The
following is an abstract of the presentation given by Mr. Bill Lehman
of Arizona Public Service Palo Verde Nuclear Generating Station. Mr.
Lehman's presentation was given at NIC in December 1995.
Background
At Palo Verde, we have a total of 71 Borg Warner pressure seal
design swing check valves installed in our primary side systems.
The pressure seal design uses a silver plated 316 stainless seal
to prevent bonnet leakage. This type of seal is typically capable
of sealing against minor surface imperfections (.002".003"
in-depth). At our site we have experienced many problems with bonnet
leakage after performing maintenance on these valves. When a leaking
valve is identified we attempt to stop the leakage by using a combination
of system pressure and bonnet bolt re torques. If those are unsuccessful
the valve is scheduled for disassembly at the next available opportunity.
In the mean time, it creates significant housekeeping and radiological
concern.
This valve is located on the discharge of one of our four (4) safety
injection tanks and had been a chronic leaker since it was first
disassembled for routine inspection in November 1991. At that time,
the inspection did not indicate any abnormal conditions with the
sealing surfaces and the valve was reassembled with a new silver
seal. When the valve was put in service at the end of the Outage
it immediately began to leak and build up contamination on the external
surfaces. This required a special decontamination effort at power
during the operating cycle.
Our first attempt to correct the problem was in the following Outage of April 1993. We disassembled the valve and
performed a very close and detailed examination of the sealing surfaces to look for any scratches or evidence of galling. The bonnet surface had some minor pitting located along the seal contact area. The pitting depth
was not quantified but estimated to be less than .001" deep. The bonnet was replaced and a new silver seal was installed. Again the bonnet leaked as the Unit was restarted, and repeated attempts to retorque the
bonnet bolts were unsuccessful.
After that failed attempt to resolve the problem, we assumed that
the cause was some type of out of round condition within the valve
body and the following two (2) options were considered for the next
outage:
Machine the body for an oversize silver seal.
This method had been previously used in another identical valve
where an out of round condition was found in June of 1990. The cost
of the machining work alone was $32,000.
Use of an alternative seal design.
Other valve manufacturers had documented success with graphite
composition in pressure seal applications, so we felt strongly that
it was an option worth pursuing. In addition to the superior sealing
properties, we expect a significant man hours savings during disassembly
and reassembly using the graphite seal. The old seal, which has
been expanded into place during assembly and subsequent system pressurization,
must be removed carefully to prevent it from becoming cocked and
galling the valve body surfaces. On large bore valves that process
can take multiple shifts (~100 man hours) to accomplish, sometimes
heat blankets and porta powers are necessary to break the bonnet
free from the seal. This problem is eliminated with the EGC Pressure
Seal.
We contacted BW/IP and requested that they look into the development
and application of a graphite seal to replace the silver seal. Working
with them, we scheduled a number of tests using Grafoil Composite
PressureSeals purchased from EGC Enterprises, Inc.
BW/IP testing
Tests using prototype seals in both a four (4) inch and a twelve (12) inch valve were performed at the BW/IP facility in Vernon, California. The valves were hydrostatically tested to 5000 psig with
the EGC Pressure Seals installed. In one test a .020" gouge across the sealing face and a .030" out of round condition was introduced to observe the sealing capabilities. Throughout the tests at varying
pressures up to 5000 psig there was no detectable leakage.
Conclusion
Based on what we saw during our testing, we decided to use the
EGC Pressure Seal kit if the valve body was not out of round by
more than .020".030", and default to remachining the sealing
surface as a last option.
Valve Inspection and Rework
When Unit 2 began the outage, we disassembled the check valve and
inspected it for eccentricity in the bodysealing surface. Based
on our findings, we decided to use the EGC Pressure Seal kit Installation
and test. Because this material is compressible, we needed to monitor
bonnet movement during the system pressurization (to ensure that
the seal assembly would have sufficient preload to maintain a tight
seal under all anticipated system pressures) and to verify that
the final bonnet height resulted in an acceptable alignment of the
disc to the seat. New torque valves were calculated based on the
highest expected system pressure and the seal was installed and
tested to that pressure (1800psig). Once at 1800 psig, a final torque
was performed and the bonnet bolt nuts were lock-wired in position.
The new EGC Pressure Seal has now been installed for the past 18
months of the operating cycle at normal operating pressures of 600-700
psig with no bonnet leakage detected.
