Information Notice No. 82-35: Failure of Three Check Valves on High

                                                            SSINS No. 6835 
                                                            IN 82-35 

                               UNITED STATES 
                       NUCLEAR REGULATORY COMMISSION 
                          WASHINGTON, D. C. 20555 

                              August 25, 1982 

Information Notice No. 82-35:   FAILURE OF THREE CHECK VALVES ON HIGH 
                                   PRESSURE INJECTION LINES TO PASS FLOW 
Addressee: 

All nuclear power reactors With an operating license or construction permit.

Purpose: 

This information notice is provided as notification of an event that may 
have safety significance. It is expected that recipients will review the 
information for applicability to their facilities. No specific action or 
response is required. 

Description of Circumstances: 

At Davis-Besse Unit 1 on June 4, 1982, a stop check valve (HP-57) in the 
normal makeup system failed to pass flow although 120 psid was applied 
across the valve. Normal opening pressure is about 5 psid. The problem was 
discovered while filling the reactor coolant system (RCS) using a small 
low-head pump following a refueling and maintenance outage. Normal makeup at
Davis-Besse is via one of the four 2 1/2-inch high-pressure injection (HPI) 
lines. Upon further investigation, HPI valves HP-48 and HP-56 also failed to
pass flow at 120 psid. Each HPI line has a stop check valve and a swing 
check valve in series (see Figure 1 for HPI system diagram). 

The stop check valves are manufactured by Velan. Similar valves may be used 
in safety related systems at other plants (an exact listing of affected 
plants is not yet available). According to the manufacturer, all Velan 2 
1/2-inch stop check valves are of the same basic design. The internals 
consist of a disc which is lightly spring loaded against the valve seat. The
disc opens to allow flow at pressures sufficient to overcome spring tension.
A valve stem, which is not connected to the disc, can be turned down on the 
disc via a handwheel to block it against the seat. In this mode, the valve 
provides an isolating function in addition to preventing backflow (see 
Figure 2 for valve internals diagram). 

Upon detection of the initial failure, special testing was performed in 
order to determine the status of the four HPI stop check valves. A portable 
hydropump was used to find the differential pressures at which the valves 
would lift open. The testing yielded the following results: 

1.   Valve HP-57 required 580 psid to open. Since the valve is in the normal
     makeup path, however, its operability was not in question before the 
     outage. 

2.   Valve HP-48 required 180 psid to open. Its operability is uncertain 
     since the last pass flow surveillance tests were conducted on the 
     valves on September 17, 1980. 

8204210414 
.

                                                           IN 82-35 
                                                           August 25, 1982  
                                                           Page 2 of 5 

.

                                                           IN 82-35 
                                                           August 25, 19  
                                                           Page 3 of 5 

.

                                                           IN 82-35 
                                                           August 25, 1982  
                                                           Page 4 of 5 

3.   Valve HP-56 did not pass flow with up to 120 psid pressure. Its 
     operability since the last surveillance test is also in question. 
     Additional differential pressure had not been used to open the valve so
     as to inspect the closed valve internals and determine the failure 
     mechanism. No damage to the seat, disc, or valve body was detected upon
     visual inspection of internals. 

4.   Valve HP-49 opened with less than 5 psid, as designed; thus, it was 
     determined to be functional prior to the refueling outage. 

5.   All valves were tested by applying up to 150 ft-lbs of torque to the 
     stems and observing the differential pressure required to open them 
     after the stem was raised. Only a valve HP-56 continued to bind, even 
     with as low as 25 ft-lbs of torque applied to the stem. According to 
     these tests, it has been calculated that RCS back pressure, while in 
     the operating mode, would have caused HP-56 to bind. 

The causes for valve failure are thought to be a combination of overtorquing
by operators and a steep valve seat angle. Wear may have also been a 
contributing factor; however, no obvious signs of wear have been detected by
visual inspection. Before the failures, surveillance testing of the HPI 
system was performed, and all the valves opened properly. Then the operators
manually closed the stop check valves to isolate the HPI system from the 
RCS. Normally the motor-operated valves (MOVs) located outside containment 
are used for HPI isolation. 

Because the stem packing of the valves was so tight, the operators used a 1 
1/2-foot valve wrench rather than the handwheel to close the valves. The 
valve manufacturer recommends that no more than 150 ft-lbs of torque be used
to close the valve. With the valve wrench the operator could have easily 
overtorqued the valve. Later, the stop check valve stems were returned to 
the open position to allow refilling the RCS. At this point it was found 
that HP-57 failed to pass forward flow when makeup was being provided by a 
small booster pump rather than by the normal charging pump. If the normal 
charging pump had been used to refill RCS, the binding valve would not have 
been detected because the high pressure would have opened the valve. HP-48 
and HP-56 also failed to pass flow except with large differential pressures. 
The manual closure of the valves resulted in the failure of the valves to 
open, although the earlier pass flow test have demonstrated operability. 

In addition to overtorquing the valves, a steep seat angle of 15 could 
have contributed,to the valves binding. Valves HP-57, HP-56 and HP-48 had 
seat angles of 15 and valve HP-49 had a seat angle of 30. This is 
because HP-49 was a later replacement valve and the manufacturer changed 
seat angle specifications after three original valves were installed. The 
licensee has since replaced the plugs and relapped the seats for the other 
three valves to help prevent binding. 
.

                                                           IN 82-35 
                                                           August 25, 1982 
                                                           Page 5 of 5 

If you have any questions regarding this matter, please contact the Regional
Administrator of the appropriate Regional Office, or this office. 



                                   Edward L. Jordan, Director  
                                   Division of Engineering and  
                                     Quality Assurance  
                                   Office of Inspection and Enforcement 

Technical Contact:  George Lanik, IE 
                    (301) 49-29636 

Attachment: 
1. List of Recently Issued Information Notices