ML20199G099
Text
.
s.____
l
. '#'TELEDYNE 1
~
ENGNEERING SERVICES i
130 SECOND AVENUE WALTMAM.uatest HUSETT502264
$$171000 3360 TWX (7106 334-7508 y -
February 21, 1985 6216-7 p
t h
t V
i Mr. Vincent S. Noonan, Director F
Comanche Peak Project U.S. Nuclear Regulatory Comnission p
i 7920 Norfolk Avenue I
Bethesda, Maryland 20814
Subject:
Prelimiiiary Consulting Report on Comanche Peak Steam Electric
[
Station - Piping and Support Design j
t
Dear Mr. Noonan:
' Attached is a copy of the subject report.
Provided is a discussi6n on the Design Process in general as well as some detailed concerns (Concerns 1 C'
In addition, there is discussion on four other specific items
.through 5).
(Concern 6) which can be construed to be a result of the existing Design All of the items in Concern 6 have been raised by others and I l
Process.
have merely provided my own opinion in these areas.
There are currently a number of other issues that are still a concern to the. staff (i.e.,
[
t U-bolts Richmond inserts, etc.).
However, it is important' to recognize i
that the majority of these concerns are interdependent and cannot be addressed as stand-alone issues.
That is, the war.ious outstanding issues (not only limited to those discussed in the attached report) must be i
addressed in combination so that the overall effect on the adequacy of piping and supports can be determined.
p If you have any questions, please do not hesitate to-contact me.
)
Very truly yours,-
I TELEDYNE ENGINEERING SERVICES
[
L b
Donald F. Landers (5
Executive Vice President L
L DFL:jej attachment 8606250072 860611 PDR FOIA PDR GARDE 85-59 l
=_,_
b*
2,,,
t c
~~
L ' g W ::... -
. ~. a v-.
~"
~
,e w.
m
,g Technical Report NNES TR-62168.
In determining the acceptability of Design QA(1), two important f
issues need to be reviewed.
The first is to determine whether a Design
- ~
Process is in place and functioning.(2) The second is to de.termine whether the existing Design Process is structured so that, if followed, reasonable j
I assurance exists that the licensing commitments for a plant are complied.
with.(3)
The second issue above is the primary purpose of developing a process to control the design. Control is intended to channel the efforts of the design groups to the goal of fulfilling licensing comitments.
This, in fact, may require some members of the design staff to do things differently than they are used to.
Also it may require approaches, f
I techniques, analyses, etc., which are significantly different than the h
last nuclear power plant project completed by the design agent simply because the licensing comitments are different.
It is important to recognize that both issues must be acceptable or questions with respect to
{
adequacy of the design may exist.
[
~
Y
- ('-
. place, supported by For example, a Design Process may be in procedures, subject to meaningful audits and verification and yet be l
flawed because it does not address the licensing committments. Similarly a
[
I Design Process which addresses the licensing commiitments may be in place 1
~ but it is not functioning properly and required audits and verifications j
are not being performed to demonstrate inadequate implementation and to f
provide corrective action.
[
t I
r p
II) Note that this terminology has been used in these proceedings. The I
author does not endorse its use in the i.ontext of the concern at Comanche Peak but will comply with current terminology.
(
(2) This is essentially a review of paper. For example, proper sign-offs exist, audits were performed appropriately, check lists were complete, C~
etc.
I ) This is essentially a review of technical adequacy. For example, does f
)
the process assure implementation of a design that complies with v
applicable Regulatory Guides and Codes,
)
m m
} } _ _. 4 f
g.
~
1
,g T
1 Report ENGINEERING SERVICES,
Two examples of the above situations were found by Cygna during the g
Independent Design Verification.
The first is the issue related to g J
g ion.
Under questioning during r. January 10, 1985 meeting with
)
Td M the NRC, Cygna indicated that d
I QA audits using the process in place would at have indicated noncompliance l'
with the licensing connitments for Comanche Peak. The second is the issue
[
~
related to Under questioning during tne same h
January 10th meeting w
. _NRC, Cygna indicated that an acceptable procedure exists,in the Design P/ocess at G&H which addresses mass point
[
d.
spacing, however,
[
=
.xra p
.n It would appear that until the Phase 4 effort.by Cygna the issue i
to determine whether it r
related to controlled design such hat the' licensing commitments were satisfie g
~
7 This opinion is reinforced by the fact that, at this point
[
in time, Cygna is revising their Phase 1, 2 and 3 conclusions related to j
i Design QA.
i.
Having established that Design QA has two sides, a paper trail side and a technical side, it is necessary to look at the process in existence fo. Comanche Peak for piping and supports.
y Pipe supports and piping are so closely intertwined and technically hI interdependent that it is difficult to separate them.
In designing a piping system the designer makes certain assumptions concerning individual j
i support configurations.
Also, a piping designer usually cannot make I
appropriate.iudgements on the adequacy of a piping system without i
I i.
w.1
?
.J, ; - ;T 3.p.y.,;; _,
?."f'[*
- 7 -- ;
igqqggq,q p m.q,~gg.;.
TN Technical Report.
ENGINEERING SERVICES TR-62168 t reviewing the piping layout with all of its supports.I4) This is particu-
{
larly important when addressing an issue such as support stability since the interaction between the support and the pipe is usually critical in staking this determintion.
For example, for a pin-pin connection, the f
displacement of the piping at the support location due to operating condi-tions (thermal expansion) can result in a reduction in the ability of the support to carry a load along its axis.
Also, the concern of the author f
with respect to support stability is directed towards anticipated water
[
and/or steam hamner events which usually result in higher loads and dis-(
placements on the' piping system than does a seismic event. To accomplish the kind of review discussed above it is necessary to have an established and functioning link between the group responsible for piping design and analysis and the group responsible for support design and analysis.
[
l.
In the majority of cases a utility constructing a nuclear power plant
[.
{
contracts with a design firm (usually one of the major AE's) to provide l
design services in the areas of piping and pipe supports (along with a number of other areas not relevant to this discussion).
elect t to"T' third party;
[
however, responsibility for, and control of, the design of both pip'ing and supports rests with the AE. This responsibility and control exists even ma
---a
-~
The AE I
when the third nartv ucae ite awn n einn nA Prnesee 1
will review and approve the process and perform audits to determine accept-ability of implementation. The above does not eliminate the requireseent b
that,the utility is ultimately responsible.
f E
[
C,
'f
} Your attention is called to Welding Research Council Bulletin 300,
" Technical Discussion on Industry Practice," Section 1.7, page 26, g
December 1984.
r I
o b
u
~~
~~*
=.
~ ^ ^ - ' ~ " -
p.g.*.l - % -- J ~. * ;y.,'. -+.j Q.,' ;9..; };-[.Q.f"}Tj.
- j L_ -
- [Q.{gj 7,3,q,
o Technical Report E N N S N ES TR-6216B - In response to questions at four meetings with the NRC(5), TUGC0 ~ indicated that the process for initial design, including issue of initial construccion drawings, consisted of the following. I (1) G8iH performed preliminary free thermal expansion analysis and forwarded these to ITT Grinnell and/or NPSI. (2) Deadweight supports were located by Grinnell and NPSI using the hanger spacing table established in ANSI B31.1. i Potential locations and directions of seismic restraints were h established by ITT Grinnell and NPSI. Guidelines for spacing these restraints were established by G8H and were based on frequency considerations. f a L (3) 68H then performed piping design and complete analysis, including location and selection of the type of pipe sup - 4 ports. This required the normal iterative process of layout, C.. analysis, support location, modification of layout, analysis, - y etc. Eventually a design evolved that analytically complied i)r with the licensing commitment. [ ? = (4) Support locations, types and load combination data were sup-plied to ITT Grinnell and NPSI. (5) Support details- (including selection of standard hardware) I were developed and support analysis perfomed by ITT Grinnell i and NPSI. Cases could arise where the location of a specific support for the specified loading was not acceptable (i.e.,' [ f an adequate design could not be reasonably developed). In such cases the support contractor would inform GliH and i C. r another iteration in the piping analysis process would occur. [ r (5) August 9,1984, January 10, 1985, January 15, 1985 and January 17, [ 1985. i .)- I .. ~.
v~ j~ - - s m..., a, v. L..- ~~re A?. y:P' Jgfffgy:;j.jQQg[};[*}g g. y... WTELED(NE Technical Report SEE 5-TR-62168 (6) Design and analysis was completed and supports were fab-( ricated and shipped to the site. Review of the support f details at G&H was snot required at t!iis time in the design I: o process. c (7) Modifications to supports required by field conditions were made by field engineering (Texas Utilities responsibility) and a Component Modification Card (CMC) was executed. i-f. (8) The CMC was forwarded to the responsible support design agent p I (ITT Grinnell or NPSI) for review and approval. L L (9) A third pipe support group (PSE) was formed which was under the technical direction of TUGCO. This group functioned just as ITT Grinnel and NPSI did although the engineering and administrative procedures differed between the three {' organizations. i. i (10) Also in this time frame, ITT Grinnell and NPSI sent support designers and analysts to the site to perfarm design, anal-ysis, modifications, and review of CMC's. These ITT Grinnell [ and NPSI personnel were administrative 1y controlled by TUGC0 ~ but utilized their own procedures in performing their L required tasks. For ITT'Grinnell these, procedures were the o same as those for the home oft' ice. NPSI developed specific h procedures to be used by their personnel at the site. t (11) Any of the three organizations who had concerns with a CMC f. informed the initiating field engineer of that concern in 'a Technical Services Design Review (TSDR) memo. j { i (12) At a point in time when the pipe was installed and Brown and Root (B&R) felt confident that the support as designed or t. ll N mw -- ,e,,% -'.-w--- ,ryg 9 g,.
is
-rww.g .e. +
= _ m.x,= a -- r a n. c =. k[$ hINki;. b _.. Ehh.hg]{.g.pf;;gg :Qg}lg,i].g;_}/,. in i .x.n. -.-.. _, c.2 W TELED(NE gggg $gg Technical Report TR-62168 u modified would be able to be installed, an as-built walkdown (. was performed by TUGC0 personnel and a package forwarded to [ EH for their review, reanalysis (as required), coments ll EH coments or concerns with as-built f and/or acceptance. ) condition were transmitted to TUGC0 in a E H memo. 0 0 [ (13) After piping reanalysis and determination of new loadings, h the responsible support manufacturer would be supplied with the new loads by EH to be used in their review, reanalysis. [ [ coments and/or acceptance, of the as-built support con-figuration. For cases where piping reanalysis was not required, the support designer would raview, reanalyze, com-ment on and/or accept the as-built configuration. p F The documentation from EH and the support design organiza-(14) tions was then forwarded to TUGC0 who reviewed the documenta- [ { tion and stamped those supports which were accepted by the [ [ support design organizations "as-built certified." [ basis until all [ ifterative (15) This process continued on an piping and supports were accepted M l s gr EH in their review of as-built. information was responsible (16) for acceptance of the niping system (piping plus pipe sup-I ports)ascomplyingwiththelicensingessnitments. t a As indicated, the Design Process at Comanche Peak was, modified as p This is not the project evolved from design to design and construction. unusual in the construction of a nuclear power plant, and a description [ of the current process exists in the Applicants Sumary Disposition on g {-
- L L:
Design QA. f-ith the process described above and [ These concerns do f with some~aspe so .mp ementafion of that process. r i i.
- ~~"
- - - - - ~ - -., _ _. -y
_...?... W* M W.'.d, k J5>.6'*6:.1 - & a.c.w;. ' w= \\- 1;2
- (5j'$ : )$.((i;' L 0 Y',
~ 4%$$ L- ~ _.x } "WTELEDYNE $_$alReport. EN@MNG SNES [ C I w not necessarily result in a conclusion that the process or implementa- { tion is sufficiently flawed to result in a design that is not in com- [ pliance with NRC safety criteria or the licensing comitments of TUGC0 b l for Comanche Peak. The concerns are as follows: g Concern 1 g r i The failure of the Design Process to require G&H to review designs [ (and modifications) of pipe supports prior to fabrication and installa- [ tion can result 'in a situation that is of concern. Piping is not a [ " stand-alone" comodity.(6) A basic premise in designing a piping F system includes (but is not limited to) the fact that support designs [ f will reflect the assumptions made in the analysis of that piping. This is of particular concern to the author as it relates to anticipated ( steam and water hamer resulting from plant operating transients. Since [ (. G&H was not required to (and therefore did not) review support designs i prior. to their fabrication and installation th are al s dealin with e i s s not e cons rued as a ement th s occurre anche [ Peak nor is it to be construed as a judgement on the adequacy (safety J l significance) of the design that exists at Comanche Peak. t I Again, my major concern is related to anticihated transients such I as steam hamer resulting from a turbine trip or water hammer resulting I from pump switching and rapidly closing check valves. With respect to [ l seismic loading it is my current opinion (based on the data available to [ ~ b l P (0) G&H agrees with this in footnote 13, page 17, of sumary disposition. t r ~ - - T - z --3
g- .~~jfyQ. app .z-;. :3..g, y T* w ,x. l W TELE 7VNE l 'E. D me) that the existing supports will be adequateN. This is based on the fact that the CPSES piping was designed using lower damping values than are currently permitted. Use of PVRC damping has resulted in ft reductions of peak accelerations of up to 50% with general reductions on j the order of 35 to 40%. Further, test data indicates that piping l systems with supports that are
- flexible, have gaps and pinned f
connections usually result in higher damping since a signif.icant amount of energy is used up in deflecting the restraint, closing gaps and e moving about the pinned connections. g 1 Concern 2 r The use of nomographs based on frequency to locate seismic { restraints usually results in an excessive number of restraints. This F f approach was used at Comanche Peak and apparently resulted in excessive i 1 ' seismic restraints. This is verified, to a degree, by the fact that a p (. majority, of the seismic restraints are very lightly loaded. Lightly ~1oaded restraints which are designed using a deflection criteria (i.e., f 1/6-inch maximum) are usually very flexible. Flexible restraints have } been a subject of concern at CPSES. Concern 3 l I The stability question has resulted in a number of analyses and {j some modifications to supports. In one area, on the main steam system, bumpers were added to prevent rotation of the support about the I t to provide pipe. [ ] C m Those restraints which are pinned vertically and have bumpers for h out-of-plane displacement control are an exception and are discussed h in Concern 3. l 4 I e v ____t,___ _ _. _ _ _ _ _ _ _ _ _
. a. 2: _ y ;; ;--:- .p 3. n :x ;- g 4,. ?- _ [^ .".;l};; _. -..-,,y]& i:' Tex':;422W: -ins-)S&h
- ;L
,g ., c,_ ~ ~ ~'yt' . - c -.n. , _ rg. l it: r.;64;tM: e:. n : j ~_ ' L-L, ; + WM Technical Report. MGIMNG SNES [ TR-6216B.I a stability.'8) TUGC0 has performed seismic analysis with the supports in (., place and with the supports removed and the rrsulting stresses are acceptable in both cases. However, the supports are still in place and, according to Cygna, will not function. M which could be that the hasbecome""g / or-id then a dynamic load is [ ~ support applied to the system. Does the tilted support provide restraint in a h direction that was not intended? Once tilted does the suppor. restrain thermal expansion? To assume that a support is acceptable because it is analytically not required may not " bound the problem" in every case. This would also apply to a support that was overstressed. To perform a piping analysis without the support in place and demonstrate acceptable stresses in the pipe and other supports is not always the worst case unless support failure is complete (or the support is physically [ removed) and does not impose a restraint on the system that was not ( accounted for. Concern 4 n ^ i a Noncon ormance eport (NCR) is not the only document for accomplishing [ i this. Examples of.other techniques used in the past are a Field Change Request (FCR) and a Drawing Change Notice (DCN). [ ace. I process underwent an evolution as plant construction activity increased. The following discussion addresses the process from its initial to its i r final stage as now understood. h f l s (8) January 10,1985 Transcript, pp. 72 and 73. a -. ?
a. ,+. t.::.- ... r: .c. 7,C",G ~;-TY*p.+ y x.:= ; &:.; g:: m:;. q s " -p %g gg g ;g. n p.,;m3 Pr w wn f-y.,. ;.. 1 m 1 -;.- - : ; ;3 - ..y_ . py WTELENNE Technical Report - MN SNE TR-62168 _ l In the initial stages (and for some time) CMC's on supports were gjf $jf h-generated by the Field Engineering Group (a subgroup of TUGC0 Pipe Support ' ~ ' ' l'I Engineering - PSE) and were forwarded to the organ.zation responsible for y 4., d that support (ITTG and NPSI). The CMC was placed in the system file by I f$$ ITTG or NPSI and would be worked on as the piping system required rework or - ktg h asTUGC0 requested.(9) This resulted in construction of the modification [4 I $ continuing without review by the responsible design organization. In some 'N I AM" cases, as-built analyses performed by G&H could have included supports i with outstanding CMC's although the appropriate CMC would be included in the as-built package. Based on the defined process, this would mean~ l s that the affected support would not have been approved by the l appropriate design organization at that time. However, the su'pport design organization was also involved in the as-built process and review i of the support would have been accomplished as a part of that process. One could suggest that a method of controlling the number of outstanding ( CMC's on a given drawing (say 3 to 5), or controlling the time that a CMC can be outstanding, would force review, approval (or disapproval) and incorporation of the CMC into the drawing. This would reduce the j turnaround time for approval and reduce the number of outstanding CMC's l [ i in a given as-built package. i Eventually, a site group was established under PSE which included ITTG and NPSI personnel. Under this organization CMC's were dispositioned by the PSE group on site. This shortened the comunication link and should have resulted in mori rapid turnaround of CMC review. Houever, no change to the process occurred. (i.e., time limit on CMC or limit on outstanding number of CMC's on a given drawing) except that the field engineer, who authorized construction to make a change to a support, had available, on site, the complete design C. resources of ITTG, NPSi and PSE. I t t III January 15,1985 Transcript, pp. 30 and 31. f t n'**'- +,mbs, _,W,d ,,e ,ym_., a
~ ^ I. :.. ~ ' '~ - 2.... - lf t '-' SW % i. m?... ,l.. in:::...% %.WC lr.?A:. ll
- n..
[2&%..
- ~7*.
- ~
i'Q.R ci Technical Report SNE TR-62168 i b When I1TG, NPSI and PSE reviewed a CMC and found an unacceptable condition (i.e., stresses too high) they generated a handwritten memo (TSDR) noting the condition. This TSDR was sent to the field engineer responsible for generating the original CMC. The field engineer would reply back to the originator of the TSDR (on the original [ f TSDR in a section set aside for a reply) noting the changes now recomended for the support can be found in the next revision of the [ CMC.(10) The support design organization was now. responsible for i F reviewing the next revision of the appropriate CMC. [ g i One area ol concern with respect to QA control is that' CMC's were f i handled by the site document control center and those individuals on the [ effected drawing distribution list received a copy of the CMC. Copies of the TSDR's were not controlled. There does not appear to be a definitive link between QA and design in the area of CMC's and absolutely none with the TSDR's. Therefore QA could only determine that h C'.- changes to design were occurring if they performed audits (which they l L did) and reviewed both the CMC's and the TSDR's. This need not be a real area of concern in the initial design stages where construction was not underway, however, once a construction drawing is issued it is j important that QA be aware of changes that are planned to that drawing. [ L This is particularly important when those changes are already being built. QA can be effective in recognizing repetitive design changes and developing trends and then modifying their audit plan and schedule to i focus on the affected areas. TUGC0(Chapman)statesiIIII f " Applicants have established a procedure, CP-QP-17.0, " Corrective Action," to review documented conditions adverse to quality for the purpose of providing corrective action to i preclude repetition of significant conditions adverse to quality. This procedure provides for Quality Engineering Staff to review design changes documented on CMCs. The i I f (10) January 15, 1985 Transcript, p. 46 and Motion for Sumary Disposition, July 3, 1984, p. 53. [ (11) Motion for Sumary Disposition, July 3,1984, p. 54 f f
[., ; L.QJ;&ggggggg.gy - .pynk:4.js]%;4kkg W TEU M NE s_galReport, ENGINEERINGSERVICES i < f t t results of these reviews are -tracked using trend analysis j h techniques as an objective method of ascertaining the need for l corrective action to preclude repetition of s1 'ficant conditions adverse to quality. Periodic reports sumarue the results of the reviews, including
- trends, and provide I:
recomendations, where appropriate, for corrective action with I respect to identified conditions which are considered to be l } significant. [ l This is appropriate, however without receiving copies of TSDR's it is not clear that trends of field engineer ing to propose inadequate h ~ changes to design are not explicitly covered.unless one assumes that the f revision to a CMC resulting from a TSDR defines that the reason for the l revision was either a TSDR or a request by the reponsible design organization. ( ? i Concern 5 EH had a Site Stress Analysis Group (SSAG) at CPSES that was l administrated by TUGC0 but reported to G&H. Mr. Ballard of EH states:(12) "SSAG was established to evaluate and approve ( proposed changes and modifications to pipe routing, pipe support locations and/or pipe suppoft type, as requested by site engineering groups. The l evaluations ara made employing the latest as-designed piping stress analysis. SSAG provides g revised design information to the applicable site f {.. organizations. All these activities are conducted [ t (12) Motion for Sumary Disposition, July 3,1984, p. 20. i ) I ~
.- +. _ _. a. a.- _w: .=l .sf }fifjQyjjf1l}.iQQ -;_a~;_ gQfQ-Q.[ g, WM f 1 Technical Report SNE TR-6216B in accordance with CPSES Engineering Instruction CP-Ei-4.6-9, Rev. 1, entitled " Performance {'. Instruction for Piping Analysis by SSAG" and Gibbs & Hill Applied Mechanics procedures previously cited. i These documents have been established to assure that f the SSAG activities are accomplished in a manner' l comensurate with the original as-design analyses." i The concern here is related to the fact that SSAG performed their j function "as requested by site engineering groups." It is understandable that a modification to a pipe routing of considerable magnitude would have been routed through the SSAG. It is assumed that this was accomplished through the use of CMC's as discussed for supports in Concern 4. However, a major modification to a support which could have an impact'on pipe stresses may not be routed to the SSAG since the indiv'idual responsible for generating the CMC may not have considered (or recognized) the change would effect pipe stresses. Concern 6 t i The following are discussions of those items which are specific in nature and yet tell us something about the design p'rocess. - 6.1 Mass participation This issue is addressed in introductory remarks (see page 2) and is important from a design process standpoint and a support / pipe adequacy standpoint. Based on the Cygna review it appears that the average mass participation of. piping systems analyzed by G&H is in the order of 40%.(13) One could expect that a seismic analysis cut-off at b-- t (13) January 10,1985 Transcript, p. 70. b i f - ~ ' ^ ' ~. 'I-. " T ? ^^:~ ~ r r - ~~
= .=- w n.,. w. n 2 = 'L h5h -;[ h:hl.. ?. ,.f:hiQh.~j%Q(('l.Q\\hQ. Q-Q y i r.. TM 1 Report gggg l 33 Hz should normally result in 90% or greater of the total system mass. For, piping stresses this would usually be acceptable. For supports how-ever the contribution of rigid mode response could be important, l particularly' for sup;iorts located close to large concentrated masses or [ where the support is providing axial restraint. In these cases the total seismic load should include a rigid mode component equal to the floor ZPA times the weight of the supported component or sepent.of [ l pipe. Based on normal expectations a mass participation of 40% is [ unacceptable. Further, the design process at EH did not control this [ effect since a procedure was not available. 4 Reanalysis by EH to include total mass participation will result in significant increase in some support loads. This efect when 1 coupled with the low support stiffness (flexible restraints) could result in the need to modify supports (see Concern 2). 6.2 Support stability In addition to the discussion under Concern 3 which addresses some specific restraints there are some gener'ic concerns. Many of the restraints and supports at CPSES utilize box beams with either pinned p struts or snubbers connecting the box to structure. This is not' a j comon design for seismic Category I nuclear piping. Box beams them-selves are not uncomon, however they are teually rigidly connected to the building structure using standard structural siapes. A second type t of support that is of concern is the trapeze style support which is com- [ posed of a structural member supported off the building structure by pinned struts or snubbers and attached to the pipe by a U-bolt 'or h trunnion. Again, this type of support is not a comon design for [ seismic Category I nuclear piping in plants licensed to operate in,the last 4 or 5 years'. (Trapeze type supports ith U-bolts can be found in y L non-seismic piping at nuc1' ear plants and ' in other facilities such as process and fossil plants.) A third concern is related to support i f )
- _s y=-
.~ ; ;i: a b
- n..w.
~ Z - =_,.2 .:v.:;llr n x .-wec., m%h> J Ti;:;Uk;,,;GB].y;n;c,;y. ;n - y.ya...:.w.:t- -~ ' m --- l ;.i y. y w ; j ys;. i t W i.j Q g;jr M n ~ .a ,fSN.,d32M3 J
- ; _._.,.,i y -
T Technical Report. N S N ES TR-62168 application. That is, the use of struts or snubbers suppcrting a pipe ( from the bottom of the pipe to a floor or platform below the pipe. [ Since these supports are pinned they are unstable vertically as soon as [ 1 horizontal displacement of the pipe occurs and system stability is pro-vided only by the end conditions of the piping system or any horizontal-restraints that exist. It has been pointed out that piping must be con-sidered in conjunction with the existing supports and therefore the [ presence of pinned supports applied in the manner described above must I be judged based on the overall support system. / 6.3 As-built reconciliation The as-built reconciliation process has two functions. The first, and most obvious, is to take dimensions, etc., of the actual r i as-built configuration of piping and supports and reconcile those with ( the as-designed doc eentation. The second is to have a qualified piping i designer walk the system to develop an understanding of the ~ overall i geometry and to determine if the installation generally reflects the I analysis. The importance of this second step is obvious, the overall . configuration is there to see and one is not dealing with a neber of i j ) different drawings trying to piece together a system. l The existing design process at CPSES required as-built, information to be gathered by TUGC0 technical services personnel and forwarded to G&H applied mechanics personnel. Already the ideal situation where the G&H analyst or members of the 55#.9 walked the system did not exist. However, this is not a fatal problem nor is it uncosunon in the industry to have "others" gather as-built data. It merely makes i the problem of system acceptance and analysis reconciliation more dif- { ficult. The as-built reconciliation program was started at the time that the piping was installed and Brown & Root determined that the i 6 t I r= .=.=.:=----:= =-.- - --=.. ...:=- :. u :: --- -
= m m w :.. f. - ~ hl6 j fs '; 3 l _ j 'u? v;: l 1.Y, _ [, s
- y.,
b i o TM r 84GINEERINGSERVICES t supports not in place could, be fabricated and installed as they were [ h designed. The number of installed supports on a given stress problem varied from 20% to 80%(14) at the time G&H started reconciliation b efforts. Having only 20% of the supports installed has two impacts, one [ that could be positive and one that could be negative. The positive l L impact is that with only 20% of the supports installed the G&H analyst I i should have had an early indication of what the support designs looked like and could have requested modification (if there was concern) prior l to fabrication and installation of the remaining 80%. That is, the undefined pressure to accept constructed supports was significantly less than one could hypothesize for the situation where all of the supports were installed. The negative impact is that the piping analyst is'not dealing with the complete as-built system and one can anticipate that a - number of iterations will be required to complete the reconciliation { process since modification to one of the remaining 80% of the supports { could impact the total system including the installed 20% of the i supports. Iterations such as this are not unconsnon but sometimes tend to result in cursory reviews of already accepted situations. One major concern with respect to as-built reconciliation is the situation where more than one piping system was supported by 'a frame, particularly frames which were pinned connected to building structure. G&H, though aware of the fact that the frame supported more than one system, dealt with the support as a single support on the piping system under consideration at that tief.(15) The support designer was supplied with the loads on the frame for each piping system No being supported and determined the structural adequacy of the frame. one was apparently respons'ible for looking at the interaction effects inherent in a pinned frame supporting a number of pipes. It is my G (14) January 15, 1984 Transcript, pp. 22 and 23. h (15) January 15, 1984 Transcript, pp. 23 and 24. ( I y.__
- A_.? ' S ^ St :~J i. ~ 4.,. .g z._. m 1 ? YYr .ll, !! k_., _..f .f e -). .h L. ' _ k
- i..
?.- 'a ll o WTELED(NE Technical Report gggg TR-6216B
- i opinion that this is the responsibility of the piping designer and G&H l
{-- acceptsthatresponsibility.(16) il r L i 6.4 Support mass h ? f Many of the support designs at CPSES result in considerable f l mass which is not acting at the outside diameter of the piping. It is l comon practice to add support mass to the piping analysis and this is usually done at the centerline of the pipe since it normally involves a j clamp. In the, case of a box beam rigidly connected to the building structure the mass is not applied to the pipe and therefore need not be j considered. In the case of a box beam pinned to the building structure 1 the mass acting 90 degrees to the direction of restraint should be applied to the pipe centerline. A specific geometry that cannot have the mass applied to pipe {' and be representative of the as-built condition is a support centerlina, restraint that is pinned to the building structure and has a beam sone distance from the pipe G and the pipe 0.D. The beam is attached to the pipe by welding a trunnion to the pipe and the bene.(17) The effect of i the offset mass rigidly connected to the pipe results in forces and moments on the pipe which will not be represented properly by modelling l the mass at the pipe centerline. TUGC0 apparently accounted for this effect on the main steam system onlyhflowever, there are sine x s concerns with the approach used in that instance. - 3 q ~ (16) January 15, 1984 Transcript, pp. 11, 49 and 50. N ) This would normally be called a trapeze restraint but if used as a f ( horizontal restraint on a vertical pipe that could be a misleading i statement since a trapeze support is normally considered to be a [ 2 i vertical support on a horizontal pipe. f Applicants Motion for Sumary Position Regardin Allegations Concerning Consideration of Force Distributions in Axi!1 Restraints, I j dated July 9, 1984. -, -.,,. ~.,- ,-m, ____,.-,,__,_,,_,_,___,m_,_m_.,, ,,,.,,,,. _ _ _. _., _,.,, _.,,,,. ~,,, _,,,,,,,.,
- "
- a...
~ r - f 7' 'h, l',
- 5
- }
J.;' g. }* gW) f eTELED(NE I T 1 Report MGemSNES '. The concern with trunnion (stanchion) type restraints is the I f {. following: r t (1) What are the stresses in the pipe wall due to the forces and moments generated as a result of the trennion? (2) What are the stresses in the trunnion? (3) What are the loads on all of the support components (i.e.,
- strut, base E,
- welds, etc.)
and are they acceptable? TUCCO has pointed out that the analytical techniques (response spectrum analysis) do not consider the relative phasing between axial and rotational motion. Howeve: for all supports (Figure 1 of summary disposition) the rotational motion is a result of the axial displacement of pipe. Therefore for a Type 2 restraint the axial loads in the strut or snubber resulting from axial pipe loads and moments generated due to the trunnions should be additive. Also the applicant states the following on Page 7: 1 i "In other words, that rotation cannot exceed the value which would occur if there were no rotation constraint." 4 I certainly agree with this and would suggest that this would require TUGC0 to either limit the stresses in the trunnion or model the support as a three-pin restraint. That is a pin at the connection with building structure, a pin at the strut / snubber connection with the [ trunnion and a pin at the location of yielding in the trunnion (most b likely at the. pipe-trunnion weld). It is obvious that such a support ~ (Type 2) would not provide any restraint in the direction of the pipe unless a significant pipe displacement had occurred. f l
- y
_y,, g
h.y $ ..[,;.[.( $ h,N5 27.1._ i,1 M9.y[hpgff,$pdyU.Q.<@.. _1..g. -....Q[ @ M.Q M ipj g .'. u.2 N j ,.2.
- g j j /.q, / f 9,6 f p f TM Technical Report gg TR-6216B,
y h l{ Of greater concern is the water / steam hanner loading which can j. result in loadings higher than that for the earthquake. For the main [ 1 steam system it is quite probable that an earthquake of the magnitude of j the 08E would result in a turbine trip. A turbine trip generates g I dynamic loads in the main steam system due to the pressure wave gen-I erated by closing the turbine stop valves traveling down the pipe. The loads due to this condition should be combined with the earthquake load-j i ing. No evaluation has been presented to demonstrate the adequacy of f i these type supports for either water / steam hanner loading or a combina-I j tion of seismic plus water / steam hammer loading. With respect to lug type supports the same concerns expressed r In attachment 1 Pipe Lug Elastic-Plastic Analysis (18) the above exist. j applicant states: "As stresses exceed-the yield
- strain, the f
stress-strain is no longer linear but changes with the increasing strain level. In a I load-unload-reload loading pattern, it is observed I that the new yield points occur at different stress I levels. This behavior is called strain hardening." h [ Here again the applicant has ignored the dynamic load associated with l steam / water hanner which does not follow the load-unload-reload pattern. Strains of the magnitude specified result in stralises which exceed the [ a'llowable requirements of N8, NC, h0-3600 or ANSI 831.1. It should be noted that in Paragraph 121.3.2.8 of 831.1 the allowable stress in welds [ attaching lugs or trunnions to pipe is limited to 80 percent of the [ allowable for the remainder of the support. For N8, NC, ND-3600, the ( l stresses in the pipe should comply with the requirements for piping as [ {'. defined in Code Case N-318-2, N-391 and N-392. [ n 1
ih - ( d,: ^. ....,i.G N )g f.[')$$ 4syt3}. ;.glM$$. ydd fM "#PTELED(NE . Technical Report NMGSNES l y. TR-6216B REC 04MENDATIONS TO THE STAFF ( { Reviewing the items discussed above, as well as those identified by { others, one finds a list of concerns related to supports, piping anal-f [ ysis and support-piping interface. Some of these are: I i t 1. mass participation, r 2. node, point spacing, i 3 3. support stiffness. f 4. friction loads, t' 5. Richmond inserts, 6. U-bolts. 7. support mass. 8. axial restraints and 9. support stability. t ( Concern with one of the above items, or'even two.or,,three, may not necessarily result in an overall concern with respect to compliance with licensing commitments. However, when the list is viewed as a whole and when the interdependence of items is considered, a different perspective resu'lts. The interdependence of the above list is an important issue since the resolution of one issue may result in failure to comply in ( The, nterdependence of mass participation and support i another area. stiffness has already been mentioned. The adequacy of Richmond inserts and U-bolts is a function,of the applied 1 ad, and' items-1, 2,' 3. 4, 7 g l and 8 can have an effect on the applied 1c. d. Support stability is, to [ I a degree, a separate issue only because it would exist even if the other I eight items were resolved. However, the applied loads and therefore ~ displacements (including rotations) of the piping system will have an i j impact on determining support stability, C r Another concern when looking at the existing situation at CPSES and } attempting to make a decision on the adequacy of the process is the h e e l ( ~'
~ a nw m N '$ N. ' .~. l =* -J YM g g g'g Technical Repor.t . TR-62168 l q h approach used by the applicant in addressing concerns, either in the p In most_of these f l form of Summary Dispositions or study-type analyses. {* cases the applicant has provided analyses which are well beyond that used in the normal design process. A typical example is that discussed t in Concern 6.4 related to trunnions and lugs. Having yerformed these i ~ i " state-of-the-art analyses" has not resolved the issue in some cases [ (i.e., trunnions and lugs, Richmond inserts and support stability). I With respect to the Design Process, any flaws appear to be limited f to interfaces with the exception of EH. The design process in place at l [ ITTG, PSI and PSE was acceptable if external interfaces are not con-The checking and verification of designs and analyses are com-l sidered. mensurate with that generally utilized in the industry. The only. [ [ l exceptions to this that exist to spy knowledge are those related to mass In the first case the l l participation and node point spacing at EH. l process did not address the issue (mass participation), in the second case checking and verifiction did not catch the failure to follow the It is not an j procedure required by the process (node point spacing). [ l essential requirement that each step in the computer modeling or For example. l interpretation of results be delineated in a. procedure. individuals experienced in piping dynamics should have recognized the f F mass participation and node point spacing problems without a procedure. With respect to ITTG, NPSI and PSE, the fact that the list of items i of concern contains five items that.are ;suppErt related requires l Many of the support designs for CPSES are not consonly [ evaluation. found in commercial nuclear power plants. This is not in itself reason [ f for concern but leads one to review the design and the supporting anal-f ysis critically since industry standards or experience cannot be totally 4 relied on. f3 L Based on the above a decision concerning_ the adequacy of the design at CPSES cannot be reached. It would be necessary to review a set of ) m I TMZ .~_..____,,_.__.._._._.__J_.Jr_1.____. . _ _._--. -_ _-. _ Z Z L,,, _. _.. _.
W TELEDYNE Technical Report ENGINEERING SERVICES c TR-6216B [ analyses which are representative of the as-built configuration, con-sidered ~ app'ropriate load combinations, evaluate all components (piping and supports) in accordance with the CPSES licensing comitments (and { provide acceptance criteria for those items such as support stability not covered by published Codes, standards or regulations). If the f sample size were sufficient and included all those items of concern. t listed, then one could reach appropriate conclusions. In addition to the above, any set of analyses performed should con- ~ form to the following: 1. use as-built geometry and hardware representative of as-built pipe and supports, i. 2. include offset support and mass effects, ( 3. include appropriate mass participation, t i 4. use acceptable node point spacing, i E 5. include actual support stiffness. r 6. use time-history analysis for steam / water hamer loading, 7. apply controls to the design process recomended in WRC Bul-latin No. 300, i [ l any new analysis shall be considered the analysis of record [ 8. rather than a study, and t (. ~ 9. comply with current licensing commitments for CPSES with respect to acceptance criteria. l l
L ~ .-._w _x ,L -a = .i s t i )' .y DRAFT 0 Technical Report TR-62MB I In determining the acceptability of Design QA, two important issues need.to be revi,ewed. The first is to determine wnether k Design Process is 1 in place and functioning.2 The second is to determine whether the existing Design Process is structured so that, if followed, reasonable assurance exi.sts that the licensing. commitments for a plant are complied with. The second issue above is the primary purpose of developing a process to control..the design. Control. is intended to channel the efforts of the h design groups to the goal of fulfilling licensing commitments. This, in a fact, may require some members of the design staff to do things differently than they are used to. Also it may require approaches, techniques, analyses, etc., which ar'e significantly different than the last nuclear power plant project com'pleted by the. desjgn-egent simply because the f licensing commitments are different. It' is important to recognize that I I both issues must be acceptable or questions with respect to adequacy of the i design may exist. I i For example, a Design Process may be in p.l ace, supported by t procedures, subject to meaningful audits and verification and yet be [ flawed because it does not address the licensing commitments. Similar1'y a [ Design Process which addresses the licensing commitments may be in place l but it is. not functioning properly and required audits and verifications are not being performed to demonstrate inadequate implementation and to f provide corrective action. t y 6 I Note that this terminology has been used in these proceedings. The author does not endorse its use in the context of the concern at Comanche Peak bLt will comply with current terminology. 2 This is essentially a review of paper. Fer example, proper sign-of fs exist, audits were performed approprihtely, check lists were complete. { etc. 4 3' This is essentially a review of technical adequacy. For example, do6s the process assure implementation of a design that complies with applicable Regulatory Guides and Codes. ye. -.a me+ -m m,nwo oo we + ~ * - - - ~ = 4**
- eo--
L -ex-v>ma
- -..i. 4
-;. L __ Technical Report TR-62168 ' Two examples of the above situations were found by Cygna during the Independent Design verification. The first is the issue related to mass ~ participation. Under questioning, during a January 10, 198,5 meeting with th i U e NRC, Cygna indicated that no procedure existed at Gibbs. and Hill (G&H) f to control this portion of the Design Process. Thereforegation and QA audits using the process in place would not have indicated noncompliance 'with the licensing commitments for Comanche Peak._ The second is the. issue related to mass point spacing. Under questioning during the same January 10th meeting with the NRC, Cygna indicated that an acceptable I procedure exists in the Design Process at G&H which addresses mass point j spacing, however, in many cases this procedure was not followed. Design _ verification and QA. audits' failed to uncover the inadequate implementation of an existina eror dure which was Eplace. Io. provide a design u=*. O comp 1ied with the licensing commiments 3 . It would appear th'at until the Phase 4 effort by Cygna the issue / related to technical review of the Desigo Process to determine whether it f controlled design such that the licensing commitments were satisfied was h not performed..This opinion is reinforced by the fact that, at this point h t in time, Cygna is revising their Phase 1, 2 and 3 conclusions related to ) Design QA. U [ Having established that Design QA has two sides, a paper trail side and a technical side, it is necessary to look at the process in existence for Comanche Peak for piping and supports. J Pipe supports and' piping are so closely intertwined and technically J l interdependent that it is difficult to separate them. In designing a piping system the designer makes certain assumptions concerning individual f support configurations. Also, a piping designer usually cannot make 4 f appropriate judgements on the adequacy of. a piping system without f r y
...y n y - W.L.y g ff @ @ :- - % ?.d=p d a s:J m +. M M 5Hhd Q,g _ p Technical Report TR-62168 l' + reviewing the piping layout with all of its supports.4.This is particu- ) larly important when addressing an issue such as support stability since ~ he interaction L,etween the support and the pipe is usually critical in making this determintion. ~ For example, for a pin-pin connection, the displacement of the piping at the support location due to operating condi-I tions (themal expansion) can result in a reduction in the ability of the f support to carry, a load alon'g its axis. Also, the concern of the author I with respect to support stability is directed towards anticipated water I \\ and/or steam hammer events which usually result in higher loads and dis-placements #on the piping syste:n than does a seismic event. To accomplish g the kind of review discussed above it is necessary to have an established f and functioning link betw'een the group responsible for piping design and analysis and the group re'sponsible for support design and analysis. p w I In the majority of cases a utility constructing a nuclear power piant contrdts with a design firm (usually one of the major AE's) to provide design services in the areas of piping and pipe supports (along with a .I number of other areas not relevant to this discussion). The AE is respon-sible for the design process interface controls and procedures required to i b develop construction drawings for piping and pipe supports. The AE may elect to subcontract a portion or all of this work to a third party;. D however, responsibility for, and control of, the design of both piping and supports rests with'the AE. This responsibility and control exists even when the third party uses its own Design QA Process and Procedures. The AE will review and approve the process and perform audits to detemine accept- [ ability of implementation. The above does not eliminate the. requirement I that the utility is ultimately responsible. 3 j-4 Your attention is called to Welding Research Council Bulletin 300, " Technical Discussion on Industry Practice " Section 1.7, page 26 December 1984. c i. ? o ..m. _ _. _ _ _,., _ _. _. _ _.. _
. =.. .: a,:.;.s-w - n w + w - m.g.n a- - - ~ ~ x.. -1 Technical Report / TR-6216B. pu M i a I h e gI is my understanding that, for Comanche Peak TUGC0 contracted with G&H to perform piping design and.also contracted with ITT Grinnell and NPSI ( to perform support. design. The only.G&H interfacing (or control) doc,ument g imposed on ITT Grinnell and NPSI was Specification MS-46A. In a contractual sense, therefore TUGC0 was acting as an AE. This role [ required TUGC0 to perform the activities necessary to assure that an j overall process was in place, including necessary interfaces, and that the [ process was one that would provide control on the design so that proper l engineering would result in construction drawings that complied with the ~ l licensing commitment. Some utilities have acted in 'this fashion, however, I l they do perform and control a significant portion of the piping and pipe support design and have significant experience in this area. It is noted y that this contractual appro&ch was not.. unique to. Comanche Peak for G&H. { They had the same type of contractual arrangement for the Fort Calhoun / project. No information is available concerning the design process for ' Fort Calhoun,nor how it compared to that for Comanche.. Peak - particulply wthrespgetto-G&Hrole'. y 5 In response to questions at four meetings with the NRC, TUGC0 indi-cated that the process for initial design, including issue of initial { construction drawings, consisted of the following. l l (1) G&H performed preliminary free thermal expansion analysis and forwarded these to ITT' Grinnell and/or NPSI. (2) Deadweight supports were loca,ted by Grinnell and,NPSI using the hanger spacing table established in ANSI B31.1. Potential r locations and directions of seismic restraints were established l by ITT Grinnell and NPSI. Guidelines for spacin'g these re-straints were established by G&H and were based on frequency l considerations. l 5 August 9, 1984. January 10, 1985, January 15, 1985 and January 17, 1985. I (
- w..
..~,-,_. ~
3 %-s%,&.: : ;., x;q.54g;p;49 _;.,;g 3;&g; 4q f pl Technical Report TR-6216B (3) G&H then performed piping design and complete' analysis, l ? including location and selection of the type of pipe. supports. This required the normal iterative process of layout, analysis, supp' ort location, modification of layout, analysis, etc. I Eventually a design evolved,that analytically complied with the licensing commitment. f 7 (4) Support locations, types and load combination data were sup-g plied to ITT Grinnell and NPSI. 1 (5) Support details (including selection of standard hardware) were developed and support analysis performed by ITT Grinnell and f NPSI. Cases could.arise where that. location of a specific f support for the specified loading was not acceptable (i.e., an p adequate design could not be reasonably developed). In such [ f cases the support contractor would inform G&H and another ~ iteration in the piping analysis process would occur. (6) Design and analysis was completed and supports were fabricated and shipped to the site. Review of th= mannet details at G&H [ r was not reaui nd at this time in the design process. [ (7) Modifications to supports required by field conditions were made by field engineering (Texas Utilities responsibility) and i t a Component Modification Card (CMC) was executed. i (8) The CMC was forwarded to the responsible support design agent (ITT Grinnell or NPSI) for review and approval. (9) A third pipe support group (PSE) was formed which was under the technical direction of TUGCO. This group functioned just as l ITT Grinnel and NPSI did although the engineering and i ( administrative procedures differed between the three organiza-tions. t y__..,,, v._
- p 3.;- ; 3;;p
- ;;;ggg:5 : 3;g;w.
- ggL ;;pqqgyg;.cgy,-
g.n ~ Technical Report TR-6216B -6~ L (10) Also in this time frame, ITT Grinnell and NPSI sent support designers and analysts to the site to perform design,' analysis, modifications, and review of CMC's. These ITT Grin ~nell and NPSI personnel ~were administratively controlled by TUGC0 but f utilized their own procedures in performing their required I tasks. For ITT Grinnell these procedures were the same as those for the home office. NPSI developed specific procedures to be used by their personnel at the site. ej 3 I t! (11) Any of the three organizations who had concerns with a CMC informed the initiating field engineer of that concern in a ( [s Technical Services Design Review (TSDR) memo. r (12) At a point in time when the pipe was installed and Brown and 4 I Root (B&R) felt confident that the support as designed or modi-p Q fied would be able to be installed, an as-built walkdown was I g performed by TUGC0 personnel and a package forwarded to G&H for b their review, reanalysis (as required), comments and/or Vacceptance. G&H coments or concerns with as' built condition were transmitted to TUGC0 in a G8H memo. 4 -(13) After piping reanalysis and pegrmination of new loadings, the responsible support =Yi!$uis. would be supplied with the new [ loads by G&H to be used in their review, reanalysis, comments j and/or acceptance, of the as-built support configuration. For [ cases where. piping reanalysis was not required, the support E designer would review, reanalyze, comment on and/or accept the ~ as-built configuration. L (14) The documentation from G&H and the support design organizations was then forwarded to TUGC0 who reviewed the documentation and stamped those supports which were accepted by the support ctesign organizations "as-built certified." t k. e -..$.-a = $'H"? !' St*"' 'C1
.y_- . e:&..a :;.,;.. ;, y r x..s. w.:.. p : g.u n.... x ;,. - - - e. ' Technical Report TR-6216B i ~ (15).This process continued on an interative basis until all piping { and supports were accepted. 0 ? G'H in their review of as-btnit information was responsible for (16) ( acceptance of the piping system (piping plus pipe supports) as complying with the licensing commitments. de As indicated, the Design Process at Comanche Peak was modified as the i project evolved from design to design and construction. This is not ] unusual in the construction of a nuclear power plant. ~ [ The author has some concerns with the process described above and with i some aspects of implement'ation of that. proc,es,s...These concerns do not r necessarily result in a conclusion th' t tfie process or implementation is k ~ a 0 sufficiently flawed to result in a design that is not in compliance with p NRC safety criteria or the licensing commi+.ments of TUGC0 for Comanche [l Peak. The concerns are as follows: in (SKfk/d5 * ""# #Y Concern 1 I iping is not a " stand- [ alone" commodity.6 A basic premise in designing a piping system includes ( (but is not limited to) the fact that support designs will reflect the assumptions made in the analysis of that piping. This is of particular i concern to the author as it relates to:. anticipated steam and water hammer resulting from plant operating transients. Since'G&H was not. required to 9 (and therefore did not) review support designs prior to their fabrica~ tion 6 G&H agrees with this in footnote 13, page 17. of summary disposition..
U7 5 12 ' Technical Report
- TR-6216B
-8 and installation T N n as des d situation is is not to be construed as a judgement that this occurre at omar.che Peak nor is it to be construed as [ a judgement on the adequacy (safety significance) of the design that exists at Comanche Peak. Again, my major concern is related to anticipated transients such as y steam hammer resulting from a turbine trip or water hammer resulting from I pump switching and rapidly closing check valves. With respect to seismic loading it is my current ' opinion (based on the data available to me) that _the existing supports wiil be adequate. This -is based on the fact that %f the CPSES piping was designed using lower damping values than are currently permitted. Use of PVRC damping has resulted in reductions of peak accelerations of up to 50% with general reductions on the order of 35 to f 40%. Further, test data indicates that piping systems with supports that i are flexible, have gaps and pinned connections usually result in higher damping since a significant amount of energy is used up 1,n deflecting the restraint, closing gaps and moving about the pinned connections. Concern 2 y /, g yz g g &GSTA%g-&fpfg Smns [- The use of nomographs based on frequency to locate seismic restraints usually results in an excessive number of restraints. This approach was used at Comanche Peak and apparently resulted in excessive seismic re-E straints. This is verified, to a degree, by the fact that a majori.ty of f/ A the seismic restraints are very lightly loaded. Lightly loaded restraints g/ (9 /5 which are designed using a deflection criteria (i.e.,1/6-inch maximum) pf/ are usually very flexible. Flexible restraints have been a subject of l f concern at CPSES. 7 Those restraints which are pinned vertically and have bumpers for out-of-plane displacement control are an exception and are discussed in Concern 3. h .m.-~ -j. r________-_______:z____________ v_v - -
= p. ac _m. ;-.,,..;. . m. .m., jm mg.yy ~
- m m.2 e c-_.
~ Technical Report TR-6216B. d glgConcern3 [5 The stabil,ity question has rea.ulted 'n a number of analyses and some y modifications to supports. In one area, on the main steam system, bumpers were added to prevent rotation of the support about the pipe. Cygna has_ ~ not accepted this design as sufficient to provide stability.8 TUGC0 has performed seismic analysis with the supports in place and with the supports ~ removed and the resulting stresses are acceptable in both cases. However, the supports are still in place and, according to Cygna, will not function. ~ My concern is that the seismic anilysis does not bound the real situation which could be that the support has become " tilted" or unstable and then a dynamic load is applied to the system. Does the tilted support provide restraint in a direction that~ was not intsnded?- Once tilted does the support restrain thermal expansion? To assume that a support is acceptable p l because it is analytically not required may not " bound the problem" in every case. This would also apply to a support that was overstressed. To perform a piping analysis without the support in place and demonstrate ' ? acceptable stresses in the pipe and other supports is not always the worst case unless support failure is complete (or the support is physically removed) and does not impose a restraint on the system that was not accounted for. { [d W 7A4L' O f C M C- ) r 7'S A A 'S_) / Concern 4 I l A design process must provide a controlled communication between con-struction activities and design. TUGC0 is right in pointing out that a Nonconformance Report (NCR) is not the only document for accomplishing this. Examples of other techniques used in the past are. a. Field Change { ~ ~ l Request (FCR) and a Drawing Change Notice (DCN). TUGC0 used a Component Modification Card (CMC) to provide this interface. However, some concerns exist with the implementation of this interface. The design process under- ] went an evolution as plant construction activity increased. The following 8 [ January 10, 1985 Transcript, pp. 72 and 73. f ML ?AM L*T M*"***T - - + = - * ' = * * " W _,TT }y'*'"
r~ ,,g _ _,w: =.2;.s :- % w m.m-n,n- # a-r.a -2 1 w_
- Technical Report
~ TR-6216B, 't ~~ discussion addresses the process from its initial to its final stage as now j understood. e In the initial stages.(and for some time) CMC's on supports were g generated by the Field Eng,ineering Group (a subgroup of TUGC0 Pipe Support ( Engineering - PSE) and were forwarded to the organization responsible for that support (ITTG and NPSI). The CMC was placed in the system file by ITTG or NPSI and would be worked on as the piping system required rework or h as TUGC0 requested.9 This resulted in construction of the modification continuing without review by the responsible design organization. In some cases, as-built analyses performed by G&H could have included supports with outstanding CMC's although the appropriate CMC would be included in the as-built package. Ba' sed on the defined pr.ocess, this would mean that the effected support would not have been a;) proved by the appropriate design organization at that time. However, the support design organization was I alsp involved in the as-built process and review of the support would have I been accomplished as a part of that process. One could suggest that a I method of controlling the number of outstanding CMC's on a given drawing (say 3 to 5), or controlling the time that a CMC can be outstanding, would force review, approval (or disapproval) and incorporation of the CMC into f the drawing. This would reduce the turnaround time for approval and reduce I the number of outstanding CMC's in a given as-built package. Eventually, a site group was established undert. PSE wh'ich included ITTG and NPSI personnel. Under this organization CMC's were dispositioned t by the.PSE group on site. This shortened the communicatio,n link and should [ have resulted in more rapid turnaround of CMC review. However, no change 1 to the process occurred (i.e., time limit on CMC or limit on outstanding number of CMC's on a given drawing) except that' the field engineer, who ? authorized construction to make a change to a support, had available, on site, the complete design resources of ITTG, NPSi and PSE. 9. January 15, 1985 Transcript, pp. 30 and 31. t . g.e .....==ip. ... -..== == ,y. 7 n... .-~---e. .m
h 3 ~. :.,:
- ,~
w.2 ~ .. =. -w a Technical Report TR-6216B, When ITTG, NPSI and PSE reviewed a CMC and found an unacceptable condition (i.e., stresses too high) they generated a handwritten 1 mem7(TSDR) noting the condition. This TSDR was sent to the field engineer [ responsible 'for generating the original CMC. The field engineer would reply back to the originator of the TSDR (on the original TSDR in a section l set aside for a reply) noting the changes now recommended for the support I' can be found in the next revision of the CMC.10 The support design organization was now responsible for reviewing the next revision of the appropriate CMC. k i One area of concern with respect to QA control is that CMC's were i handled by the site document control center.and those individuals on the I effected drawing distribution list received,a <opy-of the CMC. Copies of f the TSDR's were not contro_11ed. There does not appear to be a definitive E ~ link between QA and design in the area of CMC's and absolutely none with_ [ the TSDR's. Therefore QA could only determi~ne that changes to design were Nk occurring if they performed audits (which they did) and reviewed both the j CMC's and the TSDR's. This need not be a real area,of concern in the [ initial design stages where construction was not underway, however, once a l construction drawing is issued it is important that QA be aware of changes [ E that are planned to that drawing. This is particularly important when those changes are already being built. QA can be effective in recognizing repetitive design changes and developing trends and then modifying their audit plan and scheilule to focus on the effected areas. TUGC0 (Chapman) F states:II i " Applicants have established a procedure, CP-QP-17.0, " Corrective f Action," to review documented conditions adverse to quality for the purpose of providing corrective action to preclude.repeti-tion of significant conditions adverse to quality. This proce-l dure provides for Quality Engineering Staff to review design h changes documented on CMCs. The results of these reviews are 10 i January 15, 1985 Transcript, p. 46 and Motion for Summary Disposition, July 3,1984, p. 53. II Motion for Summary Disposition, July 3,1984, p. 54. ~
fy w e - ..wi.mttw ..= u - ~~ m --a w wa..v g. l ' Technical Report TR-6216B, [ i tracked using trend analysis techniques as an objective method of ascertaining the need for corrective action to preclude repetition of significant, conditions adverse to. quality.. Per' odic reports summarize the results of the reviews,' including f trends, and provide recommendations, where appropriate, for cor-f rective action with respect to identified conditions which are considered to be significant. l ? r This is appropriate, however without receiving copies of TSDR's it is i not clear that trends of field engineering to propose inadequate changes-to design are not explicitly covered unless one assumes that the revision to a CMC resulting from a TSDR' defines that the reason for the revision was ) f either a TSDR or a request by the reponsible des.ign organization. i . x. Concern 5 {$~c/A/}92r CAMW6565 Mr 25# # G&H had a Site Stress Analysis Group (SSAG) at CPSES that was adminis-trated by TUGC0 but reported to G&H. Mr. Ballard of G&H states: f I2 ' SSAG was established to evaluate and approve proposed i, changes and modifications to pipe routing, pipe sup- ,I I port locations and/or pipe support type, as requested by site engineering groups. The evaluations are made employing the latest as-designed piping stress analy-sis. SSAG provides revised design information to the applicable site organizations. All these activities l are conducted'in accordance with CPSES Engineering In- [ struction CP-EI-4.6-9, Rev.1, entitled ~ " Performance Instruction for Piping Analysis by SSAG" and Gibbs & [ [ s I t 12 Motion for Sumary Disposition, July 3,1984, p. 20. T
- W e e so *-== m muum,emyne p q q. m.pp e, e,
e,,9p ,. g pe ye se -us, e e- ~ me.- -- = --e. .,m..,,,
5 - y. :: ~,+ .. + =;, 2... hw. =::.=.1 .&5+ &.&S Nsv: Technical Report TR-6216B, Hill Applied Mechanics procedures previously cited. These documents have been established to assure that I th.e SSAG activities are accomplished in a manner [ commensurate with the original as-design analyses." k ~ k [: The concern here is related to the fact that SSAG performed their i t function "as requested by site engineering groups." It is understandable that a modification to a pipe routing of considerable magnitude would have been routed through the SSAG. It is assumed that this was accomplished g through the use of C.MC's as discussed for supports in Concern 4.
- However, a major modification to,a support which could have an impact on pipe i
b stresses may not be routed to the SSAG since the individual responsible for generating the CMC may not have considered]or recognized) the change would effect pipe stresses. Concern 6 t The following are discussions of those items which are specific in nature and yet, to a degree, have an impact on the design process. I b 6.1 Mass participation b L i This issue is addressed in introductory remarks (see page 2) and [ is important from a design process, standpoint and a support / pipe adequacy [ standpoint. Based on the Cygna review it appears that the average mass if participation of piping systems analyzed by G&H is in the order of 40%.13 I k One could expect that a seismic analysis cut-off at 33 Hz should normally result in 90% or greater of the total system mass. For piping stresses this would usually be acceptable. For supports however the contribution of 4 ^ 13 January 10, 1985 Transcript, p. 70. 4 5.% q. 9p.- w .n; ..g-a = rm- .. p. ep pear og,, _, _ pp,ny M
- Tw**
,?-
- ,~'
iz 1.. a '~ -i s J:.v g 4_m.g ,; y e--~ s Technical Report TR-6216B - i O rigid mode response could be important, particularly for supports located close to large concentrated masses or where the support is providing axial restraint. In these cases the total seismic load should thclude a rigid mode component equal to the floor ZPA times the weight of the supported component or segment of pipe. Based on normal expectations a mass partici- .pation of 40t is unacceptable. Further, the design process at G&H did not I control this effect since a procedure was not available. E Reanalysis by G&H to include total mass participation will result in significant increa:e in so:re support loads. This efect when [ coupled with the low support stiffness '(flexible restraints) could result [ E in the need to modify supports (see Concern 2) Mult of this I fblilem It may be appropriat.e'for the CPSES engineering Q organizations to consider the latest Code revisionsI4 ~ with respect to [ co damping. This would result in a significant reduction in peak accelera-I tions. However, should the new damping be adopted it is recommended that all of the requirements of WRC Bulletin 300 be adopted. There are signifi-s I cant discussions therein on interface control and responsibili*,ies, ~ 6.2 Support stability i I In addition to the discussion under Concern 3 which addresses some specific restraints there are some generic concerns. Many of the I restraints and supports at CPSES utilize box beams with either pinned [ struts or snubbers connecting the box to structure. This is not a common l _ design for seismic Category I nuclear piping. Box beams them: elves are not l uncommon, however they are usually rigidly connected to the building structure using standard structural shapes. A second type of support that [ is of concern is the trapeze style support which is composed of a structural member supported off the building structure by pinned struts or 14 Code Case N-411 - Adopts PVRC damping discussed in Concern 1. 6 , + - ,-en,s..=e+- w .n
g- ._a c _- a, w2 z. w.6 : <.%..m:. , _T 1 (* Technical Report TR-62168. ( snubbers and attached to the pipe by a U-bolt or trunnion. Again, this h type of support is not a common design for seismic Category I. nuclear [p t piping.in plants _ licensed to operate in the last 4 or 5 yeart. (Trapeze [ type supports with U-bolts can be found in non-seismic piping at nuclear plants and in other facilities such as process and fossil plants.) A third concern is related to support application. That is, the use of struts or snubbers supporting a pipe from the bottom of the pipe to a floor or platform below the pipe. Since these supports are pinned they are unstabl p* b h vertically as soon as horizontal displacement of the pipe occurs and system ( stability is provided only by the end conditions of the piping system or l any horizontal restraints that exist. It has been pointed out that piping P must be considered in conjunction with the existing supports and therefore Ip he presence of pinned supports applied in the manner described above must t b udgeLbased on the overall support' shstem. owever, a rea f designer would not provide-tMs-t-yp ;f sup r in a number of locations on the sar$[e' piping system and rely on pipe end conditions or horizontal [ restraints to provide stability. For example, removal of the horizontal ] restraints for maintenance work (or as a result of a snubber reduction {. program) could leave the system in an unstable situation. For the pinned box beams and U-bolt trapeze supports it _is nearly [ impossible.to demonstrate by analysis that stability exists. There are only two reasonable alternatives and they are: r: (1) Perform a test of a system representative of a system at [ CPSES (system meaning pipe and supports). The test should provide dynamic input commensurate with the plant design [ seismic and operating transient conditions (water / steam hammer). (2) Modify the supports. 4 g 9 p-=gpe. _. - ,e._==***-+==7* . w am -
- y
- - - - - - + =we-. .. ~. + - - * - + = = = = - *. - --we+-+*-
.n ,. ya :g.:
- .: w :
= x-e,.* l Technical Report TR-6216B. I [I L 6.3 As-built reconciliation ll hI' The as-built reconciliatior. procet s has two fu ctions. The p first, and most obvious, is' to take dimensions, etc., 'of the actual I as-built configuration of piping and supports and reconcile those with the ~ as-designed documentation. The second is to have a qualified piping f ~ f designer walk the system to develop an understanding of the overall geometry and to determine if the installation generally reflects the } I analysis. The importance of this second step is obvious, the overall configuration is there to see an'd one is not dealing with a number of different drawings trying to piece together a system. j. t The existing design process' 'at C{SES'r6' quired as-built informa- [ tion to be gathered by TUGC0 technical services personnel and forwarded to } G&H applied mechanics personnel. Already the ideal situation where the G&H [' analyst or. members of the SSAG walked the system did not exist. However, this is not a fatal problem nor is it uncommon in the industry to have "others" gather as-built data. It merely makes the problem of system acceptance and analysis reconciliation more difficult. [ i t The as-built reconciliation program was started at the time that j the piping was installed and Brown & Root determined that the supports not in place could be fabricated and installed as they were designed. The number of installed supports on a given stress problem varied from 20% to 80:15 at the time G&H started reconciliation efforts. Having only 20% of [ the supports installed has two impacts, one'that could be positive and one [ that could be negative. The positive impact is that with only 20% of t ~ f supports installed the G&H analyst should have had an early, indication of what the support designs looked like and could have requested modification S I L 15 January 15, 1984 Transcript, pp. 22 and 23. y, ,o. 4 .=w-e-
e .m m#, n r c. ...n. _ g ...m. r s.. Technical Report t TR-62153 f f r (if there was concern) prior to fabrication and installation of the remain- [ ing 80%. That is, the undefined pressure to accept constructed supports f was significantly less than ore could hypothesize for the situation where [ all of the supports were installed. The negative impact is that the piping [ analyst is not dealing with the complete as-built system and one can j anticipate that a number of iterations will be required to complete the [ reconciliation process since modification to one of the remaining 80t of I the supports could impact the total system including the installed 20t of h I the supports. Iterations such as this are not uncommon but sometimes tend N to result in cursory reviews of already accepted situations. One major concer.n with respect to as-built reconciliation is the situation where more than one pipiig. system wh" supported by a frame, particularly frames which were pinned connected to building structure. GLH, though aware of the fact that the frame supported mere than one system, dealt with the support as a single support on the piping system + under consideration at that time.I6 The support designer was supplied with the loads on the frame for each piping system being supported and deter-I mined the structural adequacy of the frame. No one was apparently responsible for looking at the interaction effects inherent in a pinned g frame supporting a numb.er__ of nipes. It is my opinion that this is the responsibility of the piping designer and G8H accepts that responsibiiity.II I t 6.4 Support mass Many of the support designs at CPSES result in cbasiderable mass which is not acting at the outside diameter of the piping.' It is common practice to add support mass to the piping analysis and this is usually done at the centerline of the pipe since it normally involves a clamp. In i 16 January 15, 1984 Transcript, pp. 23 and 24. -II ~ January 15, 1984 Transcript, pp.11, 49 and 50. 4 y n. ,,-,.._.-,.-.i.- ... ~ _ - - - -. ~ - -
= - . -,_ _ - m ; 2;. = = - - - o *y * . Technical Report . e. TR-6216B. i the case of a box beam rigidly connected to the building structure the mass is not applied to the pipe and therefore need not be considered, in the case of 'a box beam pinned to the building structure the mass ' acting 90 degrees to the direction of restraint should be applied to the pip 6 center-g I line. l ? A specific geometry that cannot have the mass applied to pipe i centerline and be representative of the as-built condition is a support restraint that is pinned to the building structure and has a beam some f distance from the pipe [ and the pipe 0.D. The beam is attached to the pipe by welding a trunnion to the pipe and the beam.18 The effect of the offset mass rigidly connected to the pipe results in forces and moments on the ~ pipe which will not be represented properly _by wodelling the mass at the a pipe centerline. G&H apparently accounted for this effect on the main j steam system only.19 ~ f t l f I- ~ 5 L I h ~ I8 This would normally be called a trapeze restraint but if used as a horizontal restraint on a vertical pipe that could be a misleading statement since a trapeze support is normally considered to be a verti-cal support on a horizontal pipe. 19 January 15, 1984 Transcript, pp. 61 and 61. ? s ~ ...,}}