ML20126B234
| ML20126B234 | |
| Person / Time | |
|---|---|
| Issue date: | 08/16/1979 |
| From: | Gary Young Advisory Committee on Reactor Safeguards |
| To: | Advisory Committee on Reactor Safeguards |
| Shared Package | |
| ML20126B231 | List: |
| References | |
| ACRS-SM-0167, ACRS-SM-167, NUDOCS 8003100113 | |
| Download: ML20126B234 (7) | |
Text
F0IAExemption(b)5 e
,e UNITED CTATES L.
o,,
- 8
NUCLEAR REGULATORY COMMISSION n
5 4
i ADVISORY COMMITTEE ON REACTOR SAFEGUARDS wAsmNGTON, D. C. 20555
%...../
August 16, 1979 i
i.
ACRS Members SIGNIFICANCE OF RECENT SNUBBER FAILURES AT OPERATING NUCLEAR POWER PLANTS At the request of Dr. Siess, I reviewed the status of snubber-related problems at nuclear power plants. The review included a literature search and an LER search.
I also had discussions with Dr. Siess and with Mr. Horace Shaw of the NRC Staff. Mr. Shaw has worked for several years on snubber problems and is currently working on the NRC-sponsored research of snubber i
performance.
Attached please find a short description of the operation of hydraulic and mechanical snubbers. This description is intended to provide a definition of some of the snubber-related terminology (e.g. lockup rate, bleed rate, etc.).
As a result of the LER review, I prepared a table of snubber-related problems at operating nuclear power plants from January 1978 through June 1979. The number and type of snubber problems at each reactor is included l
in Table 1.
l A total of 101 LERs were written on snubber problems from January 1978 until June 1979. The 101 LERs indicated 459 individual snubber problems.
The problems ranged from complete loss of function to slight degradation of snubber function.
The snubber problems can be categorized into three general problems:
- 1) failure to fully lock up, 2) lockup during normal use, and 3) failure to meet lockup and/or bleed rate criteria. These three categories of failure will be specifically discussed in the following paragraphs.
1)
Failure to Fully Lock Up The failure of a snubber to fully loc'K up (i.e. become rigid during dynamic events) can result from several causes. Among the causes are loss of fluid in the snubber reservoir, inverted snubber which gets air in the reservoir, and disconnection of the snubber from the support structure, i
8003100lh
This type of failure is significant only when the piping system is exposed to a dynamic. event. During normal system operation the snubber is an ir$ctive component of the support structure. Only during rapid dyrgmic events (e.g. earthquake, waterhammer, f ast valve closure, %tc.) does the snubber perform an active function in the support structure.
The failure of only one snubber to lock up during a design basis dynamic transient was investigated by Pacific Gas and Electric in a report titled " Stress Evaluation of Piping Systems Assuming Single Snubber Failures" (letter to NRC dated January 24, 1978, Docket 50-275/323). The investigation concluded that the proba-bility of a safe shutdown earthquake, a single snubber failing to snub, and subsequent pipe failure in one year of operation is 10-7 This is no more likely than other piping failures 3.38 x considered in WASH 1400.
Apparently no investigation has been done on the consequences of multiple snubber failures. Many examples of multiple snubber failure are recorded in recent LERs. One extreme example is in an LER written in July 1979 (therefore, not included in Table 1),
where Fort St. Vrain reported 365 snubbers f ailed the acceptance test and 192 of those failed to lock up as designed.
If Fort St.
Vrain had experienced a safe shutdown earthquake prior to the snubber inspection, the result could possibly have been multiple piping system failures.
- 2) Lockup During Normal Use Lockup of a snubber during normal use can result from such things as corrosion in mechanical snubbers or installation damage to either hydraulic or mechanical snubbers. Only seven of the problems listed in Table 1 (under the "other" column) are related to snubber lockup. Of the seven, only one involved a mechanical snubber.
'It should be noted that mechanical snubbers are not required to be inspected for operability by present-day Technical Specifica-tions. This could explain the small number of mechanical snubber failures reported.
An example of the lockup problem potential for mechanical snubbers was experienced at the Fast Flux Test Facility (FFTF) in the first half of 1978. The FFTF has more than 4000 mechanical snubbers.
Performance testing of those snubbers was started in 1978 and it was found that an extremely high percentage of the snubbers were locked up or otherwise inoperable. 00E then replaced all the mechanical snubbers that were installed with an improved model.
This incident has not been formally reported to NRC since the FFTF is not regulated by NRC-
g The consequence of one snubber locking up during normal use is considered f
in the Pacific Gas and Electric report.
In the report, the probability ofasnubberfailingrigigandcausingpipefailureinoneyearof operation is 1.04 x 10. This is a considerably higher probability than the failure caused by a non-locking snubber. One reason for the I
higher probabilitysis that, in pipe support design, a snubber is usually specified only when the piping system cannot take the stress associated with the use of a rigid restraint. Therefore, a locked up or rigid snubber exposes the piping system to.large stresses during normal, routine plant operation. Thermal movement is resisted by a locked up snubber and the designed-in piping system flexibility is lost.
Snubber lockup during normal use is the worst problem of the three listed in this discussion.
It can cause pipe failure during routine plant operation. The problem is compounded by the fact that plant Technical Specifications do not require periodic surveillance of mechanical snubbers, and often, when a mechanical snubber fails, it fails by locking up.
- 3) Failure to Meet Lockup and/or Bleed Rate Criteria During surveillance testing, it is quite frequent that hydraulic snubbers do not perform as required by Technical Specifications for lockup rate and/or bleed rate. The snubbers do in fact usually lock up and bleed, but not within the limits set by Technical Specifications.
The concern with this type failure is that the piping system response during a dynamic event is in question since the snubber will not perform as expected. No investigation has been made into the actual conse-quences of this type failure. However, NRC is funding research into the area of lockup rate / bleed rate versus piping response to a dynamic event. Tne research is intended to better define the acceptance criteria for snubbers which will be reflected in future Technical Specification requirements.
I will continue to follow the status of snubber related problems to keep abreast of new develcpments and to report those developments to the ACRS Members.
Please feel free to contact me if you have any comments or questions about this work.
M
@e a
Garry G. Young v
ACRS Fellow Attachments:
1.
Table 1, LERs on Snubber-Related Problems 2.
Appendix A. Summary of Operational Characteristics of Hydraulic and Mechanical Snubbers l
cc:
R. F. Fraley 1
M. C. Gaske M. W. Libarkin T. G. McCreless (for Branch circulation)
J. C. McKinley (for Branch Circulation)
+
s TI9 6LE s
Staussee Remrev Posceu sic.ENSEE kVENT kEPORT5 CN JanuAny 13% - Juss. IM9 Nutneen e v SNv56 E R Pno st ems Lockup /hSV@hWE PL-RNT NAM E Low Ftuso Leva
.Tuventce Discouxecreo ovr or seee,r,ea r,oa Ornm ARynasns - l 4
G4 esa Roa 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> B
BRcwH> Fetcy - 2 i
BrewNs f^cenf 3 G
Snu.*suocon - l 5
2 I
BROH w.tn 2
2i T
IS I
CALKT ( Che f() 2 l
l Coc,Ps"R - I J
Caysin L. Etics 3 I
D.C. Ccen - 1 3
G3 D. C. C a r n 2 I
DRcsDcn - l I
DRcspen - 3 6
DUANC ACWCLD l
E.1. linren - l 7
I 2
E. I. lin ren - 2 1
Farirstn oc x - l 4
20 Fr br: Vrms - I
'l H.0 Resiasca 2 3
finocon Necoc I I
kesu Auscc - l 3
A '5.covc 2
/
No.ui lhoa lb Nr - l l
I tkeni ANNr) I G
I I
I OceNCc I 2.
Occute 2 5
oysim Cacot - l l
PAtossets - l 2
i PcActi &rrem 2 1
4 1%ctom - l
/
4 Poswr Brn sos I I
fboHT Bt1s t ob - 2
/
Qu/1D CeT#CS /
/
Quno Crncs
.2 2
RANeolc sccc l
Y 9
i R. C. GonNA I 2
SuRaf I 5
1 27 l
bcRRf - 3 4
2 50 I
Thstc ni 1sano 1 I
I ToRsev hawt - 3 I
Tonn et k nr-4 20 Venmcar YAuscc-l I
Lica I
5
- 2 s vN 2
2I 2
2 7"oTALS
/22 B
/S 260
.50
s ej>
1 APPENDIX A k
[
SUMMARY
OFOPERATIONALCHARACTERISTICSOF HYDRAULIC AND MECHANICAL SNUBBERS
.3E Hydraulic Snubbers The mechanistic mode of operation of the hydraulic snubber centers around the control valve. The control valve converts the snubber from a free acting device to a strut with a given stiffness. When the snubber is subjected to motion exceeding the lockup velocity (usually 8-10 inches /
minute), the poppet valve closes due to the flow of hydraulic fluid (a pressure drop is created across the va1.ve) and subsequent flow is directed to the smaller bleed orifice.
The snubber is then able to carry a load because of the restricted flow. The load is resisted by both the fluid column and structural elements. The snubber endpoints continue to trans-late at the bleed velocity (usually 4-6 inches /iainute) as the load is resisted. The bleed velocity is proportional to the magnitude cf the applied load.
In general, the stiffness of the snubber and the peak-to-peak displacement it sees under dynamic load are a function of the bleed rate and lockup velocity.
The lockup velocity and bleed rate are very sensitive to the viscosity of the hydraulic fluid.
A schematic illustra-tion of a snubber in operation is presented in Figure A.l.
Mechanical Snubbers There are two types of mechanical snubbers. The first and most common type arrests motion to a specified maximum acceleration. The second type senses motion above a specified threshold and then becomes an elastic strut.
The second type ceases to translate in resisting the load.
The first type continues to " bleed" as it resists load.
The basic concepts of operation are similar for all mechanical snubbers.
Linear motion is mechanically converted into angular motion through the rotation of the ball screw shaft.
Attached to the shaft are a torque transfer drum and an inertia mass that rotate along with the shaft.
Enclosed within the torque transfer drum is a capstan spring.
Tangs on the capstan spring project through the torque transfer drum. The tangs engage the inertia mass.
Under a slowly applied load, the entire mechanism rotates freely.
- However, excessive axial acceleration will cause the inertial mass to lag behind the torque transfer drum.
The' inertia mass then catches the tangs of the capstan spring and winds the spring up around a stationary mandrel.
The t
load is then mechanically resisted.
The mechanical advantage is on the order of 50,000 to one.
A schematic illustration of a mechanical snubber is shown in Figure A.2.
NURt.G - 0%7 DATCD CONE. MP76.
4 T/9(E4 FROW m
aESERv0iR artER tocxmc eEroat Locumc j
N
'N p
~
~~
eASCO CHECK VALVE (
ppgyygyg g
~%
q Q
b av-pass eLEED ANN \\%%N\\\\\\\\\\M%\\\\\\\\\\\\T
{
Roc eusamo aco utPER 4 --
I
~
Vj PISTON Seat Fig. A.1. Schematic illustration of a hydraulic shock arrestor in l
operation. Drawing courtesy of ITT-Grinnell.
l l
I l
l
~ ~ ~ ~ ' '
l l
llfIl1 llll1Ill 1i
~
- at M
M 1
1 M
M
\\s R
M E
B
};
U T
N a% j g.
0, J.
A
?
A j
R
.j T
1
.rot s
pJe 7
3 i
.c7 e
J.
g r
w a
s u
k re.g(s p
M o
c U
h R
v=
s D
R l
NI i
h-i A n
E a
% g!
%N F
G S
c i
SU R
l[
MT a
O S
h E
H T
b Mo V
c F
e A
H\\=
\\
r T
R m
S i g._
u S o N
P T
a M7~-
W t
E t
f R
AB oc C
p L
S i
nf L
.,c=.x,,u -
tn L
oi A
it B
ae ri tc 4._
sS u
i' z_
a_v.t s.
E li R
lc D
if IL cc R
N i
E D
Y ia C
N tP IL 2
_. ~
T a
Y R
O mf C
eo G
P 1
P h
NI U
cy P
S O
Ss m
2r
_lw
.te CS x
EL u
E n
Ao T[
.c r
g
/_
=
ig
=
Fn
/a
=
i c
w I ra+s;._
G m
a N
G r
R U
D g2 A
L E
P D
=
i #a u D
L N
Ac E
'J. %
RI R
E E
D H
4.
7 N
P S
4#
L T
Vc
~
w i,
m, lI!!l,