Bulletin 79-13: Cracking in Feedwater System Piping

                               UNITED STATES 
                       NUCLEAR REGULATORY COMMISSION 
                   OFFICE OF INSPECTION AND ENFORCEMENT  
                           WASHINGTON, D.C. 20555 

                               June 25, 1979 

                                                       IE Bulletin No. 79-13

CRACKING IN FEEDWATER SYSTEM PIPING 

Description of Circumstances: 

On May 20, 1979, Indiana and Michigan Power Company notified the NRC of 
cracking in two feedwater lines at their D. C. Cook Unit 2 facility. The 
cracking was discovered following a shutdown on May 19 to investigate 
leakage inside containment. Leaking circumferential cracks were identified 
in the 16-inch feedwater elbows adjacent to two steam generator nozzle elbow 
welds. Subsequent radiographic examination revealed crack indications in all 
eight steam generator feedwater lines at this location on both Units 1 and 
2. 

On May 25, 1979, a letter was sent to all PWR licensees by the Office of 
Nuclear Reactor Regulation which informed licensees of the D. C. Cook 
failures and requested specific information on feedwater system design, 
fabrication, inspection and operating histories. To further explore the 
generic nature of the cracking problem, the Office of Inspection and 
Enforcement requested licensees of PWR plants in current outages to 
immediately conduct volumetric examination of certain feedwater piping 
welds. 

As a result of these actions, several other licensees with Westinghouse 
steam generators reported crack indications. Southern California 
Edison,reported on June 5, 1979, that radiographic examination revealed 
indications of cracking in feedwater nozzle-to-piping welds on two of three 
steam generators of San Onofre Unit 1. On June 15, 1979, Carolina Power and 
Light reported that radiography showed crack indications in similar 
locations at their H. B. Robinson Unit 2. Duquesne Power and Light confirmed 
on June 18, 1979, that radiography has shown cracking in their Beaver Valley 
Unit 1 feedwater piping to vessel nozzle weld. Public Service Electric and 
Gas Company reported on June 20, 1979 that Salem Unit 1 also has crack 
indications. Wisconsin Public Service company decided on June 20, 1979 to 
cut out a feedwater nozzle to pipe weld which contained questionable 
indication, for metallurgical examination. As of June 22, 1979 and since May 
25, 1979 seven other PWR facilities have inspected the feedwater 
nozzle-to-pipe welds without finding cracking indications. 

The feedwater nozzle-to-pipe configurations for D.C. Cook and for San Onofre
are shown on the attached figures 1 and 2. A typical feedwater pipe-to-
nozzle weld joint detail showing the principal crack locations for D.C. Cook 
and San Onofre are shown on the attached figure 3. 

On March 17, 1977, during heat-up for hot functional testing of Diablo 
Canyon Unit 1, a leak was discovered in the vessel nozzle-to-pipe butt weld 
joining the 16-inch diameter feedwater piping to steam generator 1-2. 
Subsequent nondestructive examination of all nozzle welds by radiography and 
ultrasonics 

.

IE Bulletin No. 79-13                                       June 25, 1979 
                                                            Page 2 of 4 

revealed an approximate 6-inch circumferential crack originating in the weld
root heat-affected zone of the leaking nozzle weld. The cause of this crack-
ing was identified as either corrosion-fatigue or thermal fatigue initiating
at small cracks probably induced by the welding and postweld heat treatment 
cycles. The system was repaired replacing with a piping component employing 
greater controls on the welding including maintaining preheat temperature 
until postweld heat treatment. 

Cracked weldments have been removed from the D. C. Cook Units 1 and 2 and 
San Onofre Unit 1 feedwater systems for extensive metallurgical 
investigation by Westinghouse. Based on preliminary analysis, Westinghouse 
stated the D. C. Cook failure may be "fatigue assisted by corrosion." The 
San Onofre cracking was stated to be characteristic of "stress assisted 
corrosion." 

The cracking experienced at Diablo Canyon, D. C. Cook and San Onofre would 
appear to have different cause - effect relationships which are not fully 
understood at this time. 

The potential safety consequences of the cracking is an increased likelihood
of a feedwater line break in the event of a seismic event or water hammer. A 
feedwater line break results in a loss of one of the mechanisms of heat 
removal from the reactor core and would result in release of stored energy 
from the steam generator into containment. Although a feedwater line break 
is an analyzed accident, the identified degradation of these joints in the 
absence of a routine inservice inspection requirement of these feedwater 
nozzle-to-piping welds is the basis for this Bulletin. 

Actions to be Taken by Licensees: 

For all pressurized water reactor facilities with an operating license: 

1.   Facilities which have steam generators fabricated by Westinghouse or 
     Combustion Engineering that have not conducted "volumetric examination 
     of feedwater nozzles since May 1979 shall complete the inspection 
     program described below at the earliest practical time but no later 
     than 90 days after the date of this Bulletin. 

     a.   Perform radiographic examination, supplemented by ultrasonic 
          examination as necessary to evaluate indications, of all feedwater
          nozzle-to-piping welds and of adjacent pipe and nozzle areas (a 
          distance equal to at least two wall thicknesses). Evaluation shall
          be in accordance with ASME Section III, Subsection NC, Article 
          NC-5000. Radiography shall be performed to the 2T penetrometer 
          sensitivity, level, in lieu of Table NC-5111-1, with systems .Void
          of water. 

.

IE Bulletin No. 79-13                                       June 25, 1979 
                                                            Page 3 of 4 

     b.   If cracking is identified during examination of the nozzle-to 
          piping weld, all feedwater line welds up to the first piping 
          support or snubber and high stress points in containment shall be 
          volumetrically examined in accordance with 1.a. above. All 
          unacceptable code discontinuities, other than cracking, shall be 
          subject to repair unless justification for continued operation is 
          provided. 

     c.   Perform a visual inspection of feedwater system piping supports 
          and snubbers in containment to verify operability and conformance 
          to design. 

2.   All pressurized water reactor facilities shall perform the inspection 
     program described below at the next outage of sufficient duration or at
     the next refueling outage after the inspection required by item 1. 

     a.   For steam generator designs having a common nozzle for both main 
          and auxiliary (emergency) feedwater systems, perform volumetric 
          examination of all feedwater nozzle-to-pipe weld areas and all 
          feedwater pipe weld areas inside containment in accordance with 
          item 1 above.1/ In addition, conduct an examination of welds 
          connecting auxiliary feedwater piping to the main feedwater line 
          outside containment. This examination should include an area of at
          least one pipe diameter on the main feedwater line downstream of 
          the connection. 

     b.   For steam generator designs with separate nozzles for main, 
          feedwater and auxiliary feedwater, perform volumetric examination 
          (in accordance with item 1 above) of all welds inside containment 
          and ups-bream of the external ring header or vessel nozzle for 
          each steam generator.  If an external ring header is employed, 
          also inspect all welds of one inlet riser on each feed ring of 
          each steam generator.1/ 

     c.   Perform a visual inspection of all feedwater system piping 
          supports and snubbers in containment to verify operability and 
          conformance to design. 

3.   Identification of cracking indications in feedwater nozzle or piping 
     weld areas in one unit of a multi-unit facility shall require shutdown 
     and inspection of other similar units which have not been inspected 
     since May 1979, unless justification for continued operation is 
     provided. 

1/   Welds in the feedwater system, (other than the feedwater nozzle-to-pipe
     welds) that have been examined since May 1979 need not be re-examined. 
.

IE Bulletin No. 79-13                                  June 25, 1979 
                                                       Page 4 of 4 

4.   Any cracking or other unacceptable code discontinuities identified 
     shall be reported to the Director of the appropriate NRC Regional 
     Office within 24 hours of identification. 

5.   Provide a written report to the Director of the appropriate NRC 
     Regional Office within 20 days of the date of this Bulletin addressing 
     the following: 

     a.   Your schedule for inspection if required by item 1. 

     b.   The adequacy of your operating and emergency procedures to 
          recognize and respond to a feedwater line break accident. 

     c.   The methods and sensitivity of detection of feedwater leaks in 
          containment. 

6.   A written report of the results of examinations, in accordance with 
     requests by Regional Offices preceding this Bulletin and with Bulletin 
     item 1 and 2 including any corrective measures taken, shall be 
     submitted within 30 days of the date of this Bulletin or within 30 days 
     of completion of the examination, whichever is later, to the Director 
     of the appropriate NRC Regional Office with a copy to the NRC Office of 
     Inspection and Enforcement, Division of Reactor Operations Inspection,      
     Washington, D.C. 20555. 

Approved by GAO, B180225 (R0072), clearance expires 7/31/80. Approval was 
given under a blanket clearance specifically for identified generic 
problems. 

Attachments: 
Figures 1, 2 and 3