ML20234C440
| ML20234C440 | |
| Person / Time | |
|---|---|
| Site: | Comanche Peak  |
| Issue date: | 05/01/1987 |
| From: | Richards J CYGNA ENERGY SERVICES |
| To: | Ellis J Citizens Association for Sound Energy |
| References | |
| 84042.53, NUDOCS 8707060471 | |
| Download: ML20234C440 (28) | |
Text
. _ _ -
y
=....
2121 N. Cahfornia Blvd, Suite 390, Walnut Creek, CA 94596 41s/934-5733 May 1,1987 84042.53
'r OS D'-
M N4 h (h(hf Mrs. Juanita Ellis President, CASE 1426 5. Polk Dallas, TX 75224 94
Subject:
Communications Report Transmittal No. 23 s
Independent Assessment Program - Phase 4 Comanche Peak Steam Electric Station TU Electric Job No. 84042
Dear Mrs. Ellis:
Enclosed please find some communciations reports associated with Cygna's audits of the Stone & Webster piping and pipe support verification activities.
If you have any questions or desire to discuss any of these documents, please do not hesitate to call.
Very truly yours, JdC7 4t &/ H / cc, cy b dessamyn Richards Administrative Assistant 3WR/am Attachments Mr. D. Pigott (Orrick, Herrington & Sutcliffe) w/ attachments cc:
Ms. A'. Vietti-Cook (USNRC) w/ attachments Mr. C. Grimes (USNRC) w/ attachments Mr. E. Siskin (SWEC) w/ attachments g{
Mr. W. Couns11 (TU Electric) w/ attachments M
4-Mr. L. Nace (TU Electric) w/ attachments
[
Mr. 3. Redding (TU Electric) w/ attachments g gig [
Mr. 3. Muffett (TU Electric) w/ attachments ZZZg ge (,Les I
l gg [Od I
l I
8707060471 870501
- I PDR ADOCK 05000445 i
A PDR
/,86{
I
(
San Francisco Boston Chicago N
t
1 Communications
)
k A h
ammmmmmmum
[Tetocon D Con **w M* port Company:
m proyect Joe No.
TU Electric o.w:
SA056 Seegoct LFbl5 1AP - Vna5e 4 Tm:
Z/23/87 Dual Ported Safety Relief Valves emos:
10:00 a.m.
Pano,ena:
Dick Zahorsky Crosby L.J. Weingart CES et
.t leem Commente Reid Action By Dick had the following comments regarding Crosby's position on flow imbalance in dual ported relief valves:
1.
Crosby considers no imbalance between outlets.
2.
Crosby has no test data to substantiate the equal flow assumption.
3.
Some of Crosby's competitors recommend use of a 60/40 flow split.
4 Crosby feels that any flow imbalance would be due to fabrication and installation tolerances on the dimension from the valve centerline to the valve outlet flange to the discharge elbow.
Dick recommended use of ANSI B31.1 nonmandatory Appendix 0 to determine discha ge loads.
r l
l I
1
{
4 i
_A A
a
+
!NVu b%w R~
~ ~
i+em;w;a 04,er4su+40n seo sweet
- m-
i i
l Distribution Sheet
)
Distribution:
N. Williams S. Tuminelli Project File J,, P.edding J Ellis i
A. Vietti-Cook L, Nace C Grimes W. Counsil J. Muffett J. fluss W. Horstmart D. Leong K. Parikh i
l j
PRJ:DISTLISTV l
.)
I Communications k%N k A ceasear a fosseen KCoasconesmesoci p,,,est TV Electric m ne.
e9uco CPSES IAP Phase 4 iians 3/17-19/87 suniest Audit of Stone & Webster's P1pe 5 tress has Cherry Hill, NJ and Pipe Support Reconcilation Procedures ji,,,
soston, na i
See text below for list of participants er er commens nee a acion er The audit was conducted for three days, March 17 and 18 at SWEC offices in Cherry Hill, NJ, and March 19, at SWEC offices in Boston, MA.
Participants were as follows:
All three days:
Ron Klause - SWEC Alan Chan - SWEC Stephen Tuminelli - Cygna Robert Pascaretta - Cygna March 17 and 18:
Fred Ogden * - SWEC Warren Wong * - SWEC Wyn Evans * - SWEC Tung Lu Wong * - SWEC Ozzie Bilgin * - SWEC Sean Chen * - SWEC Bill Dykstra * - SWEC T. Ginwala * - SWEC Syed Ali * - SWEC Yen Wu * - SWEC Nancy Williams - Cygna (March 17th only)
Indicates part time Gli nJ' WI)
LN M mem* ween See Attached Distribution sneet, me
f J
E Communications Repod 4L 6 i lilllilililliiiiiiiiiiiiiiiiii
[c7o7s#
nem comments v
March 19:
Dave Foster - SWEC 4
Ray Hankinson - SWEC Dan Van Duyne - SWEC L. A. Budiong - SWEC Fluid Transient - RIL 20 (Ref.1) - see also Ref7 3, App. S ; Ref. 4, pgs. 46, 52, and 87; Ref. 5, pg. 66; Ref. 7, Attach. 3-1; and Ref. 9, Attach. B.
Cygna's concern is that the current CPPP-7 procedures (Ref. 7 Attach. 3-1) do not provide adequate assurance that dynamic convergence is achieved when the fluid transient analysis is performed. Specifically, the concern involves four issues:
1.
The pipe models are built to accurately capture nodes to 50 Hz, yet analysis procedures allow modes in excess of 200 Hz to be used.
Hence, accuracy of modes beyond 50 Hz is not assured.
2.
The procedures allow use of modes beyond 50% of those available in the model.
Accuracy of these modes is not assured.
3.
During the step function analysis phase (Ref 7,
Attach. 3-1, Section 3.1.2 to 3.1.6), the analyst must determine the cutoff frequency for analysis using recognizable support forces.
The term " recognizable" must be quantified to assure that convergence is achie ved.
4 The current procedures do not require knowledge of the frequency content of the input.
The Stone & Webster response to these issues is as follows:
Page
=
Communications 4L ci Repod 181llll11111111111lllllllllli1
[eYoNy stem comments a.
Though the models are built for accuracy to 50 Hz, they are really good for at least 100 Hz for accurate bending modes (Ref. 9, Attach B). Since 100 Hz is beyond the significant frequency content of the input (see item d below), the bending modes will accurately respond to the transient. The procedures will be modified to require analyses to include all modes up to a minimum of 100 Hz.
b.
Since accurate modes may be captured up to 100 Hz, modes beyond the 50% available nunber will be local bending and axial leg modes.
Local bending modes which may be inaccurate will not respond to the transient anyway.
Axial leg modes are those where a leg mass oscillates as a rigid body supported by a support or supports at or near the elb ow(s).
Though these modes may be of high frequency and beyond the 50% nunber, they are still accurate because the pipe leg as a whole oscillates and the mass point spacing does not affect the total mass of the leg.
c.
Since the analyses will use a 100 Hz cutoff es a minimum, the recognizeable support forces only become important for those supports at or near elbows, i.e., to capture modes where a leg may oscillate as a rigid body (See, Item b).
In order to avoid the possibility that the analyst may
" recognize" support forces from branch lines connected to the leg or other effects, the analyst will be required to " recognize" at least 90% of L
the input step function value for support (s) at or near elbows.
The procedures will be modified to incorproate the 90% criterion and will define "at or near" to mean within approximately one-half the calculated mass point spacing for the line, d.
Stone and Webster has performed Fourier Transforms and/or Response Spectrum analysis for all blowdown / hammer transients and has determined that
(
none have significant frequency content beyond 70 3
12 J
f E
Communications Repod 4L i i ll1111lllllllllll11ll11lllll11
[eYoTYy nem comments Hz.
Mass Participation - RIL 1 (Ref.1) - See also Ref. 3 - App R; Ref. 4 - Pages 17, 56 and 83 Ref. 5 - Page 47; Ref. 7 Sect. 3.10.6.8; and Ref. 9 -
Attachment A.
The problems selected for the SWEC study (Ref. 9 -
Attachment A) were those that had the lowest frequency at which the Response Spectrum returns to the ZPA value. The difference between the approach of applying the missing man correction at 50Hz versus the ZPA frequency are nominal. Further, SWEC pointed out that Nl REG-1161 does not explicitly define an appropriate cutoff frequency.
(Sec./ Page 31 NUREG-1161).
Discussions between SWEC and one of the authors of NUREG-1161 indicates that they were more concerned that a missing mass correction would be applied rather than identifying the specific frequency to begin the correction at, as long as that frequency was reasonable.
Minimum Wall - RIL 27 (Ref.1) - See also Ref. 4 - Page 93; Ref 5 - Page 100; Ref. 6 - Pages 93 and 111; and Ref. 7 - Attachment 3-14 and PM-066 Rev.
1.
SWEC knows of no cases where a wall thickness below Code minimum has been accepted.
Procedures will be modified so that the SWEC Options Committee will decide if and on what basis a wall thickness below Code minimum will be accepted.
If this occurs, SWEC will notify Cygna. Otherwise, wall thickness below Code minimum will not be allowed.
Interface Anchors - RIL 28 (Ref.1) - See also Ref. 4 - Page 90; Ref. 7 - Attachment 4-10; and Ref.
9 - Attachment C.
Cygna does not sgree with the justification of the design procedures presented in Attachment C of Reference 9.
Specifically, justification for using a 4
12
Communications Report 4L 6 i 181llllllll1111llllllll1111ll1 1
Ac$oNy item Comments factor of safety equal to 2 needs to be provided.
Further, the assumption that an elbow and pipe bend always exists which will limit the moments at the anchor near to be stbstantiated. SWEC will reconsider their position.
Support Stiffness - RIL 13 (Ref. 2) - See also Ref. 3 - App E; Ref. 4 - Page 152; Ref 5 - Page 104; Ref. 7 - Section 3.10.8.2; and Ref. 8 - Encl. D and E.
The five problems used for the SWEC study (Ref. 8 -
Encl. E) were selected to include those problems with the largest variation in stiffness between adjacent supports. This is a biased sample intended to produce the largest expected difference between responses computed using generic versus calculated stiffness.
Further, TERA selected an additional five problems using different criteria.
The results from the TERA samples are similar to those for the SWEC study.
A563 Grade A Nuts Used with the High Strength Bolting RIL-28 (Ref. 2) - See also Ref. 4 - Page 156; Ref. 6 -
Page 67 and 95; and Ref. 7 - Attachment 4.5 and PM-110.
The A563 nuts also exist on high strength through-bolts. SWEC is aware of this and CPPP-7, Rev. 3 incorporates modifications to which this situation.
If the nuts are galvanized, the 0.6 factor used in the procedure would be too high. SWEC is currently investigating whether the nuts are galvanized.
Uncinched U-Bolts as a Two-Way Restraint -
RIL A3 (Ref. 2) - See also Ref. 6 - Page 52; Ref. 7 - -3; and Ref. 9 - Attachment D.
j Allowables for the static case includes friction forces while the dynamic case does not.
Moments and I
shears are computed using STRUDL models.
Interaction is the circular section interaction equation.
Allowables for Hilti Anchors Having Edge Distance Less 2 (Ref. 2) - See also Ref. 6 - Pages 72 Page of warm
I Communications t4 Ce c i Repod 11111ll1111K11111lll11111ll11
[cYo"nYy Item comments and 104; Ref. 7 - Attachment 4-4 and PM-099; and Ref.
9 - Attachment G.
In pure shear the SWEC data (Ref. 9 - Attachment G) shows side shear failure for Nelson Studs close to the edge which is similar to the Hilti data.
The overlap of the shear one due to applied load and that due to the wedge on the Hilti was discussed. Cygna is concerned that the load at failure may be affected by the stress field due to the wedge even though the l
failure modes of the studs and Hilti's are the same.
Further, Cygna is concerned that the shear /t 1
i Page of 1
Communications 4L Repod t i 1114lllllll1ll111lll11ll111111
[c7oNy item comments i
Hz.
L Mass Participr4 tion - RIL 1 (Ref.1) - See also Ref. 3 - App R; Ref. 4 - Pages 17, 56 and 83 Ref. 5 - Page 47; Ref. 7 Sect. 3.10.6.8; and Ref. 9 -
Attachment A.
The problems selected for the SWEC study (Ref. 9 -
Attachment A) were those that had the lowest frequency at which the Response Spectrun returns to the ZPA value.
The difference between the approach of applying the missing man correction at 50Hz versus the ZPA frequency are nominal. Further SWEC pointed out that NWEG-1161 does not explicitly define an appropriate cutoff frequency.
(Sec./ Page 31 NUREG-1161).
Discussions between SWEC and one of the authors of NWEG-1161 indicates that they were more concerned that a missing mass correction would be applied rather than identifying the specific frequency to begin the correction at, as long as that frequency was reasonable.
Minimum Wall - RIL 27 (Ref.1) - See also Ref. 4 - Page 93; Ref 5 - Page 100; Ref. 6 - Pages 93 and 111; and Ref. 7 - Attachment 3-14 and PM-066 Rev.
1.
SWEC knows of no cases where a wall thickness below Code minimum has been accepted.
Procedures will be modified so that the SWEC Options Committee will decide if and on what basis a wall thickness below Code minimum will be accepted.
If this occurs, SWEC will notify Cygna.
Otherwise, wall thickness below Code minimum will not be allowed.
Interface Anchors - RIL 28 (Ref.1) - See also Ref. 4 - / age 90; Ref. 7 - Attachment 4-10; and Ref.
9 - Attachment C.
Cygna does not agree with the justification of the design procedures presented in Attachment C of Reference 9.
Specifically, justification for using a Page o
Communications Report L
ii litillfillfillllilillllitillil
[c7lNy tiem comments factor of safety equal to 2 needs to be provided.
Further, the assumption that an elbow and pipe bend always exists which will limit the moments at the anchor near to be stbstantiated. 3WEC will reconsider i
their position.
Support Stiffness - RIL 13 (Ref. 2) - See also Ref. 3 - App E; Ref. 4 - Page 152; Ref 5 - Page 104;
{
Ref. 7 - Section 3.10.8.2; and Ref. 8 - Encl. D and E.
The five problems used for the SWEC study (Ref. 8 -
Encl. E) were selected to include those problems with the largest variation in stiffness between adjacent supports.
This is a biased sample intended to produce the largest expected difference between responses computed using generic versus calculated stiffness.
Further, TERA selected an additional five problems using different criteria.
The results from the TERA samples are similar to those for the SWEC study.
A563 Grade A Nuts Used with the High Strength Bolting RIL-28 (Ref. 2) - See also Ref. 4 - Page 156; Ref. 6 -
Page 67 and 95; and Ref. 7 - Attachment 4.5 and PM-110.
The A563 nuts also exist on high strength through-bolts. SWEC is aware of this and CPPP-7. Rev. 3 incorporates modifications which address this situation.
If the nuts are galvanized, the 0.6 factor used in the procedure would be too high.
SWEC is currently investigating whether the nuts are gal vanized.
Uncinched U-Bolts as a Two-Way Restraint -
RIL 43 (Ref. 2) - See also Ref. 6 - Page 52; Ref. 7 - -3; and Ref. 9 - Attachment D.
Allowables for the 'tatic case includes friction s
forces while the dynamic case does not.
Moments and shears are computed using STRUDL models.
Interection is the circular section interaction equation.
Allowables for Hilti Anchors Having Edge Distance Less 5
b
[
4
l Communications Repod t4 L 6 i 181111111111111111111111111111 l
[cSoNy item comments than 50 - RIL 42 (Ref. 2) - See also Ref. 6 - Pages 72 and 104; Ref. 7 - Attachment 4-4 and PM-099; and Ref.
9 - Attachment G.
In pure shear the SWEC data (Ref. 9 - Attachment G) shows side shear failure for Nelson Studs close to the edge which is similar to the Hilti data.
The overlap of the shear one due to applied load and that due to the wedge on the Hilti was discussed.
Cygna is concerned that the load at failure may be affected by the stress field due to the wedge even though the failure modes of the studs and Hilti's are the same.
Further, Cygna is concerned that the shear / tension f
interaction equation may not be applicable for Hilti i
bolts closer than 50 to the edge. SWEC's position is that the PM-099 procedures are conservative as long as the Hilti can be set, i.e., side blowout to the free i
4 edge does not occur.
Punching Shear - RIL 4 (Ref. 2) - See also Ref 3 - App V; Ref. 5 - Pages98-129; Ref. 6 - Page 42; and Ref. 7
- Attachment 4-13.
Tables 4-15.7 use beam equations.
Web crippling in the ttbe steel side walls becomes significant when i
beta is greater than or equal to 0.8 Web crippling doesn't have to be specifically addressed when washers are used because the washers spread the load out to the wall and crippling does not control.
Two-Bolt Baseplate Qualification Procedure -
{
RIL 39 (Ref. 2) - See also - Ref. 6 - Page 77; and Ref. 7 - PM-059.
In order to develop procedure 16, different plates
{
geometries were used varying the plate attributes as defined on page 1 of PM-059, e.g., plate size, l
attachment size, eccentricities, etc.
Each geometry was analyzed for four load cases, i.e., Fz, Mx, My and the sum of Fz, Mx and My.
Anr. lyses were performed using the SWEC nonlinear firste element baseplate program, BAP. These analyres provided stiffness, bolt loads and plate stresses.
The extent of nonlinearity i
I Page 0
c a
Communications Repod L4 (9 6 i 1111lll111:1..
?!!
UeYoNy nem comments 4
was assessed by comparing the sumed load case to the individual load cases.
Expected loads were applied, i.e., loads of a magnitude that a real plate would be expected to experience. Rigid plate theory was then used in an equalibrium solution to arrive at a basic form for the equations for bolt tension, shear and plate stress. Correlations were made between the rigid plate theory equations and the results of the finite element baseplate analyses to arrive at the factors used in the equations of PM-059.
Bolt Edge Distance - Effect on Bolts / Plate / Stress -
RIL 11 (Ref. 2)
- See Ref. 4 - Page 148.
The issue is that procedures allowed QC to accept edge distances which were different from those specified.
TV surveyed 60 large bore and 60 small bore plates and measured the edge distance (Ref. ER-ME-09, Rev. 0).
SWEC performed a study (GENX-082) to calculate the changes in plate response due to the difference in edge distance.
Using the TU report as the basis, SWEC found that the edge distance differences relative to plate size were largest for the large bore plates.
They then selected ten plates of the 60 which had the highest bolt interaction ratio using the "as designed" dimensions (the highest interaction in the sample is 0.52).
Further, they also selected ten plates of the sixty which had the largest variations in "as designed" versus "aseuilt" edge distance.
(Variations ranged from 5/8 to 1-1/4 inch).
This resulted in a total of 19 plates since one plate appeared in both categories.
These 19 were then analyzed using the "as built" dimensions.
The results showed that the maximum increase in interaction ratio is 5% and there is no increase in plate stress.
NPSI Rear Bracket Sizes - RIL 40 - (Ref. 2) See also Ref. 6 - Page 79; and Ref. 7 - Attachment 4-2 and PM-080.
A SWEC review of Grinnell data indicates that several Page of l
7 12
Communications Report A (.
i i lllllllll111111lll111111111111 flem Comments Ac o y
sizes of Grinnell rear brackets were used. SWEC will prepare a calculation enveloping the various sizes used and procedures will be modified appropriate.
Dual Strut /Sntbber Design - RIL 16 (Ref. 2) See also Ref. 3 - App L; Ref. 4 - Page 156; Ref. 6 - Page 53; Ref. 7, Attachment 4-8; and Ref. 9-Attachment E and F.
Cygna was concerned that the riser clamps were not rated for eccentric (one sided) loading. Review of the catalogues and discussions withe NPSI and Grinnell revealed the following:
Grine11: The design values contained in the catalogue i
are designed for a load inbalance, i.e.,100%
of the rated value may be carried on one side.
NPSI:
The design values asume a 50-50 load split between the two ends of the riser clamp.
Accordingly, when an eccentric load is used in the design, SWEC uses half of the stated catalogue value.
Cygna will confirm these two statements.
Sntbbers Close to Equipment Nozzles - RIL 7 (Ref.1) -
See also Ref. 4 - Pages 69, 87 and 102; Ref. 5 - Page 142; and Ref. 7 - Section 3.10.6.10.
SWEC does not have specific criteria concerning sntbbers proximate to equipment nozzles, however, each stress problem is reviewed by an experienced pipe stress engineer together with an experienced pipe support engineer prior to commencement of the actual pipe stress reconcilliation. This review is documentation in a Piping System Review Documentation (PSRD) calculation which is sometimes referred to a. a
" paper walkdown". These PSRD calculations contain directions to the pipe stress support engineers regarding the procedures to be used to reconcile the stress package, i.e., piping and its supports.
PSRDS typically contain guidance concerning:
8
$2
i Communications Repod 4L t i Illhilllllilillllllillll!!!!!
)
[Eo7*ey d
nem commena i
I o
special modeling techniques o
sntbbers proximate to rigids or nozzles o
thermal displacements i
o rigids proximate to rigids or nozzles o
support weights to be included with the piping o
support stiffness calculations o
unstable supports o
support orientation Cygna reviewed PSRD and pipe stress calculations for three stress problems.
They are:
CC-1-055B and PSRD-055B CS-1-N069 and PSRD-N069 CC-1-062D and PSRD-0620 i
Specifically, these calculations were reviewed to assess the guidance given concerning proximate snthb ers.
All sntbbers were reviewed.
i.e., those recommended for removal as well or those allowed to remain.
In all cases, Cygna concurred with the guidance given.
Force Distribution in Dotble Ported Safety Valves -
RIL 17 (Ref.1) - See also Ref. 3 - App Y; and Ref. 4
- Page 35 and 97.
No consensus could be reached concerning the application of 55/45 vs 60/40 urbalance for these val ves.
Consequently, Cygna reviewed calculations to determine if the 60/40 split would adversely impact the reconciliation of the affected systems.
These valves exist only on the Main Steam lines, all of I
which are similar, so Cygna reviewed one, calculation; MS-2-0230 Revision 0.
Increasing the split from 55/45 to 60/40 increases the torsion applied to the pipe by the valve by a factor of two.
The net vertical load remains the same and the horizontal load is negligable in both cases (See also App 0 to ASME section III).
The analysis assumes the response from all five valves occurs et the same time and that all Page of 9
12
i I
Communications
)
Report 4L i i M::
^:!!!!
[eENy item comments valves are biased so that the applied torsions due to the urbalance at each valve are additive.
Cygna agrees with SWEC that this is a conservative analysis. The effect of increasing the torsion by a factor of two increases the pipe stress but it is still within the allowable. All the torsion is reacted at the five-way restraint. An assessment of this support shows that it is easily capable of reacting the increased torion using ASME allowable stresses.
Line Contact Stress - RIL 37 (Ref. 2) - See also Ref.
5 - pages 92 and 122; and Ref. 7 - Attachment 4-6B and 4-6C.
This issue involves four separate items:
a.
Considering that e.g., 6a of Attachment 4-6B includes bending. SWEC presented a comparision of e.g., 6a results to finite element work performed by Westinghouse (WCAP-10627) for 10,24 and 32 inch diameter piping.
This comparision showed that 6a correlated very well with menbrane plus bending from the finite element work while it does not correlate to just menbrane alone, b.
Use of to = t + tn in Section 3.2.4 of Attachment 4-6B (See also Ref.10). SWEC performed a finite element study of the pads and pipe (Calculation No. 599.470.1.02-NP(B)-154) and analyzed approximately 20 configurations.
They concluded that use of tp in Attachment 4-6B equations provides an accurate prediction of pipe wall stresses.
However, tp underpredicts the pad stress by factors of 1.5 to 2.0 in a local area under the load. SWEC revised CPPP-7 procedures to use tp for pad stress calculation by revising the allowable downward to 0.75 Fy which effectively allows actual pad stress in a local area under the load to reach an ultimate stress of 1.5 Fy.
c.
Rationale for including preload only in term 10 Y2
Communications 4L n i Report lll111111lll1111!!1111111ll111
[c?oT$
Item comments Sn3 *** Attached 4-6c.
The preload is not considered to produce primary stress, therefore, it is not included in Code equation 8 or 9, i.e., -6c equations 1 or 2.
Also, it is not considered to produce fatigue stress, therefore, it is not included in Code equation 10 or 11, i.e., Attachment 4-6c Equations 3a or 'b.
- in However, it is included in term for Sp1 -6c equation 4 which is a deflection limit using all loads.
d.
Applicability of equation 9, Section 3.2.2. in -6c. -6c requires two separate evaluations to be made for pipe wall local stress.
One is due to the U4 olt around the pipe with a load distribution strap (equations 8) or without (equations 9).
The other is due to the yoke bearing pad per Section 3.4.
Data provided in RLCA/FA & 2/01-06/004 which is a finite element study stbstantiated the validity of the Attachment 4-6c procedures.
Weld Stresses Between pad and Pipe - See Ref. 7 - -6B and Attachment 4-2; and Ref.10.
Current procedures consider only the loads from the trunnion for the design of the weld.
SWEC has determined that the existing procedures can permit configrations which will result in high weld stress due to pressure. A study is still in progress to determine impact if any, on the existing designs. The result will be documented in calculation 15454-NZ(B)-GENX-035, Revision 2.
11 32 N'
r Communications Repod 4L i i 181llllllll1111111111111111111 AcYoNy item comments
' REFERENCES 1.
Pipe Stress Review Issues List, Cygna Energy Services, tevision 3, January 6,1987.
2.
Pipe Support Review Issues List, Cygna Energy Services, tevision 3, Decenber 29, 1986.
3.
3eneric Technical Issues Report, Texas Utilities / Stone &
dsster, Revision 0, Status Date, June 27, 1986.
4 Transcript of meeting between SWEC/TU/Cygna at Cherry Hill, New Jersey, Novenber 13, 1986.
5.
franscript of meeting between SWEC/TV/Cygna at Cherry Hill, New Jersey, Novenber 14, 1986.
6.
'franscript of meeting between SWEC/TV/Cygna at Glen Rose, Texas,
)ecenber 15 and 16,1986.
7.
)esign Criteria for Pipe Stress and Pipe Supports, CPSES Units 1 knd 2, Stone & Web ster Engineering Corporation, CPPP-7, Rev. 2, April 25,1986.
8.
.etter from W.G. Counsil (TV) to N.H. Williams (Cygna), TU Log iXX-6133, Decenb er 4,1986.
9.
.etter from W.G. Counsil (TV) to N.H. Williams (Cygna), TV Log f XX-6280, Feb ruary 18, 1987.
10.
Audit to Review Stone & Webster Procedures for Local Pipe Stress 1 valuation, Cygna Conference Report, S.C. Tumminelli,
')ecenb er 30, 1986.
12 Page T2
Distribution Sheet Distribution:
N. Williams S. Tuminelli Project File J. Redding J. Ellis
" A. ' Vietti-Cook L. Nace C. Grirres W. Counsil J. Muffett J. Russ W. Horstman i
D. Leong K. Parikh E. Siskin l
i PRJ:DISTLISTi
k' Communicatisns k%d t A ammmmmmmum CES gym am%
conwear:
Propet Comanche Peak TV Electric Joo m 84056 CPSES IAP Phase 4 Does:
3/5/87 ampst Cable Tray Support Review has:
9:32 a.m.
Stability of Pipe Supports M
Walnut Creek,CA D. Rencher TV Electric y
Cygna J. Richar ds et M
Commenes meea Acton Dy Dave called and ask if we could telecopy letter 84042.074 (2/19/85) to him. Af ter the conversation, the document was transmitted.
)
\\
l
,4 t\\
!i b Nl))llfM4LAS b*
1 h
~
See Attached Distribution Sheet.
~
J
Distributirn Sheet Distribution:
N. Williams S. Tuminelli Project File J. Redding J. Ellis A.' Vietti-Cook L. Nace C. Grimes W. Counsil J. Huffett E. Siskin J
i PRJ:DISTLIST i l
I 1
Communications k%N k &
I Ossapear X Teeseen
. Centeronse Repeft p,,, set.
TU tjectr1c o*vao mg CPSES I AP gj gef g eiasset Aud.its of SWEG/5upport dL1Trness
- yes, AA:au a.m.
weinuc u een Alan Chan, SWEC af Jim Muffet, Owen Lowe TUE S. Tumminelli, N. Williams, R. Pascaretta, L. Weingart,
Cygna
"n ceawnense n.e a as, n ey Cygna provided the list of items for future audit and/or discussion:
Audit
- 1) Sr.thbers close to equipment nozzles.
- 2) Dual Ported Safety Valves - Cygna needs to check stresses and evaluate the impact of inbalance loads.
3)
Pipe minimum wall below code allowable 4) 2/18/87 SWEC stbmittals
- 5) Support stability
- 6) Support stiffnesses
- 7) Line contact stress 8) 2-Bolt base plates l
- 9) Use of litC 316 tolerances All parties agreed for Cygna to be in the SWEC offices in Cherry Hill on March 17, and in Boston on March 18 and 19, l
to perform the required audits.
Steve Tununinelli explained Cygna's interpretation of the SWEC Generic Stiffness Study.
Based on three simple piping models generic stiffness values (Kg) were developed so that pipe system frequencies were near the flat region on the pipe frequency vs. support stiffness curves. The flat region is chat portion of the curve where the support stiffness no le'ger affects system frequency.
Then minimum stiffness values (Km) were derived such that the maximum reduction of frequency from that using the generic O
8.
(
A l
f e
n...wa n4,+
4w.,+4m.
cwmm+
r
Communications t4 L ci Repod lll1111111111llllll11lll111111
)
[e*tfoTYy nem comments values would be no more than ten percent for any of the three simple piping models. Tnese two stiffnesses values (Kg and Km) were implemented in production as follows:
1 1)
If a support stiffness is greater than Km use Kg.
2)
If.a support stiffness is less than Km use the actual stiffness.
3)
If a support stiffness is less than 20% of Km consider that support for removal.
Then, to verify the procedure, SWEC studied the responses of five prablems analyzed using the production rules above and reanalyzed using the actual stiffness values for all supports.
SWEC agreed with Cygna's understanding. Steve then explained the following Cygna concerns / questions:
1.
Was the sample in the study composed of all hot or all cold lines?
In the case of a hot line, the thermal loads / stresses will tend to be a stbstantial part of the overall design loads / stresses. This would tend to minimize the impact of increased seismic loads (the sample analyses showed 19% of all supports examined had seismic load increases greater than 10% when comparing the analyses using the generic stiffness vs.
the analyses using the actual stiffness) 2.
What is the basis for extrapolating the percentages shown in the study to encompass all piping systems in the plant?
SWEC explained that the purpose of the study was to show that changes in support loads, as a result of using the generic stiffness method, were small.
The selected problems in the sample analyses were considered to be among those with the greatest stiffness variations among supports. This selection conservatively biased the study.
SWEC intends to address the Cygna concerns in writing.
This should be ready for review when Cygna is in Cherry Hill.
Cygna pointed out that the SWEC response must address the effect of a potential eroded margin for the pipe supports.
Page of 1020 0t b
e,*
- Distributian Sheet Distribution:
N. Williams S. Tumminelli Project File J. Redding J. Ellis
' A.: Vietti-Cook L. Nace C. Grimes W. Counsil J. Muffett E. Siskin PRJ:DISTLIST i
E Communications Report Comoeay-
%Toiseen O Centevense Report prosest TV Electric m ee 64056 CPSES IAP Phase 4 g,,,;
4/8/87 W
Audit Schedule Twa:
1:20 p.m.
mes*
Boston A. Chan SWEC S. Tuminelli Cygna W
m Commenes neo action ey We have scheduled an audit to review remaining pipe stress and pipe support issues.
It will be held on Friday, April 10,1987 at 9:30a.m. in the SWEC Boston
- Office, l
i In. 3A h
T.,
IIr Y $l],lk LA A W
W h
See' A'ttadh'6d Distributidh Sheet.
unus
F
- " Distribution Sheet 2
Distribution:
N. Williams S. Tumminelli Project File J. Redding J. Ellis A. Vietti-Cook L. Nace C. Grimes W, Counsil J. Muffett E. Siskin PRJ:DISTLIST i i
m Communications 6 i Report n
[Te4 econ o conkronce Report company:
Protect.
Joe No.
TV Electric oew:
84056 subrect CF5t5 1AF - Phase 4 Tsrne:
2/23/87 Dual Ported Safety Relief Valves n ce:
10:00 a.m.
Participants.
og Crosby j
Dick Zahorsky
\\
L.J. Weingart CES of item Commants Reg'd Action By Dick had the following comments regarding Crosby's position on flow imbalance in dual ported relief valves:
i 1.
Crosby considers no imbalance between outlets.
2.
Crosby has no test data to substantiate the equal flow assumption.
)
3.
Some of Crosby's competitors recommend use of a 60/40 flow split.
4.
Crosby feels that any flow imbalance would be due to fabrication and installation tolerances on the dimension
{
frcm the valve centerline to the valve outlet flange to i
the discharge elbow.
{
Dick recommended use of ANSI B31.1 nonmandatory Appendix 0 to determine discharge loads.
I i
I t
i
{
lNMJ blhnw
'~
c'
\\
s e e item;nein4serssution sneee o
t - u - -- ---- --- J
.., 6 Distribution Sheet Distribution:
N. Williams S. Tumminelli Project File J. Redding J. Ellis
. A. Vietti-Cook l
L. Nace C. Grimes W. Counsil J. Muffett J. Russ W. Horstman D. Leong K. Parikh PRJ:DISTLISTv
-