ML20056B498

Responds to NRC Request for Addl Info Re General Relief Request for Pump Vibration Submitted 900315.Relief Request Changed to Insure Data Taken Over Range That Encompasses All Main Potential Noise Contributors

ML20056B498
Person / Time
Site:	Catawba
Issue date:	08/22/1990
From:	Tucker H DUKE POWER CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
NUDOCS 9008280365
Download: ML20056B498 (11)

Text

' 'l {i, g .

Duke heer Cornpany lin 11 h w tice Prendent -

PO Box 33198 Charlotte, N C 28242 Nuclear Production (704)373-4511 DUKE POWER August 22, 1990 U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, DC 20555

Subject:

Catawba Nuclear Station Docket Nos. 50-413 and 50-414 Response to Request for Additianal Information General Relief Request for Pump Vibration This letter is .in response to the NRC Request for Additional Information on the General Relief Request for Pump Vibration submitted by my letter dated March 15, 1990.

Each question in the Request for Additional Information deals with deviations between the Relief Request and the requirements of OM-6. Some confusion has resulted from the wording of the original relief request. It was not Duke Power's intent to adopt OM-6. The intent was to measure vibration velocity, as required by OH-6, because this ptovides a better measure of pump mechanical condition than vibration amplitude. Some parts of the relief request are based on OM-6, and as a result it was referenced when the relief request was originally submitted. The relief request does not specify compliance with OH-6 because there are fundamental problems with certain OM-6 requirements such as: 1) the vibration instrument accuracy statement is not clear or reasonable over the required frequency respense range 2) inaccessibic vibration measurement points are specified, and 3) vibration acceptance criteria are to restrictive for smooth running pumps.

This letter also makes several changes to my original submittal. These changes are:

1) The accuracy specified in tha relief request has been updated to reflect the use of a special National Institute of Standards and Technology (NIST) calibration and Computational Systems Incorporated newest data collector (CSI 2110). Per CSI the closed loop accuracy of the CSI 2110 is:

X=(9+C +K )*

Where: X = Closed loop accuracy (maximum error) in percent reading C = Maximum error of the integrator. The signal is integrated once (converted from Gs to in/sec); therefore the maximum error of the integrator is +/-3%.

K = Maximum error of the sensor. This value is 5% of reading.

9008280365 900822 jo t

PDR ADOCK 05000413 p PDC

'l1 d

f . .

U. S. Nuclear Regulatory Commission August 22, 1990 Page 2 Therefore, an accuracy of +/- 6.56% of " reading" instead of 6.56% of'

" full scale" is specified in the relief request.

2) An editorial change has been made to the table in the relief request that-specifies the vibration ranges for smooth running pumps. The statement in the first column, second row, now reads "For All Pumps When Vr </=

0.075 in/see". Some of this statement was inadvertently omitted in the -

original relief. request.

3) The. relief request has been changed to insure that data is taken over a range that encompasses all of the main potential noise contributors. It now specifies that vibration velocity will be measured over a rango from 1/3 minimum pumi: shaft rotational speed to 1000 Hz.

Attachment I contains my reply to the NRC request for additional information.

Attachment II contains the-amended relief request.

Very truly.yours,

h. .

g H. B. Tucker HHH/60/td cc: Mr. S. D. Ebneter Regional Administrator, Region II U. S. Nuclear Regulatory Commission 101 Marietta St., NW, Suite 2900 Atlanta, GA 30323 Mr. IIerb Rockhold EG&G Idaho, Inc.

Idaho National Engineering Laboratory 1520 Sawtelle Street Idaho Falls, Idaho 83402 Dr. K. N. Jabbour Office of Nuclear Regulation U. S. Nuclear. Regulatory Commission One White Flint North Mall Stop 14H25 Washington, DC 20555 Mr. W. lf. Orders NRC Resident Inspector

, Catawba Nuclear Station l

l

-n%pnge,mw w r,y ~ wm a

e

(.. 4 i

a .

i

..I.'..' 'g v.' p i( .

j u

,.4 f 4

{

w e ,

) -.

h p

l f'

(- ATTACIIMENT I  ;

f., .

Y'

,t -

b

(

F D

f r  !

p:

k s' >

r r

E 7

I.

b b

hE i.

1,.

b

}' :'.

F e

1,.

b r

f' n

I.

ta s.

p p..

E.

f' y 7-f~

I r

p I

t t

s We% b Anvua- o--

m

• 7 REPLY TO REQUEST FOR-ADDITIONAL INFORMATION ON CATAWBA GENERAL PUMP RELIEF REQUEST

1. OH-6 Section 4.6.4 Paragraph (d) states "... the reference points must be ,

c1carly identified on the pump to permit subsequent duplication in both >

location and plane." Does Catawba intend to mark the' measurement locations, especially in the cases where there will be a deviation from OH-6 because the pump bearing housings are not accessible?

Answer Relief has not been requested from IWP-4160 which requires position sensitive instruments to be permanently mounted or provisions made to duplicate their position for each test. Therefore,. all vibration measurement locations will be marked.

2.' -OM-6 Table 1 states Acceptable Instrument Accuracy for Vibration-measurements to be +/- 5%. Using the total / absolute accuracy statement of +/- 7.3% uncertainty for the accelerometer and the data collector given in the relief request and applying it to the square root of the sum of the squares method to determine the total loop accuracy gives an overall maximum accuracy of +/- 10.32%, twice the allowed value from n OM-6. While you have indicated that the proposed instrumentation is the ,

best that can be reasonable obtained, there is a question whether this amount of error could mask a degraded pump or cause a good pump to be declared inoperable by stretching or compressing the envelope allowed.

• In view of the above, provide a basis for the proposed instrument accuracy addressing the need for the test to assess operational readiness and detect degradation. Also, address whether instrumentation that meets Code accuracy can be obtained at grtater expense.

Answer:

An' accuracy statement of +/- 6.56% of " reading" is in most cases more accurate than the vibration accuracy requirement of IWP. IWP requires vibration instrument accuracy to be +/- 5% of "fu11' scale" and requires the full-scale range to be 3 times the reference value or less.=

Therefore IWP allows accuracy to be +/- 15% of " reading" at the reference value and greater than +/- 15% of " reading" at less than the reference value.

A comparison between the accuracy specified in the relief request and OH-6 is not as straightforward as the above. The relief request specifies accuracy as percent of " reading". OM-6 does not specify that its +/- 5% accuracy statement is relative to " readings" or " full scale".

Therefore, comparisons between the accuracy specified in OH-6 and the relief request can not be clearly made.

I.

The basis for the proposed instrument accuracy is that large changes in vibration levels are the best indicator of pump degradation. For most pumps vibration will have increased at least 250% to enter the Alert Range (>2.5V to 6V ) or 600% to enter the Required Action Range (>6V ).

Since such h{gh percent changes in vibration usually occur before a pump enters the Alert or Required Action Ranee, a +/-6.56% of reading accuracy statement-is insignificant.

Some pumps may enter the Alert or Required Action Range based on &6 fixed vibration limits of >0.325 in/sec-to 0.70 in/sec for the Alert Range, and >0.70 in/sec for the Required Action Range. These fixed vibration units were extracted from vibration severity charts. While the~

charts are in general agreement, there is not agreement on specific numbers. The fixed units in the Relief Request are the units required by-OM-6. These fixed units are necessary because for a pump with a high V an alert range of 2.5 V to 6 V and a required action range of 6 V, is#

r not acceptabic.

The accuracy statement in OM-6 could probably be met if unlimited resources were available. However, any improvement in vibration instrument accuracy would have a negligible positive impact on the effectiveness of vibration monitoring. As stated above, a +/- 6.56%

accuracy statement is insignificant; therefore, any small improvements in accuracy are even more insignificant. Instrumentation changes would be expensive and would disrupt the well established and effective vibration monitoring program in place. Following are examples of how better accuracy statements would be obtained.

- More accurate and more expensive accelerometers could be used, however, due to limited frequency response ranges measurements would probably have to be taken with two different accelerometers at each monitoring point. This would increase monitoring times as well as radiation

-exposure and manpower needs. It would also require data splicing to get an overall reading.

- Expensive and very accurate laser vibrometers could be used. They must be totally isolated from vibration sources; however, so they are not suitable for field applications.

3. OH-6 Table 3 shows " Acceptable", " Alert", and " Required Action" ranges based on the reference value for vibration velocity. Catawba proposed assigning absolute vibration velocity " Alert" and " Required Action" limits for their smoothly running pumps (thosc pumps with vibration velocity reference values less than 0.075 in/sec). This deviation from OH-6 may not be conservative for every pump in this classification since it could allow a large ene ge in vibration velocity measurements, from the reference value, prior to reaching the assigned limits. Alternate methods such as using different absolute values of acceptance criteria or setting minimum reference values snould be adequately justified.

m, .. . j ,

c .-

Answer:

Fixed acceptance limits have been assigned for smooth running pumps to

~

insure that they do not fall'into the alert or required action range thtle still running smoothly. The alert and required action ranges in OM-b are relative to reference values; therefore, smooth running pumps could-full li.to the alert or required action range. For example, when V

= .005 li.lsec, action is required at .031 in/sec. This is int.ppropriate#

sin':e -pump vibrations of .031 in/sec are of no concern. Also, any corrective actions taken could possibly increase vibration levels.

The fixed acceptance limits that have h en assigned for smooth running >

pumps ace acceptable because they are conservative. Vibration levels within.the limits are acceptable for all pumps and any change up to the limits are considered acceptabic.

The limits were established, in part, by substituting 0.075 in/sec into the relative acceptance criteria of OM-6. Therefore, "very smooth" running pumps (V #

< 0.075 in/sec) will have the same alert and action limits that OM-6 establishes for " moderately smooth" running pumps (V # =

0.075). ,

It should be noted that IWP-3210 allows alternate acceptance criteria to -

establisned and used.

4. OM-6 recommends measuring pump vibration displacement for low speed pumps (i.e., <600 rpm). Clarify how this relief request addresses the OM recommendation to measure displacement for low speed pumps.

Answer:

None of the pumps in the Catawba Inser tco Testing Program operate at ,

less than 600 rpm; therefore, there is no need for-the relief request to address the OM-6 recommendation to measure vibration displacements for l pumps that operate at less than 600 rpm.

I 5. OM-6 Section 4.6.1.6 states that the frequency response ranpa "...shall l be from 1/3 minimum pump shaft rotational speed to at least 1000 Hz."

Provide justifleation to assure that pump degradation of low speed pumps

!' will be observed with the proposed instrumentation. Address this issue l for the low speed pumps at Catawba in terms of likely degradation mechanisms with responses at less than RPM frequencies. Also, address whether instrumentation that meets Code accuracy can be obtained at greater expense.

r" t

Answer:

The; slowest speed pumps in the Catawba Inservice Testing Program are the Nuclear Service Water Pumps and the Control Area Chilled Water Pumps.

They run at 715 and 1770 rpm respectively. -1/3 pump speed for these pumps corresponds to 3.95 and 9.83 Hz respectively. -The relief request specifies that vibration instrumentation will-be calibrated from 10 to 1000 llz. Therefore,.the nuclear service Water Pumps are the only pumps that will be monitored significantly below the range instruments are calibrated over.

The fact that instruments will not be calibrated below 10 Hz does not mean that meaningful measurements can not be taken over the range from 3.95 to 10 Hz. The vibration instruments used are repeatable in this range. Also, the instrument manufacturers specify that their accuracy statements are valid over this range under certain canditions (this can not be: verified due to instrument limitations).

The range from 3.95 to 10 Hz constitutes less than 1% of the range over which measurements will be taken; therefore, any additional inaccuracy over this' range will have an insignificant impact on overall vibration measurements. Hence the lack of an accuracy statement over this range is insignificant.

The discussion of the available instrumentation in response 2 also applies here. ,

f

/

-ugg ggyp -.

,4<v y~ q e ,

< ll 1

.i yc l

, -5 lr i,

k i

5 ATTACIIMENT II 3 3

1

~

-t 4-k

.f t

d k

i

&>. n. .

2 f

+ DUKE ~ POWER COMPANY Catawba' Nuclear Station General Relief Request PUMPS: All pumps included in the IST. program.

TEST REQUIREMENTS:- IWP-3100 and IWP-3300. require vibration amplitude to be measured. IWP-3210 specifics the allowable ranges of vibration amplitude measurements. IWP-4110 requires-the accuracy of vibration amplitude measurements to be +/- 5% of full scalo.-

- IWP-4120 requires the full-scalo range of vibration instrumentation to be three times

- the reference value or loss. IWP-4510 requires displacement vibration amplitude to be read at one specific location-during1cach test. IWP4520(b)' requires the frequency.

response range of vibration instrumentation to be from one-half minimum speed to at; least maximum pump shaft speed.

BASIS FOR RELIEF:- Experience has shown that measuring vibration as required by IWP is'not the!most effective way to determine the mechanical condition of a pump. In order to better determine the mechanical. condition of pumps, multiple vibration velocity measurements will~be.obtained/cvaluated and-supplemented, when necessary,1 with acceleration / displacement measurements and spectral 1 analysis. In order to facilitate this. testing, digital vibration instrumentation will be used.

IWP does not provido adequato gu.t :nce/ requirements for performing the better/ alternate testing.

ALTERNATIVE TESTING: In lieu of the vibration requirements of IWP-3100 and IWP-3300, peak vibration velocity will be measured. In most cases, vibration velocity gives the best indication of machine mechanical condition.

In lieu of IWP-4520(b) vibration instrumentation will be calibrated over a range of 10 to 1000 Hz. This is the range that the state of the art instrumentation used can be adequately calibrated over. In lieu of IWP-4520(b) vibration velocity will be measured over a range from 1/3 minimum pump shaft rotational speed to 1000 Hz.

p,- ,.

r

^

.(Measurements at other frequencies will1bc

'taken.as-i cessary.) This range will encompass most potential noise contributors.-

In lieu of the. vibration instrument accuracy requirements of-IWP-4110, the loop accuracy of vibration instruments.will'be +/- 6.56%

of reading. This accuracy will be used because IWP does not:specify an accuracy for vibration velocity.- This accuracy is the best that can be reasonably obtained from tho' state of the art? instrumentation used.

(The requirements of-IWP allow vibration j inaccuracies of greater than +/- 15% of l reading.) ,

i In lieu of the range requirements' imposed'on l vibration instrumentation by IWP-4120, there will be no vibration instrumentation range requirement (digital vibration -!

instrumentation is auto-ranging). It is not1 '

necessary to have a range requirement l because the accuracy statel above and the- '

readability of a digital gauge are not s dependent upon instrument range.

In lieu of the vibration ranges specified in ,

IWP-3210,,the following ranges shall.be used. These ranges will be used:because IWP  !

does not specify ranges for vibration velocity. These ranges are based on current ,

vibration standards (vibration severity j charts). l l

Acceptable Alert Required l Range Range Action Range j i

For All Pumps 0 to 0.19 >0.19 to 0.45 >0.45 When V </=0.075 in/sec in/sec in/sec  :

in/sec#

For Centrifugal </=2.5V # >2.5v to oV >6V or Pumps When V. or >05325 tor >0.70 In/sec

>0.075 in/seb 0.70 in/sec For Reciprocating Pumps When V </=2.5V # >2.5V # to 6V # >6V

>0.075 in/se6

_ __ _ _ _ U

~

ITi ,,

b. -

u h '

In lieu of IWP-4510, peak. vibration velocity Me -measurements shall be taken during each

-test. -!

- Hon contrifugal' pumps,. measurements shall~

.be taken'in a plano approximately perpendicular to.the rotating shaft in two orthogonal directions.- These measurements shall be taken on cach accessibic pump bearing housing.- If no pump bearing-

-housings are accessible, these measurements shall be taken at the-accessible location that gives the best indication of lateral' pump. vibration. This location shall be one ,

of tho'following.- l Pump' casing Motor bearing housing j

-l

-Measurements also shall be taken in the j axial direction. This measurement shall bc  ;

taken on each.accessibic pump thrust bearing i housing.' If no pump thrust bearing housings are accessible, this measurements shall-bc-taken at the acce.ssibic-location that gives .

.the best indication of axial-pump vibration.. ]

This location shall be one'of the following.

Pump casing Motor thrust bearing housing j

1 Motor cacing i

On reciprocating pumps, a measurement l shall be taken on the bearing housing of the  ;(

crankshaft,.approximately perpendicular to both the: crankshaft and the line of plunger travel.

Jaws \iwp\vibrrevl. doc >

__ . . _ _ _ _ _ ._ - . _ _ _ _ _ _____ ________ ____.. ___ ____._ _ _ _ _