ML20235J984
| ML20235J984 | |
| Person / Time | |
|---|---|
| Site: | 05000000, Kewaunee |
| Issue date: | 04/08/1968 |
| From: | Specter H US ATOMIC ENERGY COMMISSION (AEC) |
| To: | Rosen M US ATOMIC ENERGY COMMISSION (AEC) |
| Shared Package | |
| ML20234C970 | List:
|
| References | |
| FOIA-87-40 NUDOCS 8707160093 | |
| Download: ML20235J984 (6) | |
Text
-_
- t Morris Rosen, Chief APR 8 1968 Muelear & Systems fechnology Branch Division of Reactor Licensing Eerschal Specter
,l,
Nuclear & Systens Technology Birksh-Division of P.eactor Licensing KBfAUMEE CONTAINMENT FRESSURE TRANSIENT ANALTSIS DEL:N68TB R$
RT-581 CONCLUSIONS The key to the successful opsration of the Kewaunee containment system after a loss of coolant accident is the reliable operation of the Shield Building Vent Systes. Staff calculations indicate that only minor energy i
additions to the annulus air will raise the internal pressure and leakage from the annulus. An addition of only 6700 BTU is sufficient te raise the annulus pressure to +1.0 inches of water at which the leak rate is 407/dsys Since the energy addition to the annules during the first 20 minutes folloving a loss of coolant accident is several hundred thousand I
BTU's, there is a potential sufficient pressure differeace to drive all the air in the annulus out through the Shield Building.
Consequently, unless the Shield Building Vent System operates as designed, it would be as if the containment only coasisted of the primary containment shell.
The acceptability of this design is therefore a matter of reliability ecre tha.n f easibility. The applicant must demonstrate that:
i e) The p rir.c ry n: t : crn be rapidly isolated f rom che secondary q
system af ter a loss of coolant accident, with minimal leakage through the isolation valves.
b) Th3 valves in the vent system will open, rapidly and reliably.
c) The fans will start in time, and will achieve a negative pressure within a prescribed time limit based upon dose esiculations.
d) That the Isakage to the environment f rom the vent system itself would be very low, particularly f rom the ductwork upstream of the filtara.
e) That the condition of the filters is always such that they can perform as designed.
.s,\\ '
N.*l-I l
THOMAS 87-40 PDR l
j l
1
e E Rosea APR 8 1938 D.
,Recoessendations in order to construct & a highly reliable containment system it is recommended that all componects be thoroughly tarted prist to start up and that there be a comprehensive surveillance program to insure that there is ne deterioration of the systan.
The Shield Building Yeat System proof teste should include manometers at numerous locations within the annulus, meteorological effects such as barometric changes and high wind velocities should be cons @ted in the selection of the f an especity, the selection of all components should be that they have a very high reliability of continuous opera-tion, and that the adequacy of the fans to reach a negative pressure t<1 thin a specified time be established. The initial acceptance tests of the Shield Building Vent Systwa should be performed (with heat addition to the annulus air at rates and amounts equal to that which would be rejected by the stasi primary containment vsssel during the design basis loss of coolant accident), unless it cat, be demonstrated that an isothermal flow test ie an adequate approximation 46 aca.ident ccadittens.
In addition to producing a highly reliable Vent System, the cvsrall containment system would be significantly improved by having the primary and secondary containment leak rates reducsd to as low as p ra ctical.
hower leak rates will produce icver doses and wesid also permit a relaxation in the start-up and performance requirements of the fans and the butterfly valves in the vent systems. Practical leak rates should be established and put into the technical coccificetions.
ii.
Facility Description and Operations, The Kevaunae containment system consists of three major parts:
a steel pressure vessel primary containment shell surrounded by a con-crete Shield Building which acts as a secondary containment structure, and the Shield Building Vent System.
The steel containment vessel is a cylindrical shall with a hemispherical doma and an 2111psoidal bottom. The cylindrical section is 1.5 inches thick,105 f aet in diameter and about 142 feet in height. The thick-
' ness of the steel in the dome is 0.75 inches. The containment vessel is designed for a maximurn internal prsseure of 46 psig at a temperature 3
of 268'y.
The net free volume exceeds 1,320.000 ft. A containment vessel leakage rate of 1% of the free volume per day at 46 psig has been used as the basis of the Kevaunee dose calculations, llowever, the construction of the vessel and its penetrations vill be such as to produce a significantly lower leak rate.
..____________._____________________w
M. Basen APR 8 E8
]
The Shield Emilding will be a rainforced concrete structure with 2.5 feet think walls and a 2.0 feet thick does. There will be an annu!.ar air space of approximately 400,000 ft3 between the steel
{
centsiament vessel and the Shield Building. The Shield Building j
leakage rate used for the off-site dose analysis is 101/da1 at a differential pressure of 0.25 inches of water. This differential pressure is the differemais between atmospheric conditions outside of the Shield Building and the anaulus air pressure. Actual Shield
{
Suilding leakage rates are espected to be well below the 10%/ day.
j value used La the dose calculations. A Shield Building leak rate of i
about 0.5%/ day [at 1/4"' Eg0 differential) has been calculated for a construction utilising gastated bulkhead personnal doors.
An important part of the containment system is the Shield Building Vent System. The Shield Building Vent System provides a negative l
pressure within the annults immediately following the loss of coolant accident, enhances mixing of the air within the annulus, and provides long term filtration of the annulus air.
1 During normal operation of the plant the shield Building Vent System is not operating. The fans and valves of this system are activated upon a signal of high pressure within the containment vessel. The applicant has calculated that the vent systen will be operative as early as 12 seconds with off-site power.
If there is a loss of off-site power and one diesel of the auxiliary power system fails to start, then the time for the vent system to reach full capacity would be 46 seconds. The analysis of the dose effect for a delayed fan start assumes full capacity in 60 seconds.
CGculations based upon a ground release model and a 600 meter exclusion radius indicates that it would take over 7 minutes of outleakage from primary containment shall (at 1%/ day) before 10 CFR 100 limits would be exceeded. Consequently, a fan full-capacity time of 60 seconds is acceptable.
The Shield Building Vent System consists of fans, filters, an' exhaust vent, ductwork, and valves. A suction header is placed above the top' of the outside of the containment vessel, while an inlet vent system header is placed in the annulus near the bottom of the containment vussel. The inlet vent system header will be placed where the bulk i
of the fission products are expected to leak into the annulus following a loss of coolant accident.
After e loss of coolant accident the temperature and pressure within the primary steel containment ressel will rise rapidly, part 6f the energy associated with the accident will be deposited,into the steel vessel, raising its temperature. and in, turn some of the anarre ca r
]_
M.'Rossa APR E EB 1
1 it.t a the steel vessel will be transmitted to the air in the auculus.
Concurrent with the energy addition to the air in the annulus there would be a leakage of fisaba products from the primsry sentainment into the naa a a.
The high pressure signal from the primary somtainment causes faela-ties valves to elese and thereby separate the primary contaissent from the annulus, opens butterfly valves in the vent system, and
)
starts the retirewistion and exhtest fans of the Shield Suildiwg Vest System. The air in the annulus is pulled through the suction header, put through the filter system (demister-particulate-obsoluto-i aharcoal) and then recirculated back through the inlet vent system header. The design objectivs of the suction and inlet vent system j
headers is to induce airing and promote hold-up of the fission products within the aannius.
Within minutes af ter thefans in the vent system are operating t.he pressure in the annulus will be negative, relative to atmospheric conditions. At this time there will only be in-leakage into the Shield Building and there will be no contribution to the off-site de,se from the ghield Building itself.
In order to maintain a negative pressure within the annulus, the pressurizing effect of the hast addition from the steel containment is offset by discharging, after filtration, a portion of thJ circu-lated air. At the energy from the steel shall that is transmitted to the annulas air decreases with time, the discharge flow also decressas. Af tar about 14 mirutas the discharga flow is just that required te offect the sir mass addtion due to shield building in-leakage which is 168 f t /mipute. AEC staff evaluations have indicated 3
that the applicant's discharge flow values are consistent with the annulus heat up rate ar.d pressura leskage relationships.
5.
Desita Basis Accident The appif. cent has reviewed varf ous kinds of less of coolant accidents in order to ' determine the situation that produces the highest eentain-ment prassure.
Fros. previous analyses it has been shown that the 3 ftI break size resulta in a blowdown period of sufficient duration to allow & maximum not energy tyr.nsfer from the core and reactor ves-sel internals to the coolant. This imparts a maximum blowdown pren-sure to the containment. Consequently, all of the containment analyses are based upon 3 ft2 break sise.
Pressure transients vere calculated for three different situations:
i i
i
L' Rosea _g A.
3 ft1 break,' with sore cooling 3.
3 ft2 break, ne aere eselias, essentially all emergy transferred to the sentainment in first 39 seesuds C.
3 ft2 break, no sore eeelias, some of the initial energy is kept is sore to start the Er-E 0 reacties.
2 These three cases saa be compared as follows:
Case
. Peak Pressure, PSIG Time of Peak Pressure, Seconda A
42.2 15 A
45.5 39 C
44.9~
270 All values are below the containment design pressure of 46.0 psig.
Casa A which has tho' largest margia to the sentainment design pree-sure represents the most probabla peak pressura if one has constructed a highly reliable emergemey core cooling system.
Caseo B and C, which are consistent with Criterion 49, appears to be a conservative analysis of energy releases for the no ecta cooling situation, based on present core heat up and metal-water resetion
- models, t>md upon the energy sources and sinks supplied by the applicant, staff calculations indicate a peak presours of 43.5 pois for Case A if all the primsry coolant is blown down.' The applicant assures a aamil amount of water remains in the vessel and this produces slightly lower peak pressures. The agreement between AEC calculations and the applicant for Case A is considered adequate.
j It should be pointed out that any sacculation of the peak contain-l ment pressure is subject to sa " error band".
Present analytical models that predict blowdown rates tend to overestimate mass flev rates which in turn, overestimates peak pressure. Assuming an instantaneous break area, rather than a more gradual area producing mechanism also produces high mass flew rates and high peak pressures.
On the other hand, errors in the initial and/or final soolant inven-(
tory, the heat transfer coef ficient during blevdown, the condensing I
coef ficients, etc., could produce changes in the peak pressure. An
]
important area that is subject to much guestion is the description l
of the geometry of the core during a major metal-water resetion, and what mechanism wuld actually terminate this reaction. If large s
s
____1___________._________________.__._
6-APR 8 E8 9
M. Rosen segments of the e, 's ault and collast in the pressure vessel, will the vessel melt through and rapidly import a large mass of het metal and fuel to the water en t.as reactor btL1 ding floort Another possi-bility is the questtoa of delayed core cooling.
If, initially, there is no core cooling sad the cladding heats up to temperatures abeva 1800*F and then the secumulaters turn se, what would be the thermal shock 9ffsets en the fuel and would there be a sudden increase is netal-vater reactions and a sharp pressure risef Since tM se vaksowas exist it seems that the only fair centement that can be made about the adequacy of the Kewaunee plant to with-stand a loss of coolant accidaat is this: Based upon present models and svailable data, the calculated peak containment pressure is slightly less than the containment design pressure. Since the containment has the structural capabilities of withstanding com-siderably higher pressures than the design pressure, it has the ability to absorb some errors in the calculated peak pressure.
At this time an error analysis on the containment peak pressure has not been performed.
Leakage, and therefore dose considerations, vill be more limiting es the peak pressure than structural limits. If leak rates increase rapidly for pressures above the design pressure, the dose limits may be exceeded long before there is any permanent damage to the primary containment vessel.
1 Distribution:
Suppl.-
%is ERL Reading N& STB Reading AD/RT Reading 4
bec:
S. Levine H. Spectar l
omer >
......--. DRL..
sURNAMc>
.MS.P.pc.t.RKxt.b.
h.ff.[.kh.-
~
1 DATE>
R)rm AEC-818 (Rev,9-63) u.$ sovrRNMENT PRINTING OFFICE;1W 214429 f
I