IR 05000382/1989020

From kanterella
Jump to navigation Jump to search
Insp Rept 50-382/89-20 on 890705.No Violations or Deviations Noted.Major Areas Inspected:Svc Failure of Shaft in Chilled Water Pump B
ML20247K582
Person / Time
Site: Waterford Entergy icon.png
Issue date: 07/25/1989
From: Barnes I
NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION IV)
To:
Shared Package
ML20247K579 List:
References
50-382-89-20, NUDOCS 8908010004
Download: ML20247K582 (5)


Text

y ,y ,

-

- - - - - - -

-

.

.,M -

hb '

9 (' g; 1 h[ g . '

-

,

.

,

APPENDIX

,

>

U.S.' NUCLEAR REGULATORY COMMISSION

REGION IV

'

1 <

, ,

. NRC Inspection Report:. 50-382/89-20 Operating License: .NPF-38

' -

Do~cket: 150-382

  • LLicensee: Louisiana Power i 8. Light Compan: (LPAL) -

l317 Baronne Street

'

New Orleans' ,- Louisiana 70160 E Facility Name:" Waterford. Steam Electric ^. Station, Unit 3:(W3SES)=

Inspection'At: W3SES, Taft, Louisiana Insption Conducted:' July. 5,1989 o .

Inspector:: L b=ag 7-2r-#9'

lT. Barnes, chief Materials and Quality Date

+ Programs.Section, Division of Reactor Safety Approved: 8 - => 7-ar- 4"/

. I. Barnes, Chief. Materials and Quality Date Programs Section Division of Reactor Safety

' Inspection Summary-Irispection Conducted July 5,1989 (Report 50-382/89-20)

.1 Areas Inspected: Nonreutine, announced inspection of the service failure of

, ' the shatt in Chilled Water Pump Results: . Examination of the failed pump shaft indicated that fatigue was the i failure mechanism, with the origin appearing to be a square keyway corner. A laboratory failure analysis is being perfonned to provide information en the root cause of the shaft failure. One unresolved item was identified ((paragraph-2.6) in regard to review of the findings of this analysis and

! .- licensee response action 'i l

-l

-!

8908010004 890727 PDR ADOCK 05000382

.

Q FDC

-- . _ _ _ _ _ . _ _

___ _ _ _ _ _ _

O- A .

.

-

-2- ,

r DETAILS Persons Contacted LP&L

  • J. R. McGaha, Plant Manager
  • R.: Brian.. Electrical System Engineering 1 Supervisor

.*C. P. Talarac, System Engineer

  • L.'W. Laughlin, Site Licensing. Supervisor
  • P. V. Prasankumar, Assistant Plant Manager, Plant Technical Services
  • G. F; Koehler,' Operations Quality Assurance Audit Supervisor
  • T.' H. Smith, Plant Engineering Superintendent
  • J. B. Perez. Engineer, Mechanical. Engineering' Design-

.

  • D. E. Baker, Manager, Nuclear Operations Support and Assessments
  • D..F. Packer, Assistant Plant Manager, Operations and Maintenance
  • G. M.~ Davis, Event Analysis Manager

'

NR *T. R. Staker, Resident Inspector

  • Denotes 1those personnel attending the exit intervie . Followup'on Chilled Water Pump B Shaft Failure 2.1 Background On June 28, 1989, a failure occurred of the shaft in ChilledEater Pump B

>

(Buffalo Forge.Model 4013 CRE, Serial No. 365214-001) during an 8-hour run following preventive maintenance to replace packing. The failed shaft was initially installed in the pump in October 1986 because of pump leakage that was occurring due to excessive wear of the original pump shaf t in the packing area. As a result of incorrect (i.e., different design to those

, used by LP&L) spare shafts having-been furnished by the pump vendor, the shaft which ultimately failed was manufactured locally using material furnished by LP& .2 Examination:of Failed Shaft Review of component photographs taken after pump disassembly showed that the shaft fractured at a location just within the impeller hub. A keyway 1 .was noted to have been used for locking the impeller in position on th shaft. -The impeller to shaft assembly was completed by insertion of 1/2-inch bolt through the impeller and into a drilleo and tapped hole on the centerline of the shaft. The diameter of the shaft at the failure location was 11/8 inches. Visual examination of the fractured shaft showed essentially smooth fracture surfaces with little surf ace detail evident. The final fracture area was small indicating the shaft nominal applied, stress was low. The fracture area adjacent to one side of the s

l

'

/

_ - _ _ _ _ -

- _ _ -

.

'

-

-3-keyway was dark in appearance, which was indicative of the presence of a surface oxide film, and thus suggesting that a crack had been preexistent for some period prior to final failure. Examination of the smc11

.

(approx %tely 1 inch) f w+ ; red piece of shaf t that had been inserted in the impeller hub showed additional cracks were present. Cracks were observed which propagated along one corner of the keyway and then changed direction to approximately normal to the axis of the shaft. The inspector also noted at the end of the shaft that a crack ran from the keyway corner to the bolt hole on the centerline of the shaft. Examination of the keyway showed that the keyway had been cut with square corners with no radiu The inspector concluded from the limited visual examination that shaft failure was by a fatigue mechanism, with initial crack initiation occurring at the stress-concentrating notch created by the square keyway corner. LP&L personnel informed the inspector that the failed shaft would be sent for laboratory failure analysis to confirm the root cause of failur .3 Review of Original Pump Shaft Requirements Review of Drawing 1564-4833 Revision 1, showed the shaft material to be ASTM A-276. Type 316, but did not designate which of the two conditions listed by this specification was to be used; i.e., solution annealed or solution annealed and cold worked. It was noted from review of Ebasco Specification LOU-1564. "HVAC Water Pumps," Revision 7, dated February 6, 1984, that Section 17.06 required impellers to be keyed to the shaf This section also specifically identified that square keyways were not acceptable and that manufacturers should have corners of keyways contoured to a radius to distribute loading and minimize the possibility of failur Section 17.07 of this specification also required dynamic and hydraulic balancing of impellers and shafts as subassemblie .4 Examination of Incorrect Spare Shaft Furnished by Buffalo Forge As a result of the specific keyway requirement noted in the Ebasco specification, the inspector examined the irspeller keyway in one of the ;

incorrect spare shafts furnished to LP&L by Puffalo Forge. The corners of {

the keyway in this shaft also appeared squ6re with no radius being i visually apparen .5 Procurement of Failed Shaft Review of Condition Identification Work Authorization (CIWA) 020218 and Gulf Traveler L-1038 showed that Gulf Engineering had fabricated the shaft from 2 1/16 inch SA-182. Grade F304 bar stock furnished by LP8L. The original shaft was also provided to Gulf Engineering for measurement of dimensions and preparation of a drawing. Documentation was available of a

,

telephone contact with Buffalo Forge regarding approval of the I substitution of SA-182, Grade F304 stainless steel for the ASTM A-276, ,l l

!

<

_ _ _ ._

_

. - .

w

Type 316 shhf t material used by the pump manufacturer. No documentation was noted which would confirm that LP&L had been unable to obtain the applicable drawing for the shaft (i.e., AN 15957) from the pump manufacture In response to a- request from the inspector, LP&L contacted Buffalo Forge and was able to obtain a copy of Drawing AN 15957. As a result of the lack of legibility of dimensions on the microfiche copy of the Gulf Engineering shaft drawing, that was rtviewed by the inspector, a precise comparison with Drawing AN 15957 could not be made. Review of the failure region did indicate, however, that an apparent 0.161 inch depth keyway was listed on the Gulf Engineering drawing versus a depth of 0.145 inch specified by. Drawing AN 15957. An average depth of 0.157 inch was measured by LP&L subsequent to the inspection as being present. The inspector did not confirm that this dimension was taken at the edge of the keyway. The inspector also noted that the 1/2-inch bolt hole on the centerline of the shaft was indicated on the Gulf Engineering drewing to be tapped to an apparent depth of 11/8 inches versus a depth of 7/8 inch specified by Drawing AN 15995 A certified material test report (CMTR) was not obtained by Ebasco for the original pump shaft, thus precluding comparison of actual mechanical properties with those listed on the CMTR for the failed SA-182 Grade F304 replacement shaft. Specification minimum strength properties are, however, the same for both grades of material. Comparison of reported mechanical properties for the failed shaft material against properties reported on a CMTR for the incorrect spare shafts furnished by Buffalo Forge showed higher strength values for the failed shaft material; i.e.,

SA-182 Grade F304 - 55,600 psi Yield Stress, 89,400 psi Ultimate Tensile Strength; ASTM A-479, Type 316 - 40,100 psi Yield Stress, 83,300 psi Ultimate Tensile Strengt No reference was noted in the CIWA to dynamic balancing of the l impeller /SA-182, Grade F304 shaft subassembly prior to installation in Chilled k'ater Pump B. Review of the data from the last two quarterly ASME Section XI Code tests performed on the pump prior to failure showed, however, acceptable vibration value J l

2.6 Summary Examination of the failed pump shaft indicated that fatigue was the failure mechanism, with the origin appearing to be the stress-concentrating notch created by a . square keyway corner. The depth of the keyway was also noted to exceed .that specified by the pump manufacturer's shaf t drawing. The !

presence of an impeller keyway with square corners was also observed on an incorrect spare shaft furnished by Euffalo Forge. This observation, when considered in conjunction with the similarity of mechanical properties for the two grades of material, raises a potential concern regarding the condition of the shafts in the remaining two pumps. LP&L is currently j having a laboratory failure analysis performed which should prov',de l

!

_ Y

- yy. ,, _

'

.1'-

'

j y'f(l -

,

. ;, . . ; y y ' '

.

'

,

r l' ([ .'t

_

,

11 a, ;

    • '

.- '5

. .

<

.-

~.

-q, p ,

'

L

-

additional ~ information on.the root cause of~ shaft failure. Review of the.

L sis and planned LP&L response actions ar . findings.of.thislaboratoryanaly(382/8920-01)

considered an unresolved _ item. .

'

3.- E_xit Interview i>

L An exit interview was held on July 5, 1989, with those individuals. denoted-

~

in paragraph 1 of this report. At this meeting, the scope of the inspection and the findings were summarized. The licensee did not identify as proprietary c any of. the infonnation provided to or reviewed by the inspecto ' ~

,

.

..<

i l

_ _ _ _ __ __ -.