ML20128K200

Forwards Responses to 840316,19,20,21 & 22 Handwritten Questions Re Supports Inside Containment & U-bolt Allowable Loads.Questions Raised Via Telephone to D Rencher on 840330 Under Review.Responses Will Be Sent by 840427

ML20128K200
Person / Time
Site:	Comanche Peak
Issue date:	04/19/1984
From:	Popplewell L TEXAS UTILITIES ELECTRIC CO. (TU ELECTRIC)
To:	Williams N CYGNA ENERGY SERVICES
References
NUDOCS 8507100513
Download: ML20128K200 (14)

Text

r

) ,. ..

TENAS UTILITIES GENERATING CO.11%NY

\ DDft

,' r.o. sox ioo cia nosz. rzxas as iRU April 19, 1984 .

k* .. -

h

.'C: no D51/ G u CL.

R.T LCCGED:

' M /5'V

LOG nc.

47 Cygna Energy Services y , /, f Af , g 101 California Street !m.

f Suite 1000 // Ar - dc / 6 San Francisco, Califomia 94111 [cPoss arr. FILE Attention: Ms. Nancy Williams ProjectManager 00mNGE PEAK STEAM ELECTRIC STATION .

Gentlemen:

In response to your handwritten questions provided to WGC0 on March 16, 19, 20, 21 and 22, enclosed is a copy of the questions followed by WGCO's response. Several questions are still under review and will be answered shortly. In addition, questions asked via telephone to Dave Rencher on March 30, 1984, are presently being reviewed and responses will be for-warded by April 27th.

If there are any further questions or comments, please contact me or George Grace (Site Ext. 500).

Very truly yours, TEXAS UTILITIES GENERATING COfPANY ENGINEERING DIVISION

. 7J L. M. Popplewell Project Engineering Manager gggg A/N W M l IJfP/cp h0N d.tu.abga) c YC C. (skMf In.%Lmar l

Evok Awfut FA(FAc~

k (

8507100513 840419 PDR ADOCK 05000445 [h[ j'/>

/

'l f

1 A PDR l yJ it ' , p W 4 .. .. . .. .u . .r ru . . , r, u r. . . c, n e c r. ,=, . =

t

-*-.t -<

,.,,..--.- ..-<r-

March 16, 1984 ,

CYGNA COPMENT:

1. In reviewing supports MS-1-002-006-572R and MS-1-004-006-S72R.

CYGNA has noted use of A563, Grade A nuts on 1-1/8" diamet'er 197 B7 threaded rod. Per the ASTM specification for this nut.

Please pmvide Grade A is suitable only for A307 material.

documentation justifying this material usage.

TUGC0 RESPONSE:

1. It is common practice to use A 194 2H (or equivalent) nuts on A 193 B7 rods. However, use of double A563, Grade A, nuts will allow sufficient thread area that the use of high strength rod allowables is justified, since the tensile allowable load for high strength bolting material is approximately twice that .

of standard material ( A563).

CYGNA COMMENT:

2. In reviewing a number of supports inside containment (for example, MS-1-003-008-C72K), CYGNA has noted tube steel with bolts through the steel into inserts. In all cases, the mtation about I in the In most cases, the rotation "STRUDL" model is released at the bolts.

about X is also released. While this is conservative for the member, the subsequent calculations do not account for the possibility of prying by the tube / washer on the concrete. Attached is a calculation Has the prying by CYGNA which indicates potential design impact.

effect been considered? If so, where is this documented 7 TUGC0 RESPONSE:

2. (This question has been assigned to NPSI. A formal response will be issued by the Secaucus office by 4/23/84).

March 19, 1984 CYGNA COMfiENT:

1. In reviewing the calculations for U-bolts used in supports, CYGNA has noted instances where:

a) A 2" p U-bolt is called out, with no vendor specified, nor any design calculation supplied (MS-1-003-001-C725. Rev. 3).

b) A PUH-340 (21s" p) U-bolt is called out, but this does not appear in the NPSI catalogue. The designer then qualified the bolt based on a pure tensile check (.6Fy). ( MS-1-002-001-c)

S72R, Rev. 3)(2-3/4" p) ITT U-bolt is called out. Since no A Fig. 1375N allowable is supplied, the designer qualifies the bolt based on a combination of tensile and side load (because of swing angle). When the side load causes failure, a memo from IfT is used to accept side loads (due to swing angles up to 5 )

.- without tensile load reduction. The designer then qualifies the bolt based on tensile alone. (MS-1-002-004-572R. Rev. 2)

f

' ' page 2 i

1 .

l* d) Again, in MS-1-003-006-572R, a designer qualifies a Fig.

137N 2-3/4" p U-bolt based on tensile stress only, although l the designer does use the root area of the bolt (conservative).

In all cases, except where no check was done, there was no attempt  !

j to consider the complex stress pattern of the U-bolt as a bent beam t i

with a distributed load.

i TUGC0 RESPONSE:

)

1 1. U-bolt' allowable loads i

a) TUGC0 answer will be provided later.

b, c, d) ITT-Grinnell is preparing a statement based on a test

! program they are conducting. Release of that statement in 4 .

the week of 4/23/84 will supplement this statement. Vendor  !

supplied Load Capacity Data Sheets (LCD's) contain U-bolt *

allowable load information based on MSS-58 which limits allowable stress to 9000 psi. This low allowable is desira-ble to use for original design and quick checking, but is certainly not required by NF.

At the designer's discretion and in the absence of published '

j 4

allowables, a pure tension (or combination of tension and i shear) check is acceptable, comparing calculated stresses l with NF allowables. Attempting to consider the U-bolt as a 1

bent beam with a distributed load is a very complex task and is not realistic. The above tests will indicate clearly that the methods of analyses being used are conservative.

The vendors (ITT and NPSI) have not provided documentation l

on qualification methods for non-standard U-bolts. (Note:

ITT is reviewing calculations pertaining to item (c) and i will issue a supplement to their calculation package.)

l i  :

i CYGNA COMMENT:

i l

2. CVGNA has noted instances in the CCW system (CC-1-028 003-A33R, ,

4 CC-1-028-001-A33R) where no gap is specified for a pipe within [

a frame. Has TUSI done a' study to show that, below a certain l

! l j

pipe temperature, pipe growth effects are negligible, since it  !

is standard practice to use a 1/16" gap?

J i

l TUGC0 RESPONSE:

j i

2. The 1/16" gap is required in a rigid frame to allow fme axial

' thermal expansion of the pipe. For rigid frames at CpSES all vendors specify the appropriate gaps in accordance with their design procedures. On the two supports in question, the frame

' is on a pinned connection and hence the entire trapere will i displace as'the pipe moves; no axial support is introduced into  !

l the system, therefore no gap is required.

l

_ - ~ _ _ _ ________

t Page 3 4

i .

1 Diametrical thermal expansion of the pipe can induce additional

!' loads into the support frame (and especially the welded connections). f 1

These loads need to be combined only with nomal and upset mechanical loads from the pipe onto the frame and are self-relieving in nature.

The allowable for this load combination from NF is increased by a factor of 3 over nomal allowables. A study has been conducted on l

j j

this issue and is summarized below:

Frame - 6x4xis tube steel (4" contacting pipe)

Pipe - 24 Component Cooling (3/8" wall)

Resulting 1.oad F Resulting Additional Pipe Stress Temp. Pipe 454 lbs 7104 psi 200* F 1886 psi 130 F 163 lbs

! The detailed calculations are available at the site. The additional '

- load is quite low compared to the NF allowables increased by a factor of 3. '

CYGNA COMMENT:

! 3. In reviewing the MS pipe support design calculations, CYGNA has j

seen many instances where components are not checked. In some In there is reference to previous cales which are not included.

j others, there is no reforence. As an example, CYGNA has chosen i

MS-1-003-013-C72K.

As noted on the attached sheets, most of the l

! items are acceptable.

i ,

Question a) Why are there no calculations or documentation for the standard supports?

! b) Who is responsible for insuring that the data on drawings ,

is correct? l c) What size U-bolt is in the field, Ik" p or 1-3/4" p?

i d) How does NPSI/ITT ensure that all parts of a design have i'

been checked, if a package does not contain or reference

! all the data?

! TUGC0 RESPONSE: ,

I Before responding to the specific questions some general comments

~

3.

are in order. It is not uncommon for a designer to use standard catalog callouts for convenience, even thoughCommon the item is not examples are vendor supplied, but is fabricated at site.

l thmaded rods (RAT RET, RFT), washer plates (RWP), and U-bolts (PUS,PIM). Explanatory information in the bill of meterial over-i For example,  !

3 rides standard size designations from the catalogue.

an RET may be specified with an extended thread length; an RWP may show a larger thickness; and a PUS may be specified with a larger rod diameter.

1 l

l

Page 4 Specifically, (a) ,many standard support items are verified in the original design calculations and not re-checked as part of the vendor certification process when the as-built load decreases from or stays the same as the original design load. (b) The design engineer is responsible for assuring that data on t.he drawings is correct; QC is responsible for assuring)that the For the product in the field matches the design drawing. (c support in questions, a 1-3/4" 9 U-bolt is installed in the field.

(d) All design changes to a support must be reviewed and approved If the cumulative changes to by the original design organization. the a support are minor (from original design to final product),The otiginal certification calculation will only verify the changes.

calculation package will contain data to assure acceptability of the rest of the support. For substantial design changes, entire net; This methodology is ad-calculations are generated for the support.

dressed in the procedures of the support vendors.

CYGNA COMMENT:

4. In reviewing certain U-bolt and beam combinations, CYGNA has noted instances where the local effect of the U-bolt on the beam is not For example, in MS-1-003-001-C725, the U-bolt has an considered. As shown in applied load of 23.7K, and is attached to 6x6xh" TS.

the CYGNA calculations, the local bending stress in the top portion of the tubesteel is quite high. How does TUSI/NPSI/ITT determine if connections can carry the applied load? Are standard tables available?

TUGC0 RESPONSE:

4. Local stresses in this type of connection are normally checked standard on a case by case basis when considered significant;This unreinforce tables are not in use.

not used where a large load is present and should not have been used here. This is an isolated case that was overlooked in the support certification. This is a spring hanger which has already experienced its full load and a field observation hasthere Thus, clearlywill indicated that there is no local deformation.In addition we feel not be ultimate failure of this member.

CYGNA calculations are incorrect since they have not taken into effect the rounded corners of the tubesteel Out nortotheout correct loaded ares under the nut (see sketch on attached sheet).

on the nut almost covers the entire flat portion of the top of the tube steel. Pursuant to good, conservative engineering practice, however, washer plates will be added.

CVGNA C00 MENT:

5. In reviewing installations with 0" gap around U-bolts, CVGNA has the following questions:

a) On Page 11 of CP-CPM-9.10. Rev.11. snug tight is defined twice.

, , . . . - ~ . . . . ..r...,.. . ..r. . 3 7....

'~"

,w a a A a_.--. -J - .-- A

,, .............~..4---~~*~*"- ' *-l* *

=

e 1*

h~ ku

, i h Nd NW i

seu ,

/

/

r

. I i l__

N!

ut 2I t

I -

e 4 W i.

a h

\

~

ql i t A at E

h .

O s

e*

g a6ed ,. ..-

~

page 6 l l '

i.

" NOTE: Snug tight is defined as the tightness attained by a few impacts on an impact wrench or the full effort of.a man using an ordinary spud wrench."

~

"When U-bolts are specified on the design document as hot having any clearances, the U-bolt shall be snug tight so ,

4 that the U-bolt cannot be moved by hand. The nuts on the  :

U-bolt shall be installed on the side of the plate as shown on the design drawing."

I i

Which is used for installation of U-bolts?

l '

b) In support CC-1-028-007-533R, a PUS-240 U-bolt is used to '

l provide support stability. If " snug tight" is taken as the  ;

i first definition, this could mean--80 ft-lbs of torque since:

1 i

' T = .2FD '

i T = 80 ft-1b = 960 in-lb l

" D = 1" for a PUS-240

' .F = 4800 lb/ leg i 5

For piece 15 (1/4x4x4 (,)

rf 4800 (2) , , p, i

4(.25) 6 How does the installation procedure insure that the). will I

not be deformed?

c) In reviewing support MS-1-004 003-572R, CYGNA has noted no gap shown on the drawing and both nuts on the same side of the beam, implying the U-bolt is " snug tight."

1 Have the local pipe stress and increased U-bolt load due

! o l i

to .500 F thermal expansion been considered? ,

l TUGC0 RESPONSE:

i i 5. (a) For the installation of U-bolts, snug tight is defined as i

" ... the tightness attained by a few impacts on an impact wrench When

or the full effort of a man using an ordinary spud wrench."

QC inspects U-bolts, if the nuts are tight and the U-bolt cannot  ;

be moved by hand, then their inspection criteria for snup and tight i

is satisfied. (b) The expression "T=0.2FD" is empirica may not be applicable to U-bolts, Visual inspection of the support in the field has shown no defor- l t

mation of the clip angle. QC, as part of their normal support  !

l inspection, assures that structural members (including these angles) 4 We have performed calculations that show that clip l

are not deformed.

i angles used for these type supports may be highly stressed, r perform their function.

l and as a matter of geod conservative practice, we are in the process i l of modifying these supports to remove the clip angles as support for i

. the U-bolts. (c) No gap shown on the design drawing with both nuts on the same side of the beam is the proper configuration for snug

• -==~=";~1~~~~~~"** -

Page 7 tight installation. Local pipe stresses and U-bolt stress increases due to diametrical' pipe expansion will be addressed at the completion of the ongoing,U-bolt test program. Our original assumption was that thermal stress effects on both the U-bolt and pipe would be negligible because the pipe and U-bolt were in contact and wrapped in. insulation, yielding a low temperature differential between the two. Calculations have been performed that substantiate our original assumptions. .

March 20,1984 CYGNA COMMENT:

1. For the MS springs, CYGNA has not found any design for hydrotest loads. The G&H stress analysis did not include any hyrdotest loading, but an approximation would be:

F hydro - normal (W water W. pipe + W pipe) -

" - 1.7 normal Since the springs are pinned during hydro, they do not need checking. What calculations have been done for the base plates, welds, rods, and bolting to ensure acceptability during hydrotest? If temporary supports are used, are there cales. for design?

(

TUGC0 RESPONSE:

1. As part of the design process for pipe supports on steam filled lines, a structural check is made to determine if the support i

can accomodate the hydro loads. If it cannot, a note is put on the drawing to the effect that a temporary support is required for hydrostatic testing. Temporary supports areThe " designed" in MS lines have i

accordance with site instruction CP-E!-4.0-30.

been successfully hydro tested indicating that adequate mechanisms do exist to assure system acceptability.

y CYGNA COMMENT:

2. In reviewing the frames, springs, sthuts, and snubbers, CYGNA has noted that in most cases, seismic displacements are not .1" 4

^

chec ked.'

While these displacements are typically small ( (

w/o SAM), (4, is" with SAM), has TUSI considered these in allowing for frame gaps or spring travel?

NOTE:

See attached sumary of displacements for MS inside containment and the CC line.

TUGC0 RESPONSE:

2. TUSI, and the pipe support design organizations, have not consistantly considered seismic displacements in their support '

designs. However, spring supports are designed within their

• • *

• e een e e mpee o.am e + e >e

-e * . m,, sana . e e.y e. e e- se e s *=-+-

- , _ - _ . . _ _ _ _ _ _ _ _ _ . _ _ _ . _ . , _ _ . . . _ _ , - _ _ , ._ ,.--,_-,_..-_,_,.__._,_,__.r_. , , , _ _ _ . - . _ , , , _ , _ _ _ , _

• Page 8 i l

working range (generally near mid-travel); a minimum k" travel exists beyond working range limits to reach a fully For most springs, more than extended or retracted position.This is sufficient to account for the small k" is available. For box frames, the seismic movements seismic displacements.

for the supports listed were checked against the designs.

In all cases, adequate clearance existed to allow combined '

thermal and seismic displacements.

i CYGNA COMMENT: \

3. In reviewing certain MS supports, CYGNA has noted instances ,

where beams with small gaps are used to provide stability, l instead of tightening U-bolts (supports MS-1-004-003-572R, l for example). In these cases, no analysis is done on the

" stability bumpers." CYGNA has performed calculations which show that the load on these " supports" could be quite high, assuming one accepts the instability of this structure during dynamic loading. Has TUS! used Also, this design for any supports where is the documentation with static compressive loads?

for the integrity of this support arrangement in general?

TUGC0 RESPONSE:

3. The " bumpers" supplied on MS 1-004 003 572R (in lieu of a snug U-bolt) for stability are designed to take an oscillating, momentary load (for a system at 20Hz, the applied loadthe Hence, onto the bumpers will act less than 0.05 seconds).

nature of the cyclic load assures stability in that there is j not sufficient time for a constant applied force to push the The calculations i bumpers back and allow the pipe to lift up.

which consider a static upward load are therefore erroneous.

The stress in the bumper steel stays within its elastic limit under the impact load and hence will return to itsThis original design l position when the load reverses and pushes down.

has not been used on any supports which would experience a static compressive load. Structural acceptability of the bumpers is based on the momentary load of less than 0.05 second duration. Size of the members, welds, were judged adequate by inspection.

March 21, 1984_

CYGNA COMMENT:

1. In reviewing MS 1-004-001-C725, CYGNA had the following questions:

a)

There appears to be a 7/16" flare bevel weld between itemsDoes  !

15 and 22. This does not seem possible due to Item 34.

TUSI haveL,ikewise, of this weld? documentation conforming the weld 5/16") between 26 (the andsize 157 and co r

• e

• 2*i' " ' E , *L*: ~: L'_'I

-_ , -- - __ Y~L 2 i *K"*

• Page 9

~

The model uses a fixed point at the embed plates (3. 6.12 b) 15 joints). Spec. 2323-55-30, Rev.1, requires these to be treated as pin joints, unless the embedment is stiffened.

Where is the stiffener in this calc, and has structural accepted this configuration?

c) Per ASME Appendix XVII, Para. 2442, shear loads on connections with both welds and bolts must be taken by -

the welds alone. Item 16 is attached to team #21 with both welds and Hilti's. The weld sizing calculations What is done use the bolts to share the shear load.

TU51's standard practice in this type of connection? Is this weld acceptable?

d)

The weight of the constant support itself is not included in the support design load for the frame. A #53 constant weighs - 600 lb, or 5% of the design load.

• What is TU5!'s standard practice for spring anchorage design?

TUGC0 RESPONSE:

1. (a) In the NPS! original design,is" well tube steel was used and the 7/16" flere bevel weld was possible. Modifications made The at the site changed most of the steel members to 3/8" wall tube.

weld between items 15 and 22 is a flare bevel weld such that the groove is filled and ground flush to facilitate installation of tem 34 CYGNA is correct in their observation that a 7/16" flare bevel weld does not exist, the size was inadvertantly not removed durinf) the revision process. The stiffeninti effect The of Item 34 on th's joint assures structural acceptabil'ty.

5/16" flare bevel weld between items 26 and 15 is an acceptable weld since the tube steel thickness is 3/8" (greater than 5/16").

Documentation confirming the size of these weids is on file with (b) See comments dated 3/22, QC in their inspection package. Items 15 and 19 are connected to response 1. (c) beam 21,notitem16.) (EditorialThe connection to the beam via embe plate (f.e., by welding) and As via baseplate such, each is (i.e., by capable of bolting)ing resist is at separate locations. using shear. It should be noted that Hilti joints are designed bolt shear allowables based on ultimate test loads divided by 5.

This is not the standard engineering approach to design a bearing or friction joint using code allowables for the bearing or friction condition. Using our design approa'ch, the Hilti joints, since they are pre torqued, would perfom as a friction joint within their working loads. At ultimate loads all joints (bearing or friction) would act as bearing joints (i.e., slip would occur in the friction connection). (Following comments, in response to CVGNA's comment on TU$1's standard practice are provided here for infomation only).

Connections of the type shown below have been used to some extent where space is limited, py"

- Fillet Weld Base plata --

'O

- "Hilti" Bolt )

Embedded plates gN#

- - - ~ * ~ ' " '

.-- n . . . .

page 10 Generally, the weld on the embedded plate is modelled as a pinned connection such that prying action in the baseplate is *naximized. Since the Hiltis have a lower capacity than the welds, it is assumed for qualification purposes that the Hiltis resist some shear. This is conservativeApplying in that it maximizes stresses in the weak link of the system.

the total shear in the connection to (d)the CVGNA weldrequesteddoes n6t cause an overloaded condition in the weld. '

information on supports where the weight of the support assembly could be a significant part of the design This load and question could effect the adequacy of the pipe support.

is limited to supports that were not analyzed using "$TRUDL" (i.e.,determinantstructures)sinceSTRUDLanalysisincludes all dead weight loads. The worst case condition would occur in supports having a relatively small design load and a heavy assembly, for example, a constant or spring support as indicated by CYGNA.

We have examined 20 as-built supports of various types and found that in no cases did the assembly weight exceed 6% of the design load used in the design calculations.

(CYGNA indicates dead load as approximately 5% of the design load.) In addition, the weight of the hardwareTherefore, is propor- the tional to the load it is designed to support.

variance in the design load due to the support assembly dead weight will remain within the 6% limit.

A 6 % variance in the design load is negligible when compared to the inherent design conservatisms of the variousTherefore, support components and construction techniques employed.

the design support is considered adequate.

CYGNA COMMENT:

2.

In the spring design loads for the main steam system. CVGNA cannot match the loads in the design calc to the loads in the data transmitted by G&H. How has the design load for a spring been calculated?

TUGC0 RESPONSE:

2. Design load for a spring is based on the latest information that the responsible pipe stress organization (Gibbs Piping & Hill or Westinghouse) Supplied to the support The designer.

or an insignificent load increase on a spring support.

piping analyst, using his discretion, may not transmit these load changes to the field support designer due to their in-The load differences in the examples consequential effect.

provided by CVGNA are indi:stive of this practice (i.e., the differences are negligible).

• so- -

• * * **'= e* *** e e-o e 4 e e., e amun, ammmenas.as em ao som + e sees e se e o.a= 4 -= e.

f Page 11 t

March 22,1984 CYGNA COMMENT: '

1. In CC-1-O31-008-S33R, CYGNA has noted another instance of a

" fixed" joint at an embed plate, rather than a " pin"ijoint (see comments on MS-1-004-001-C725, dated 3/21). In this case, a " fixed" assumption is needed, since self-weight was, considered. The civil Spec. (S5-30) requires stiffeners be provided..

a) Is piece 7 considered the stiffener?

b) Has the civil group provided guidelines for stiffeners?

c) What direction has TUSI given the support designers in using embed plates as " fixed" connections?

d) The designer appears to use conservative cales in determinir.o olate loads. Then, he uses the highest plate allowables. What is TUSI's direction for

• determing embed plate allowables?

TUGC0 RESPONSE:

1. a) For support CC-1-031-008-533R, Item 7 is considered as the stiffener. (b) The site Civil Group has not provided formal guidelines for stiffeners. (c) Any attachment to an embedded plate (whether it's a rear bracket, flat plate, tube steel member, etc.) will effectively stiffen the embedded Thus, the transfer plate in the local area where it attaches.

of moments into the embed through a stiffened area is a good assumption. (d) Embedded plate allowables are extracted from Specification 2323-55-30. Loads applied at stud locations permit use of a higher plate allowable while a lower allowable is used for loads acting at mid-span between studs. Field ~

verification of Nelson Stud location with respect to the applied load may be used in determing allowable limits.

CYGNA COMMENT:

2. In reviewing CC-1-028-022-533K, CYGNA has not found any checks of the hardware, other than references to previous design cales. CYGNA has checked piece 3Using (24" special riser the " max clamp) against the Grinnell 0 load sheets.

operating temp" of 200 F from the line list, the emergency allowable for the clamp is (94' c-c):

E=1.25(16,800)

= 21,000 lb Since E act = 21,964 lb, this support may fail, assuming the above calc. is OK.

Since the cale states the original design loads were higher, a) what is the correct procedure for calculating clamp allowables?

b)

Has TUSI considered that one snubber may be more highly loaded than the other, due to pipe rotation?

+

s

. Page 12 TUGC0 RESPONSE:

2, a)Reviewoftheogig(inaldesfcncalculationsshowsadesign temperature of 128 F not 200 which puts the clamp within its L.CD allowables, even with the larger design loadiconsidered.

The original calculations indicate there was a conscious effort made to determine the actual temperature in order to correct,1y design the clamp. b) Later.

-.---... .. .- ... . .. . .. .