ML20027E277
| ML20027E277 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 11/02/1982 |
| From: | NRC |
| To: | |
| Shared Package | |
| ML20027E275 | List: |
| References | |
| NUDOCS 8211120504 | |
| Download: ML20027E277 (5) | |
Text
c..-
onew
~~
o UNITED STATES g
, [
NUCLEAR REGULATORY COMMISSION y.i(gdd.h.;j,[; g o c
{
WASHINGTON, D. C. 20555
~
%a g.
...../
SAFETYEVALUATIONREgRT'
'O THREE MILE ISLAND UNIT 1 PURGE AND VENT VALVE OPERABILITY Requirementi Containment purge valves that do not satisfy the operability criteria. set forth in Brancti Technical Position CSB 6-4 or the
~
~
Staff Interim Position of October.23, 1979 mus_t be sea. led closed as defined in SRP 6.2.4, Item II.6.f during operational conditiens -,.-
h 11.
1, 2, 3, and 4. Furthermore,.th,ese valves must be verified to be
'2-l_.-
closed at least every 31 days _.. Applicants must be in compliance
~ ~'
r
^:
with this position before-thegeceive their operating license.
O Syst'em Descriptiofi-r-
.s-2
~~~
=
~ ' "
1-1 2
- The vahes identified in the Gen ~eral Public Utility (GPU) sub-
~
mittal of March 22, 1982 are as follows:
1.
Spring closed valves AH-V-1A and AH-V-1D 2.
Motor operated valves AH-V-1B and AH-V-1C These valves were described as 48 inch Pratt model R1A.
The pressure rating for these valves was not ident'ified.
The in-stallation configuratior.s for these valves were not described in
~
the submittal.
,/
- Qualification Approach A complete stress analysis was not provided in the GPU submittal
.of March 22, 1982.
A summary stress table was provided along with a list of considerations used in the stress analysis.
The con-siderations were as follows:
1.
Valve closure time during a LOCA will be less.
than or equal to the no flow time demonstrated" during shop tests, since the fluid dynamic effects tend to close the valve.
~.
2.
The maximum differential pressure was applied against the valve for all angles of opening.
3.
Worst case is determined as a single valve closure with containment pressure on one side of the valve and atoris-pheric pressure further downstream.
4.
Containment back pressure will have no effect'on closing
~
the valve.
l s
5.
The subject valves do not use accumulators.
8211120504 821102 PDR ADOCK 05000289
(
p PDR
~
~
r l
s There are no torque lithiting d' vicer fo'r the air 6.
p operated valves AH-V-1A and AH-V-10.-~The-s,ettings of the torque limiting devices for the electric motor operated valves AH-V-18 and AH-V-1C are compatible with the torques required to operate the valve during the design basis conditions.
3 7.
The effect of downstream piping is ignored as a l
conservative approach.
./-
8.
The vdiva disc and shaft orientation does not affect
~
torque calculations.
~
P>
t
, y. '
t*
s.
, [-
-e
,,.,g
~.:
9.
The qualification was performed for the valves closing _
~
from 4 maximug opening position of 30*.
- +
l'*?-
Seismic loads' we're comtiined 'with LOCA idads in the' stress...
~i. <
~~
10.
- .~."
analysis:
i
~
~
ASME Boiler and Pressure Vessel Code (hereafter referred to as the code) stress allowables were used in-some cases but 90% of-yield strength was used for some valve internals.
For some cases the faulted condition stress allowables were used (1.5 Sm).
This was equivalent to the yield strength of the material.
This was done for a valve interval component (shaft), as well as, the pressure boundary components. While the code provides these rules for the design of pressure boundary components they are not intended to assure operability of a component such as a valve.
A description of the torque coefficient tests performed by Pratt was not submitted by GRU.
However, during a meeting on August 20, 1,981 with members of the NRC staff, Brookhaven National Lab, and H o ry Pratt Valves, the Pratt model test was described to consist of 5 inch model valves representing in shape and aspect ratio the Pratt line of disc designs.
The test installation was configured to establish straight line approach flow to the valves.
Torque data was recorded,in order to establish torque coefficients.
Asymmetric disc designs were flow tested in both directions.
The dynamic torque equation is used with appropriate dynamic torque coefficients, media differences, and size factors to determine the maximum torque which would be experienced in the subject valve.
In evaluating the structural integrity of the valve operators the calculated torque curing a LOCA was compared with the maximum torque rating of the opeFator per the manufacturer's data.
The analysis.was performed assuming the valve travel was limited to 30* and therefore was required to close from no greater than the 30 position.
r I
a e6 e
n -
\\- '
-~
i s
~
Evaluation
[.
r..
The stress analysis was performed using the peak containment N
pressure at all angles of opening from 30 to fully-closed.
This s
is a conservative method of applying LOCA loads.
The stress analysis did not, however, show the 48 inch purge valves are capable of closure from the 30* position under LOCA loads. The stress analysis showed -
shaft combined stresses of 30,748 psi for valves AH-V-1A and AH-V-1D,
3 and 30,734 psi for valves AH-V-1B and AH-V-1C.
The stress: allowable ~ ~~
submitted for this valve is 1.5 Sm =-30,000 psi.
This val @ is -
~
equivalent to the yield. strength.
It-was GPU's judgement that this..
L
- small overstress would not create a failure situation.
C
.... ~ -
s-
~
This judgement -that.it -is acceptable to go beyond yield stiength.
J' for this condit' ion'is'not acceptable.
Even if there was no
-f
- f.ailure of-the 'coniponerit in' the.overstressed' condition, yielding
-d.
of the shaft could affect the valve's ability to operate.
Based,.
on this overstressed condition it is the staff's conclusion that~.
the 48 inch purge v'alves foi Three Mile Island,-Unit 1 have not-been shown to be able to close against dynamic loads in the event of a LOCA.
In addition to the overstressed condition in the shaft the following areas were noted of concern during the review:
1.
The allowable-loads under shear for the shaft key and disc pins'were much higher'than acceptable.
While these components did not appear to be over-stressed the margins were much less than shown.
The allowable shear load was indicated to be 90% of yield strength.
2.
The materials listed in.the GPil report were incon-sistent with the Pratt'drawi 3 and bill of materials.
For example, the body material is listed in Table 2 as ASTM-A-36, in the Bill of Materials as ASTM-A515 Grade 70 and on the Pratt drawings as ASTM-285 Grade C.
3.
GPU's submittal does not contain information on instM 'alion con. figurations.
Upstream bends or ens separation distances, shaft orientations
%t ve to upstream bends, disc closure direction wt
- aw direction all have an effect on torque loads.
rc, warison between actual installations and the Pratt test configuration.were made to justify the torque loads used in the analysis.
The straight pipe configuration believed to be the applicable test configuration does not normally yield the most conservative loads.
4.
GPU did not discuss the detai7s of how it plans to limit the valve opening en a long-term basis.
A tolerance of plus 1.75 for valves AH-V-1A and AH-V-10 s
i I
t J
l
\\.
y 4_
i
~
and 3.29 forAH-V-1B'andAH-h-1CWas' proposed.
No
~
justification was provided for extending ~t.he tolerance beyond the 30, especially in view of the fact that the valve analysis shows the valve to be overs'ressed at 30*.
t 5.
GPU addressed a leakage problem for these valves at out-side ambient temperatures below 60 F.
GPU's response to -
this issue is that the valve would be above 60 F Jn the
~
event of a LOCA.
a Whild the val.ve temperature-would-increase to above,60*F,..
in a LOCA event this should not be expected to occu(~
~TC
,. ;?-
instanYaneously and some leakage'would therefora ber-
-. ~ ~
~
expecte'dc In addition, valve leakage may occur unde -
~
sirably 'at times other than LOCA's.
GPU should take'
~
-ri
.6,
/. positive. steps._to el.iminate.sthis. concern.- The present
[
method appear.s to, be adjusting or replacing seati when;.
the valve fai]s a leakage test.
This solution does not'.
preclude'a le~akag'e problem between periodic leakage -
A checks.
g 6.
Periodic inspection of valve seats was discussed but not the maintenance program for the anti-friction roller bearings used on these valves.
This area should be addressed in assuring operability.
7.
The air actuators are expected to have solenoid valves to vent the air.
The qualification and air quality of these valves was not discussed.
Solenoid valve failures are often the cause of air operator valve failures.
8.
Environmental and seismic qualification of the valve operators was not addressed.
Evaluation Summary It is the staff's conclusion that the operability of the Three Mile Island, Unit 1 48 inch butterfly purge valves has not been t
demonstrated in the event of a LOCA.
The major reason for this determination is the overstressed shaft of these valves in the event of a LOCA.
There are a number of.other concerns that also contributed to this determination as discussed above.
Unqualified valves should be sealed closed as defined in SRP 6.2.4, ltem II.6.f during operational conditions 1, 2, 3 and 4.
Further-more, such valves must be verified to be closed at least' every 31 days.
l
.1
s-
. ~
Sealing these valves closed will not eliminate-the leakage problem due to temperatures of 60*F or fielow.-~The-4;icensee should take action to eliminate this concern.
O h
e o
me g
.m.
9 6 -
9 w
w-
- g
-r-o i
3-
- O
...r.-
I.
- s..'..
.o
+
e
.~.
s -
4.-
- ~
~
j
-~'
j '- '-
~.
L a_
-w
,m,
-*e
.. one e.
e e
e J
e e
.g 4
1 e
1 i
i i
ew-m-
-,,m--
m
-m 4
+.
,w-y
.