Information Notice No. 82-09: Cracking in Piping of Makeup Coolant Lines at B&W Plants

                                                            SSINS No.:  6835
                                                            Accession No.: 
                                                            8202040131 
                                                            IN 82-09 

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

                               March 31, 1982

Information Notice No. 82-09:  CRACKING IN PIPING OF MAKEUP COOLANT LINES
                                  AT B&W PLANTS 

Description of Circumstances: 

On January 21, 1982, Crystal River Unit 3 commenced shutdown to investigate 
an unidentified 0.9-gpm primary leak.  During power reduction the leak rate 
increased to about 1.0 gpm and the plant proceeded to hot standby 
conditions. 

A visual inspection inside the reactor building at this time revealed the 
leak was associated with a 2 1/2-inch check valve (MOV-43) in the makeup 
line to the 26-inch reactor coolant (RC) loop A inlet line.  This line is 
used for normal makeup of reactor coolant but is also part of the redundant 
high-pressure injection system.  After the insulation was removed from the 
affected valve a 140o circumferential crack in the check valve body near the 
valve-to-safe end weld (i.e., valve end toward RC inlet nozzle) was found.  
The leak was nonisolatable and the plant promptly proceeded to cold shutdown 
conditions in accordance with plant technical specifications. 

The check valves was removed and liquid penetrant testing (LPT) was 
performed on the accessible inside diameter (ID) surfaces including 5 inches 
into the 2 1/2-inch line on the inlet side of the affect valve.  This 
inspection disclosed an extensive network of heat-check type cracks around 
the safe end ID surface.  A similar conditions was observed inside the valve 
body from the discharge side up to the disc seat area.  The valve inlet side 
and connecting piping were not affected.  The most severe cracking in the 
safe end appeared to have penetrated up to 25 percent of the wall thickness.  
A visual inspection also revealed the thermal sleeve inside the 
high-pressure injection (HPI) nozzle was loose and showed evidence of wear 
in areas of contact.  Some cracking of the thermal sleeve was also observed.

As a result of the Crystal River 3 findings, Duke Power Company initiated a 
radiographic examination of the RC inlet nozzle connections on the two HPI 
lines used for normal makeup at Oconee Unit 3 to determine the thermal 
sleeve conditions.  This examination disclosed that in one of the makeup 
nozzles the thermal sleeve was loose, the four thermal sleeve retaining 
button welds on the safe end side were missing, and the thermal sleeve was 
slightly displaced in the upstream direction of flow.  Action was then taken 
to remove the pipe extension to replace the affected thermal sleeve.  
Further findings and expanded inspection as a result of this action are 
summarized below. 
.

                                                            IN 82-09 
                                                            March 31, 1982 
                                                            Page 2 of 3 

Investigation and Findings: 

A.  Crystal River 

    A metallurgical investigation of the affected valve body indicated two 
    crack initiation sites.  One was inside on the valve body at a machine 
    mark (i.e., weld counterbore area) and one was on the outside diameter 
    (OD) at the valve-to-weld transition (geometrical discontinuity).  The 
    cracks progressed through the wall on a slightly different plane and 
    merged about mid-wall of the valve body.  Scanning electron microscope 
    examination of the fracture features disclosed the cracks propagated 
    transgranularly and exhibited clearly defined grain structure striations
    characteristics of cyclic fatigue failure.  Cracks in the thermal sleeve
    and safe end sections exhibited similar fracture morphology.  No 
    evidence of corrosion interaction from chemical attack was identified. 

    During the design phase, Babcock and Wilcox (B&W) performed the stress 
    analysis on the primary system up to the affected check valve which is 
    the design code (USAS B31.7-USAS B31.1) interface boundary.  Gilbert 
    Associates, as architect-engineer, performed the balance of plant 
    design. The B&W design calculations for the HPI lines included a pipe 
    section that was not installed during plant construction.  The potential 
    thermal discontinuity at this point is believed to be partly responsible 
    for the cracking and is currently being evaluated by both organizations. 

    Based on the above findings, the mode of cracking was tentatively attri-
    buted to thermal cycle fatigue. However, the synergistic 
    thermal-hydraulic effects contributing to the failure mechanism are yet 
    to be determined. Contributing factors being investigated include 
    operational design limits and setpoints with regard to makeup water 
    temperature and flow rate, minimum bypass flow, and system 
    thermal-hydraulic parameters around the HPI nozzle used for makeup. 

B.  Oconee 
 
    When the pipe extension at Oconee 3 was removed to gain access to the 
    thermal sleeve in order to repair it, liquid penetrant testing (LPT) 
    disclosed cracks on the ID surfaces of the makeup/HPI pipe extension 
    and nozzle safe end.  Crack features were similar in nature to those 
    found at Crystal River.  Reportedly, the cracks penetrated up to 20 per-
    cent of the thickness of the pipe wall.  The other makeup nozzle 
    assembly was examined by radiography and a special ultrasonic testing 
    (UT) technique developed by B&W for this purpose.  No indication of 
    cracking or degraded thermal sleeve conditions was observed.  Further UT 
    and radiographic testing (RT) of the two remaining HPI nozzle assemblies 
    indicated a loose thermal sleeve in one of the nozzles (Nozzle 3B1). 

    At Oconee 2, results of the UT and RT indicate the thermal sleeve in one
    of the makeup nozzles may be loose and the retaining button welds on the
    safe end side are missing.  Cracking was also found in the safe end and 
.

                                                            IN 82-09 
                                                            March 31, 1982 
                                                            Page 3 of 3 

    pipe extension.  The other makeup nozzle showed no indications of a 
    degraded thermal sleeve or cracking.  Examination of the two remaining 
    HPI nozzle assemblies indicated a loose thermal sleeve (i.e., retaining 
    weld buttons missing) in one and a crack in the rolled area of the other 
    nozzle thermal sleeve. 
    
    At Oconee 1, examination of the four HPI nozzle penetrations to the RC 
    loop inlet line showed no evidence of degradation. 

Discussion: 

In B&W design plants the line(s) for normal makeup of reactor coolant are 
also part of the redundant high pressure injection system.  These plants do 
not have a regenerative heat exchanger in the makeup coolant circuit.  
Therefore, during operations, the potential exists for the makeup coolant 
temperature to be much lower than the reactor coolant temperature in the 
loop.  Fluid temperature fluctuations resulting from mixing in the HPI 
nozzle coupled with hydraulic effects are thought to be primary contributors 
to the cracking problem at Crystal River and at the Oconee plants.  Although 
the cracking location is within the scope of the LOCA (loss-of-coolant 
accident) safety analysis, the existence of cracking in an area not 
routinely included in the program of ISI represents an unacceptable 
challenge to system integrity. 

An evaluation of the cracking problem and its resolution has been requested 
of the B&W Regulatory Response Group. 

Pressurized-water reactor systems of the Combustion Engineering and 
Westinghouse designs do have a regenerative heat exchanger in the makeup 
coolant line which is a separate, dedicated system.  During normal power 
operation the makeup coolant enters the nozzle at temperatures on the order 
of 50-150 F below the temperature of the reactor coolant loop 
respectively. However, transients may occur in which the makeup flow rate is 
greater than the letdown flow rate.  Depending on the frequency and duration 
of these transients, the makeup coolant might not be heated to the expected 
temperature.  Therefore, the potential may exist for large temperature 
fluctuations in the makeup nozzle to cause problems similar to those 
discussed above.  Past experience has shown similar thermal fatigue problems 
with nozzle-thermal sleeve assemblies in other systems of both BWR 
(NEDO-21821, 1978) and PWR (WCAP-7477 and NEDO-9693-1980) designs. 

This Information Notice No. is provided as an early notification of a 
potentially significant matter that is still under review by the NRC staff. 
If NRC evaluation so indicates, further licensee action may be requested.  
In the interim, we expect that licensees will review this information for 
applicability to their facilities. 

No written response to this information notice is requested.  If you need 
additional information, please contact the Regional Administrator of the 
appropriate NRC Regional Office. 

Attachment:
Recently Issued IE Information Notices