Information Notice No. 91-74: Changes In Pressurizer Safety Valve Setpoints Before Installation

                                UNITED STATES
                        NUCLEAR REGULATORY COMMISSION
                    OFFICE OF NUCLEAR REACTOR REGULATION
                           WASHINGTON, D.C.  20555

                             November 25, 1991 


NRC INFORMATION NOTICE 91-74:  CHANGES IN PRESSURIZER SAFETY VALVE SETPOINTS 
                               BEFORE INSTALLATION


Addressees

All holders of operating licenses or construction permits for nuclear power 
reactors. 

Purpose

The U.S. Nuclear Regulatory Commission (NRC) is issuing this information 
notice to alert addressees to excessive changes observed in the setpoints of 
pressurizer safety valves (PSVs) after setting and acceptance testing but 
before installation.  It is expected that recipients will review the 
information for applicability to their facilities and consider actions, as 
appropriate, to avoid similar problems.  However, suggestions contained in 
this information notice are not NRC requirements; therefore, no specific 
action or written response is required.

Description of Circumstances

The Duke Power Company is investigating excessive changes in Dresser PSV 
setpoints that it has frequently identified during post-operational tests.  
The investigation has revealed that the setpoint changes may have been 
present before installation.  Duke Power began the investigation when 
testing of as-found PSVs removed from the Catawba and Oconee Nuclear 
Stations in refueling outages early in 1991 revealed setpoints significantly 
outside the technical specification (TS) tolerance of �1 percent.  Each of 
five valves tested failed to meet the TS limits, and the average deviation 
was over 5 percent (four failed high and one low).  While performing 
subsequent investigative tests on two valves that had been acceptably set 
and tested to the TS criteria, but not installed, Duke found unacceptable 
setpoints.  One valve setpoint was high by 2.2 percent and the other was 
high by 5.6 percent, as compared to the �1 percent TS limit. 

In its investigation, Duke initially examined its use of a pre-installation 
"jack-and-lap" procedure following setpoint testing as a possible 
contributor to the setpoint changes.  This procedure involves partial 
disassembly of a PSV, maintaining its spring in compression, and polishing 
the seat surfaces to remove minor irregularities that may cause seat 
leakage.  Licensee personnel questioned the adequacy of the controls 
previously applied to the jack-and-lap process.  Procedural controls were 
added and an acceptably set and tested PSV was processed using the revised 
jack-and-lap procedure.  Calculations indicated that any setpoint change 
caused by the new jack-and-lap procedure should 

9111180310 
.

                                                       IN 91-74
                                                       November 25, 1991
                                                       Page 2 of 5

not exceed 0.5 percent.  However, retesting following performance of the new 
procedure revealed a setpoint that was 2.2 percent high, which is greater 
than a 1 percent change from the previous setting.  This is the 2.2 percent 
high setpoint previously referred to in the above paragraph.  The setpoint 
increase beyond the 1 percent TS limit found in this test suggests that the 
jack-and-lap procedure has not been wholly responsible for the unreliable 
setpoints at the Duke plants.  However, the fact that the setpoint deviated 
less than was recently experienced indicates that the previous jack-and-lap 
procedure may have contributed significantly to the problem. 

The valve referred to in the second paragraph above which tested 5.6 percent
high in the pre-installation test had successfully completed the sequence 
normally used to prepare the valve as a replacement to be installed during 
the next refueling outage.  It was obtained from storage at the Catawba 
Nuclear Station.  Therefore, the licensee had not controlled the 
jack-and-lap procedure as closely for this valve as for the one which tested 
2.2 percent high.

Duke has identified a number of possible causes of the excessive changes in 
PSV setpoints and considers the following two the most probable: 

     Valve leakage during setpoint testing that is eliminated by the 
     subsequently performed jack-and-lap procedure:

     If a valve is leaking during setpoint testing, its huddle chamber may 
     become partially pressurized, increasing the valve seat area and 
     reducing the pressure required to cause lift.  The pressure required to 
     produce lift will be increased above the original setting if personnel 
     adjust and test the setpoint with this leak and later repair the leak 
     (through the jack-and-lap process) without verifying the setpoint 
     again, as was the licensee's practice. 

     Inadequate control of the jack-and-lap process

     The jack-and-lap process involves disassembling the valve partially and 
     polishing the seating surfaces.  This may introduce errors that can be 
     limited with close controls but that may be inadequately quantified if 
     the setpoint is not reverified afterward.

Duke considered other possible causes of setpoint change including the 
following:  

     �    failure to recognize adverse trends in the three setpoint pops 
          used in the verification test, 

     �    performance of ring adjustments without reverifying the setpoint,

     �    temperature effects (as described in NRC Information Notice 89-90, 
          "Pressurizer Safety Valve Lift Setpoint Shift," and Supplements 1 
          and 2), 

.

                                                       IN 91-74
                                                       November 25, 1991
                                                       Page 3 of 5


     �    spring performance variables (e.g., corrosion or inadequate 
          lubrication of spring/spring washer sliding surfaces), 

     �    seat adhesion, 

     �    handling or transportation shock, and 

     �    various test process parameters (e.g., steam quality, 
          pressurization rate, valve installation variables, equipment 
          calibration errors, and personnel errors).

Duke reported that, based on its investigation, the refurbishment and test 
sequence it plans to specify for its PSVs each refueling outage is as 
follows: 

     �    Mount the valve on a test stand and strap on electric heaters and 
          thermocouples.  Two thermocouples to be placed 120� apart at each 
          of the following locations:  valve inlet flange, lower bonnet, and 
          upper bonnet.

     �    Supply the valve with saturated steam at 90 percent of the set 
          pressure for heat-up.

     �    Allow the steam to heat the valve's internal components while the 
          heaters and thermocouples are used to maintain the proper valve 
          temperature.

     �    Leak test the valve with steam by placing a mirrored surface probe 
          at the valve discharge and observing its surface for condensation.  
          Use the following acceptance criterion:  "No leakage shall be 
          allowed."

     �    If leaking is found, repair the valve by the jack-and-lap process.

     �    Determine the initial actuation setpoint by raising the steam 
          pressure at a ramp of 100 to 200 pounds per square inch per second 
          until the valve stem lifts. 

     �    If the setpoint is not within the specified �1 percent limits, 
          notify the licensee and adjust the valve setpoint and retest the 
          valve until it is within the specified limits.

     �    Require three consecutive successful tests within the specified 
          setpoint limits for acceptance.

     �    Leak test the valve with steam at 93 percent of the nameplate 
          rating by passing a mirrored surface probe around the disc-seat 
          interface and then visually inspecting the mirror for 
          condensation.  Use the following acceptance criterion:  "No 
          leakage shall be allowed."

     �    If leaking is found, repair the valve by the jack-and-lap process.  
          Following repair, install the valve on the steam header and verify 
          the setpoint.  
.

                                                       IN 91-74
                                                       November 25, 1991
                                                       Page 4 of 5


     �    If leakage is not found following the setpoint verification, 
          package the valve and transport it to the station for storage 
          until the next refueling outage.

Subsequent to the investigative tests mentioned above, Duke retested a PSV 
for Catawba that had previously been refurbished and set in accordance with 
the above sequence and then placed in storage.  The setpoint was determined 
during five actuations, and each lift was within 1 percent of the original 
setpoint.  This indicated that the more strictly controlled refurbishment 
and testing may help reduce the magnitude of setpoint changes.  
Unfortunately, the valve failed the test because of minor leakage.

Discussion

The industry has continued to find that the setpoints of PSVs and main steam 
safety and safety relief valves can be unreliable.  Recently, the NRC staff 
reviewed data from the Nuclear Plant Reliability Data System and found that, 
historically, as-found PSVs in U.S. plants have failed over 40 percent of 
the setpoint tests performed.  The NRC staff believes that the number may 
actually be higher because of unreported failures.  In performing the 
investigation described herein, the Duke Power Company identified possible 
causes of setpoint change that may have not been previously given adequate 
recognition.  This investigation indicates the need to ensure that setpoint 
testing is performed after all operations that may cause changes.  In this 
example, jack-and-lap procedures used to correct leakage before installation 
may have caused the setpoints to change.  Many licensees prefer not to test 
the setpoint after performing procedures to correct leakage (such as the 
jack-and-lap) because the setpoint test itself appears to often lead to 
subsequent leakage, resulting in a repeating cycle that may be difficult to 
end.  Leakage itself cannot be tolerated because even small pre-installation 
leaks may lead to steam cutting and increasing leakage during operation, 
which could cause the setpoint to change during operation.  The licensees 
may need to maintain closer controls on certain maintenance, setting, 
testing, and other operations performed before the valve is installed.  

The Duke Power Company investigated setpoint changes on PSVs manufactured by 
Dresser Industries, which is understood to have supplied similar valves to 
approximately a third of the nuclear utilities in the United States.  
However, similar problems may exist with PSVs supplied by other 
manufacturers due to similarities in design, setpoint testing, maintenance 
practices, leakage correction practices, and other factors. 

Related Generic Communications

Previous information notices issued on related topics include IN 86-56, 
"Reliability of Main Steam Safety Valves"; IN 86-92, "Pressurizer Safety 
Valve Reliability"; IN 88-68, "Setpoint Testing of Pressurizer Safety Valves 
with Filled Loop Seals Using Hydraulic Assist Devices"; and IN 89-90 and 
Supplements 1 and 2, "Pressurizer Safety Valve Lift Setpoint Shift."
.

                                                       IN 91-74
                                                       November 25, 1991
                                                       Page 5 of 5


This information notice requires no specific action or written response.  If 
you have any questions about the information in this notice, please contact 
one of the technical contacts listed below or the appropriate Office of 
Nuclear Reactor Regulation (NRR) project manager.




                                   Charles E. Rossi, Director
                                   Division of Operational Events Assessment
                                   Office of Nuclear Reactor Regulation


Technical contacts:  Edward H. Girard, RII        Charles G. Hammer, NRR
                     (404) 331-4186               (301) 492-0791

                     Francis Jape, RII            Mary S. Wegner, AEOD
                     (404) 331-4182               (301) 492-7818


Attachment:  List of Recently Issued NRC Information Notices
.