ML19305D748
| ML19305D748 | |
| Person / Time | |
|---|---|
| Site: | Clinch River |
| Issue date: | 10/31/1979 |
| From: | WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| To: | |
| Shared Package | |
| ML19305D743 | List: |
| References | |
| WARD-D-0218, WARD-D-0218-R1, WARD-D-218, WARD-D-218-R1, NUDOCS 8004150480 | |
| Download: ML19305D748 (18) | |
Text
'
\\
l WARD-D-0218 Category 2 p @h\\
U Revision 1 O
Oct. 79 Clinch River Breeder Reactor Plant STRUCTURAL RESPONSE OF CRBRP.
SCALE MODELS TO A SIMULATED HYPOTHETICAL CORE DISRUPTIVE ACCIDENT O
OCTOBER 1978 Prepared for the United States Department of Energy under contracts EY-76-C-15-2395 and EY-76-C-15-0003.
Any Further Distribution by any Holder of this Document or of the Data Therein to Third Parties Representing Foreign Interest, Foreign Govern-ments, Foreign Companies and Foreign Subsidi-aries or Foreign Divisions of U.S. Companies Should be Coordinated with the Director, Division of Reactor Research and Technology, United States Department of Energy.
O
@ Westinghouse Electric Corporation ADV ANCE0 RE ACTORS DIVISION "W
BOX 158 M ADISON, PE NNSY LV ANI A 15663 s.99413o+3 0
O WARD-D 0218, Revision 1. Affected Pages 1.
Replace p36 (Figure is poorly reproduced).
2.
Replace Figure 3.3-18 (original drawing is poorly reproduced).
3.
Replace Figure 3.3-20 (original drawing is poorly reproduced).
4.
Replace p99 (" lower" should be " higher" in line 5).
f 5.
Replace Table 4.1-3 (some data are in error).
6.
Replace Figure 4.1-3 (update Title).
7.
Replace Figure 4.1-42 (the " pretest" profile is incorrect).
8.
Replace pl91 ("Section 3.3.3.3" in line 4, paragraph 3, should be l
"Section 3.3.3.2").
i O
l l
I i
+
1 i
l l
l O
l l
3.3.3.2 The SM-4 and SM-5 Models and Instrumentation
?\\
V Figures 3.3-16 and 3.3-17 show the SM-4 and SM-5 models respectively with their instrumentation. While these models are identical, SM-5 r
has over twice the instrumentation of SM-4.
In addition to the SM-2 and I
SM-3 features, each of these models has a three-rotating-plug head, a more prototypic core support structure, a vessel bottom head, a thermal r
I liner, a horizontal baffle, an in-vessel transfer machine (IVTM) column and a vessel wall whose thickness varies in a more prototypic manner.
The three rotating plug head is perhaps the.most important additional feature since it provides direct experimental confirmation of the ability of the three-plug head to withstand the scaled SMBDB loads. The under-head shielding j
was modelled accurately enough so that the effect of shielding flexibility on the load transmitted to the head was taken into account. The core support
[
plate thickness was accurately modelled and blind holes were drilled to f
account for the inlet module penetrations. The variation in core barrel wall j
thickness was also accurately represented. A vessel bottom head was f
added since this structure and the liquid above it will affect the dynamic response of the core support plate. The core plate response in turn can f
affect the kinetic energy imparted to the liquid slug. The thennal liner i
is important for prototypic response since it has the effect of shielding l
the vessel wall from the direct pressure loads in the upper plenum.
It j
will also affect the load felt on the head. The IVTM column was included since it could potentially produce an asymmetric response of the UIS.
f Finally the vessel wall thickness was made more prototypic, being 119 mils at elevations below the overflow and make-up nozzles and 140 mils at higher elevations. The thickness of the vessel flange transition section was also modelled more prototypically than in SM-2 and SM-3.
Figure 3.3-18 shows the sealing arrangement used in SM-4 and SM-5 to prevent fluid escaping from between the head plugs. Similar, though less complex, arrangements were used in SM-2 and SM-3.
A gas space above the margin ring,
(
t 35
_._g p
e-._.y.#.
+--p
WARD-D-0218 Rev. 1, 10/79 which had a volume equivalent to that of the large riser annulus, was incorporated so that an estimate of the pressure in this region could be made in the SM-5 test. The 0-ring seal adjacent to the large margin ring was added for two reasons. First, it guaranteed that the ring would be in its correct position immediately prior to slug impact. Second, it allowed for a specified constant gap connecting the regions below and above the margin ring. The 0-ring acts to block direct communication across the circumferential gap as shown in the diagram opposite. However. it does radial gap f,
,pgojenatall vess not block communication through the radial gaps between the ends of the J._f_M j1$
r b er "
margin rings. These latter gaps
/
/
fe have a constant (and known) orifice margin ring gap area throughout the transient. This margin ring gap allows a well defined geometry to be margin used in the MAXPRES calculation ring 1
simulating the cover gas response (see Section 4.3).
The cross-sectional area of the racial orifice gaps for gas communication is equivalent to about 10 percent of tSe fully open margin ring circumferential gap.
The amount of instrumentation on SM-4 was slightly less than on SM-2 and SM-3.
There were two pressure transducers (P), P ) in the core, three on 2
the vessel wall (P, P and P with a penetration through the thermal liner 3
4 5
for P ), and one on the head (P ).
Seven strain gauges were mounted on 4
6 the vessel wall (SG) through SG and SG10) and three on the vessel hold-6 and A ) were down bolts (SG, SG and SG ).
Three accelerometers (A), A2 3
7 8
g mounted on the head, one on each plug. These are shown 6n Figure 3.3-15.
Two water surface cauges (WS) and WS ) were mounted under the head shielding 2
to measure the upward movement of the water slug. Thus SM-4 had a total of 21 instrument channels. These are listed in Table 3.3-4 where they can be compared to the instrumentation of the other tests.
In SM-5, two pressure gauges were mounted in the core (P), P ), five along the vessel wall (P5 2
through P ), four in the vicinity of the head (P through P13), three in g
10 the gas gap between the large rotating plug and the vessel flange (P)4 through O
36 i
f'
\\
+
l SG16.17 HOLDDOWN COMPLE X HEAD
! A A
4 8
A (72 PLACES) 3 COVER GAS A
A lA t
3.45 in.
p g
A h.7,,; 5 ' W,4 16 7 A bA 15 REGION q
6 2
WATER SURF ACE
.,SG9,g pg'_ P _
{
o g
p -
10 12~
"d f
'"E
e P
19,20 7.80 i
0.119 in.
1.35 en.
I h
! ((
lf kI 9 P
(3 -- en.
VESSEL j.a p, w3 w3 WS, 3
2 WALL i
,8 SG11,12,13 p
Is A
8 ql UPPERN N SG21 ll l
14.10 in.
30.75 I
INTERNAL
,9 i
STRUCTUR E N,
d i
f l [PX.
0.078 in.
4
,a THERMAL 21.85 34 83 LINER l
UPPER CORE 12.86 in.
- j j
l g
j i
(Ni 200)
FORMER s' SG5.6..
3
- I !
P P
7 RING 25.675 f,
s.
w - h - a-=-f7
~7 DI APHR AGM -
e+I[II,l
!ll"lli CHARGE
.e
'l 35 CANISTER
,Y
"~
m;;,- Z ALUMINUM LZ Z t _. s i
-5h_
lP GARR L 6
i l N M [~lO[
\\
0.100 in.
~-
l l
THICK SG1,2 l SG3,4 i
SUPP T STRUCTURE l CANISTER S
SUPPORT
-- [
l SUPPORT
,$ PS
{
j STAND STAND
/
( Wl l
I l
p<
mw"M%%19;A%W%+itJrsW4Etnw M A-3929-141 A s
m-Figure 3.3-17 SM 5 WITH INSTRUMENTATION 67
... \\x s
s'd\\_y 'N N N _s
/
//
/.
/
Rubber *
.E, l
G e
s q
g
'qq y/]emgW V
pgp
'sy Ay%27 QL gi:
De il "C" Detail "A" Detail "B" NC Pressure Gauges P15 and P16 (At 0 and 180 ).
Detail "B"
Gas Space Simulates Volume of Riser Annulus I
~.
Detail "C" Detail "A"
' / >t--;
l d'L_
ir.
Tl L-i (i
4 y
Pressure Gauge P14
-p, _
. f l.,_.._.
, 4_b_f gg L
, ;=____zI o
ej=hk
<g
?
- "0" Rings Adjacent to l-
~?
.1 Margin Rings Serve to 5
l l
ll
$2 Assure Correct Positioning of Margin Ring During Test Figure 3.3-18 SM-5 Vessel ' Head Sealing Arrangement
P,,
p
~'
P P,4 w :
i
,.}
N
~i u
f
=
i lr-T t
M9E/
I.
SM4 50 50 40 40 I
g j30 l
l' 6
30 in h10
. 51 1
g0 Q \\
I I I Il O, Il I
I I
20
\\ \\, U y 3
1 N
hh 9 I
I 10 i
f I *~ \\l J'
\\
Up_
E
.t 1
i It E
o I_R M ni
.hh 5,0
[-
O i
l N
-10 V
-10 O
2 4
6 8
10 0
2 4
6 8
10 TIME - msee TIME - mste P FLANGE RING AIR GAP (180*)
P FLANGE RING AIR GAP (O')
15 16 700 50 600 40
..E E 500 n--
- k hi l 400 f
30 520 j'j s3*
/
\\
^#
10 l'#
f hII II
.MMM.___
{ 100 0
0
- f
['n
-10
-100 0
2 4
6 8
10 0
2 4
6 8
10 TIME - miec TIME - msee P FLANGE-AIR GAP P HEAD-AIR GAP AT CENTER y4 13 M A-3929-241 Figure 3.3-20 SM-5 COVER GAS, INLET PLENUM AND UIS PRESSURES o
71
WARD-D-O'218 REV. 1, 10/79 ggg,
.4 g j r -.
f1" 5
700 - i 8
600
. p. p t
i 500 M
l g
r p
'l} 9 l l l
-L_.%__rdT J l
400 m.
w g 300
- -%:- 'g} ! l l
,y j h2
', rm
~'
- I y 200 u
+
4 o,
=}
{ gg 1
I " ' V%
7 M...i.n_d.. -/
0 U
P
-100 S
6 b
TIME - msee P VESSEL WALL OPPOSITE UIS SM 5 7
1400 i
3500 1200 30c0 c 1000 lE2500 I
2000 800 w
m
$ 600 g 1500 jfi t
$1000
? (
N
- l. 11e w
490 y
D/
T 7 cc y
a 500 N -g g
a 200
,'%f j
\\ _ j g
o L
_Soo 0
1 2
3 4
5 6
0 2
4 6
8 to 1
TIME - msec TIME - maec P LOWER USS P VESSEL WALL 3
8 700 700 600 600 g 500 f 500
-f l 400 j 400
$ 300
- .,0
\\^
a 200 1
8
~~DM 200 100
),4 M 100
--4 ;
0 1
O f
l
-100
-100 0
2 4
6 8
10 0
2 4
6 8
to TIME - msec TIME - msec P OWER VESSEL P UPPER USS S
4 M A-3929 -239 l
Figure 3.3-20 SM-5 COVER CAS, INLET PLENUM AND UIS PRESSURES (Concluded)
O 72
WARD-D-0218 Rev. l. 10/79 three-dimensional model of the head may thus be necessary to pick up the more p
complex LRP response. The accuracy of the predicted IRP acceleration V
response lies somewhere between those of the LRP and SRP. However, if the analytic IRP response is integrated to give the plug displacement history, the resulting profile is found to follow, although at higher j
amplitude, the experimental SM-5, IRP response.
This can be seen in Figure 4.1-34 which compares these results. The experimental curve is taken from that location on the plug which experiences highest deflection.
Thus use of such an analytic model would appear to result in a conservative pre"ction of head doming.
Figures 4.1-35, 4.1-36 and 4.1-37 compare the experimental and analytic acceleration frequency spectra for the three plugs.
These spectra were derived using a Fast Fourier Transform of the acceleration histories discussed above, and from them, three conclusions can be drawn.
- First, both the analytic and experimental data indicate all plugs have fundamental excitation below 500 Hz.
Second, the IRP and SRP responses indicate primarily rigid body motion. This conclusion is drawn since the experimental frequency spectra for these plugs follow quite closely the predicted spectra in which the flexural modes of excitation are not V
modelled. Third, the experimental LRP response shows significantly greater higher-frequency content than does the experimental response.
This indicates a quite complex plug response which is not found in the analytic model. The SM-1 test indicated,that the LRP flexibility was significantly greater than the flexibility of either the IRP or SRP.
This is consistent with the conclusion drawn from these spectral responses.
Figures 4.1-38, 4.1-39 and 4.1-40 show the predicted total forces on the 5
5 margin rings. The peak forces are 2.25 x 10, 1.25 x 10 and 4.72 x 10 lbs across the LRP, IRP and SRP margin rings, respectively, while the local yield forces at the corresponding margin ring junctures are 2.0 x 5
5 4
10, 1.75 x 10 and 5.25 x 10 lbs*. Hence the analysis predicted some slight plastic deformation at the LRP juncture while none was predicted at the IRP and LRP junctures.
In the actual SM-4/SM-5 tests no plastic l
defonnation.was observed at any of the margin ring junctures.
i' h'
- These effective yield forces were taken from finite-element analyses to v
determine the static force-deflection characteristics of the margin ring junctures. The results of this analysis are consistent with the results of the experimentally derived force-deflection characteristics discussed with respect to the SM-2 test.
99
Table 4.1-8 sumarins the informatien abova. As in the SM-2 comparison, the predicted peak head accelerations do not correlate too well with the experimental values, since the experimental peaks are extremely narrow and reflect very complex plug interactions.
However, as discussed, the predicted IRP displacement history was similar in shape, though somewhat greater in amplitude than the experimental values from SM-5.
In addition the pre-dicted frequency response spectra of the plug accelerations showed similar characteristics to the experimental spectra.
4.1.2.3 Conclusions from Pre-Test ANSYS Analysis From a comparison of the pre-test predictions of head response with the experimental results, the following conclusions can be drawn:
The response characteristics of the SM-2 head resulting from the e
pre-test analysis were in good agreement with the experiment. The predicted peak acceleration resulting from slug impcct was, however, significantly higher than the experimental value (see Figure 4.1-26).
No residual deformation was predicted around the SM-2 margin-ring o
and none was found in the experiment.
The predicted responses of the SRP and IRP of the SM-4 and SM-5 tests e
matched the experimental histories in their general characteristics.
However, because the analytic model did not' account for the flexibility of the LRP, the simulation was unable to generate the higher frequency response of this plug (see Figures 4.1-31 through 4.1-37).
While the SM-4/SM-5 pre-test analysis predicted some slight plastic e
deformation at the LRP juncture, the tests showed no plastic defonnation.
4.1.3 Assessment of Differences Between the Predictions and Test Results A post-test quantitative and qualitative assessment of the importance l
of some of the input factors for the REXC0-HEP analysis is provided in l
this section. The predicted load on the vessel head changes slightly as l
a result of this re-analysis. Hence an updated vessel head analysis, using the ANSYS head model, was performed and is also reviewed in this section.
100
WARD-D-0218 F.ev. 1, 10/79 O
Table 4.1-3 j
Comparison of Predicted (2) and Experimental (I)
Energies and Slug Impact Velocities i
i SM-2 SM-3 SM-4/SM-5 Exp.
Comp.
Exp.
Comp.
Exp.
Comp (.3)
Slug Impact Velocity, ft/sec 91.5 65.6 62.5 57.4' 62.4 52.9 l
Slug Kinetic Energy, KJ 9.49 3.67 4.43 3.05 4.85 2.56 t
Core Barrel Strain Energy, KJ
.46 3.76
.42 4.08
.24 3.48 l
O Reactor Vessel Strain Energy, KJ 6.19 9.40 3.18 9.80
.83 6.67 i
i i
(1) These parameters are computed from measured parameters; refer to Reference 4-7.
t (2)
REXC0-HEP Pre-Test analyses.
(3) Calibrated pressure source REXCO-HEP case.
O 109
Table 4.1-4 SM-2 HEAD ELEMENT CONNECTIVITY AND REPRESENTATION ELEMENT N0DAL NUMBER CONNECTIVITY COMPONENTS SIMULATED 1
1-2 Plug Weight, Margin Ring Force-Deflection 2
2-1
. Bearing Ring Force-Deflection 3
4-2 Vessel Weight & Force-Deflection 4
2-3 Flange Weight, Bolt Force-Deflaction 5
3-2 Ledge Force-Deflection O
SM-2 HEAD ELEMENT REAL CONSTANTS Stiffness Damp. Coef.
Mass Margin Ring Yield Force
/LB EC
[LBSECb k [106 C
3 Number
(
k g\\
GAP (IN)
Fy (10 LB)
Element
/
/
1 21.07 1099 0.3583 0.0061 231.3 2
34.95 1415 0.0
-0.0001 78.34 3
5.481 76.91 0.6745 0.0 99.99 4
4.300 30.91 0.1389
-0.000005 104.5 5
14.52 56.81 0.0
-0.01359 0.0 O
110
4 O
+ 6.075 lN.
HEAD I
i W 3.79 E AIR IN.
N b__
- P 1.35 IN.
SG,PS7 8
VESSEL 22 g
(0.119 IN)
WATER 21 20
\\
19 SG.P46 UPPER INTERN ALS o
12.86 IN.
x
>SG 'P2S LEAD SHOT AND
//
12 WATER CORE BARREL 11 M
g 10
~~
N r
u CORE 8
/
7 10.055 lN.
ALUMINUM g sg,kn y
p, p
2
/
6
\\
p STEEL 5
l (E L ASTIC-
\\
PLASTIC) g y
4 3
2.0 IN.
2 3 4 5 6 7 8 U~~
1'igure 4.1-2 RIiXCO-illiP \\1ottel for S\\l.1 l're-~1est Anal) sis 0526-2 l
117
WARD-D-0218 Rev. 1, 10/79 y
6.075 IN.
t HEAD 8
3.79 IN *,
g 8,
AlR N
m \\
\\
_ _ _ _N _
=
1.35 IN SG,Pg P10 7
26
{-
VESSEL + LINER
--g 25 (0.197 IN) 24 j
WATER m 23
\\
12.86 IN 21 UPPER INTERNALS s
20 STRUCTURE
\\
jg p
N
?
WATER M> SG,
Py
/
S
/
N
//
LEAD SHOT AND WATERm 15 i
CORE BARREL m
N
//
I (0.1 IN)
\\
l 13 CORE BARREL m
[
(0.2525 IN) h
/
12 CDRE m 11 9.52 IN ALUMINUN
'Pj,P2 10
/
SHIELD m
\\
//
8 CORE SUPPORT m
/
7 (EL ASTIC PL ASTIC) g
//
U CORE SUPPORT ^
CONE (0.235 IN)
'\\
/
6 1.2IN.
/ /
/
5
/
WATER m
/
4 OPS
/
6.86 IN.
i 3
2
/
I
"~~~
/ 8 7
l 2 34 5 6 l
Figure 4.1-3 REXC0-HEP Model for SM-4/SM-5 Pre-Test Analysis (SM-5instrumentationshown) 1 0526.1 118
~
I l
WARD-D-0218 Rev. 1, 10/79 t
a STRAIN (PERCENT) 7 6
5 4
3 2
1 0
STRAIN (43 msec)
[
1 l
I I
I I
I,s*,,/ I 2
4 5 4 3 2 1 0
l
,/,f 3
l l l 1 l/l
/
8 I
4
/
/
1 i.
~
l VESSEL
/
~
I
\\
\\
6 g
WALL
/
5 N
\\
7 lg g
\\
N
(
4 N
8
's%
~
h 3
[
i 1
i s
2 N
j -
10 j
g k
I
's 11 l
l, i I, i Ti 4, 0
I s
1 12 1
\\
240 180 120 60 0
i
)
13 i
5 R (mils) l
)
~
I4 1
STR AIN (46 msec)
Il 15 4 32 1 0
i TOP II-16 I i gjF l
g l
h-
/y' 8
11 I
l 18
[t 7
~
'S l
l1
~
0
\\(f 20 l
m
/
5 Rt.R R E L
.I 21 4
l
[
)
22 3
?
(
23 2
s N 1
24 s
s-li,li i.,1,,li II' 0
25 6R( l}
R (mi s LEGEND:
--- REXCO POST TEST l
REXCO PRE TEST EXPERIMENT f
r i
Figure 4.1-42 Comparison of SS1-2 Pre.md Post-Test liEXCO-ilEP 1
Deformation Profiles 0526-39 O
157 l
STR AIN (PERCENT)
LEGEND:
4 3
2 1
0 AVERAGE DATA
, ' ~ ~
1 STR AIN (%)
SM-4 SM-5 2
1 0
/-
2
's
[
3 l 13 l
REXC0 POST TEST REXCO PRE TEST N, \\(
4 12 r
5 11 s,
6 to
's X 7
9 l
CORE BARREL s
s 8
8 b R tmih) b R (mils)
\\
9 7
180 120 60 0 300 240 180 120 60 0
l 4
10 6
4 11 5
l lllI lllIll j
4 12 4
4 13 3
STR AIN (% 6 msec)
STR AIN (%-3 msec) r q'
14
, 2
- l 4 3 2 1 0 7 6 5 4 3 2 1 0 J
85 i i l I I I I I I li f t l l 5{
16 1
8 8
8 sfl'
'I
[-
18
,0 o'
7 l,,l T
s J
7 I
120 60 0
gf 6
'/
l 6
s f- ~
6 R (mils) h
[s{}-
5 5
,e ',
fg(
l- - 20 s
,s_
4 if N
t\\
4
\\
gj 21 tg#d
\\
l 2
J-
!n 3
g s g
3 r
22 e
t.g 2
t I
(,,'_(' M l
r l23 1
p j
L 24 g
i
-0
- 25 l
26
's
-- 27 l
l
'g 28 29
[
leilivliil, 240 180 120 60 0
i i
6 R (mils) i r
I l'igure 4.1-4.1 Comparium of S\\1-4,S\\1-5 l're-ami l'ostJIesi RI.X( O Ill l' I
1)eformatiim l'rol'iles 0526-40 9:
158
WARD-D-0218 Rev. 1, 10/79 4.3 Cover Gas Response and Its Leakage Implications V
In Section 3.3.3.3 the cover gas pressure response from the SM-5 test was discussed. This section compares these results with results from simulations of gas response below and above the large margin ring.
The SM-5 test had pressure gauges at the lower elevation of the head shielding (P ),
g in the dip seal region (P)4) and in the gas space above the margin ring (P
and P16). The first of these pressure histories is used in conjunction 15 with the MAXPRES-2 code (Reference 4-14) to determine the pressures at the locations of P and P These are then compared with the corresponding j4 15 experimental loads.
The MAXPRES-2 code models a series of annular regions of differing cross-section. The liquid and gas remain, throughout the transient, as two distinct fluids separated by one planar interface. As the liquid moves up the channel, the gas above it is assumed to compress adiabatically.
While throttling effects are simulated, shock wave effects are neglected.
Figure 4.3-1 shows a schematic of the region, the corresponding MAXPRES-2 geometry and its dimensions. The gap across the margin ring was assumed to be constant throughout the transient for the reasons discussed in Section 3.3.3.2.
An overview of the region together with the locations 1
of the relevant pressure transducers is given in the SM-5 drawing in Appendix B, while details of the margin ring iegion can be seen in gauge (see Figure 4.3-2)
Figure 3.3-18. The pressure loading from the Pg was used to drive the column of water up the annular space. The first 2.6 milliseconds of P data were removed from this figure since it takes that time g
for the transient loading to be transmitted to the lower shielding elevation.
l Two calculations were performed.
In the first no fonn losses resulting in energy dissipation or reflection were assumed so as to provide a bounding calculation. As a result, the pressure calculated in the dip seal region, as shown in Figure 4.3-3 was severe and quite non-prototypic.
In the second calculation, realistic form losses were introduced between the lower shielding elevation and the dip-seal region, and as a result, t%,
191
th:re is a drastic reduction in peak pressure. The responsa with fom losses is shown in Figure 4.3-4 where it can be compared with the experimental pressure detemined from the dip seal region (Pj4).
It should be emphasized that the magnitude of the fom loss used was arbitraily chosen to give an experimental result whose peak pressure approximates that of the calculated value and to show that with such a loss coefficient the calculated response characteristics are similar to the experiment. The most significant conclusion from this analysis is that, by not including energy dissipating fom losses, predicted pressure can overestimate the actual pressure by at least an order of magnitude. This is of considerable importance in leakage assessments.
Figure 4.3-5 compares the predicted and experimental (P15) pressures above the margin ring in the gas space simulating the large riser annulus.
The analytical curve resulted from the second of the two calculations above, in which form losses were included below the margin ring. The observed overprediction may be the result of two effects.
First, energy dissipating form losses were not assumed to exist through the complex geometry of the margin ring gas space. As discussed above this is an extremely conservative assumption.
Second, the gas gap across the margin ring region was modelled with uniform cross section which may not have simulated the actual gas gap closely enough. For discussion of the actual gas gap geometry across the margin ring, see Cection 3.3.3.3, while for the dimensions of the analytical model, Figure 4.3-1 should be consulted.
Even with these conservatisms, attenuation is significant enough to reduce this pressure to less than 100 psi.
Current design analysis techniques do not generally include energy dissipating form losses because these are difficult to determine. However, by excluding them, gas pressures associated with short tem dynamic loads may be significantly over-predicted in the convoluted annular regions close to and above the vessel head.
192 l