ML20215G101
| ML20215G101 | |
| Person / Time | |
|---|---|
| Site: | Trojan File:Portland General Electric icon.png |
| Issue date: | 06/12/1987 |
| From: | Cockfield D PORTLAND GENERAL ELECTRIC CO. |
| To: | NRC OFFICE OF ADMINISTRATION & RESOURCES MANAGEMENT (ARM) |
| Shared Package | |
| ML20215G104 | List: |
| References | |
| TAC-65471, NUDOCS 8706230126 | |
| Download: ML20215G101 (14) | |
Text
.
Portland General MCoiitswry
- David W. Cockfield Vice President, Nuclear June 12, 1987 Trojan Nuclear plant Docket 50-344 License NPF-1 U.S. Nuclear Regulatory Commission ATTN: Document Control Desk Washington DC 20555
Dear Sir:
Main Feedwater Restraint Failure Engineering Analyses During.the 1987 refueling outage, a failed seismic restraint (SR-8) was' discovered inside containment on the main foodwater line to the "B" steam generator. In the course of evaluating the failure, several-analyses and calculations were performed to identify the potential cause of the failure. Initially, a calculation was performed to determine the load required to cause the observed failure. A
]
description and the results of this calculation are contained in Attachment A.
1 The failure occurred following the 1986 refueling outage during which f
a Containment integrated Icak rate test (ILRT) was performed.
Since the failed restraint is in close proximity to the Containment penetra-tion, a calculation was performed to determine the effects of the ILRT on the restraint. This calculation and the results'are described in Attachment B.
i
)
Bechtel Corporation was contracted to model the feedwater line and i
perform analyses to identify the potential cause of.the restraint failure. These analyses are described in Attachment C.
a Additionally, following the discovery of the failed restraint on the "B"-feedwater line, inspections.of other lines were conducted. During the inspection, it was observed.the. equivalent restraint to SR-8_on the "A" feedwater line (SR-4) was under load while in the cold condi-tion. This condition was evaluated and the results are provided in 1
Attachment D.
121 S W Salinon Street, Pomand, Oregon 97204 1
E Portland General ElectricCoiwsry Document Control Desk l
June 12, 1987 Page 2 l
These analyses are being submitted for review by the Office of Nuclear-I Reactor Regulation (NRR) as agreed upon in discussions with NRR and the Region V Office of the Nucicar Regulatory Commission. Representa-
]
tives of Portland Ceneral Electric Company will be meeting with the NRR staff to discuss these and other analyses the week of June 15.
Sincerely, l
1 i
Attachments c:
Mr. John B. Martin j
Regional Administrator, Region V
]
U.S. Nuclear Regulatory Commission
]
l Mr. R. C. Barr i
NRC Resident Inspector Trojan Nuclear Plant Mr. David Kirh, Director State of Oregon Department of Energy 1
1
s Document Control Desk
- .._ Trojan Nuclear Plant.
June 12, 1987 Docket 50-344 Attachment A License NPF-1 l
TO:
R. E. Fowler TEB-55 -87M FROM:
T. E. Bushnell
- i DATE:
June 3i 1987
SUBJECT:
TROJAN NUCLEAR PLANT Main Feedwater Restraint Failure Action Plan Following is the presently available information which-was requested pursuant to the subject action plan.
TEB/SGH/KJM/mc 8368C-Attachment c:
T.
D.
Walt A. N.
Roller C.
P. Yundt C. A. Olmstead D. W.
Cockfield F.
R. Rogan - Pacific R.W. Fosse - Bechtel i
1
. ]
r I
l i
l
.--,-.,,.,.~_.~,._-my,
i:
TEB-155 -87M -
June 3, 1987 r;
.Page 2
. Item A.1.a.1 An evaluation of the restraint installation is provided following the; description of the as-found condition of the main feedwater EBB-3-1 Seismic Restraint SR-8.(SR-8).
_I, tem A.1.c the following.
Upon inspection of SR-8 after disassembly,(see attached' sketch of; design conditions were. identified configuration. Attachment 1):
x 13'(W4) diagona1>
The' concrete under the base plates of the~W4 and the W6 x 20 (W6) vertical members was' broken out in an area by 24-in.
in plan dimensions of'approximately 33-in. east-west
.The average depth.of'the failurefsurface was-north-south.
in along the outside edges;ofLthe-approximate 1'y 1/4'to 1/2 increasing to approximately 2-5/8_in. deepfunder:
case' plates the W6 base plate'and 3-7/8-in.fdeep.under the W4 base plate.
A large section of irregular-shaped concrete was found pulled under the front' half of the base plate of the W4. diagonal.
The section measured 12 in. in width (north-south), 20 in. in out and was a maximum of 3-7/8 in. deep.
.The-shell's of the, length,two anchor bolts (5/8-in.-diameter Phillips' Redhead.typef front anchor bolts) under the W4 diagonal base plate were still thirdiof this concrete section'which shows engaged in the frontthe concrete failed as opposed to shell conclusively that Due to the conical shape of-the failed concrete pullout.
section and the relatively clean fracture'surfacelof the area crushing-orvgrinding) fit under the section (ie, no apparent appears.the concrete section failed due to tension and shear loads imparted from the W4 diagonal to its base plate.
The broken concrete under the base plate of the?W6 vertical was found under.the-W4 member was in smaller pieces than what There appears to.have:been some diagonal member base plate.
the concrete under the.W6. base-plate: indicating:
crushing of that the failure was most likely caused by shear and compcession loads.. The southwest anchor bolt' ('right ' f ront bolt ~
of W6 base plate in' plan) remained anchored:below the failure ~
.The1 anchor. bolt had loosened.
surface of the concrete. be removed by. hand-appliedJ tensile.
laterally but could not This.istevidence'that this~anchortboltLdid not force.
to:cause failure'in a:
experience tensilesloads-sufficient tensile' concrete cone' failure mode or in a'shell tensile-pullout mode.
. -The writer is led to believe.that, based on'theLevidence-summarizad above.,the sense of load causi~ng.the failure.'in SR-8:
either11nitially.or as a result' ofjcyclicJprogression.
- was,
~
.~ni.,
--m ut
- ~ 9s
- 2r.* c med s z m' --
~
TEB-55 -87M June 3, 1987 Page 3 I
tension in the strut assembly which produced combined compression and' shear on the W6 vertical member base plate-and combined tension and shear on the W4 diagonal member base plate.
Restraint SR-8 may, however, also experience significant loads in a sense of compression on the strut assembly.
Load reversals should not, in this section of main feedwater line, be unexpected if initiated by a transient.
condition.
i The structural steel assembly of SR-8 was visually inspec.ted.
There was no evidence of' damage to welds or indication of permanent deflection in W6 or-W4 members or their base plates.
There'was possible indication (paint scraping) that the pin.had come in contact with the pin bracket plate, but with no clearly j
definable sense of load magnitude or direction.
Further nondestructive examinations are to be performed on SR-8 assembly welds.
Item A.1.a.1 From results of the inspections, the anchorage installation of i
)
SH-8 appears to have been in accordance with the design intent.
The anchor bolt shells were found to be set at the proper depth, as best as could be determined, per manufacturer's installation standards.
The anchor bolts obviously developed the concrete shear cone type failure mechanism without premature pullout as evidenced by the observed concrete failure surfaces and configurations.
This type of concrete shear cone failure mechanism is a common mode for anchor bolts of this type and of ultimate load development is provided for in establishment of design allowed loads.
The presence of base plate leveling nuts on the anchor bolts below the base plates is not considered germane in the SR-8 installation.
Anchor bolt shells of the type used are set by impact which expands the plug on the bottom of the shell causing shell engagement and development with the concrete..
Leveling nuts.would not inhibit the shell development.
Proper engagement of expansion-type anchors, which are set by the end of the bolt by use of bolt expanding a sleeve at torque, would be affected by the use of base plate leveling nuts, but as explained above, that is'not the case in the SR-8 l
anchorage.
anchorage was concluded to be, installed in In summary, the SR-8 accordance with the design intent.
- - - _.... sus
- hew e w w
.m 9 m
,,%,a m._,
~
TEB-55 -87M June 3, 1987 Page 4 Item A.1.d 54.4 for EBB-3-1, the Upon inspection of Pipe Whip Restraint following conditions were observed:
At first observation it appeared that there was an approximate 1/16-in. gap between the base plate and the grout at.the bottom of the upper whip restraint base plate.
A similar gap appeared between the grout and the concrete at the top of this same base plate.
There was no evidence of damage to the anchor bolts, welds, base plate, or structural steel members.
There wasEno evidence of damage to the whip restraint bottom base plate, grout, welds, or anchor bolts.
The grout chamfer was removed around the perimeter of the upper i
whip restraint base plate.
It.was evident that the chamfer had been added recently and, therefore, would not be-a reliable indicator of structural performance of the restraint.
A portion of the original grout under the outside edge of the upper base plate was removed for closer inspection.
Two flat bar steel shims were located behind the top and bottom of the.
base plate and were tight up against the concrete and base plate.
A feeler gauge was inserted around the perimeter of the base plate and the deepest penetration parallel.to the base plate found was 2-3/4 in.
Two layers of thin, brittle, and corroded metal sheets were found between the upper base plate and the original grout.
From these observations it was concluded that the gaps first observed were from the installation and shrinkage of the grout-placements, and possibly from deterioration of the metal sheet layers.
Further examinations were performed of the area around the upper base plate following additional grout removal and removal of damaged concrete around SR-8.
Hairline-sized cracks were discovered on the concrete ledge directly behind the upper base plate of Restraint 54.4.
Sounding of the concrete.in the general area indicated a looseness of the concrete directly behind the upper base plate measuring approximately 30 in.
along the length of the ledge by 15 in. perpendicularly into the ledge by 8 in. deep.
The depth of this zone extended'to the approximate centerline of the 2-in.-diameter rock anchors of the upper base plate.
~w+m.
- s. a s a,
y
,,s
1 TEB-55 -87M June 3, 1987 Page 5 f
Item A.2.a2 Of the six (6) concrete anchor bolts (1-in.-diameter wedge-type) in the base plate of Support EB8-3-1-SR-4, one (1) was found.to have evidence of lifting approximately 1/16 in.
and No other evidence of support damage or failure was found, the bolt was able to be retorqued to design values.
This.
)
support was reinstalled to its present configuration in 1979, using the plant installation standards current at' that time.
This required inspections of anchor bolt i'nstallation to ensure proper installation.
Due to the fact that the gap was approximately 1/16 in. and the gap was.found on only one of the six anchor bolts, it is possible that the dynamic load from a feedwater line water hammer or transient jolted the. support, a
causing slippage of this anchor.
At this point, the anchor probably took hold and performed its function.
Since the anchor bolt could be retorqued, and there were no indications of a concrete cone pullout, the retorquing of the bolt reestablished the design capacity of the support.
4 it is recommended that the support be reinspected
- However, during future refueling outages to "erify its integrity.
Item A.3.a1 Feedwater Support EBD-6-2-SR-17, 71-ft Elevation in Main Steam Support Structure (MSSS).
Inspections of this support revealed the following:
The top two anchor bolts were out from the base plate approximately 1/8 in, while the bottom left anchor bolt.was no-longer snug.
The pipe clamp was rotated such that the centerline of the clamp at the strut was 1-1/4 in. below the centerline of the-pipe.
The strut was bound up tight at the' clamp and painting on the pipe and clamp showed that the clamp had not moved since the painting was completed.
The paint between the base paint and the wall.had cracked. showing that the base plate had moved away from the wall.
The preliminary indication is that the probable cause for the as-found condition (more investigation is being performed as of June 4, 1987) is as follows:
The vertical movement of the pipe at this location.is approximately 3/4 in.
With the clamp rotated and-the strut locked tight (in the cold condition), a vertical load and moment will be imparted into the strut at the clamp-(as well as the thermal tensile load).
This vertical' load and moment in the strut will cause a moment in the base. plate due to the eccentricity of the pin from the base plate.
This.would in i
TEB-55 -87M June 3, 1987 Page 6 turn induce extra tensile forces in the top two bolts causing them to slip.
The lower bolt was probably pried loose at this point.
With the thermal load off the pipe, the base plate would be' pushed back to the wall.
Since the paint on the clamp indicates that it has been rotated for some time.
The clamp was probably either installed that way or was rotated by_someone climb.ng on it when the pipe was i
cold and set in its position when the pipe heated up.
After our further investigation has been completed, the clamp position will be corrected and the support repaired to meet design loads.
Item B.2 Following is a summary of the analysis performed in response to the request for an initial evaluation of the range of loads that could have caused the observed failure in Restraint SR-8.
First, it is important to state that the analysis was based on static load calculations, not dynamic loads and effects.-is now Although the loading condition causing failure of SR-8 believed to have been dynamic in nature, rational translations between static and dynamic load regimes are available.
These translations are described further in this response.
Until noted otherwise, the following discussions are based on static principles.
1.
Significance of sense (strut assembly tension or compression) of failure load.
As described in the response to Item A.l.c, the evidence indicates that the SR-8 anchor bolts were adequately developed to cause, primarily, a tension failure mode in the concrete due to vertical reactions.at the base plates.
Vertical equilibrium, static or dynamic, must be maintained (ie, equal and opposite vertical reactions at the W6 and W4 member base plates), and the failure load should be essentially the same for either tension or-compression in the strut assembly.
Thus, within the context of other analytical considerations used and described herein, the sense of load (tension or compression in the strut assembly) which resulted-in the anchorage failure of SR-8 should not significantly affect the overall' conclusions.
i l
a l
1
~*$
5 TEB 55 '-87M June 3, 1987 Page 7 2.
Distribution of. horizontal load imparted by'the. strut assembly to members of SR-8.
Restraint SR-8 was' designed to resist loads. transmitted'by the feedwater line through the strut assembly in a horizontal direction.
When the strut load is imparted to the SR-8 framing, two bounding case load paths are possible:~
(a) considering effectively complete rotational-resistance at the welded interface of the W6 vertical and W4 diagonal members (likely), and-at the member base '
plates, and (b) considering effectively no' rotational resistance at these locations.
In case (a), calculations
)
i demonstrate that 92 percent of'the horizontal load is resisted by 'the W4 diagonal, and in case (b), 100 percent of the horizontal load is resisted by the W4 diagonal.
For purposes of this evaluation, the difference is by no means significant.
3.
Concrete Strength 1
The actual versus design concrete strength, which is a l
parameter in anchor bolt development, was of necessity due l
to lack of any other specific data, treated by judgment.
The design concrete strength (measured in terms of 29-day ultimate compressive strength) was 5,000 psi.
Concrete strength generally, but not always. increases with age, j
For these calculations, a concrete strength of 5,000 psi was considered reasonable.
The possible variations in concrete strength should-not also result in intolerable differences in conclusions since both concrete and anchor. bolt strength parameters are.a j
square-root function of the concrete compressive strength.
4.
Effects of anchor bolt performance with respect to.
shear-tension interaction and base plate flexibility:
i Shear-tension interaction considerations - The effects l
a.
of combined shear and tension on'the-ultimate failure load capacity of an anchor bolt of the type studied are, at best, unclear to the writer.
Tests (Hanford and others) have shown that sheer in: combination with tension could contribute to premature anchor-bolt tensile failure.
The interpretive difficulty is that the test data failure mode (anchor bolt shear pull-out.
or bolt shear versus concrete failure cone development)
I
'TEB-55 -87M June 3, 1987-j Page 8 l
does not seem to lead to any clear conclusion regarding
{
capacities at concrete failure, which is the case of present interest.
(
b.
Base plate flexibility. considerations:
Because the SR-8 framing members are welded to their base plates and.are of such dimension (particularly the W4 diagonal) as to effectively stiffen the base plate, potential base plate flexibi'lity is not considered to be a.significant contribution to lower bound failure of the anchor bolts due to base plate prying action.
5.
Static Failure Load Evaluation As stated previously, the static failure load evaluation was based on a linear elastic analysis.
The statically applied strut load predicted to cause anchorage failure was determined based on the anchorage configurations (base plate and anchor bolt geometrics), bolt type and embedment, estimated concrete strength, and the type of failure observed (concrete cone-type failure).
Considering.the above, anchor bolt failure loads wire.(1) taken from applicable industry test data (Pittsburgh Testing Laboratory) and (b) theoretically derived from failure load models (which are also based on test data) taken from American Concrete Institute (ACI) Standard 349, Appendix B.
By both methods used and some var,iations in assumptions considered reasonable, the'statix strut failure load was calculated to be on the order of 40,000 lbs.
6.
Dynamic Effects j
4 As is well known, the effect of a dynamic load can be translated to an equivalent static load to allow a static failure analysis if the character of the dynamic loadfcan be estimated or bounded.
Use of a dynamic-load factor j
(DLF) is a convenient means to perform'this type of translation. -The DLF is simply the factor by'which the peak value of a dynamic load.is multiplied to determine the static load which produces the same.effect.
For an 1
impulsive. load, the DLF can vary between 1.0 and 2.0, ie,
]
for a DLF of 2.0, the peak value-of an. impulsive-load.
i causing failure of SR-8 could have been as' low as 20,000 lb.
l As another example, although simplistic, if the dynami'c load were to have'resulted in a system resonant ~ condition of sinusoidal character, for low system damping-(for example, use 5 percent), as few as five resonant cycles could have resulted in a DLF on the order' of 8.
-)
a
.m,
a
.g TEB-55 -87M June'3, 1987
- Page 9' In summary, the static' equivalent strut assembly load-resulting in SR-8 anchorage failure ~could have been on the-order of 40,000 lb.
If the. load causing failure were the j
dynamic in character, which is believed to be.the case, value could have been much less than 40,000 lb.
l
- 7.. Pre-existant Anchorage Degradation
-}
As is obvious,.a pre-existing weakened condition in the-SR-8 anchorage could have a dramatic effect.of;the-failure
?
load.
At~present, there is no clear evidenceLknown to the writer which.would indicate that the SR-8 anchorage was or l
was not degraded. prior to failure load application.
j J
Item C.1.a J
1 Repairs to' SR-8 are presently being performed.
The repair design concept was to provide a new anchocage system which.
f
~
would develop the capacity of the W6 and W4, welds,~etc.
This i
design capacity for SR-8 repaired, with normal. factors'of safety included, is greater than 40,000'1b.
If desired, a final capacity calculation can be performed when SR-8s.as-built information is received.
Please advise of any questions you may have concerning this memo.
TEB/SGH/mc 8368C I
1 la a
1 1
t
'l
.i i
q y
w.-._,.._-
e
~r
.h ~.)s 1
't Eg',$
PQT size essenirtion v or.
71 29A
Z $. *.! -/C.17';';.,'? A HIMi!.Y(S& 'fA/; 7~ D
/
/
~
!4~/*:? J',*F ! s'A v.l::Afr,v r (1,f' s S
/
MLS3 3
/
C>
<~It%l'-I';6, 4
/
5 1 4 v.c.' ?.x,'.1 i / 4-
- f ~*
/
4 ')
f2* \\,2 4 s /'- O".) !<!)~.': a. h. ? ;ff.7 Si ?; :: f *)
20.T L.
s f, :
f/g : : v:*! ~f rdl7: e.C'.'1 7
i
=/
.ft :.\\.:' '-y S.'?
7,'C, 3'72 ' ; st i m'. n.';'S E
/&
_9 2.
C. f
- *-l
- s' !}' s : '. ~: J f; s: 2.'s *j 7: ;. yL}
L4'
.s.n... s
.-... )
- e _.,-
- <f" x /:" &.x O' f.G(W.A71.*A.?&
//. f I
I
/O
/
c5 5-ecc (SE3"2 DE7~ '2 ' CW&.' / ~.5A )
l wa c esy.o*
g Em.m*
A '. Q "
ll s:
x
~~-
I s,
- a:
.a
- l. l l
,(
)
!h 4, /
-'g
- f.
f, (T,
- l g g
--' --- - =
t l
i i
i l ('..
U.1.E.
{
l l \\m
..1;.o t
i
. -..?.' '. :
.c gf. 3,,:, a l',
<. l. '.
'e.o g...'-,,,....
- y.A._
. ~..
o1 r_
,,.l
- I
- x
~
,,e-..-
d.k-I i
/.,/
1.m
,.g y
~
,3 w
o ll Q
/.
N O
[//((
's.
v l
'K.
o 2. p.Q
]l l
k *g C.j,,.
7'4 5
" -~ ~ E, ~~~p
~
,,4 :rfy/~
(ld.J.I L j d'i I..
h fd O/C.:.... _,,
.., i.:
--.t...
._.n'-
n-.
//JNt 9 '/.:.
/[. $ ~
e n
y a
- L A '. o '
e t
j r : r.--,r, i
I t
=
_ _,,_t,. -
l 4
,:, 3 i ! I i
O/
0 O
~,";
f.', E VJ ~~'C '.' l A s
i e
is-W is:54: !!T 3:u7
'I.
,sh
/
- 4 a
n s
\\
4 ---
,fjv,'.,;,.
l
[
.ia.
,g A l k' l
i o
6!
N i
d j : t.- ~ >. ~~ ~'
/*
..Ljf'.{./
a
- g./.cr *.
- : -
4,,,,
.L 1:
l
- v. 3..:... ! 14 2A
- l i.A,i.. )
A,"
p.
>= :
5 l.~,.s~7.u/1 l /d2 # l LOCATION PLAN e
CUSTOMER ORDER No.
if.7-/
// H,'# W M /r','~~ l
.~ 5. 9 X 2 ~5, 4,c < w e s WRIGHT SCHUCHART. HARBOR INC.
rieme sysrs.
wn.a<.wr -
,n PIPE REF. >f
- 4,*
jl l
l BECHTEL CORP.
arr en ewes owc. sTt P.tr. '.. //,e t j
s.-...
j l
TRO.lAN. Nile,t FAR PI ANT _ M3..~, u c.r.n if!- !_- U-A
/'
u
Trojan Nuclear Plant Document Control Desk Docket 50-344 June 12, 1987 License NPF-1 Attachment B Page 1 of 2 LOAD AT SR-8 DUE TO CONTAINMENT INTEGRATED LEAK RATE TEST Summary For an integrated leak rate test (ILRT), Containment pressure is raised to 60 pounds per square inch gauge. This pressure exerts a radial force on
'the Containment wall.
During the 1975 ILRT performed at the Trojan Nuclear Plant, data had been taken indicating a radial displacement of the containment wall at the feedwater line penetration of 0.140 inches due to the ILRT pressure. LAs can be seen from Figure 1, SR-8 is positioned parallel to the Containment penetration such that it~would resist motion of the feedwater line as the penetration displaced radially.- In view of this, the load at SR-8 resulting from the ILRT was calculated.
Task
. Calculate the load at SR-8 during an ILRT assuming a radial outward movement of the Containment penetration of 0.140 inches.
)
i Descriptjon A mathematical model of the "B" feedwater line from the containment penetration to the steam generator nozzle was developed.. The steam-j generator nozzle and containment penetration were modeled as anchors with no movement assumed for the steam generator nozzle and an outward radial movement of 0.140 inches for the penetration. The pipe was assumed to be at ambient conditions. The analysis was performed using the computer program NUPIPE, Version 1.6.3-CYBER, 1985 edition.
Result The maximum load on SR-8 during an ILRT was determined to be:1626 pounds.
Conclusion The load on SR-8 during an ILRT is well below the calculated pull-out load of the restraint (36,000 pounds) and below the design seismic load for the l
restraint (3484 pounds). Therefore, the observed damage could not be expected to have been caused by the performance of the ILRT.
SAB/kal 1387P.687 a
a eoneaaDa nounnOW De0Y s
5 4 MGuiD Zc WM*
s n
- .E ccnm wN. cew n
COnX@e OO s
r s
>ne=n Baen o MOO 30O Zm ~
3 mmoo N om N o
EN 5 s I
s-L ss s R
4 t
l ET g
4 ee*
A 5
M g
S R
t 7
t F
y
{
E 7
F M
t "B
.4 s
n a
t e
s p l
SS c/
a 3
I' n
3 I
3 M
R 0
A O
E T
1 T
A e.
a S
R E
"B N /
E C
s-R 5
s S
4 E
N 45 IL
-RF NO R
'1 T
E I
2 A
T RT A
1 m
\\,
E W
N E
D E "o
E R P
C#
l 9,
1 T
U E
l
/g F
N E
'2 G
M 3
I "B
F N
IAT 2
M 2
O 3
C D
N 45 A
R "A
F 4
N; Rs N
e
~
M R
A O
g E
T
+
g T
A p
S R
g E
"A N E
C 2
4 5
3 R
I 1
F E.
'o
~ #
I
- *p g
J.
I 1 7' 1@
(
4 S
2 5
1 PbP 1 g
R n
,'5
. W s 3 i
s -
9p 1
i5 L
s 5 1
c R
41 3
E
,l 9
4 E
T A
I 4
M R V 5
H L
1 o
s A
1 tE L
R C W
}2 E
A F
I E
H T
g 5
F C
R V
5 I
"A T
E R
V EV r3
~
r fl f,i'
- {
4
{i li!
!I
<