten inch check valve tested by Duke Power Company in March 1984. Taken from: W.M. Suslick. Proposed Technique for Monitoring Check Valve Performance. Presented at the INPO Check Valve Technical Workshop,

October 30-31, 1986.

Duke Power Company [3] installed an acoustic sensor on top of a 10-inch cold leg accumulator discharge check valve. A schematic representation of the installa- tion is illustrated in fig. 2. After initially charging the accumulator to 100 psig, the motor-operated discharge valve was cycled. The acoustic sensor output during this cycling was processed and displayed on a strip chart recorder. The resulting acoustic signature (fig. 2) shows that the sensor detected the metal-to-metal con- tact occurring at the end of both the opening and closing strokes.

Duke Power Company has also carried out check valve acoustic emission testing under controlled flow loop conditions and with the introduction of various implanted defects which simulated severe aging and service wear [4]. Accelerometers were strapped to the bodies of three check valves in a manner depicted by fig. 3. The following discussion summarizes the results obtained from those tests.

Tapping of the valve disc against its backstop was easily detected and distinguished from background flow noise as shown in fig. 4. In addition, by using two (or

"--'] -- FFT OSCILLOSCOPE

~ -7 CHANNEL TAPE RECX)RDER -- ' ] - - AUDIO SPEAKER

~ - - PRE AMP

l I CHECK VALVE /

g x~ 1~ - - AC~ELEROMETER

Fig. 3. Acoustic emission equipment (schematic representa- tion) used by Duke Power Company in 1987 flow loop tests. Taken from: W.M. Suslick, H.F. Parker, B.A. McDermott. Acoustic Emission Monitoring of Check Valve Performance. Presented at the EPRI Power Plant Valves Symposium, Octo-

ber 11-12, 1988.

286 1-1.D. Haynes / Evaluation of check valve monitoring methods

more) valve mounted acoustic sensors, Duke Power was able to approximately locate the source of the tapping based on a comparison of the "time of arrival" of the acoustic signals acquired from the two sensors. An example of this technique is shown in fig. 5.

By using the acoustic emission check valve monitor- ing method demonstrated by Duke Power Company, it appears likely that the following check valve opera- tional conditions can be determined: - Valve rapid opening (backseat impact) - Valve disc tapping during reduced flow - Hinge arm tapping during reduced flow - Valve rapid closing (seat impact). Although a fully open check valve could be assumed by the existence of flow noise without the presence of tapping, the absence of detectable tapping noise is itself no guarantee that the check valve is fully open since the valve disc may be oscillating without tapping in mid-stroke or have fallen off and be in a position that prevents it from impacting the valve body at any location.

Several tests were carried out by Duke Power Com- pany on an 8-inch check valve in new condition and with simulated degradation. Hinge pin diameters and disc/hinge arm clearances were both varied during valve cycle tests that generated acoustic emission signa- tures during opening and closing.

cn ..,~

ud

/-, ~A ] ,

I

I

tX \~^ ;.I ~,IL~ \

- i

d,t = 134 ~ TIME

TAPPING SENSOR RESPONDING LOCATION TO PRESSURE WAVE FIRST

HINGE PIN THE ONE NEAREST THE HINGE PIN BACKSTOP THE ONE NEAREST THE BACKSTOP

Fig. 5. Time-of-arrival technique. Taken from: W.M. Suslick, H.F. Parker, B.A. McDermott. Acoustic Emission Monitoring of Check Valve Performance. Presented at the EPRI Power

Plant Valves Symposium, October 11-12, 1988.

Ir tg l

O

UNSTABLE OPERATION

.tL _ _4 L , t r ~ ] I r r T - . ~

~2 u~ o_ (n

8

STABLE OPERATION I

TIME

I

i t Fig. 4. Acoustic waveforms for a check valve during unstable (tapping) and stable (flow noise only) operations. Taken from: W.M. Suslick, H.F. Parker, B.A. McDermott. Acoustic Emis- sion Monitoring of Check Valve Performance. Presented at the EPRI Power Plant Valves Symposium, October 11-12,

1988.

Valve closures with new and artificially worn hinge pins are illustrated in fig. 6 and show that, with the worn hinge pins, an acoustic transient preceded the seat impact. This transient may result from impact between the hinge pin and hinge arm surfaces as a result of the increased clearance between these two parts.

A similar transient event occurred as a result of increased clearance between the disc stud and hinge arm as illustrated in fig. 7. Also shown is a closure of a check valve having both a worn hinge pin and a loose disc/hinge arm connection.

3.3. Qualitative leak detection

Acoustic emission techniques have long been used to detect fluid leaking through a valve. Philadelphia Electric Company (PECO) has been utilizing acoustic techniques to detect valve leakage in their nuclear power plants since 1974 [5]. Their test procedure con- sists of acquiring two sets of valve acoustic emission readings, one while unpressurized and one with a pres- sure difference across the (closed) disc. The acoustic

H.D. Haynes / Evaluation of check valve monitoring methods 287

noise associated with a leaking valve is then deter- mined based on the difference in readings.

PECO has had good success with a portable, bat- tery-powered data acquisition unit for leakage monitor- ing. The acoustic data collected from baseline (unpres- surized) and pressurized tests are downloaded into a computer for analysis, trending, and archiving.

3.4. Specialized acoustic leak detection equipment

Leak Detection Services, Inc., of Annapolis, Mary- land, has developed an acoustic valve leak detector for use aboard U.S. Navy submarines which has also been used to detect internal vane leakage at several com- mercial nuclear and fossil power plants. The device permits the operator to observe the acoustic emission signals on a meter and to record them on an X -Y plotter. The device provides capability for acquiring

u'J o k.

8 O

co (/)

Z f,/)

t- ,0

,(

i

[- i

NEW HINGE PIN

l i , ltl Iillh.._:: tlHIf"i -- 7,11~l P"'I I

Iiil, til-ll, ........ t..,...lillllii i i l l '~,_ _ _ le i, ' iT' ' :

t,=,,.~LI ~ . . . . . . . t i l l l iHi i , I

!~,,.,_tlltill[lN ........ ,. il,.ltl'l - " ~ ~,, TI'- 111 Hi111'~72

' t l l f11" / .... t'~1' /

TIME (20 ms/div)

O.20-1NCH UNOERSIZED HINGE PIN

0.40-INCH UNDERSIZED HINGE PIN

Fig. 6. Check valve closures with new and artiticially worn hinge pins. Taken from: W.M. Suslick, H.F. Parker, B.A. McDermott. Acoustic Emission Monitoring of Check Valve Performance. Presented at the EPRI Power Plant Valves Sym-

posium, October 11-12, 1988.

ACOUSTIC EMISSION WAVEFORM OF SWING CHECK VALVE CLOSURE WITH WORN DISC STUO CONNECTION

' I I,, I I I I ~ , ,. ~ '

' i I Imml l i iM l | I I ._ t , . ' l a ln .~- , I I

Ir IBBHIB I I IBU = . . . . I I l lmBl l lmt I I

8 , t ? " 'T"F I I * , a

TIME

ACOUSTIC EMISSION WAVEFORM OF SWING CHECK VALVE CLOSURE WITH WORN HINGE PIN AND DISC ARM CONNECTION

z l

9. l i_d ~l.. J l i l i / i _. i I T " IF r i IBT" I I

' H i l l ,,= |?,r

TIME

Fig. 7. Check valve closures with two artificial degradations. Taken from: W.M. Suslick, H.F. Parker, B.A. McDermott. Acoustic Emission Monitoring of Check Valve Performance. Presented at the EPRI Power Plant Valves Symposium, Octo-

ber 11-12, 1988.

acoustic signals from two sensors simultaneously, one sensor mounted on the valve and the other mounted on the pipe about 10 pipe diameters away from the valve [6]. The two channel responses are then adjusted and the background noise signal (acquired by the pipe-mounted sensor) is electronically subtracted from the valve-mounted signal. A positive difference signa- ture is a qualitative indication of a leaking valve.

Other vendors that supply specialized acoustic emis- sion monitoring equipment for valve leak detection include Canus Corporation and TAPSCO, Inc.

4. U l t rason ic inspect ion

4.1. Basic principles

Ultrasonic inspection involves the introduction of high-frequency sound waves into a part being exam- ined and an analysis of the characteristics of the re- flected beam. Presently a device (CHECKMATE TM) is commercially available from Henze-Movats, Inc., that can determine disc position of a swing check valve using an ultrasonic inspection method.

288 H.D. Haynes / Eualuation of check ualue monitoring methods

t'.

i H3ZA7VNV ]VNDIS ~SBA~INV4~t'~3N30~I 03SVB-El31nd~lO0 "IVNgI8 01NOSYI:II7n

u3onasNvul C)tNOSVU17n

Fig. 8. A simplified depiction of the CHECKMATE TM sys- tem.

Figure 8 provides a simplified drawing that illus- trates the basic operation of the CHECKMATE TM system. One ultrasonic transducer is used (pulse-echo type) that provides both transmission and receiving (sensing) capabilities [7]. The ultrasonic signal passes through the valve body where it is reflected by an internal part (e.g., disc, hinge arm, etc.) back towards the transducer. By knowing the time required for trans- mission of the ultrasonic signal from and back to the transducer, the transducer location, and other valve geometries, the instantaneous disc position may be determined.

CHECKMATE TM utilizes signal processing cir-

cuitry that filters out undesirable ultrasonic signal re- flections present in the raw received signal so that the resultant processed signal provides a more easily inter- preted valve disc position signature.

4.2. Detection of calue disc mo~'ement

Figure 9 shows ultrasonic signatures taken from a swing check valve installed in a flow loop at two disc positions: full open and partially open. It is noted that the disc was fluttering (unstable) in the partial opcn position and the flutter was clearly detected using the ultrasonic method.

Figure 10 illustrates the similarity between disc mo- tion signatures acquired with the ultrasonic sensor and with a specially installed rotary variable differential transformer (RVDT) attached directly to the hinge pin to provide a direct measurement of disc position.

4.3. Diagnostic capabilities

Ultrasonics can be used to detect the following operational modes

Operational mode

Full open or full closed Free flutter Backstop tapping

Seat tapping

Signature characteristic

Steady signal Smooth cyclic signal Similar to free flutter, but with flattened upper peaks Similar to free flutter, but with flattened lower peaks

0.21

0.00

~ -o.21

-0.41

-0.62

CHECKMATE V2~T6

DISC AT STABLE POSITION

0.91 DISC AT UNSTABLE POSITION

0.60 i CHECKMATE

/\ >~ 0.30 / f~, o.oo j.,

-0.3C I ~ I i I i i ~ i , 0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0

TIME (s)

Fig. 9. CHECKMATE TM signatures during stable and unstable check valve operations. Used with the permission of HENZE- MOVATS, Inc.

H.D, Haynes / Evaluation of check valve monitoring methods 289

0.16

0.00

-0.16

-0.33'

-0.49

0.21

~" o.o0 1,- ,_,1 ~ -0.21

-0.41

-0.6

CHECKMATE V2W2

CHECKMATE VS. RVDT ON HINGE PIN

RVDT r~ ' ,/d V2W1

V I I I I I I I I I I

0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 TIME (s)

Fig. 10. Comparison of CHECKMATE TM signature with that from a RVDT installed on the hinge pin of a check valve. Used with the permission of HENZE-MOVATS, Inc.

In addition to disc position indication, ultrasonic signa- tures can be used to detect missing and stuck discs, loose hinge arm/disc connections, and worn hinge pins. The CHECKMATE TM data analysis program can also provide estimates of hinge pin wear rates and fatigue damage of valve internal parts.

5. Magnet ic f lux mon i to r ing

5.1. Basic principles

Research carried out by ORNL as part of the NPAR phase 2 study of check valves has led to the identification of another new check valve diagnostic technique, Magnetic Flux Signature Analysis (MFSA). MFSA is based on correlating the magnetic field strength variations monitored on the outside of a check valve with the position of a permanent magnet placed on a moving part inside the check valve (fig. 11).

Proof-of-principle tests have utilized a Hall-effect gaussmeter probe outside of the check valve to detect the magnitude of the magnetic field produced by a small cylindrical or rectangular (bar) permanent mag- net attached to the hinge arm. The Hall-effect probe detects both stationary and varying magnetic fields, and thus continuously monitors both the instantaneous position and the motion of the check valve disc.

5.2. Detection of valve disc movement

MFSA provides the ability to monitor disc position through an entire valve stroke using one externally mounted sensor. A comparison of disc position mea- sured mechanically (by an angular displacement trans- ducer attached to the hinge pin) with that obtained by MFSA is shown in fig. 12 for a 3-inch swing check valve

~ GAUSSMETER

, , J

Fig. 11. Magnetic flux signature analysis (MFSA) principle of operation.

290 H.D. Haynes / Et;aluation of check valve monitoring methods

70"

Z 60, Q i::: 50- o3

N lO.

0 '

-10

3-INCH SWING C, HEC, K VALVE - DISC MOVED MANUALLY

f 0.24'

0.21.

~N o.18-

~ o.12 ~ o.o,. ~ 0.06.

0.03-

0 .~T

-0.03 0 5 10 15 20 25 30 35 40

TIME (s)

1 MEASURED MEO.IANICALLY (N~CCAR DISPLACEMENT TRANSOUCER) = GAUSSMETER PROSE INSTALLEO ON VALVE

Fig. ]2. Comparison of magnetic field strength and disc angu- lar position measurements.

whose disc was moved manually. MFSA has been ap- plied to several swing check valves ranging in size from 2- to 10-inches.

0.5

0.4

0.3

0.2

0.1

0.0

-0.1

0.5

04

0.3

0.2

0.1'

0.0

-0.1

r . . . . FLUTTER MAGNITUDE [

t

PUMP STARTED - LOW FLOW RATE (INSUFFICIENT TO OPEN VALVE FULLY)

DISC ON BACKSTOP

PUMP STARTED - HIGH FLOW RATE (OPENS VALVE FULLY)

TIME (2 SECONDS/DIVISION)

Fig. 13. Use of MFSA to detect disc instability (flutter).

MFSA also provides indication of disc flutter. This was demonstrated by tests carried out by ORNL on a 2-inch swing check valve that was installed in a water flow loop. The acquired magnetic flux signatures (see fig. 13) showed that at a low flow rate (insufficient to open the valve fully), the disc fluttered considerably in mid stroke, whereas at a higher flow rate, the same valve achieved a fully open and stable condition.

0.70

0.60

0.50 n- UJ

0,40

0.30 I ),..

0.20

0.10

i

0.001,, 0.00

SMALL PIN INSTALLED-~ .','1'~, . '' (MORE SCATTER DUE ~,, .~ : " ; , " TO INCREASED HINGE ~.';;.~...~'" VALEVNE

ARM ROCKING~. : O

.. ' " , '~ I~ NORMAL P

. , , . . . . . . L . . . . . . . . . i . . . . . . . . . | . . . . . . . . . 0.10 0.20 0.30 0.40

X GAUSSMETER

X GA!.JSSIqE]ER ]

Fig. 14. Detecting worn hinge pins using MFSA.

H.D. Haynes / Evaluation of check valve monitoring methods 291

5.3. Detection of worn hinge pins

Experiments carried out at ORNL have shown that MFSA techniques can be used to detect hinge pin wear. Figure 14 illustrates a technique for detecting worn hinge pins that makes use of two Hall-effect gaussmeter probes, mounted so that each probe pro- vides an independent measurement of instantaneous hinge arm position. When both probes are mounted on the valve cap at locations equidistant from and perpen- dicular to the projected hinge arm travel line, both gaussmeters should provide identical signatures when the hinge arm moves in a purely swinging motion as the valve opens and closes.

In addition to swinging, the hinge arm moves in a side-to-side rocking motion as well, as a result of flow turbulence and the clearances between the hinge pin and hinge arm. As this clearance increases (e.g., due to hinge pin wear), the propensity to rock increases. Thus, the increase in hinge arm rocking is detected as in- creased deviations from the single line (pure swinging) relationship between the probe output signals as shown in fig. 14.

6. Comparison of check valve monitoring methods

The preceding sections of this paper have provided descriptions of three check valve monitoring methods

that are useful in determining check valve position, motion, and leak rate. These methods, based on acous- tic emission, ultrasonic inspection, and magnetic flux monitoring, function according to different principles of operation and thus provide different (and comple- mentary) diagnostic information. At present, the esti- mated capability of each monitoring method to detect various check valve operational conditions is given in table 1. The methods are rated according to the follow- ing scale: P = poor or none, F = fair, G = good, E = excellent.

6.1. Combined diagnostic techniques

As indicated in table 1, while no single technique has the capability to detect all check valve operational conditions well, a combination of acoustic emission with either ultrasonic inspection or magnetic flux sig- nature analysis can yield a monitoring system that succeeds in providing sensitivity to detect all major check valve operating conditions. Both acoustic/ ultrasonic and acoustic/magnetic combinations have been tested.

The result of using a combination of an acoustic and an ultrasonic sensor is seen in figs. 15 and 16 [8] which compares data from the two sensors obtained for a check valve tapping its backstop and its seat respec- tively. In both tapping modes, the acoustic signature detected the tapping but not its location. The ultra-

-0.71

BACKSTOP TAPPING

""r I ~4.89

"%.0 o., o'.. ,:, ,'.o TIME Is)

BACKSTOP TAPPING

21, 218

BACKSTOP TAPPING

I I

3.6 4.0

Fig. 15. CHECKMATE TM and acoustic signatures for a check valve undergoing backstop tapping. Used with the permission of HENZE-MOVATS, Inc.

292 H.D. Haynes / Ecaluation of check calce monitoring methods

Table 1 Diagnostic capability of three check valve monitoring methods

Check valve Acoustic Ultrasonic Magnetic flux operational emission inspection sig. analysis condition

Full open P E E Mid position

fluttering F- E E Tapping

detecting E G G locating F E E

Leakage E P P

sonic signature did not unambiguously detect the tap- ping, but, in conjunction with the acoustic signature, identified its location.

The combination of acoustic emission and magnetic flux monitoring also provides a useful monitoring sys- tem for check valves. This is illustrated by fig. 17, an example of test data acquired by ORNL on a check valve whose disc was moved manually to simulate disc fluttering at different portions of the stroke.

Recent developments in commercially-available check valve monitoring systems include the addition of acoustic emission to the Henze-Movats CHECK-

DISC MOVED MANUALLY

A VALVE OPENED SLOWLY B MID-STROKE FLUTTERING 0'81CI (HINGE ARM ROTATING

ON HINGE PiN) D 0.6 VALVE CLOSED SLOWLY

TAPPING ON SEAT T' P,.GONB'=STOP t HINGE ARM ROCKING ON ~g 0.2] .,NGEP,N :'CO ~ 0.0

E

,I Jtl.l

-0.4

-0.6 1 ACOUSTIC SENSOR MOUNTED ON VA VE CAP

-0.8 I , =

0.4 E

OPEN

0.2 F QCO .Jw

-> c D u.p- O.1 ~5 k-r.U ~ o o

TIME (s)

Fig. 17. Magnetic flux and acoustic signatures for a check valve under several simulated operational conditions.

CHECK MATE 0.38 r V2~T$

~ -0.75

-1.13

SEAT TAPPING SEAT TAPPING SEAT TAPPING

12.53 I

Z] 0.0

/ 0.4 0 a 1.2 1.6 2 0 2.4 2 a 3 2 3 6 4.0

"nl~ (s)

Fig. 16. CHECKMATE TM and acoustic signatures for a check valve undergoing seat tapping. Used with the permission of HENZE-MOVATS, Inc.

H.D. Haynes / Evaluation of check valve monitoring methods 293

DUAL SENSOR

ASE- -~; ~CONDrnONINq I COMPUTER i ' - . . . . . . . . . "1 I BASED / DIGITAL / I S GNAL

AUDIO - - - B / TAPE I----'~ ANALYZER

I R~CRDER 1 I

, ! ' ,

r l -

Fig. 18. Simplified depiction of the QUICKCHECK TM sys tern.

and Canus Corporation - acoustic emission. These tests, which began in late January, 1990 and are sched- uled to be completed in April, 1990, are being directed by the Nuclear Industry Check Valve Group (NIC) and are being carried out at the Utah Water Research Laboratory located on the Utah State University cam- pus.

Approximately 12 check valves will be utilized in the tests which will include the use of hinge pins and disc studs in new condition and with simulated degradation, upstream flow disturbances, and reduced flow rates. This test program, as well as other activities sponsored by NIC, will likely make a significant contribution towards demonstrating diagnostic capabilities for de- tecting and monitoring check valve degradations in nuclear power plants.

8. Conclusions

MATE TM system and a combination acoustic/mag- netic check valve monitoring system from Liberty Technology Center, Inc. called QUICKCHECK TM [9].

QUICKCHECK TM, depicted in simple form in fig. 18, utilizes a combined acoustic/magnetic Dual Sensor to monitor simultaneously the structure-borne acous- tics that result from flow noise and internal part im- pacts, and the position and motion of an encapsulated magnet that is permanently installed on a check valve internal part (e.g., hinge arm, disc, etc.). Data acquisi- tion hardware includes the Dual Sensor(s), signal con- ditioning electronics and a digital audio tape recorder. Recorded signals are then processed, displayed, and analyzed with a computer-based system that provides detailed analysis capabilities for both acoustic and magnetic signals.

7. Nuclear industry tests of check valve monitoring methods

Three commercial suppliers of check valve monitor- ing equipment are participating in a comprehensive series of tests designed to evaluate the capability of each monitoring technology to detect the position, mo- tion, and wear of check valve internals (e.g., disc, hinge arm, etc.), and valve seat leakage. Those vendors are: Henze-Movats - ultrasonics/acoustic emission, Lib- erty Technology Center - magnetic/acoustic emission,

This paper has presented a description of three check valve monitoring methods: acoustic emission, ultrasonic inspection and magnetic flux signature anal- ysis. These methods were shown to be useful in deter- mining check valve condition (e.g., disc position, disc motion, and seat leakage), although none of the meth- ods were, by themselves, successful in monitoring all three condition indicators. However, the combination of acoustic emission with either ultrasonic or magnetic flux monitoring yields a monitoring system that suc- ceeds in providing the sensitivity to detect all major check valve operating conditions. All three methods are still under development and all should improve as a result of further testing and evaluation. The test program being carried out by NIC and EPRI, should lead to greatly improved understanding of the capabil- ity of these non-intrusive diagnostic techniques in de- tection of degradation in check valves in-situ, and should result in increased reliability of these important components of nuclear plants.

References

[1] EPRI NP-5479, Application Guidelines for Check Valves in Nuclear Power Plants (January 1988).

[2] United States Nuclear Regulatory Commission, Nuclear Plant Aging Research (NPAR) Program Plan, NUREG- 1144, revision 1 (September 1987).

[3] W.M. Suslick, Proposed technique for monitoring check valve performance, Presented at the INPO Check Valve Technical Workshop, October 30-31, 1986.

294 H.D. Haynes / Evaluation of check vah;e monitoring methods

[4] W.M. Suslick, H.F. Parker, B.A. McDermott, Acoustic emission monitoring of check valve performance, Pre- sented at the EPRI Power Plant Valves Symposium, Octo- ber 11-12, 1988, Charlotte, NC.

[5] J.W. McElroy, Light water reactor valve performance sur- veys utilizing acoustic techniques, Presented at the EPRI Power Plant Valves Symposium, August 25-26, 1987, Kansas City, MO.

[6] J.G. Dimmick, J.M. Cobb. Ultrasonic leak detection cuts valve maintenance costs, Power Engineering (August 1986).

[7] Letter from J.N. Nadeau, Henze-Movats, to H.D. Haynes, dated March 1, 1990.

[8] Private conversation between H.D. Haynes and Henze- Movats, Inc.

[9] Letter from D. Manin, Liberty Technology Center, to H.D. Haynes, dated March 1, 1990.