ML20210B073
| ML20210B073 | |
| Person / Time | |
|---|---|
| Site: | Dresden, 05000000 |
| Issue date: | 07/31/1973 |
| From: | COMMONWEALTH EDISON CO. |
| To: | |
| Shared Package | |
| ML20209J032 | List:
|
| References | |
| FOIA-87-40 31, NUDOCS 8705050224 | |
| Download: ML20210B073 (18) | |
Text
.___ ____
DRESDEN STATION SPECIAL REPORT NO. 31 TORUS RING HEADER SUPPORT MODIFICATION AEC Dockets 50-237 50-249 COMMONWEALTH EDISON COMPANY
INDEX PAGE
1.0 INTRODUCTION
1-1 2.0 DESCRIPTICN OF RING HE'. DER AND SUPPORT SYSTEM 2.1 Rine Header 2-1 2.2 Header Support System 2-1 23 IIctbod of Fabrication and Assembly 2-2 24 Header Sapport Desirn 2-2 30 DESCRIFTION OF DS9'ICIENCIES FOUND 3.1 Fabrication and Assembly Deficiencies 3-1 3.2 Design Deficiencies 3-2 1.0 00itRI::CTIVE ACTION 4-1 4
50 EVALUATION 51 Static Loadins 5-1 52 Dynamic Loading 5-3
6.0 CONCLUSION
S 6-1
~w
1.0 INTRODUCTION
During startup testing of relief valves for Quad-Cities Unit 2 on Sunday, May 28, 1972, it was discovered that four pipe hangers for the 24-inch torus suction header had failed.
The reactor was pronptly shut down for investigation and repair of the failed hancers.
On June 29,1972 Dresden Unit II and on July 1, 1972 Dresden Unit III were similarly inspected.
The 2h-inch soction header encircles the torus and provides a manifold for the auction of the various ECCS pumps.
The header is connected to the torus by four 20-inch OD pines and is supported by fourteen hangers which ara connected to the torus shell.
The pipe hangers consist of donbic strap horizontal and vertical nembers bolted to "ussets whiah arc welded to the torus shell and the 24-inch pipe.
Investigation revealed a sienificant number of fa'crication deficiencies of the followine types on both units.
1.
Bolt holes in cusset: and straps cut with a torch.
2.
Doubl holes in same straps 3
Hole locntions too close to end of straps it.
Bolts with short shank sections and thus chear loads on tnreaded section of bolts.
1 I
l-1
~_
}
2.0 DESCRIPTION
OF BING HEADER AND SUPPORT SYSTEM 2.1 Bine Header l
The Dresden Unit 2, and Unit 3 containment vessels each consist of a light bulb shaped drywell and concentric l
torus shaped suppression chamber.
The suppression chamber is filled with water to approximately mid-depth E7d pump suction header (ring header) is attached to the outside of the torus.
This header is a 24-inch OD pipe 4
and is connected to the torus by short lenghts of 20-inch OD pipe.
In addition to the four 20-inch header inlet nozzles, there are six outlet-nozzles which are used to 1
withdraw water from the header f or the Core Spray System, High Pressure Coolant Injection System, and the Low Pressure Coolant Inj ection System.
The headers are shown i
schematically in Figures 2.1 and 2.2.
l 2.2 Header Sunnort System The 24-inch ring header (suction header) is supported by the four 20-inch pipe connections and fourteen hanger 1
i assemblies.
The hanger assemblies have horizontal and j
vertical members attached to gussets on the torus shell and the 24-inch header.
Details of the assemblies are shown schematically in Figure 2 3 The gussets consist of 1/2" thick plate welded to 1" thick pad plate which is then welded to the torus shell and a 1/2" thick collar type plate welded to the 24" pipe.
The torus diameter is
(
very large resulting in a radius to thickness ('R/t) ratio i
that is very large (300).
The R/t ratio is a measure of
(
the flexibility of the shell and its ability to withstand concentrated loads.
The pad plate is used at the torus shell attachment in order to spread the applied load over l
l sufficient area to minimize stresses and deformations.
I The smaller diameter pipe is stiff snough (H/t ratio of 30) to resist the applied load by transfer through the yoke-type collar.
l The horizontal and vertical double hanger straps originally installed were 2-1/2" wide by 1/2" thick with a 13/16" diameter hole at each end for attachment to the gussets with 3/4" diameter A307 bolts.
Figure 2-1 shows the ring header supports, inlet nozzles l
and outlet nozzles along with their approximate orientation.
i l
l l
2-1 i
~
2.3 Method of Fabrication and Assembly As discussed in part 2.1 and 2.2 the ring header and hanger assemblies are shown on CBI drawing Nos. 224, 228.
The hanger clip assemblies were welded to the torus shell prior to the shell being assembled in the basement of the reactor building.
With the torus assembled in the basement the shop built header subasramblies were attached to the torus pene'tration at the field weld joints.
With thase four subassemblies welded in pla ce, the remaining seguents of the header were welded together utilizing temporary supports for positioning.
The completion of the entire header welding was followed by the attachment of the collar type gussets to the 24 pipe header.
Because of the allowable fabrication tolerances on both the major and minor diameters of the torus as well as the ring header and the positioning of the header, the attach-ment of the horizontal and vertical hanger straps could not be completed with the as-built straps.
The alignment of the horizontal and vertical straps required some adjustments and strap mosifications.
The modifications required to complete the alirnment and assembly of the pipe han2ar straps consisted of some torch cutting of bolt holcs in the collar gussets and hanger straps as well as torch cutting the straps length to s uit each installation.
The original holes were drilled.
The hangar straps wero connected to the gussets with 3/h",
10UUC cap screws (bolts threaded full length).
Due to the 1
l torch cutting of some of the hole s in the vertical straps, there was some misalignment of the bolts and uneven distribution of the load.
2 4 Hender,,Su2 port Nsian i
The hancer sitpport assemblies were originally desicned for the static dead load of the 24" header plus horizontal and vertical seismic loads.
The horizontal acceleration, as well as the vertical accelerations used in the Dresden analysis are the same as used in the Quad Cities Special Report No. 5.
The oricinal desien was based on a computed maximum hanter load of approximatcly 8,000 lbs. with the load beine approximately equal for all hangers.
9 i
2-2
e 0
e f/G-2/
4fmr 2 Rws-HrnosR 4
6
+
o-4 I
A
... m
- z...
t q$D,?0, yay,
- occ,
GO
,{.
11-Ibs 13 It b
s if s%c N
/s
[ vcencm. ana-m v
q l
,,Olf{
~"k.??.;.il a czc.'
3 HN.I
e 9
9 Re-
- 2. 2 ON/T3 RiDc-
//EADGK i
\\
l
~
h" S*
6 y
+
I 7
I
+ 3 h
9 2
,tg con (OhN,
4 44 ClRGi', g j
. /5
/
b
/[
p 1%
$/
osavau nw cc o
/*o d
[
9 "') uni.3e m,n
\\s,
~.
LPCI g p, N: E.r.*..
(gf
/! Ell
Flce 2.3 1
h Qo, kk m
y 'o t/E(in"c. A L CE*
t tAPP&{W HAN &GR 0%
e
$ *g g
18 L.
c3
-1 1
R+g O*U'~
0 NO g
flCRI % U T*0L
.ruspeer H EA DEGt.
HAN&ER T 9 9tc.f L S 4CTIo id H M90t it StAPPORT Rsati.j[
~
30 DESCRIPTION OF DEFICIENCIES FOUND 3.1 Fabrication and/or Assembiv Deficiencies It can be seen from Section 2.3 that the methods used to correct the misalignment problems encountered during assembly were less than desirable.
Burnine holes for bolts, regardless of the craftsman, results in a non-uniform bearing surface for t he bolts.
Although the contract drawines did not call for bolts with a clean shank, the original stress analysis was
~
based on an unthreaded shank, 3/4" diameter bolt.,/k" The effective cross-sectional area 6f an un' threaded'3 bolt shank is.4416 square inches versus.302 square inches for the threaded shank.
This would have increased the calculated failure load in double shear from 27,180 lbs.
to 39,760 lbs. for the bolted connections.
The use of 3/4" diameter bolts with an unthreaded shank would therefore have been desirable.
Since the original hanger system did not have provisions for adjustments, the exact distribution of loads on the support system were different from the theoretically calculated loading.
The loading on each vertical hanger in the original installation was determined.
The measurements were made by utilizing the original straps and then applying a force with a hydraulic jack to each hancer point until the bolt in t he connec+ ion was loose indicating that the jack was carryinc all of the load.
The wei ht in pohnds was determined from the pressure of the hydraulic fluid in the jack.
The measured loads are shown in Table 3-1.
TABLE 3-1 DRESDEN UNIT NO. 2 ORIGINAL S tPPORT SYSTEM HA'TGER LOADS Bay No.
Load Bav No.
Load 1
Wk 9
P. k 2
12.8k 10 9.8k 3
7.hk 11 0.Ok 4
13.5k 12 2.4k 5
0.0k 13 15 7k 6
14 7
0.0k 15 15.7k 8
16.7k 16 lo,eg
- No hanger at this location 3-1
4 TABIE 3-2 DRESDEN UNIT 3 ORIGINAL SUPPORT SYSTEM HANGER LOADS Bay Nb Load Bay No.
Load 1
7.2k 9
8.7k 2
17 9k 10 10.8k 3
0.0k 11 10 3k 4
13.4k 12 16.1k 5
8.9k 13 11.0k 6
14 7
10 3k 15 9.4k 8
6.7k 16 9.0k
- No hanger at this location It is obvious that the distribution of load is not uniform in that one support was carrying a load of 17,900 lbs., whereas some supports were carrying no load.
This inequity in load distribution indicates that some more precise method of hanging the ring header should have been used.
32 Desien Deficiencies Computing an average of the loads shown on Table 3-1 and 3-2 i
results in a value of approximately 9,000 lbs. per hanger.
The original maximum computed static load to which the hangers were designed is approximately 8,000 lbs.
The inequities in the load distribution shown in Tables 3-1,2 and the average load per hanger indicate that the method used in designing the hanger assemblies was inadequate from a purely static condition approach.
The wide range of load distribution indicates that the hanger design perhaps should have called for some method of adjustment to make up for the i
differences in the dimensions that result from allowances for fabrication of the torus and the 24" pipe header.
f f
l 3-2 i
-. _... - - _ - - _ _... _ _ _ _ _ _ _. _.. _ - _ - -. _, _ _ _ ~.., -. - _ _ _ _ _ _ _. _ _ _
~
40 CORRE3TIVE ACTION The following corrective action was taken to correct the header support deficiencies:
1.
The bolt holes were drilled out to 1 1/16" diameter for 1" bolts 2.
All horizontal and vertical hancer straps wer6 replaced with 1/2" t"ick by 3" wide, 36,000 pai yield material with 1 1/16' diameter bolt holes.
The approximate failure load for these straps (in pairs) is 54,375 lbs.
3.
New 1" diameter A325 high streneth bolts with 1 1/2" unthreaded length were installed with positive lockine technicues included (double nutted).
The approximate double shear failure load for these l' bolts is 70,686 lbs.
4 The lengths of the replacement s traps were adjusted to provide a unifonn distribution of static loads between the hangers, in the approximate 1
range of 8 to 10 thousand pounds.
5 Over sized holes and double holes in the gusset plates on the rinq header were filled with wcld metal dye penetrant tested and redrilled t
4-1
~
5.0 EVALUM 51 Static Loading As part of the bolt and hanger atrap replacement, a hydraulic jack with dial pressure gauge was used to measure the actual loads at each support point.
The jack was raised enough to pick up the header load at each poirt.
Af ter the loads were reviewed, some modification in the strap lengths effected a change in the load distribution pattern of the entire system.
The modifyinr of strap lengths resulted in the static dead load distribution shown in Table 5-1.
TABLE 5-l*
DRESDEN UNIT NO. 2 FINAL SUPPORT SYST34 HANGER LOADS BAY NO.
LOAD BAY NO.
LOAD 1
8.9k 9
6.9k l
2 8.1k 10 8.lk 3
8.6k 11 6.7k 4
9.lk 12 5.9k 5
9 4k 13 9 4k 6
4 7
7.8k 15 8 4k 8
7.9k 16 7.9k
- No hanger in this location TABLE 5-2 DRESDEN UNIT NO. 3 FINAL SUPPORT SYST31 HANGER LOADS BAY NO.
LOAD BAY UO.
LOAD 1
8.9k 9'
8.6k 2
8 4k 10 8.hk 8.6k 11 9 3k 8.0k 12 6.ltk 5
8.9k 13 9.3k 6
lh o
7 8.hk 15 9.4k o
8.7k 16 8.6k 5-1
I Chicago Bridre and Iron Company' suggested the load distribution given in Table 5-3 and 5-4 TABLE 5-3 DRESDM UNIT 2 CB & I SUGGESTED LOADING FOR TORUS SUCTION HEADER HANGERS BAY NO.
LOAD BAY HO.
LOAD 1
lik 9
6.9k 2
7.9k 10 9.2k 3
7 4k 11 12.8k 4
7 4k 12 5 5k 5
7 7k 13 16.7k 6
4 7
7.7k 15 12.0k 8
7 7k 16 6.2k
- No hanger located at this point TABLE 5-4 DRESDEIT UNIT 3, CB & I SUGGESTED LOADING FOR TORUS SUCTIOil HEADER HANGERS BAY NO.
LOAD BAY NO.
LOAD 1
12.8k 9
7 4k 2
9.2k 10 7.9k 3
6.9k 11 12.8k 4
7.7k 12 6.2k 5
7 7k 13 12.0k 6
4 7
7.7k 15 16.7k 8
7 4k 16 5 5k
- No hanger at this location The final loadings on the hanners after the modifications on Dresden Units 2 and 3 were more uniform then the CB & I recommendations.
C7 & I was contacted on this matter and stated that the more uniform loadinic were acceptable.
5-2
52 Dynamic Leadine In order to analyze the header support system for the maximum postulated load on the vertical hanger, it is necessary to combine the static dead load with the seismic load as well as the maximum re31ef line discharge load.
This procedure was followed for Quad Cities and exact values were Catermined.
For the Dresden Units where there are fourteen torus suction header hangers versus the twelve at Quad Cities and whereas the Dresden seismic accelerations on the ring header would be.067g vertical and.20g horizontal as compared to Quad Cities where the seismic accelerations used were.08g' vertical and.40g horizontal it was determined that application of Quad Cities seismic loads to Dresden hangers would result in conservative estimations of actual loads at Dresden.
In addition since both Dresden Units 2 and 3 and Quad Cities Units 1 and 2 are basically alike, the electromatic relief valve operation induced hanger loading from Quad Cities were used in the Dresden analysis to evaluate the total load anticipated on each hanger with static, seismic and electromatic i
discharge loads acting simultaneously.
Table 5 5 gives a summary of the respective values used in evaluating these loads.
TABIE 5-5 COMPONENTS OF TOTAL LOAD PER HANGER ASSUMING RELIEF VALVE ACTUATION OCCURRING SIMULTANEOUSLY WITH MAXIMUM SEISMIC LOADING Total Static Vertical
- 800 pounds
- 10,200 load per Load Per
+
vertical seismic
+
pounds
=
vertical Hanger load OBE relief hanger valve actuation induced load
~
Total Static
+
- 4000 pounds
' 3,400 load per
=
horizontal horizontal horizontal seismic
+ pounds relief hanger load per hanger load OBE valve (0) actuation induced load
- These valves were taken from the General Electric Generic Beport NEDO - 10859, Steam Vent Phenomena and Structural Response of the BWR Torus (mark 1 containment).
OBE Operating Basis Earthquake.
5-3
TABLE 5-6 Dresden Unit 2 Total maximum load on vertical and horizontal torus suction header hangers including static, seismic, and relief valve actuation induced loads.
Vertical Horizortal Hanger Hanger Bay No.
Load Load 1
27,460 16,600 2
26,720 3
27,210 a
4 27,700 5
27,950 6
7 26,350 8
26,470 9
25,500 10 26,700 11 25,250 12 24,544 13 27,950 14 15 26,960 16 26,470
'No hanger at this location.
S 5-4
e TABLE 5 7 DRESDEN UNIT 3
~
Bay No.
Vertical Hanger Horizontal Load Hanner Lead 1
27,456 16,600 2
27,042 3
27,210 4
27,358 5
27,456 6
7 26,964 8
27,459 9
27 160 10 26,,964 11 27,948 12 26,964 13
- ,948 27 14 15 27,948 16 27,210 No hanger at this location D
5-5 m
From Table 5 6 and 5 7 the maximum load on any hanger is 27,950 pounds.
Since the failure load for the vertical hanger is 54.375 pounds with the failure point being at the strap bolt holes the safety factor on the worst case hanger is 1 95 v
s 5-6 e
e
6.0 CONCLUSION
S The Quad Cities test program verified that the relief line discharge sequence introduces substantial loads into the torus and header systems.
The relief line loads, the torch out holes and the poor initial installation load distribution necessitated the repair and redesign of the torus suction header hangers on Dresden Units 2 and 3.
Quad Cities test programs were used to determine the stresses which would be imposed on the Dresden hangers.
The revised support system is adequate for a maximum postulated total hanger load of 27,950 lbs.
The approximate failure load has been determined to be 54,375 lbs. with the failure point being the pin holes in the hanger straps.
These postulated loads are based on maximum relative displacement data taken during the Quad Cities relief valve testing as well as design seismic and static loading.
The postulated loads were taken conservatively high for the purpose of evaluating the revised support system.
The torus suction header hanger modifications on both Dresden Unit 2 and 3 were initiated on July 2,1972 it was completed on Dresden Unit 2 on November 8,1972, and on May 25, 1973 on Dresden Unit 3.
e B
6-1 i
e
_