ML20199H036
Text
,
o
/>7)
[
UNITED STATES y
g NUCLEAR REGULATORY COMMISSION g
'j W ASHINGTON, D. C. 20555
%,... * * / -
JUN 141985 1
MEMORANDUM FOR:
S. Hou, CPSES TRT FROM:
R. Masterson, CPSES TRT
SUBJECT:
CATEGORIZATION OF " PUNCHING SHEAR" AS AN ALLEGATION You recently assigned to me Allegation. AP-45 Category 54, Punching Shear, as described in the CASE Document entitled " Informal Comments by CASE Witness Jack Doyle as dictated by phone. to CASE Pre'sident Juanita Ellis" dated December 18, 1984.
On page 4 of. this document the punching shear concern is identified however, this is not a new item, since it was originally addressed by D. Terao as part of the Motion for Summary Disposition on AWS vs ASME welding.
I have contacted D. Terao and he provided me with a copy (attached) of.
his response to the CASE concern.
His response was the result of his review of the CASE document as required by his association with the Motions for Summary Disposition.
Since the TRT scope of -review has not previously included the Motions for Summary Disposition, I recommend that the CASE concern for punching shear not be included in the TRT review responsibility..,
m/
R. Masterson CPSES TRT
Enclosure:
As stated A
cc:
- 0. Terao C. Poslusny' go\\k I
_B6070 _30065 860623 GARDE 86-A-18 PDR
-mr-ma
'h,,
>. ' /!/' ? tt !'
l; p
- M y 1:
- i. y 6 7 - = Shung c
sI'eNr 101 California s?eet. suite 1000, san Francisco, CA 94111-5894 415.397 5603 February 19, 1985 84042.035 Mr. J. 8. George Project General Manager Texas Utilities Gent ating Company Comanche Peak Steam Eiectric Station Highway FM 201 Glen Rose, Texas 76043
Subject:
Stability of Pipe Supports Texas Utilities Generating Company Comanche Peak Steam Electric Station Independent Assessment Program Job No. 84042
References:
(1)
N.H. Williams (Cygna) letter to V. Noonan (U.S. NRC),
"Open Items Associated with Walsh/Doyle Allegations,"
84042.22, dated January 18, 1985.
~
~
(2)
N.H. Williams (Cygna) letter to V. Moonan (U.S. NRC),
" Revision to Open Items Schedule," 84056.055, February 14, 1985.
3(3) Affidavit of John C. Finneran Jr. regarding ' Stability of Pipe Supports and Piping Systens, dated June 17, 1984.
(4) Cygna Phase 3 Final Report TR-84042-1, Rev.1, November 20, 1984.
Dear Mr. George:
As comitted to in Reference 1 and subsequently revised in Reference 2 Cygna tias completed an evaluation of the pipe support stability issue-This evalua-tion considered the support designs reviewed by Cygna as part of Phases 2, 3 and 4 as well as TUGCO's position described in Reference 3.
Since stability is a very complex issue, we will sumarize our
'Sition in six parts:
(1) Definition of Stability, (2) Dynamic Versus Static Stability, (3) System Stability, (4)'
Commentary on TUGCO's Position, (5) Classification of Cygna Review Scope, and (6) Conclusions.
F01A-85.59 s.n..._ e _ c _,o _._
p ra
.[MNQ fNitG7 imges Mr. J. B. George February 19, 1985 Page 2
?
a Definition of Stability Prior to performing an evaluation of this issue, criterfba were developed to define what constitutes an enstable pipe support.
Individual pipe supports cars be classified into two broad categories:
(1) supports which, in the total absence of the pipe, are stable, and (2) supports which., in the total absence of the pipe, are unstable.
Implicit in our definition for the second category is the fact that the instability is a rigid body type whictn may be completely removed or acconsnodated by proper attachment to the pipe That is, by restrain-ing certain degrees of freedom at the attachment to the : pipe, such as with a pipe clamp, the instability my be removed. Alternatively, by limiting the motion following instability through the pr'esence of the pipe and adjacent supports, the instability my also be eliminated. Since there is no stability i'ssue with respect to supports of the first category, omly supports of the second category need be discussed.
In order for a support of the second category to be stabile, there are two
~
~
requirements to be met, one involving force transfer between the pipe and suppo'rt and the other involving the geometric relationstnip between the pipe and!
support. The force requirement is met if adequate forces, which develoo instantaneously and can be relied upon by design, exist between the pipe and the support hardware to resist the factored load. The following definitions are provided for clarity:
de'velop instantaneously (imediately): Resistcing forces are e
activated at the same instant that piping loadis are applied. An example of forces which cannot develop imedfartely are binding forces which require a rigid body motion of ttne support (rotation, translation) to become effective.
e by design: The mechanism for and magnitude of the resisting forces are calculatable and known, or have been evaluated exten-sively by test or by use in the specific application.
e factored load: Applied load times a safety fa' ctor.
In addition to the above described force requirement, tfhe geometric relationship between the support and the pipe must remain within set limits during the operational life of the plant.
If sufficient clamping forces between the pipe and support are not present, small pipe movements nay cause large changes in tire position of the support relative to the pipe.
Piping system vibration occurring during start-up, normal operation or shut-down can cause the support to move (rotate, translate) relative to the pipe. This support movement is unfav'orable if, for a support initially perpendicular to the pipe, the direction of, pipe i
e 9
. m. _
.>m_..-==.m--------_----=
m L
1 an fs e r us Mr. J. B. George February 19, 1985 Pa ge 3 i
movement in th'e a'bsence of the suppcrt is such that-the d:fisplaced centerline of the pipe intersects the are nade by the rigid body motion) of the pipe center within the support. The new position of the support on tihe pipe nay be well outside the displacement (eccentricity) envelope for which it was designed and for which stability has been assured.- Since the support did not restrain the movement of the pipe during this process, adjacent supports must now resist an additional load f;e'r which they any not be adequate. Therefore, a sufficient condition for inotvidual pipe support stabili.ty of the second category is a design in which, upon the application of the factored load from the pipe, adequate resisting forces can be developed immediately and the position of the support attachment on the pipe does not move relative to the pipe with time.
Considering the definition presented above,' we will now dtiscuss some specialized situations in which the instantaneous development of resisting forces required for stability does not occur. For these designs momentary instability (of the rigid body type) could be tolerated, provided that it cant be demonstrated that sufficient forces eventually develop to completely remove the instability (i.e.,
stop the motion and allow the support to function as desfgned). For example, when considering the instability of a support which requftres the development of binding forces to ultinately'mintain stability, one could assume the support does not act and then determine the resulting pipe deflection in the released direction.
If that deflection is a sufficient multiple (say 4) of the deflec-tion required to develop the necessary binding forces, it then becomes
- appropriate to further investigate the ability of the susqport to resist both the binding force and the applied load. During such an investigation, it is essential to' demonstrate that the binding force mechanism possesses both sufficient strength and stiffness.
In other words, while certain designs nay exhibit sufficient strength to develop and resist the neesssary binding forces, they may not posssess sufficient stiffness to limit the rigid body displacement and thus resist the applied load. The alternative to this approach is to limit the consequences of the instability. This could be accomplished by strowing that the piping and rennining supports are acceptable' in the albsence of thee. unstable support.
In either approach, before the design can be considered satisfactory.
pipe stresses and other support reactions must be checked for the new displace-ments occurring at the support and the pipe must be. checked for the effects of the binding forces.
Dynamic Versus Static Stability The preceeding discussion addre'sses only stability due to statically applied loading. The question arises as to whether a support cotald be unstable statically under the application of naximum load, yet stable when tne same load is applied dynamically.- This is a very complex analytic problem to resolve
. which is further complicated by the fact that the naximun loading on a pipe
~
e e
e e
$QQ.
l i
r._~_----
F
, v ~ n.
l
)
Mr. J. B. George
[
j-February 19, 1985, Page 4 l
I i
support is generally some combination of static and dynamic loads. Cygna is; i
i~
unaware of any established precedent for the acceptance of statically unstatame j
supports based on dynamic arguments.
In some cases dynamiciloading can contrri-i bute to pipe support instability rather than helping to preclude it. The tiz:ne i
phasing of static and applied seismin. (random) forces can either exacerbate car l
l alleviate individual support instability.
Therefore, to demonstrate analytically that a statically unstable support is' dynamically stable would i
require an extensive evaluation using large nonlinear dynamic models and timee-l history analyses. Add to this the variety of possible geometric configuratic ns l
and input motions that must be considered, as well as the existence of static i
system preload (dead load plus thermal), and the problem becomes extremely j
costly.to evaluate. This is a particularly unfavorable approach in view of the
)
potentially inconclusive nature of the results.
j For many of the same reasons stated above, any testing program developed to prove dynamic stability would also have to be very extensive. Tests which ar e severely displacement limited and sinusoidal (non-random) in nature can only prove that a support is stable under small amplitude displacement sinusoidaI c;
input. Such tests would not necessarily demonstrate stability under conditicms l
which reflect the real nature of the random input motion.
l l
System Stability i
. Generally, the term system stability is associated with the arrangement of a:
l structure's restraint configuration such that it is not possible for the strtuc-ture to undergo rigid body motion. We will refer to this as geometric j
stability. With respect to piping systems, geometric stability is assured ween I
a pipe stress computer analysis is successfully executed. This computer analysis would have detected a system of supports which does not restrain eact of the three translational and three rotational global degrees of freedom.
l Encountering such a geometrically unstable system is an extremely rare ~ situation since almost all piping systems contain some type of anchor (e.g., equipment:
)
, nozzle, penetration, structural anchor, etc.).
When discussing system stability as it relates to pipe support stability, thee najor concern is the ability of the piping system to provide the appropriated l
stabilizing restraint for each support. This type of global stability can araly be assuied if each support is individually stable in its own right, either through its design (supports of the first category) or by adequate attachment' to the pipe (supports of the second category).
If individual support stability is not assured, system stability is not guaranteed. The instability of one support can trigger the progressive instability of adjacent supports by causing the limits of the forces and displacements to which the adjacent supports were originally designed to be exceed.ed. This may result in the formation of plastic S
k
~~..n-.=..
=..
. = ~.. - - -.. -..
w
.- a
=-.
y m[-.n
. E5_b-Mr. J. B. George February 19,.1985 Page 5 hinges in the pipe (due to overload) which in turn may develo~p into a collapse
~
mechanism. This situation would not, however, prevent successful execution of a linear, elastic pipe stress computer analysis.
~
Demonstration of system stability by removal of an unstable support from the system and subsequently showing that each remaining support can resist the new forces is not sufficient by itself.
In addition, it should be shown that removing the unstable support does not affect the stability of other supports.
That is, overall system stability should be reevaluated in the absence of the
- removed support.
Commentary on TUGCO's Position Cygna has reviewed the Reference '(3) Affidavit using the criteria. described above. The Affidavit (pages 2-8) discusses system stability and its relation to individual support stability.
In it, TUGC0 states:
"In-addition, if the total support schene does not provide proper T
multidirection support required by the piping configuration, the i
. analyst will be unable to successfully rurt the piping) analysis computer program (see Tr.12025 (Bjorknan.dtestimony).
In sunmary, the piping analyst assures the stability of the piping system by limiting deflections, which negates any need to assess stability.
separately."
\\
Cygna agrees with the first statement' since this is our basic definition of geometric stability. The second statement, however, does not follow. A piping analyst does not limit oeflections to those required to assure system stability, since, 'in general, these deflections are not known. Rather, the analyst inputs each support as a restrained node and reports the resulting deformations to the designer for consideration. Therefore, the issue is not piping system stability, but rather the stability of the individual support itself.
The key point is whether the individual support can resist the applied load within the initial eccentricities and displacement limits imposed upon it.
~
The stabiiity issue is best-illustrated in Figure 1(c) of the Affidavit, (page 4). The concern is not whether an adjacent support can provide a horizontal reaction component (since it is already known by analysis that it can and the system is geometrically stable), but rather whether the clamp'(U-bolt) can provide sufficient resisting forces to prevent rotation of the clamp (U-bolt) about the pipe or slippage along the pipe axis.- If the clamp (U-bolt) cannot provide sufficient resisting torque, the individual support is unstable and system stability as well as progressive support instability,must be re-
~ evaluated.
0 4
l
--... ~..
....... -... -. ~ ~.. - -.- --
NM Mr. J. B. George February 19, 1985 Page 6 l
Of. the specific support. configurations discussed in the Affidavit, the most unique is the box frame with zero-inch gap attached to a single strut or snubber (Affidavit, page 9). This is ~ unusual because it relies solely on the ret.ative thermal expansion between the pipe and frame during normal operation to create clamping forces. The resulting frict'ional forces which resist support restation around the pipe and translation along the axis of the pipe would stabilire the support. The lower bound'value of stabilizing frictional force which extsts over the operational life of the plant was never determined either analyt:1cally or by test. Furthermore, since clamping forces do not exist at ambient condi-tions, it is possible for the support to move (rotate and translate) reistive to the pipe. This movement of the support could be caused by normal vibratten during start-up, operation or shut-down, combined with pipe thermal transGation compatible within the rigid' body displacement envelope of the support.
~
Subsequent to this movement the support may be in a position on the pipe which is outside of the displacement range for which it was designed and for wtrrich stability could be assured. Furthermore, due to the compatible rigid badly motion of the pipe and support, the support would be unable to restrain I::he thermal movement (load) for which 1,t was designed and adjacent supportsMould have to resist this load -- a load for which they were not designed. Thts a
situation may also develop 'at temperatures above ambient since the mainteenance of zero gap over the life of the plant could be difficult to achieve. Fear these reasons, Cygna classifies these supports, without modification, as unstatale.
In Figure 4 of the Affidavit (page 13) three methods are 'shown which have been utilized to modify the box frame supports to improve their stability. TWo of these methods, " indexed lugs" and " additional struts" only provide rotatSonal stability. They do not prevent translation of the support along the axis of the pipe with time. Therefore both of these modification schemes result in scupports which must still be classified as unstable. The third modification scheme, the addition of cinched U-bolts, can prevent both rotation and translation. of the support provided it can develop sufficient lower bound clamping forces.
Since the final evaluation on the use of cinched U-bolts has not been completedt, the acceptability of supports with this configuration remains an open issue act this time.
Cygna classifies all single struts with U-bolts and a thermal gap (Aff.idavit, page 15) as unstable since the stability of this type of support has nevesr been analytically or experimentally demonstrated. Cygna understands that ali of these supports have been modified in an effort to ennance stability (Affndavit, page 18). These modifications consist of either cinching the U-bolts or adding supplementary steel that would prevent the rotation of the U-bolt crosspffece.
Cygna believes.we have addressed those supports for which supplementary ssteel was added to create " stability bumpers" in Reference (.4) Observation PS-012.
Cygna found these bumpers unacceptable since there were no calculations tuo e
y-1 0
my
-s-
- NieGF unm -
Mr. J. B. George Februa ry 19,- 1985 Page 7 demonstrate that they possessed sufficient strength and stiffness to naintain I
stability. The stability of the supports which were modified by cinching the U-bolts remains open as part of the U-bolt analysis / testing program.
Double strutted frames (Affidavit, kge 19) supporting two or more pipes were not encountered during any of the Cygna review phases. However, Cygna did find i
examples of double strutted frames supporting a single pipe and double strutted trapeze supports with U-bolts, which are configurations similar to those discussed in the Affidavit. As previo'usly discussed for single strutted frame's, both the double strutted frames and trapeze supports with uncinched U-bolts suffer.from the problem of not having the demonstrated ability to amintain their relative position on the pipe over time.
In addition, the double struts cannot be relied upon to resist compressive load until the frame (U-bolt) has rotated about an axis parallel to the struts and has bound itself in a cocked position against the pipe. Neither the stiffness requirements of the frame (U-bolt) necessary to naintain a stable position nor the binding forces and displacements required to restrict the unstability have been evaluated. Cygna therefore' classifierthese supports as unstable.
In the case of double strutted trapeze supports with cinched U-bolts, the most likely mode of instability is that due to rotation of the support about an axis parallel to the struts.
If the frictional resistance between the pipe and the
- trapeze crosspiece is not sufficient, the frictional bond will be broken and the entire destabilizing twisting moment must be resisted by the bending strength (and stiffne'ss). of the U-bolt binding against the pipe. Since neither the frictional forces nor the U-bolt have been evaluated for their capability to resist this nonlinear destabilizing moment, Cygna classifies this configuration as unstable.
The stability of a single strut or snubber with a cinched U-bolt (Affidavit,-
page 27) is directly related to the resolution of the issue of U-bolts used as pipe clamps. Until the resolution of that issue, which includes the satisfac-tory determination that lower bound preloads can provide the clamping force necessary to resist the factored piping loads, Cygna considers.all such supports to the unstable.
Classification of Cygna Review Scope Cygna has examined the 226 pipe supports within the ' Phases 2, 3 and 4 review i
scope. Thirty-seven. supports were identified as supports which, in the total absence of the pipe, are stable. Of the remairiing 189 supports which in the-absence of the pipe would be unstable,124 possess sufficient positive
. attachment to the pipe to ensure stability. The 65 potentially unstable supports may be classified as follows:
I 4
a.
- ~.
-,a.c _ m..A
._m.
~
x_uw,
wn= :.
- v. =.
- . =. =._ ;. a _
NiBTD 3-,
umn Mr. J. B. George February 19, 1985 Page 8 Single strut with box ~ frame or cinched U-bolt -(23) j.
Double strut trapeze with cinched U-bolt (25) i Multi-strut box frame.(8)
Single strut with uncinched U-bolt, stability bumpers; (2) t Double strut, double trunnion with cinched U-bolt (1)
Double strut trapeze with box frama (2)
Double strut trapeze with uncinched U-bolt (3) l Triple strut box frame (1)
There are two reasons for classifying these supports as unstable: 1)) the unconventional methods used.to develop the restraining forces betweens the pipe and the support, and 2) the lack of any demonstration that the restradning l
forces developed by these supports are sufficient to mintain the sup; port.'s stability. Supports which are designed with cinched U-bolts to proviide,the i
necessary positive connection to the pipe nay be reclassified as stabHe if the U-bolt testing / analysis program and the a'pplication of the results tco the individual supports in question is.found to be acceptable.
It shoulaf be noted, however, that this program does not address the stability of supportss which do
~
not use U-bolts, nor does it evaluate the twisting strength of U-boltcs used in trapeze supports.
Conclusions Throughout tihis letter, Cygna has applied a very rigorous definition of rigid body instability. Cygna recognizes from a practical standpoint that rsany of these potentially unstable designs any actually perform their intendeed function. However, we also recognize that the inability to quantify. the actual behavior which nay help stabilize the support in practice necessitates that stability be viewed under more idealized conditions. For that reason 1 the individually unstable supports identified above, and any similar conftigurations l
throughout the plant, should be evaluated using one of the followinguapproaches:
Modify to, provide adequate restraint at the, pipe /suppourt connection Demonstrate system stability in the presence of the urastable supports i
Quantitatively show that the individual supports are sstable
)
~
l i
)
1
~ 2
i
' \\Ur3-c
.--=an=
q.:nos 4"re" Mr. J. B. George February 19. 1985 Page 9
~
Please call to discuss any questions or clarification necessary since this is a complex subject.
Very +
yours.
m 4.1 N.H. Wi lams Project Manager
.... /a j b NHW cc:
S. Treby (U.S. NRC)
- 5. Burwell (U.S. NRC) 1 V. Noonan (U.S. NRC)
D. Wade (TUGCO)
J. van Amerongen (EBASC0/TUGCO)
R. Ballard (G&H)
J. Ellis (CASE)
D. Pitjott (Orrick, Herrington & Sutcliff)
J. Finneran (TUGCO) e
\\
j 4
1 S
e l
9 0
6 e
--.m
-ry._._.-c..,,-.
m..
~
a.
uw wu-
~a.
u.
..,,=
n.,r c,.
c,
' t A 73 e avag#e UNITED STATES N
NUCLEAR REGULATORY COMMISSION i
I wasMmetoN, D. C. 20066 y.nas yx,a l
MEMORANDUM FOR:
Vincent S. Noonan, Director Comanche Peak Project FROM:
Chet Poslusny, Program Coordinator Comanche Peak Project
SUBJECT:
FEEBACK INTERVIEW WITH ALLEGERS A-44 and A-73 On March 23, 1985, the TRT conducted a feedback interview with allegers A-44 and A-73 at the Granbury Inn at Granbury Texas. The purpose of this meeting was to discuss the TRT evaluation and conclusions concerning these allegers' concerns in the mechanical / piping, quality assurance / quality control, civil / structural, and miscellaneous areas.
The TRT attendees were as follows:
C. Poslusny, Don Landers, Dave Terao, John Fair, Paul Chen, Ron Lipinsky, C. Hofmayer, Bob Bosnak, John Beck, and Howard Levin.
The following is a sumary of each addressed allegation and the allegers' coments on each.
1.
AD-6--It was alleged that stability of pipe supports is not adequately considered in their design. One alleger noted one cause of the e
engineering problems is the Applicant does not have an effective technical audit system.
In general, both allegers agreed with Terao's suggestion that a sumary disposition writeup on piping stability was
'needed.
2.
AD-11--It was alleged that forces generated by friction between the pipe and the pipe support are not properly considered in the pipe support design. One alleger took exception to the Applicant's assumption that friction forces could be neglected where piping motion is less than 0.0625 inch.
3.
AD-15--It was alleged that design loads on some pipe supports are in error because the nonnal and upset loads are greater than the emergency loads. This allegation is still under study and was not discussed.
4.
AD-17--It was alleged that the Pipe Support Engineering Design Group used the wrong section property values in the design of pipe supports. One alleger said that the crossectional shape of the tube steel, and therefore the mechanical properties, was changed around 1978-1980, and that the Eighth Edition of the AISC Manual listed the new properties.
Steel fabricated before that changeover should use the Seventh Edition properties. Whether Comanche Peak steel was purchased before, or after, or both, is unknown.
_m_,.v,_
.. ~.,,.
5.
General: Safety factors, generic stiffness, and the effects of bolt hole tolerances were not; discussed.
6.
AD-3--It was alleged that Richmond inserts and Hilti bolts are not properly analyzed when used in pipe support designs. The allegers say this is especially true when calculating general stiffness. Also, A36 steel is used for bolting material; it is not recommended for dynamic applications. Also, the analysis assumes equal load sharing among bolts, which is not the case.
7.
AD-7--It was alleged that characteristic of U-bolts are not being considered when used in the pipe supports. The primary concern was that the Applicant assumed the U-bolt designed as a one-way restraint did not act as a two-way restraint. Also, the Applicant does not address bending stress on the U-bolts.
8.
AD-5--It was alleged that different thermal expansion, seismic displacements, and concrete creep are not properly considered in wall-to-wall, floor-to-ceiling, and floor-to-floor wall pipe support designs. The allegers pointed out several related factors which must be considered in the analysis.
9.
AD-18--It was alleged that pipe supports are designed with support pads welded over pipe girth welds. This had been identified earlier and corrected.
- 10. General: The alleger said the upper lateral restraint analysis was performed unsatisfactorily.
- 11. AE-17--It was alleged that the control room ceiling has several deficiencies (field run conduit, drywall, and lights). The redesign was discussed with the allegers, who had not seen the damage study or the redesign. No new comments were made.
- 12. AP-48--It was alleged that the heating, ventilating, and air' conditioning (HVAC) supports have no vertical bracing to accomodate seismic loads. The TRT engineers wanted a specific location, but A-73 could remember only the general location.
- 13. AC-75--It was alleged that no gap was provided at the doorway between the containment and the Safeguards Building (seismic decoupling). The TRT questioned the alleger as to specific location, but he was unable to recall the location.
==
Conclusion:==
A-44 stated she appreciated the effort expended by the TRT at this feedback meeting, as well as the TRT's total investigative effort.
Chet Poslusny, Program Coordinator Comanche Peak Project cc:
D. Eisenhut B. Hayes L. Shao J. Calvo CPP CPP M. Kline CPoslusny V5Noonan Docket Files 50-445/446
/ /85
/ /85,
- n. n..re m Ch