ML19256G270
| ML19256G270 | |
| Person / Time | |
|---|---|
| Site: | Davis Besse  |
| Issue date: | 11/15/1979 |
| From: | Scholl R Office of Nuclear Reactor Regulation |
| To: | Lainas G Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 7912280633 | |
| Download: ML19256G270 (13) | |
Text
,0E c F0/-
pn asay,
+f
'o, UNITED STATES
' j NUCLEAR REGULATORY COMMISSION 3,..,
. E WASH WGTON. D. C. 20555 a
u.- /
NOV 151979 MEMORANDUM FOR:
G. Lainas, Chief, Plant Systems Branch, D0R THRU:
D. Tondi, Section Leader, Plant Systems Branch, D0R FROM:
R. Scholl, Plant Systems Branch, 00R
SUBJECT:
TRIP REPORT OF VISIT TO DAVIS-BESSE UNIT 1 ON OCTOBER 18 AND 19, 1979 The enclosed trip report is fomarded for your informatier.
2.
1ALLJI h.
I R. Scholl Plant Systems Branch Divisior, of Operating Reactors cc w/ enclosure:
E. Jordan, IE V. Thomas, IE F. Rosa J. Smith, Reg. III D. Tondi M. Chiramal J. Beard I. Ahmed R. Wilson D. Garner R. Scholl 1650 223 7912280b P
ENCLOSURE TRIP REPORT OF VISIT BY DANIEL GARNER AND RAYMOND F. SCHOLL, JR.
DAVIS-BESSE UNIT 1 ON OCTOBER 18 AND 19, 1979 INTRODUCTION As a result of the confusion surrounding the reactor trip and station blackout of Davis Besse Unit 1 on October 15, 1979, the authors.,ere dispatched on the site on October 18, 1979, to assist in improving comunications between Headquarters, Region III personnel on site and Toledo Edison Company (TECO).
DISCUSSION Thursday, October 18, 1979, was devoted to traveling from Washington, D.C. to the site in order to interview the Region III personnel. Friday, October 20, 1979, was spent interviewing TECO personnel and returning to Washington.
The following are the salient points resulting from our interviews:
1.
The basic sequence of events is presented in a draft document,
" Unit Trip Report" which is provided as Enclosure 1 This document was prepared by members of the Davis-Besse 1 plant operating crew that were involved in the incident. Although it is a draft, may not be complete, and contains errors; it represents the best available infomation as of 2:00 PM on Friday, October 19, 1979.
Identified errors are discussed below.
2.
Three minutes before the event started, the process computer failed. This is one of several unrelated but almost concurrent failures that gave raise to the concern that some comon mode failure mechanism had been encountered. The significance of this failure is that the exact sequence of events for the reactor trip is not available.
However, the computer was repaired three minutes after trip and was available during the loss of offsite power.
As best as can be detemined, the following plant conditions existed just prior to SCRAM:
a) the plant was operating under a liniiting condition for operation of Technical Specification 3.1.3.3 because of the failure of an absolute md position indicator; b) the high power scram setpoint was 69%;
1650 224 c) the power level was 60% by heat balance and 64% by nuclear power range indication; d) the plant was experiencing a low frequency power oscillation which has a peak value of 1.5% of power at this time in core life and occurs between 55 and 65% of power.
(This oscillation starts at 35% of power and subsides above 85% of power. ) This is not a new phenomena and the Region III inspectors stated that it is common in B&W plants; e) all switch yard breakers and disconnects were closed and the crimary windings of both startup transformers were energized.
(The disconnect switches are motor operated, powered from Class IE safety busses and controlled from the control room. ) This required alignment is specified in Technical Specification 3.8.1.1; and f) a capacitor in the interface between the Integrated Control System (ICS) and the Turbine Electro-Hydraulic Control System (EHCS) failed (short circuited).
3.
The short circuited capacitor bypassed a control relay contact pair and started to increase the turbine generator load limit.
The increased stoam demand reduced steamline pressure and resulted in an ICS response to withdraw the control rods to recover pressure.
Reactor power level increased and the reactor scramed on high flux.
4 At this point in the interview, Region III personnel stated that they had a concern with Technical Specification 3.1.3.3.b because it was not clear if, upon declaring a second position indicator inoperable, the licensee had to be out of Mode 2 before the end of 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of if orderly shutdown should be started after 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.
A second problem is that it is not clear when an absolute position indicator is inoperable. The issue here is that a noisy indicator may still be capable of indicating the proper position although the noise is indictive of incipient failure.
Noise was the cause of declaring the position indicator inoperable in this case. TEC0 has already obtained replace-ments for the affected cable and connectors from the vessel head.
This cable was obtained from Florida Power and has a higher temperature rating. The use of this cable has cured a similar problem for Florida Power.
1650 225
. Finally, it was pointed out that similar capacitor failures have previously occurred. However, in previous events, the 50 psi error limit was reached and the ICS and EHCS auto-matically transferred to manual before the high scram point was reached. An inspection of the EHCS indi;ated that a wiring error in the EHCS caused transient voltage on the order of 300 volts. Accordingly, a capacitor with a 1000 volt dielectric rating was used as a r eplacement and the wiring error corrected.
It was also suggested that the Region III inspectors check contacts in Class IE systems to be sure that arc suppression systems are either not present or are clearly shown on plant drawings.
The Inspection and Enforcement personnel agreed.
5.
Following the reactor trip, the main turbine tripped and the main generator output breakers opened as designed.
Because the unit electrical hotel loads have been on startup transfomer 1 since late 1978, there was no loss of electrical power at this time.
6.
Some 23 minutes after SCRAM the control room operator, in accordance with operating procedures, opened the generator disconnect and attempted to close the generator output breakers. A muffler on one phase of main generator output breaker 34560 blew out and short circuited the "J" bus.
This resulted in a loss of electrical power to the emergency busses.
A similar muffler failed on Thursday, October 18, while a breaker was being opened so that the muffler could be modified by installation of reinforcement straps to reduce the likelihood of future failures. This modification has been recormiended by General Electric. Consequently, damaged mufflers are being replaced with an improved model. Region III stated that they will prepare a circular on these mufflers because they have failed several times.
(The authors have samples of the internal mesh that were brought back from the Thursday event.)
7.
Wherf the station blacked out, both diesel generators started and one successfully picked up all of its loads. The second generator picked up most of its loads automatically. However, Service Water Pump 3 (that had been swung to bus D1 to replace Service Water Pump 2 which was out of service for maintenance) and component Cooling Water Pump 2 had to be manually started.
This failure was traced to a faulty weld in a crank rod that mechanically connects the main contacts to the auxiliary contacts in Breaker AD101. As a result the secuencer for these two loads did not start.
1650 226 There are approximately 28 of this type of breaker in the plant. All will be inspected before the plant is restarted. As of Friday morning, 24 breakers had been inspected and no similar failures were found.
(The failure was caused by the fact that one of two fillet welds on a connecting am was missing.)
8.
After the necessary electrical loads were verified to be operating, the operator attempted to parallel the diesels with the secondary of station start-up transformer 02.
However, this action was not imediately successful because of an improperly set synchronizing switch (operator error) and failures in the two breakers on the secondary of SST 02 (HX02A and HX028).
The cause of the failure of HX02A had not been determined by Friday because it was being used to power the plant while HX02B was under inspection.
It is believed that HX02A failed for the same reason as HX02B. The specific cause of the failure of HXO2B to close on the first try was that the limit switch that detects charging of the trip spring was not actuated.
Tapping the mechanism permitted HX02B to close.
Prying on a similar mechanism permitted HX02A to close. Westing-house personnel were onsite to work on these breakers.
Six breakers of this type were tested to assure operability to allow the plant to meet Technical Specification requirements.
All other similar breakers in the plant will be inspected on a schedule to be issued by TECO. TECO is considering increasing the normal surveillance frequency of these breakers in addition to any Westinghouse recomendations that will result from the Westinghouse work.
It should be noted that there was no automatic transfer of loads from ET01 to SST02 because of the design that actually only provides a transfer on transfamer differential trips.
This situation is contradictory to the design presented in previous submittals such as:
a) the degraded grid report page 8 (TECO letter serial 543 dated October 9,1979), (Enclosure 2),
b) FSAR Appendix 3D page 3D-15, last paragraph, and c) FSAR Section 8.3 l.1 page 8-6 second paragraph (Enclosure 3).
1650 227
. TECO agreed to submit a letter which clearly identified:
a) the bus transfers that are available now, b) the bus transfers that will be available after the next refueling outage this spring, and c) the bus transfers that will be available after a safety grade anticipatory reactor trip system is installed.
9.
After offsite power was restored via SST02 it was decided to restart the reactor coolant pumps. One pump started, but the other three were discovered to have blown fuses in the loss of pump cooling interlock circuitry. The fuses were replaced and two additional pumps were started. The last pump blew the new fuses. Further investigation revealed some welded relay contacts in the interlock circuitry. The relay was replaced, new fuses installed and the pump started. TECO has reviewed the fuse sizing because the event indicated that the fuses may be too small (and therefore blew as a result of currents induced by the electrical transient) and too large (therefore resulted in relay contact damage). The review detemined that the fuses and relays were all sized correctly.
Further inspection of the Component Cooling Water interlock cabinet in the Auxiliary Building revealed corrosion of many of the pins on the backs of the relays. The corrosion allowed a shunt path to ground and this is the most likely cause of the overcurrent that blew the fuses. The relay associated with the RCP that encountered no problem in started showed no corrosion. On overall inspection of the cabinet, it appears that it had been sprayed by water (presumably during construction).
Further investigation will be pursued to determine why the failure did not occur until this event.
The fact that SS101 was not in service complicated the start of these pumps because of voltage drop problems, In order to start the reactor coolant pumps, it was necessary to remove the under voltage bus protection relays.
(Nomally one transfomer is used to start these pumps while the rest of the hotel loads are on the other transfomer).
The seismic qualifications of such plug-in relays were discussed.
The Region III and TECO personnel were sure that letters of certification of seismic qualification were available on site for the switchgear.
1650 228 At this point, a draft I&E Bulletin on plug in relays was discussed. TEC0 stated that they would object strongly to a request for failure data on relays that were not used in safety systems.
Finally, TECO reported that the revised automatic transfer scheme will provide push buttons to hynass the undervnitaos relays, and that a Technical Specific'ation change to allow the use of the bypass had been submitted to the NRC fnr approval.
- 10. The next failure to be reported was the failure of the base plate for a main steamline support in the turbine building approximately halfway between the turbine centerline and a moisture separator / reheater. This was being repaired while the authors were on site. The plant trip had resulted in the shearing of base plate bolts on the support. The corresponding support on the other main steamline was inspected and base plate bolts were found loose.
Both supports were upgraded by adding a third support member, by increasing the size of the base plate, and by increasing the size and number of base plate bolts.
11.
The final equipment failure was reported by the local telephone company.
It seems that the " red thone" showed up on the trouble board during the event but, by the time the telephone company checked out the lines, they could not find any problem. TECO investigated the loss of the telephone and detennined that the ' phone requires 110 VAC power in order to function. The loss of all offsite power will render the
' phone inoperable. At the present time, the indication is that this problem may not be peculiar to the Davis-Besse Plant.
I&E is investigating the situation.
CONCLUSION
- 1. is the best available statement of the sequence of events.
2.
The numerous failures that occurred are not the result of a common cause but several were triggered by the stresses associated with the loss of offsite power (except for the loss of the red phone which was directly caused by the loss of offsite power).
1650 229
. 3.
The program that has been agreed to between Region III and TECO appears to be both reasonable ad prudent.
4.
Electrical transients were initially responsible for the blown fuses in the reactor coolant pump interlocks but the low voltage associated with the heavy loading of SST02 may have caused the failure of the relay contacts in the four pumps.
5.
The present design satisfies GDC 17.
RECOMMENDATIONS 1.
The low frequency power oscillations in B&W plants should be investigated further.
2.
The Technical Specifications should be modified to clarify:
a) the definition of operable; and b) the action statement of 3.1.3.3.b.
3.
The degraded grid study evaluation should be delayed until TECO clarifies the design of the automatic bus transfers.
4.
Region III should follow the resolution of the concerns listed below and advise DSS and DDR of the results of their investigations:
a) pump interlock fuses; b) Westinghouse " mechanical relay" limit switches; and c) General Electric muffler problems.
5.
The Executive Director should be advised of the failure of the red telephone and advised of the final resolution. He should also be requested to consider if a set of standards on emergency telecommunications needs to be developed.
g y.U_.
aj
^ r:
Daniel Garner, Project Manager Ray d F. Scholl, Jr.
Operating Reactors Branch #4 Plant Systems Branch Division of Operating Reactors Division of Operating Reactors 1650 230
$'}O><
T-Nd CATZ CF Tl!?: Oc:cbar 15, 1979 tl
,av, s-s_ ass a..on
.v-_....
s
- DIST 7:CATICN OF OCCURRENC2: -Rasc:or Trip and S:ation 31ackou:
INIT!.iL CONDIT!0NS:
The unit was in Mcda 1, with ?cwer (MWT) = 1665, and load (Gecss MWE) = 530.
OESCRI? TICS OF CCCURRENCE: At approxi=ately 1227 hours0.0142 days <br />0.341 hours <br />0.00203 weeks <br />4.668735e-4 months <br /> on October 15, 1979, a caps-ci:or f ailad in :ha In:agrated Control Syste= (ICS) pulser circui: :o the turbina alactro-hydraulic centrol systa=.
This capacitor failure caused tha turbina con-
which rasui:ad in a reactor protectica syste= reactor crip at the reduced high flux se: point ci approxi=stely 63.8 percent of full pcwer.
Main sta=: saf e:ias 1, 2, and 4 on nu=ber one =ain stea= line and =ain s:ca= saf e:ies 1 and 2 on the nu=ber :Vo = sin stem line lifted to relieve pressura. The icwas:
safa:y valve rase: pressura racorded was 970 psig on Main Stea= Line 1-1 and 963 psig on Sain Stea-_ iina 1-2.
Pressurizer level dropped to a icw of 42 inches and Rasc:or Coolan: Sys:e= (ACS) average te=pera:ure dropped to 55307. Feedwater flow reduced to the =ini=u: flow within one =inute of :he trip. RCS pressure recovered
- o its nor=al value within five =inu:as and the plant responded nor= ally :o the
- ansient.
While raciosing :he generator ou:pu: breaker 34560 a: approxi=ataly 1250 hours0.0145 days <br />0.347 hours <br />0.00207 weeks <br />4.75625e-4 months <br />, "J"
bus : ripped which resul:ad in a da-energi:ation of the star:up transfor=er 01 and a station blackout. Both e=argency diasal generators au c=a:ically started. The Stea=
and 7eedwater Rup:ure Con:rol Systa= (SFRCS) actuated fro = the loss of all four reac-tor coolan pu=ps, and natural circulation was established in the RCS. The loss of tha reactor coolan: pu=ps placed the uni: in viola:1cn of Tachnical Specification 3.4.1 and the loss of "J" bus placed :he uni in the Ac: ion S:sta= ant of Technical Specifica:1on 3.S.1.1.
3 Tha Cc=penan: Coolin; Water (CCW) ?u=p 2 and Service Water (SW) Pu=p )[ failed to autc=a:ically star when :he diesels were s:arted. The operators =anually s:arted both pu=ps :o provida cooling :o che Diesel Generator 2.
Manual centrol of :he auxiliary feedpu=ps was utilized to increase the stea= genera-cr lavals and reduce RCS ta=peratura and therefore pressure since the turbine bypass valves and :he pressuri:ar spray wara no available. When =anusi control of the a:-
cospharic vents was available, the sa: point of the at=ospheric ven:s was lowered :o cen:rol :ha increasing RCS pressura. The highes: RCS pressura recorded was 2232 psig.
Difficui:7 was experianced in ra-energizing "A" and "3" bus frc= Startup Transfor=er 02.
This was caused by s:icking relays which racuired several a::a= pts'a: cicsure f or 'cus "A" and a sligh: :ap for bus "3".
Therafore, it was not until 1316 hours0.0152 days <br />0.366 hours <br />0.00218 weeks <br />5.00738e-4 months <br />
- hat bus "A" was re-energized and a: 1333 hours0.0154 days <br />0.37 hours <br />0.0022 weeks <br />5.072065e-4 months <br /> bus "3" was re-energi:ed fro =
S:artup Transfor:er 02.
The diesel ;anerators were shutdow= and res: ora: ion of the ncn-essentially powered systa=s began.
1650 23i
A lr
, - -, g g.,
m; (
..... a..
.m.
i v....
c.
o,
--v/>
y s..
,..e
, c.: 3
.-6 s
acc.c:or Ccolan: ?=p 1-2 was rastar:cd a: 1415 hours0.0164 days <br />0.393 hours <br />0.00234 weeks <br />5.384075e-4 months <br />. Oue to difficui:ias wi:h bicwn fusas in :ha CC's interlock, :ha remaining rasc:or coolan: pu=ps could not be restarted un:11 1915 hours0.0222 days <br />0.532 hours <br />0.00317 weeks <br />7.286575e-4 months <br /> (?u=p 2-2), 2000 hours0.0231 days <br />0.556 hours <br />0.00331 weeks <br />7.61e-4 months <br /> (?u=p 1-1), and 0135 hours0.00156 days <br />0.0375 hours <br />2.232143e-4 weeks <br />5.13675e-5 months <br /> on the nexn day
?,
g....,e
- /.
. A".'is :S C CCC'. C'INCE: The causa of the trip was the failure of the capacitor in o n s... a.. _a.,s oe.po<.. ( u-e
._.otac._a
.es....a.<...
._...,e.,..,. a.a.,.., c..aa.
- f..
~
a.
e f
--.m.
plc.n: Orip sat;cin: was li=i:ad :o lass than 70*' of full power by Technical Speci-fica:ic 3.1.3.3 since an absolu:a posi:ica indicator was inoperable.)
~ a causa of :he sta:ica blackcu: was a bicwing ou of :ha in:arnals of the rufflar a
en ;enerator cu:put braaker 34560 waich caused a faul: to ground when the breaker uss racissed.
The cc. usa cf the failure of the CCW and SW ?u=p 2 to au:c=atically s:ar: was dua to a failure of :he linkage rod asse=bly in the Diesel Genera:or 2 cu:la breaker.C101.
~.e causa of :he blown fuses in :he CCW in:arlock to tha reactor coolan pu=p s: art c
- s. c..>.
<s s.<
.. ca,. < ves.<-.<,...a..
a, o..
Adequata cora cooling axis:ed throughout the occurrance and at no ti=a did RCS condi-
- icns apprcach the sa uration poin:.
CCRRICT"lE ACT*CN:
aa failed pulsar circuit was repaired and proper operatica of
~~
- he ICS was verified. Since :his was :he third occurrence of problems with :he caps-ci: ors, 3!.W is investiga:ing :he cause of :ha failuras. No previous failures resul:ad s
...,s.. > >.
~.a sufflars on both generator cu:put breakers will be replaced if da aged and a crainforcing band ins:allad to strang: hen :he sufflers.
The linksga rod in the Diesal Generator 2 will be replaced. The linkage red on
iasal Genera:c: I as well as several other braakar linkage rods will be inspected.
The 3:icking "A" and "3" bus ralays were cleaned and adjusted.
All bicwn fusas wera replaced in :ha CCW 1a:arlock circui:.
. a s.,an..,as w,
.s.a.:,ac a a 33.
.,s:a..,.
..ao
......-..S.L.s.....
.a.. a o
.a
.3.-...
--c.
c s;& ica parsonnal.
1650 232
/
.s 5
- a A\\r-CIT T?l? RI?CRT si i
Cc:obar 15, 1979 J
?aga 3 of 3 t
.1...,.
, w. n...u.v. :
1.
Alarn prin:Out (nota: cc=putar was inoparable for saveral sinu:es before and af:a: tha : ip) 2.
?cs: trip review and sequence of events 3.
Raac:ina:e: prin: cut and graphs 4.
Cca:rol Roc.. ch:::s (la:e: - in mas:ar fila only)
-- responsa s::ip charts (la:a: - in = aster file only) 3.
La e
e e
1650 233
./
-g.
, select an alignment of ei:her s:ar:up ::ansforner 01 or 02 to bus A and likewise'for bus 3.
~he reserve source selector svi:ches can be set so that scly sne s artup transforner
'.s ali;ned :o botn 13.3 kV buses A and 3.
( Wen power is being supplied to a 13.8 kV bus by a startup tranfor=er, A.
T the reserve source selec:or svi:ch for that bus can preselect the alternate h
y
]W startup transfor=er source. This permits a fast bus transfer to the alternate startup transfor=er source should the original source fail.
l.
The fast bus transfer scheme from the unit auxiliary transfor=er 11 to
[
the startup transfor=ars 01 and/or 02 is further enhanced by fast acting synchrecheck relays. Opening of the 345 k7 breakers 34560 and 34561 during nor=al power operation vi:5 :he electrical auxiliary systems being supplied power from the uni auxiliary transformer Ill vill result in these relays nonitoring the phase angle difference between the secondary voltage of the unit auxiliary transfor=er 11 and the startup transformer (s) secondary vol: age during the 6 cycle bus transfer. Should the angle exceed 35 degrees the transfer vill be blocked to prevent out of phase reenergization and the associated overvoltage stresses on the plant's electrical auxiliary These fast acting synchrocheck relays are also functional during syste:n.
the 30 second delayed transfer schene described in See: ion II.E.1.
3.
4.16 k7 System During nornal operation and accident condi:1ons vi:hout a loss of offsite power, the safety related 4.16 kV buses C1 and 31 are supplied power from the bus tie transformers AC and 3D (Ref. Figure No. 3).
Each safety related 4.16 k7 bus is provided with a fast transfer (6 cycles) scheme which will transfer the bus from its normal source :o an al:ernate source of power. As an exanple, the nor=al supply of power :o the 4.16 kV i
r650 234 A
s O
6..:
.. e _.,.:. PC4 _.:, ev.sTwe n
w 8.3.1 AC pck23 SYs T.i 8.3.1.1 Tescrittie:
The statien distributien syste= ccesists of varicus auxiliary electrical syste=s designed to provide reliable electrical pcVer during s'.1 =cdes cf statien cperatics and shutdevn ecaditicas (see Figures 8-ua and Lb). The syste=s are designed with sufficient ;cver scur- *s, redu. dant tuses and
..- s.4..ch4
,. c. s,,, e,. e.,.4 _. b.4.,.,.,r. r 4 ee.a.
- c.,e..,.
4
..<.s
...q,a 4
.......~......
are arra: sed so tr.at a less of a single bus sectics, fer any reasen, results in caly single lesses of auxiliaries leav.ng redundant auxiliaries :c perfer=
the sa=e functic.
~he static = distributics syste= is capable cf starting and accelerating the d..
-= s '.
' -a d
., ar.- a. s. -..d..-. d.'
d.v.a... v_.d.* x *
- e.
.a.. d.. c '. *. w...... e.*. e d_ _ _ '.. ~.
the rese je_ crer._ sci:rce fast tra-e"--
. / ervdceA " he svsta vill a
f.e._l. levi-r
--+3 e.g-nerate - - -aa-;c% ria w. i. th_cu..t._ the less of auxiliary t
5%
-/
?-etective relaying is arranged fer selective tripping cf circuit treakers, thus li=iting the Icss of pcVer to the affected area.
8.3.1.1.1 Unit Auxiliary and startup Transfc=ers During =cr=al operatic cf the statien, the 52/69 M7A unit auxiliary pcVer transfer =er, connected to the generater isclated phase bus, prevides the scr-
-a1 scurce of electrical pcver for statics auxiliaries.
Tvc startup transfer:ers, each 39/52/65 MVA, are supplied frc= different 3h5 kV svitchyard bus sectic=s. -ach startup transfer =er provides pcver for startup, shutdev=, and post shutdev require =ent s. The tvc transfer =ers vill also serve as a ec=plete reserve pcVer scurce for the statics auxiliaries in x 4 3 4,,,...,. _. s t.
.s....,, a.4 4..,.. c., w.,,n 4 "cr-*y each startup transfer =er is the reserve power scurce cf c:ly ::e 13.( _k'.' tus. 6ever, if either transfer:er is cr.i ef' se'rvice',' tee 're=aini=g-
-st'artup transfer =er is available (by manual preselectics) :c aute=atically supp'.y teth 13.5 kV tuses should the =cr=al scurce (auxiliary transfer:er)
/
y fail.,_.
j' The seccadaries of all three transfe:::ers are of lev-resistance.rcunded, vye
.,...a.
.s.
1C.4. n.
.._._._..e
..a..~._. n x _ _4 _..e.,... e-4,
- a. a
' ':.. ' V k as d - ' ' d.. '. -. -....'.* *
'.'..-..-._'.,e-a_.*.-.
o,.
,_,....2
..4
...,2._t
.....:.....u.,.......s..
e........:. u.
.. <....g.as,e;.
s.....a.,
^^
.c.
A L
1650 235
'