ML20114C053

Throttled Valve Cavitation & Erosion

ML20114C053
Person / Time
Site:	Brunswick
Issue date:	12/31/1991
From:	Casada D OAK RIDGE NATIONAL LABORATORY
To:
Shared Package
ML20114C041	List: ML20114C053
References
ORNL-NRC-LTR-91, ORNL-NRC-LTR-91-25, NUDOCS 9209010129
Download: ML20114C053 (21)

Text

. .. . . . . - - .. . - - _ _ . .. . -. . .

e ORNL/MiC/L't R 91/25 Throttled Valve Cavitation and Erosion Don Casada Nuclear Plant Aging l'rograra Oak Ridge National Laboratory Oak Ridge, Tennessee December, 1991 NRC Technical Monitor: Bill Farmer L

i e

_I

gro!8as3583)!E4 P

'J- ,

4 r

--ORNL/NRC/LTR 91/25 i

. Throttled Valve Cavitation-and Erosion 4

4 Don Casada-

?'

4 Nuclear Plant Aging Program =

. Oak-Ridge National Laboratory 4

- Oak Ridge,' Tennessee December, -~ 1991 i

i NRC Technical Monitor: Bill Farmer i

s

(

1

. -9 4.

P ,

7

ORNL/NRC/LTR 91/25 4

.h -}

i e,

i "

1 .

. E j.*

~

Throttled Valve Cavitation and Erosion i

i .- .

f d

h

! Don Casada '

Nuclear Plant Aging Program Oak Ridge National Laboratory -

Oak Ridge, Tennessee

December 13,1991 -

l l:

) Research gonsored by the Office of Nuclear Regulatory Commission under Interagency Agreement DOE

1886-8082-8B with the U.S. Department of Energy under contract No. DE.AC05-840R21400 with the

! Martin Marietta Energy Systems, Inc.

i I

i i- The submitted manuscript has been authored by a contractor of the U.S. Government under contract No, i _- DE-AC05-840R21400. Accordingly, the U. S. Government retains a nonexclusive royalty free license to

- publish or reproduce the published form of the conuibution, or allow others to do so, for U.S. Government

,_ puposes.

I r

r.

i-d i -

. . . _ . _ . . _ _. -.__...,.......___._._._...._...__.._....a.2._...__..__;.:...._.,..._...__ . - _ . -

l

1. Background In November of 1988, Brunswick plant maintenance personnel discovered significant localized erosion of the valve body of a Unit 1 Residual Heat Removal (RHR) valve,1- I i

El1-F017B (the maintenance was being performed to repair the valve stem and back seat).1 The F017B valve is a 20-inch Rockwell angle globe valve that has been historically used to throttle RllR flow. Excessive throttling of the valve had resulted in cavitation-induced I crosic jamage to areas immediately downstream of the seat. Subsequent investigation )

indicated that crosion of valve bodies was a generie concern for the other RHR valves used in the same service (F017A on Unit 1 and F017A and F017B on Unit 2).

Brunswick personnel expanded the investi gation to include another set of RHR valves that had also been used in throttling service. The F024A and F024B valves (from both Units),

which are used for Suppression Pool Cooling, were also found to have been damaged by cavitation erosion. These valves are 16-inch Anchor Darling globe valves.

Testing of the F017 valves indicated that cavitation was most prevalent in lower flow ranges, but existed throughout the range of 4,000 to 16,000 gpm. Testing of the F024 valves indicated cavitation present thmughout the range of 4,500 to 15,000 gpm, and was most prevalent at higher flow rates. It was also noted that the location of the cavitation moved throughout the body as flow changed.

Subsequent investigauon of seven other valves used in safety-related throttling service, including Core Spray (CS), High Pressure Coolant Injection (HPCI), and Reactor Core Isolation Cooling (RCIC) systems, revealed that one other valve, the HPCI system full flow test isolation valve F008, had experienced notable erosion.

The Nuclear Regulatory Commission (NRC) issued Information Notice 89-01 in response to this event. The Reactor Operations Analysis Branch also performed a review of previous NRC reports and equipment failure data (Licensee Event Report and Nuclear Plant Reliability Data System databases) and issued a technical report documenting the results.2 Because erosion is an aging r: lated concern, the NRC requested that Oak Ridge National -

Laboratory perform an assessment of the significance of valve body erosion, with the principal focus of the review to be the identification of valve :ypes and applications susceptible to erosion pmblems.

2. Valve Erosion Mechanisms Four principal sources of emsion in valves have been identified:3 abrasive particles, high liquid velocity impingement,

. erosion-corrosion, and cavitation.

1

Depending upon the nature of the service conditions, one or more of these erosive sources may exist in nuclear plant valves.

Abrasive particles are primarily a concern in service and other raw-water systems.

Abrasion is not viewed to be a mejor source of concern relative to wall thinning, although some plants have noted increased valve seat and disc wear problems due to high sediment content of the water.

Iligh liquid velocity impingement and erosive-corrosive wear occurs when high velocity fluid impinges on valve or pipe surfaces. In the case of crosion-corrosion, the protective corrosion layer that naturally forms is continually removed as it is formed. These phenomena have been largely responsible for pipe and valve body failures that have occurred in secondary plant (steam and feedwater) systems.

The last erosion mecharnsm, cavitanon, is the principal factor of concern relative to the wear of valve bodies and downstream piping in conjunction with thmttled valve service.

Cavitation is a two stage phenomenon involving:

(1) flashing of the liquid due to the pressure of the liquid dmpping b: low the saturation pressure, and (2) subsequent collapse of the vapor back into liquid due to pressure recovery.

In a valve (see Figure 1), the pressure is reduced as the fluid passes through the minimum flow area (and thus the region of highest velocity), which is typically near the seat. If the pressure reduction is sufficient, saturation conditions can be reached, even for relatively low temperature service conditions (<100 F), and at least a wrtion of the fluid is vaporized. As the fluid exits the seat area, the flow area is rapidly restored, and as a result, a portion of the pressure drop is recovered. If the downstream pressure is greater than the saturation pressure, the vapor pockets are rapidly collapsed bnck to the liquid phase.

v A jL _ _

n _______.

a ______

P2 - P2 PV

... . .. ... .... P, ......... . ... .....

PVC Pvc Figure 1. Pressure Recovery Profiles for Globe and Butterfly Valves 2

_= . -_ . . . - - - _ _ _ .

The pressure profile shown in Figure 1 indicates that the fluid pressure, initially at an upstream pressure of P1, drops in pressure to Pye as it passes through the valve's vena contracta. Pye is greater than the fluid vapor pressure (Pv)in the case of the globe valve (on left), while Pye for the butterfly valve (on right) is less than the vapor pressure,- '

resulting in fluid vaporization. As the fluid passes on through the butterfly valve until the effective flow area is larger, its pressure is recovered to P2, which is greater than the vapor pressure, resulting in collapse of the vapor back to liquid phase. Note that the recovered pressure for the butterfly valve, that is P2- Pye, is greater than for the globe valve.

13utterfly valves, along with conventional ball valves, are commonly referred to as "high pressure recovery valves" due to the fact that the pressure recovered is relatively high (compared to, for example, globe valves).

It is primarily in the collapse of the vapor that damage is incurred. Localized pressures as high as 100,000 psi have been reported in conjunction with cavitation.3 It is theorized that when the bubbles collapse near a metal surface, such as the valve body wall, a

" hammering" of the metal occurs which locally fatigues the material. Other explanations have been proposed; however, regardless of the mechanism, the fact remains that local cavitation removes metal.

3, Cavitation Results As noted above, cavitation can result in erosion of valve parts, including the valve body, and for some valve designs, such as butterfly valves, the erosion can occur in downstream piping. The location of cavitation damage is dependent upon a variety of factors, including valve geometry and flow conditions.

Incipient cavitation is characterized by a relatively low-level " hissing" sound. Fully developed cavitation is much louder and sounds lil.e marbles or gravel flowing thmugh the system. The cavitation may also result in considerable system vibration, depending upon cavitation l< vel, how well the piping is supported, and other factors. System vibration effects may be manifested in other active components, such as other system valves or even the system pump (s).4 The vibration levels encountered under severe cavitation have been high enough to result in a variety of vibration-induced failures, includmg drain / vent line cracking, valve leak-off tubing failure, and loosening oflimit switches, packing followers and other valve components.

Cavitation erosion damsge is characterized by a very rough, sometimes pock-marked appearance (as opposed to the smooth wear patterns often associated with high velocity wear). The damage may be localized or may cover a relatively broad area (a few square inches).

At Bruntwick, the crosion was found immediately downstream of the valve seats. For the angle globe valves (F017 set), the erosion occuned in two regions on either side of the center guide just above the seat (flow comes from under the plug, ard makes a 90 turn above the seat). In the case of the straight globe valves (F024 set), the erosion was again just above the seat (flow comes from under the seat); however, it occurred in several regions around the full circumference of the seat.

3 l - - - - -

Erosion damage associated with cavitation from throttled butterfly valves may be -

manifested in the piping downstream of the valve,instead of the valve body itself, due to the minimal axial length of butterfly valves. Erosion damage associated with pump

. recirculation valves (particularly for high energy pumps, such as main feedwater pumps) may occur farther downstream (the result of high velocity impingement in some cases).

Erosion can eventually result in through-wall failures. These failures can progress rapidly n if a system transient results in a sudden hydraulic or mechanical load (internal pressure spike, foi example). Alternatively, the failures may show up as pin-hole leaks that can be corrected before the consequences become more significant.

4. Valve Characteristics and Cavitation Susceptibility The classical pressure drop and flow relationship for single-phase, non-compressible liquid slow through valves is expressed as:

a q=C,.g P (1) where q = Fluid flow rate C, = Valve flow coefficient AP = Valve pressure drop gr = Fluid specific gravity Figure 2 provides a typical relationship between valve flow and pressure drop for constant vapor pressure and upstream pressure conditions.

2 P K= AP p (2)

C P, - P

[DN* 3 -P

~'. .....:

l vc l <-Fully Choked -+

a c Cavitation >< - Flashing-+

P>P 2 y P2 sPy

/

where:

P3 = Valve upstream pressure P, = Fluid vapor pressure P2 = Valve downstream pressure Ke = Cavitation index Pyc= Pressure at valve vena contracta Ft= Valve pressure recovery factor AP = Valve pressure drop Km= Valve recovery coefficient = F g Figure 2. Flow Rate vs. Pressure Drop 4

1 As is the case for other fluid system components, the flow through valves is generally proportional to the square root of the pressure drop, This relationship is valid so long as l the fluid is not vaporized. However, as the pressure drops below the vapor pressure, l

, departure from the relationship occurs. As can be seen, the departure from the square root  !

proportionality occurs at point a of Figure 2. The dimensionless parameter Kc, called the-cavitation index, which is calculated at this point, has been used to indicate this point of departure It provides a somewhat limited indication of the point of cavitation initiation; measurements that have been made using acoustical and vibration instrumentation have indicated that cavitation begins at pressure drops lower than that associated with the valve coefficient departure from the characteristically linear relationship of flow and 6P.

A characteristic of importance when considering the susceptibility of a valve to cavitation is the valve pressure recovery factor,5 FL (alternatively may be expressed as Km which equals the square of FL), which accounts for the influence of the valve geometry on its capacity at choked flow conditions, and is calculated at the intersection of the extension of the constant flow line corresponding to choked flow for the given upstream pressure and the extension of the square root proportionality curve (see Figure 2). Note that F.t can be calculated based on readily measurable parameters, except for the vena contracta pressum, Pyc.

Because the vena contracta pressure is not readily measurable, a relationship between the vapor pressure, the fluid critical pressure, and the vena contracta pressure under choked flow conditions has been derived, and can be expressed as: 5 Pye = FrPy (3) where:

Fr = Fluid critical pressure ratio factor FFcan be predicted by the following:5 F= F 0.96 - 0.28 b (4)

Pc where:

Py = Fluid vapor pressure Pc = Fluid critical pressure  :

Equations 2 and 3 can be combined and rearranged to the following expression:

(5)

APm= F[(Pl - FrPv)*

where:

• Note that for most practical applications involving water, Fr is equal to 0.96, since the vapor pressure is I negligible compared to the critical pressure.

)

5 l

l

APm= The maximum AP for which increased flow results for a given upstream pressure (i.e., corresponds to the point at which flow becomes choked)

, Equation 5 is referred to as the Choked Flow Equation. It has been used to determine allowable pressure drops for valves to avoid cavitation damage. The BWR Owner's Group has established a criterion of valve AP of < 0.8

• APm as the fira cut its a set of evaluation criteria for valves used in throttled service, if this criterion is no: met further monitoring to detect the presence and extent of cavi:ation is required. Based on conversations with utility personnel, this has proved to be a reliable criterion for globe valves, ahhough it was noted that cavitation conditions can exist even if the triterion is met.

Typical valve pressure recovery factors are:5 Valve type FL Single port globe 0.9 Contoured plug angle 0.9 Segmented ball 0.6 90-degree offset seat butter 0y 0.6 P is important to note that high pressure recovery valves have low pressure recovery iactors. It can be seen from equation 5 and the above pressure recovery factors that globe and angle valves can be operated with greater pressure drops than can ball or butterfly valves. More specific information on allowable pressure drops for different valve types is pmvided below in Section 5.

5. General Valve Type and Material Considerations 5.1 Valve Types It was noted in Section 2 that some valves, such as butterfly and ball valves, are refeind to as "high pressure recovery valves"due to the fact that a significant portion of the pressure drop that occurs from upstream of the valve to the valve's vena contracta is recovered in the discharge.

As can be seen from Equation 5, the greater the pressure recovery factor Ft, the greater is the maximum AP. To illustrate the effect that the type of valve has on the maximum allowable pressure drop for the BWR Owner's Group guideline of 0.8

• APm, consider the ,

results provided in Table 1, which are based on representative valves from a commercial I valve supplier's ulve sizing program.6 Conditions that are typical of valve applications  ;

for emergency service water (ESW) and RHR valves are used. '

6 l

Table 1. Comparison of Maximum Pressure Drops for Valve Types System , Valve Type  % Open APm AP.et /APm ESW Globe 69 73 0.68 Butterfly 19 57 0.88 Full-Bore Ball 29 65 0.77 Qtrim Ball 43 74 0.67 RHR Case 1 Globe 66 227 0.66 Butterfly 18 178 0.84 Full-Bore Ball 25 201 0.75 Qtrim Ball 39 232 0.65 RHR Case 2 Globe 71 225 0.67 Buttedly 21 175 0.86 Full-Bore Ball 31 199 0.75 Qtrim Ball 46 227 0.66

• AP.ci = Actual AP The assumed conditions for Table 1 are:

Application q, gpm d,in. P i, psia P2. Psia T, F ESW 5000 16 80 30 80 RHR Case 1 3000 10 250 100 150 RHR Case 2 10000 16 250 100 -150 It can be seen from Table i that the maximum allowable pressure drops across simple ball and butterfly valves are less than those for globe or Qtrim ball valves (which have a unique valve trim which provides several stages of orificed pressure drop, thus helping avoid cavitation problems). Note that,in the case of the conventional butterfly valve in all three applications in Table 1, the actual pressure drop exceeds that suggested by the BWR Owner's Group as the limit for pitrposes of requiring further investigation.

In order to provide a perspective on the significance of upstream pressure, Figure 3 presents a plot of AP.c:/APm for a spectrum of pressure conditions for four valve types.

The comparative abilities of the valve types to handle the pressure conditions can be readily seen.

7

.= .

1.0 -

-gt- Buucrfly

-A- Globe

, o- Full Bore Ball 0.9 e- Q-Trim Ball BWR Owds Group First Cut Criterion d

g . 0.8 - -- - - --- - - ------ ---

0.7 !

Assumed Condiuons:

5000 gpm

(

. P2= 30 psia 0.6 - . T = 80* F l I I I I 60 70 80 90 100 110 Pi (psia)

Figure 3. Fraction of APm for Various 16. inch Valves as a Function of Upstream Pressure 5.2 Valve Materials In general, harder materials are more resistant to cavitation-induced erosion damage.

However, hardness is not a perfect indicator of susceptibil_ity to erosion, and a variety of factors have been reported to affect erosion resistance, including alloy composition and heat treatment.7 Table 2 provides relative cavitation resistances of several materials used as valve plugs under cavitating conditions (the higher the number, the more resistant the-material is to erosion damage). The resistances of Table 2 are referenced to 316 stainless steel.

Table 2. Resistance of Materials to Valve Plug Cavitation Erosion 3 hiatertal Relative Resistance Stellite No. 6 over Type 316 stainless steel 20 316 stainless steel 1.00 Castiron (A126 Class C) 0.75 Carbon steel (WCB) 0.38 Carbon steel (AISI Cl213) 0.17 Brass 0.08

Aluminum. 0.006 8

l

l l

l l

The erosion rates of several materials (relative to type 308 stainless steel) subjected to a cavitating jet are provided in Table 3. Note that for Table 3, the lower the value, the higher I the cavitation erosion resistance. I l

Table 3. Resistance of Materials to Cavitating Jet Erosion 8 l 1

l Matertal Belative Erosion Rate J 316 stainless steel 0.8 l 304 stainless steel 0.7 Stellite 21 weld overlay 0.1 308 stainless steel weld 1.0 Carbon steel 1.6 Aluminum Bmnze weld 3.7 The cavitating jet crosion rate of some intermetallic compounds of nickel aluminide has been reported to be 4 to 10% of that of 308 stainless steel weld material; however, the abrasive wear resistance (using dry sand) of 304 stainless steel has been reported to be superior to that of nickel aluminide.9 Abrasive wear resistance would primarily be a 4

concern for nuclear plant applications only when considering the adoption of crosion-resistant materials ; service water systems which have a heavy silt level. Some iron-based cobalt-free alloys, ueveloped specifically to provide utilities with weld overlay alternatives to cebalt containing alloys for use in nuclear applications (and thereby reduce radiation levels resulting from 60Co), have been shown to have relative cavitating jet erosion rates of 14 to 19% of 308 stainless steel.10 4

6. IIistorical Failure Experience l

Failure data from the Institute of Nuclear Power Operation's Nuclear Plant Reliability Data System (NPRDS) database was reviewed to provide at least a qualitative indication of

, historical erosion and cavitation problems. Two separate sets of failure data were l compiled. The first set was for through-wall failures of valves and pipes in all NPRDS

systems, and the second set was for failun:s in the RIIR system. Discussion of the results follows.

l 6.1 Review of Through-Wali Failures A review of the NPRDS database records of through-wall failures of pipes and valves in which erosion

• was indicated as the cause of the failure was completed. The results of the

• The scarch was conducted by acquiring those failure records for which the narrative desenption included the term crosion (or other words based on the root word erode). Failure records which involved body-to-bormet

! , or packing leaks were manually climinated during the review process. It should be noted that some known e:osion related failures of valves did not appear in this type of scarch. due to the use of a more generic tenn, such as wear, to describe the cause.

9 t

l

1 i

f j

l search and review are provided in Table 4, which indicates the distribution of reported degradation and failure, by system. Italf of the valve failures and almost three fourths of the pipe failures occurred in the Service Water system. Many, but not all, of the failures i, were clearly relateu to valve throttling. Note that this table is not a tabulation of all j cavitation or pipe / valve body crosion related problems from thc NPRDS data, but only a j listing of those which resulted in externalleakage due to body / pipe failure. It should te 1-recognized that the number of failures reported should not be considered to te an accurate quantitative rrpresentation of actual historical experience. In fact, discussions with a utility

, valve expert indicatul that his utility alone hart probably expenenced as many Jtmugh wall j failures as those shown below. Ilowever, the distrihmlon of failures provides a relatively l

accurate indicator of failux distribution among systent valve types, etc.

i l

l Table 4. Through+all Pipe or Valve !!ody Fallures 1

j Sntna Valves Plocs Component Cooling Water ((GW) 0 1 Condensate /Feedwater 4 30 Chemical Volume & Contml 1 2 Service Water (SW) 7 90 i Main Steam 2 2

) Total 14 125 i

4 q Nine of the 14 through-wall valve failures occurred in globe valves. It is important to recognize that erosion induced by the throttling of valves (and panicularly butterfly valves) would often be manifested in the piping downstream of the valve, and not in the valve body 4

itself. Many of the pipe failures recorded in the data were clearly associated with throttled a

valves, although the valve type was not identified. For some of Jg service water system 4

failures, the erosive / corrosive pmperties of the water (n oppossi to, or in addition to valve

, opemting conditions) were also cited as factors, particularly for plants with high silt content j water. Clearly, the systems of primary concem are condensate /feedwater and service water. Almost 80% of the valve failures and over 95% of the pipc failures which resulted i in thmugh-w.ul leakage occuntd in these two systems.

j There were no reported through-wall failures of the fmnt line safety systems, such as high j or low pressure coolant injection / residual heat removal, which were directly attributed to crosion. Probably the most significant factor in the absence of through wall failures in j these systems is the fact that they are typically operated only a small fraction of the time, 6,2 A More Detailed Review of Cavitation Related Valve Failures

in recognition of the fact that all valve cavitation problems would not necessarily result in

through wall erosion, a qualitative review of a portion of all valve failures in the NPRDS database was performed, and discussions with utility personnel were held to identify the 10 1

n w, - - . . - ,..v , ,, .,

m- - ~ . . - w,,- - . , , . . - .., - ,-,-w -w.. .

l valve applications most susceptible to cavitation and erosion problems and the ways in which the problems were manifested. The discussion in Section 7 pmvides information on the valve applications observed to be most often affected. Examples of symptorns of

. throttled valve cavitation are:

Erosion of the valve tuly, seat, obturator, guides, cage Vibration induced cracks in valve packing leak-off hnes or other small-boir lines (such as vent / drain 'ines or instrumentation tubing)

Loosening of valve packing follower bolts, handwheel key set screws, limit switch attachment screws, and other threaded fasteners Excessive fluttering of downstream check valves it was detennined that a review of all the NPRDS failure data for valves to assess the extent to . .:ich these symptoms wcre present would not be feasible because of the extensive numtcr and variety of valve failures in the NPRDS database. As an alternative,it was decided that a more detailed review of all reported valve tailures in a single system would be conducted. The RilR system was selected for three reasons:

it is a critical safety-related system.

the crosion identified at Brunswick was in the RilR system, and

+

it is operated more often than the other standby safety-related systems (though not as much as some of the normally operating safety-related systems such as service water).

6.2.1 RilR System Functions The RHR system at most BWR and PWR plants is designed to be used for a variety of functions. The system is most frequently used in upport of decay heat removal during plant shutdowns. At least one train is normally in service at all times when the plant is in the bot or cold shutdown condition or refueling.

The R11R system is also used at most plants to provide the engineered safety feature function of low pressure injecrimi of water to the reactor in the event of a loss-of-coolant accident (LOCA). It may aise be used for post lJX'A recirculation of containment sump or suppression poolinventory to the reactor, For some plants,it may also be used for other functions, such as containment cooling, suppression pool cooling, etc. Finally, the RHR system is periodically tested at Gow rates ranging from recirculation only (in some cases, this is less than 10% of pump best efficiency point Dow) to Dows rates exceeding the design basis flow rate.

6.2.2 Distribution of Throttling Valve Types in the RilR System A search of the NPRDS database to detennine the extent of use of two commonly used throttling valve typs, globe and butter 0y, iri the RiiR system was conducted. The numbers of these valve types ranging in nominal size from 4 to 40 inches are provided in Figure 4.

11

I-'

j lt is noteworthy that the General I'lectric (GE) plants have more butterfly and globe valves in senice in the RilR syrtem than ai the other three nuclear stearn supoly system (NSSS) l

. plants combined, in fact, there are more globe valves in GE plant RilR systems than globe  :

! plus butterfly valves at the other three NSSS supplied plants combined. Because Babcock l & Wilcox (BW), Combustion Engineering (CE), and Westin;;,eui,c (W) plants all use borated water in the RilR system, most of the RilR valves have either stainless steel l* bodies or air aainless steel lined. Most of the RIIR valves in GE plants are carbon steel.

l

! The material of mannracture is important from an erosion standpoint, since stainless steel is, as discussed in , an $, more resistant to cavitation damage than carbon steel.

] ,

a j

l 500 i  : i 400 h

! i h'

MO '

1  :  ;.

200 - . ^.

: W M

rtN 100 0 * ' --

BW CE GE W D Carboi, Steel Carbon Steel g Stainless Steel Stainless Steel Butterfly Globe Butterfly Globe l Figure 4. Number of Globe and Butterfly Valves Used in the RilR System i

l 6.2.3 Discussion of RIIR System Valve Failures Brunswick is a BWR plant, and the valves identified at Brunswick as having incurred throttling-related emsion damage are routinely operated in a throttled position at other BWR plants as well. Some valves in the RHR system at PWR plants are also routinely thmttled (such as the heat exchanger outlet and/or bypass valves). Since, as noted in Section 6.1 above, there are no recoided through wall failures of valves in the RIIR system, all failures of RiiR valves were reviewed for other manifestations of cavitation damage, such as s sion/ wear related degradation of internal valve parts and problems related to vibration.

Failure reports from the NPRDS data for four-inch and larger valves used in tne RiiR system were individually reviewed for failure symptoms similar to those discussed i

, previously. The results of this review follow, 12 l

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

4 i

j A total of 83 failures of glote or butterfly valves which appeared to be at least panially the result of cavitatio:: were found. Ihe results, by NSSS vendor and method of discovery,

! are provided in Figure 5. The number of reponed valve failures at GE plants is j ,

substantially higher than for the PWR plants. This is not unexpected,in light of the larger i population of valves (see Fig. 3). Also, note that the principle means of failure detection at 1 BWR plants was by leak testing. Since the extent ofleak testing performed at BWR plants j . is typically greater than that at PWR plants, this particular result is also not unexpected.

30

e

~

l o-h 20 - -

fz  ;

! p2 -

f gs  :

g-)

y, f 10 -- .

2 a:

0 = 0 U ' R BW CE GE W-3 Operational Abnomulity 0 Other Testing / Inspection G Leak testing 3 Maintenance O Unknown Figure 5. Cavitation and/or Erosion Related Valve Failures in the RIIR System: Methods of Detevilon '

i The principle failure mechanism identified for these failures is indicated in Figure 6.

. Clearly, crosion and wear are the dominant problems found among the valve failures. It should 'se recognized that some vibration problems might result in failures of other components, such as auxiliary piping, pipe supports, etc., which were not considered in this panicular pan of the review.

50 ; ,

g 40 ~

.d  :

y 30 -

g g 20 5 -

~

5 10 --

.M .. .. . _ _ . . . . . . .

BW CE GE W 3 Erosion / wear 0 Vibration - O other Figure 6. Cavitation and Erosion Related Failures by Mechanism 13 .

l l

l t

...._,.n , ,. . . . , ,--.---,,-..,.-.,n-, ,...,n ,w ---._,...n, ..

nw .,.._,,.,,4._..,.~.,..,_...e

., - . . , , , , --,.wn,., y ,ng,

1 I

, 7. Yalve Application Considerations a

j Most valves used in nuclear power plant systems are nonnally positioned in either the full open or full closed position (i.c, not throttled). In next cases, these valves would not be

] subjected to the erosive mechanisms previously discused. Cenain system conditions,

such as high Duid velocity and/or operation near saturation could result in craitation even ,

for fully open valves. For valves that are normally fully closed, a slight seat leak,  ;

particularly for steam system or operation near saturation, can result in steam cutting that

progressively worsens. This is not uncommon for steam isolation or pressure relief vah es.

l i

There are relatively few valves that are normally operated in the throttled position. Those valves that are throttled are primarily those used for flow, pressure or other control l purposes. Examples of commonly throttled valves are:

Outlet and/or bypass valves for heat exchangers such as:

+

CCW heat exchanger (SW side) i

+ RIIR heat exchanger (on both the RiiR and CCW or SW sides)

Emergeng d!wljacket water cooler a

Containment fan toolers (located outside of containment, and throttled to rnaintain line pressure greater than containment pressure following 4 LOCA

+

Main renerator hydrogen i;as cooler (non safety) j .

Main turbine lube oil cooler (non-safety)

Chemical and volume control system control valves, such as letdown pressure control j valves

• Pump test line valves for systems such as:

• Iligh Pressure Coolant injection

+

law Pressure Coolant injection 4

Core Spray .

Steam generator blowdown flow, pressure, and flash tank level control valves Secondary system level control valves (e.g., for the condenser hotwell and heater drain tank) l Feedwater contml valves There are variations in throttled valve applications from one plant to another that are controlled by such factors as system pump capacity, pipe sizing, and the use of other flow control devices (such as orifices). Some noteworthy areas of variation are:

The valves used in a given application may be either manually or automatically operated. For example, heat exchanger outlet and bypus valves may be automatically temperature controlled at one plant, while the equivalent valves et another plant require manual positioning.

4 '

Valve applications that require throttling at one plant may not require throttling at another. For example, the emergency diesel jacket water cooling outlet valves (normally a service water cooled heat exchanger) are normally throttled at some plants  ;

and full open at others, depending upon overall system balance. '

14

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

. , , . . .-. ,_m.-.r.___ ,._,r.. . . . . , - ,,, .. ., , m,_

i i

l

'Ihe type of valve used foi a particular application varies. For exampie, butterfly valves j are used as control valves for compor.cnt cooling water hrat exchangers at some plants, while ball valves are used at others.

The extent of temperature swings of open systems (such as service water) has a ,

i substantial impact on the amount of throttling that is required. Plants located in milder 4

. climates see less variation in service water temperature, thus requiring less flow adjustment during the course of the year to maintain proper temperature control of

) systems cooled by the service water.

• Design provisions are made at some plants to minimize contml problems, such as the i

j use of par,tilel lines of varying sizes to provide flow control without excessive throttling of any valve thmughout a broad range cf flow rates, while others depend totally upon individual valve throttling.

• The amount of entrained air can have an effect on the severity of cavitation problems.

All else being equal, a system containing water which has teen aerated (for example, a service water system drawing suction from a cooling tower basin) will experience

fewcr cavitation problems than a relatively de aerated system (for example, service l water system drawing its suction from a lake).

' =

Upstream pressure, panicularly for open systems, can play a significant role in senice conditions, as can be deduced from Figure 3.

The extent of throttling required for a given valve can also vary substantially, depending

upon operational demands. For instance, the outlet valves on the CCW (or SW, depending upon plant specific design) side of RilR heat exchangers which require little throttling shonly after shutdown when decay heat levels are high may be substantially throttled weeks (or months, for a pmtracted outage) later when the decay heat level has diminished.

Figure 7 shows the extent to which the CCW side outlet valve on an RilR heat exchanger was found throttled during a visit to a plant that was in a protracted outage. As can be seen, the butterfly valve is throttled to about 10% open (there was significant flow through the valve - over 2000 gpm- when the picture was taken).

l l The system / valve applications that are most likely to experience significant crosion problems are those that are routinely operated under throttled, cavitating conditions. This, in part, helps to explain why there have been significantly more through wall falhtres in the service water system than in other systems. Even though valves in some other systems may sec harsher conditions when the system is operated, such as the RHR valves it Brunswick and other plants, the fact that the system is normally not in operation substantially reduces the effective rate of degradation. Some valves operated under particularly severe conditions, such as feedwater recirculation valves which normally see flow only briefly during plant startup and shutdown, may experience failure in a relatively short period of time (i.e., weeks or months) if operated under continuous flow (during i protracted testing, such as power ascension testing, or inadvertendy, due to seat leakage).

, The same valves would remain intact for several years when used infrequently (which would nonnally be the case).

I 15 i i

\

! i E

i t

i i .

,, my. .

wm,-wn m w,, l s

i 4 J.hr 6 .,

- **4aem$r r +

.%'4 M 4y<.,

9cJ Df'~~'

Ny '

i j

l }' 7'"r 6.?; it

(

, .s .

,, s p. . = . y l

y.

g ..; L;a

.p.,..

y,. y ,

}(,

l 9- n (,. ,

t .

a g) * ., , .

3,Q 2, /

~

l .

...w gewwegr.tegge.. s 2. .

k ' ' #

l . f .,'l $. 1..'.

4- .,

l},

o

, . ' 6 <-

g a # - '

?

i , ,1 d'

.. g_4 - s-f e

i r s'

y , h , <' m 4 r

l

• .. i s/h l l

, t.. Q j *

.$ ,, d [-

..f ,, s_ ,,

', ~1 4 gl } l '

A&qw

~

g ' '

., 7 , l Q y.

j? *

• i:

{6 G .

1

4. g; - t s.

3 .

.g':n a.

..U'- .u

-f ..

i l

Figurc 7. A lleavily Throttled Hutterfly Valve

8. Corrective Actions Substantial cavitation-induced vibration and some erosion of valves used in certain applications has been experienced. As a result, some plants have taken or are taking actions to eliminate or minimize the problem by the use of anti-cavitation or cavitation j control trim. These types of trim are designed toward minimizing cavitation and/or directing the cavitation such that it does not occur in close proximity to the valve material surfaces.

i Cavitation can be minimized by dropping pressure in stages by directing flow through a series of restricted paths. A comparison of the pressare profiles of a multi stage valve and a single-stage, high pressure recovery valve are shown in Figure 8, i

16

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

wg ;-  ? **fy* ** **MY' M %

9{

4 .' f ' f.#' '

• d t* -

Q: , $.Y.:  % _. iff. 3.y-s Q:; ' ,

, wa . . ' y

. . , , , v.

%[ .[ f (, ,. *s f' b g .. . * [i' - [ ~

I. . [4' ~M\

. . .. , m r w - . ,

k:>

' AyI .

3 n;;

s.

• y, p

.Ra _',

l ..

..y

, , -R .

j _

.;f=

.: L.,\

% g,

~

n . yg+ yg ;;"4 p-i

• :.e .

6 ~&

, [91 lbN 'q .q?.

'. .. j. E

{

if "

o =-

' ' !i w. . , $+

v .

V

_ e%s s

e a x_

Figure 7. A IIcavily Throttled ButterHy Valve

8. Corrective Actions Substantial cavitation induced vibration and some crosion of valves used in certain applications has been experienced. As a result, some plants have taken or are taking actions to eliminate or minimize the problem by the use of anti-cavitation or cavitat on control trim. These types of trim are designed towaid minimizing cavitation and/or directing the cavitation such that it does not occur in close proximity to the valve material surfaces.

Cavitation can be minimized by dropping pressure in stages by directing now through a series of restricted paths. A comparison of the pressure profiles of a multi stage valve and a single-stage, high pressure recovery valve are shown in Figure 8.

W 16

l'g .

P2 P

Single stage trim Pi P2 P,________________

Multi-stage trim Figure 8. Pressure Profiles for Single and Multi. stage Trim Valves For the multi stage trim valve, the pressure breakdown occurs in steps, with minimal pressure recovery. As a result, the minimum pressure occurring in the valve is greater than for the single stage design, resulting in less susceptibility to cavitation.

An alternative, or supplemental design feature that can reduce the consequences (if not the severity) of the cavitation is to change the valve trim design such that the cavitation does not occur in the immediate vicinity of the valve body or seat. Such a design may use a trim with niultiple opposing orifices such that the flow fmm one orifice collides with that from the one diametrically opposite it, thereby avoiding impingement on valve metal surfaces.

The use of crosion resistant materials is another design feature that can minimize the erosive effects of cavitation. Many valves originally installed in services that were expected to be rather harsh utilized relatively erosion resistant materials; however some did not. For the latter, use of improved materials may be used to extend valve life, in addition, piping downstream of the valves may be a candidate for replacement with impmved materials.

The installation of smaller bypass lines around valves avoids operation that would otherwise involve extreme throttling under some modes of operation. This system design feature is commonly used in main feedwater and steam systems to previde better control during startup evolutions. Some plants have implemented this type of design (often with an orifice installed in the bypass line to reduce the burden bome by the valve) to alleviate cavitation problems.

Another type of system design change which has been used has been the use of two valves

, in series to distribute the energy dissipation.

17

N One plant contacted had added globe valves downstream of existing butter 0y valves in the containment fan cooler return lines. Previously, the butter 0y valves had been throttled, with signincant cavitation and erosion resulting. The globe valves are now being throttled

, (to rnaintain pressure in the lines greater than hypothetical containment pressure following a LOCA)instead of the butterfly valves, with signincantly reduced or non<xistent cavitation.

In conjunction with the types of design corrective actions noted above, or in some cases as a stand <alone measure, changes in administrative controls (e.g., operating procedures) are made to ensure that valves are not throttled inappropriately in order to minimize valve damage.

9. Summary Based on a review of historical operating data, discussions with utility personnel and on-site observations, it appears that some valves are operated under conditions that art beyond i the intended design use of the valve (for instance valves that are throttled beyond what the vendor recommends for continuous service). The result is that the valves and the associated system reliability is less than desired. The number of reportal through4 wall failures relative io the number of valves used is small, in most cases, the through wall failures have been manifested as pin hole leaks which, from a rate of leakage standpoint, are no more significant than gross packing or bonnet seal ring failures. However, where there has been sufficient crosion to result in a pin hole leak, the general integrity of the valve or piping may be questionable Changes in valve and phe materials can minimize the crosion rate 01 valve trim and bodies and downstream piping. Ilowever, it is important to recognize that emsion is not the only negative result of valve cavitation, and in fact may not be the principrJ concern. Cavitation-induced vibration can have a negative impact on not only the valve, but its operator, adjacent ccmponents and piping supports. Changes to anti-cavitation or cavitation control trim, modincation of system design, and/or implementation of administrative controls can substantially mitigate these consequences of severt duty applications.

18

- . _ __ - . _ _ _ _ . _ _ _ _ _ _ . . ._ _ __ _ . - ___ _._ ~__. -, _

References

1. Letter from J. L llamess to the U. S. Nuclear Regulatory Commission, Dmnswick

, Steam Electric Plant Unit 1 Supplement to Licensee Event Report 188-033, April 28, 1989.

2. Technical Review Repon AEOD/E908, September 29,1989. 4

3. Instrument Society of America Handbook of Control Valves,2nd Edition,1976, J.

W. liutchison, editor. Published by the Instrument Society of Amer:ca.

4. J. Orol end M. Horbaczewski, "Purrp Full. Flow Test \'alve Cavitation Problems and Their L r,vns," EPRI Third Valve Symposium and Equipment Exhibition, pp. 395-429, August 20-23,1990.

5. Instrument Society of America. " Flow Equationsfor Sizing Control Valves," Standard S75.01, 1985.

6 Neles-Jamesbury Nelstre, version 2.05U, February,1991.

7. A. Akhtar, A. S. R10 and D. Kung, " Cavitation Erosion ofStainless Steel, Nickeland Cobalt Alloy Weld Overlay Afaterials," Coatings and Bimetallies for Aggressive Environments: Proceedings, Conference on Coatings and Bimetallies for Energy Systems and Chemical Process Environments, American Society for Metals, pp 125-142,

8. P. A. March," Evaluating the Relative Resistance of Afaterials to Cavitation Erosion: A Comparison ofCavitating Jet Results and Vibratory Results," Pro:ecdings of the 1987 ASME Cavitation and Multiphase Flow Forum, FED Volume 50, American Society of Mechanical Engineers,1987a.

9. M. Johnson, D. E. Mikkola, P. A. March, and R. N. Wright, "The Resistance of NickelandIron Aluminides to Cavitation Erosion and Abrasive Wear," Wcar, V01ume 140, pp. 279 289, November 1990.

10. P. A. March,"A Preliminary Assessment of the Cavitation Erosion Resistancefor EPR//AAfAX's NOREAf Alloys," Report No. WR28-4-900 233, TVA Engineering Laboratory, January 1989.

e 19 l

. _ _ _ . .. .. .- - -. . . - . - - ..