ML20008E433
| ML20008E433 | |
| Person / Time | |
|---|---|
| Site: | Yankee Rowe |
| Issue date: | 07/19/1963 |
| From: | YANKEE ATOMIC ELECTRIC CO. |
| To: | |
| References | |
| NUDOCS 8101070171 | |
| Download: ML20008E433 (89) | |
Text
.
- =
I LOO:1 3/31/61 1
1 4
ACCIDEh"fS AND FJ2ARDS O.
400 GENERAL It is generally recognized that pressurized water reactors exhibit a high degree of inherent stability, due primarily to the large negative temperature coefficient of reactivity associated with the change in density of the coolant moderator with temperature, superimposed cn the scuewhat smaller negative coefficient of the fuel itself. The instrumentation and centrol system is designed to take full advantage of the inherent stability.
The Doppler coefficient which represents the major component of the fuel coefficient serves to effectively minimize fast transients. This coef-ficient is particularly important when the disturbance introduced is in the form of reactivity insertion. The prompt Doppler vill immediately tend to counteract the reactivity increase through a rise in fuel temperature and a consequent reactivity decrease. Because of the long time constant associated with transfer of heat frca the fuel to the moderator (of the order of 5 sec),
the void and moderator coefficients vill not be effective initially but will i
provide long term damping and stability for the cases involving reactivity insertion.
At 485 mv thermal the mixed mean temperature of the coolant leaving the reactor is 531 F and that leaving the hottest channel is 589 F.
Since saturation temperature is 636 F at 2,000 psi gage, even the water leaving the O) hottest channel is about 47 deg below saturation temperature, with the result V
that there is no bulk boiling and only a negligible amount of nucleate boiling in a small portion of the core at the hottest channels. Temperature changes, c;upled with the negative temperature coefficient of the reactor, act to limit smaller transients, while bulk boiling in the hot channels operates to reduce and limit reactivity.in. larger transients where the power increases slowly.
Twenty-four movable control rods are provided for regulating the power level of the reactor, ccupensating for fission product buildup, counter-acting the effects of fuel burnup, and for scramming the reactor either manually or autcmatically. The control rods are capable of shutting down the hot reactor to about 8% suberitical after operation at power has continued long enough to reach equilibrium xenon and samarium concentration. Additional control to bring the reactor to cold shutdown is provided by the chemical t
shutdown system through which negative reactivity can be introduced at a rate
.of 0.6% per minute.
Reactor accidents are of two general types:
those associated with reactivity insertion, and those resulting frcm mechanical failures of plant
. equipment.-
Several accidents involving reactivity insertion have been consid-m ered:
those resulting frca continuous control rod withdrawal at start up and at full power, and those involving changes in water temperature and in boron concentration. In each of these cases'the specific instrumentation v
voI01*N
a h00:2 12/23h3 h
and control features and the administrative procedures which serve to protect the reactor against each condition are outlined. In each case, also, a compound failure of protective devices and/or procedures is assumed and analyzed.
One type of mechanical accident is the complete loss of power to the main coolant pumps which reduces the coolant flow through the core and might cause a harmful increase in core temperature. Another mechanical accident is the sudden loss of load on the turbine generator and the resultant transients experienced by the reactor system. The third type of mechanical accident involves a rupture of the main coolant system and loss of large quantities of water into the vapor container.
All of these accident conditions are analyzed in detail, a "marimum credible" accident and a " hypothetical" accident are defined and the effects of these conditions on the reactor plant and environs are described.
The analysis of each of the various accident conditions have been based on operation of the reactor at both the design power out;ut of the initial core, 392 W thermal, and at h85 W thermal which corresponds to the plant rating. In most cases operating power level is found to have little effect on the severity of the accident. Where the effect is appreciable the
(]
more serious case is considered.
Lj Additional studies of the various accident conditions have also been performed to more accurately define the capability of the second core, based on operating experience and experimental data obtained over the life of Core I.
The results of these studies are described in the following Sections 1/11-h03 Accidents and transients will be reanalyzed in subsequent cores if hot channel conditions are more severe than previously studied, or if the reactor power level is increased.
i
. _ _. - _ ~ _ _ _ _
401:1 3/ 31/61 401 REACTIVITY ACCIDEN'IS O
Startup Accident Startup accidents have been analyzed in considerable detail for both research and power reactors.
In the case of the pressurized water type reactor, j
extensive analytical work has been done for this type of accident using analog cceputers. The pattern of the accident is, therefore, well understood.
A reactor of the Yankee type, with a sizeable prompt negative Doppler coefficient, is much safer during rapid startup accidents than a highly en-riched pressurized water reactor. The pattern of the accident is such that the power range is entered on a very short period and thus termination of the accident must be frca prmpt effects. Although a shorter fuel-to-moderator heat transfer time constant would reduce the long term transient more rapidly, a more important criterion for safety in a startup accident is the self-limitation of the initial burst in power to tolerable levels prior to external corrective action.
Cold Startup - The cold startup accident is hypothesized on the following basis:
-5 1.
Reactivity is inserted at the rate of 15 4 x 10 Ak/sec,whichis I
greater than the rate corresponding to movement
- of the high worth l
(safety) rod group at its maximum incremental worth position, and 4
which is equal to the experimentally observed maximum rate due to
.O the movement of the next highest worth group at its design speed V
of 12 in./ min. Although all rod groups are presently geared to operateataspeednotgreaterthan6in./ min.,thedesignwith-drawalrateof12in./ min.wasusedinthisanalysissinceitis possible that in the future it may be desirable to increase the speed of all groups except the safety group to this limiting value.
-5 I
2.
The experimentally ' determined prcept neutron lifetime is 19 x 10 to 1.6 x 10~5 sec; the delayed neutron fraction based on calculated end-of-life concentrations is 0.0061. The minimum cold shutdown
~
reactivgtyis-5%4k/k. The source corresponds to 1.2 x 10-5 to 6 x 10- mwt/sec.
3 The. initial system temperature is 70 F, the lower limit of temperature (ccupared with the normal minimum startup temperature of 250 F),
which determines core reactivity and tempcrature coefficients. This temperature gives the lowest reactivity coefficients, most reactive core, and hence the worst' condition.
4.
The initial system pressure is 200 psia.
-5 Instrument settings are those for 485 mwt operation.
Uncontrolled withdrawal of more than one group at a time is believed to be incredible because it involves simultaneously shorting the'" withdrawal" W
- contactors of'other groups at the time'the operator is withdrawing a Q.
control group.
~
F t
g
- r
,e,-
h01: 2 5/8/61
('
The responne to a otartup accident in strongly influenced by the
()
reactor power reactivity coefficient, which tende to increase in magnitude as power increases. Since a low magnitude of power coefficient leads to the most pecalmistic solution, the power coefficient van calculated at 485 mvt, a power lower than that at which damage could be expected to occur. The parameters used in determining this coefficient were 600 Btu /f t hr-F
- Average clad-to-fuel gap conductance
=
2.6 Btu /f t -hr-F
- Average fuel conductivity
=
344 F Fuel surface temperature
=
490 F Fuel nyerage temperature
=
A calculation baced on these parameters showed that the cold power coefficient is about one-half of the hot power coefficient.
Hence the cold power coeffi-cient for this analysis was assumed to be one-half the lower limit of the hot power coefficient determined experimentally in the reactor at full power.
The moderator and pressure coefficients are the same as were used in YAEC-109.**
The startup accident is guarded against by incorporating protection in the design of the equipment and by administrative controls. The following briefly outlines the protective measures:
1.
Source Range Instrumentation - horn and light at 1 decade per minute, g
and rod scram at 5.2 doendes per minute.
(s' 2.
Intermediate Range Instrumentation - horn and light at 1 decade fer minute, rod stop at 1.5 decade per minute and rod scram at 5 2 decades per minute.
3 Power Range Instrumentation - Level scram set at 35% power.
4.
Administrative control limits rate of power ascension to less than 1 decade per minute.
5 Administrative control dictates that intermittent rod withdrawal with continuous count rate monitoring vill be used for startup.
6.
Administrative control dictates that criticality will not be at-tained before boron is reduced to 200 ppm.
7 Administrative control dictates that boron dilution vill not begin until temperature equilibrium is reached at 250 F.
Thus, keff can be brought to a value less than, or equal to, 0 97 at any time during plant heatup before low power criticality. Moreover, at all times throughout core life, the moderator temperature coefficient is sub-stantially negative at this temperature with 950 ppm boron.
Maximum expected valucc.
YAEC-lO9, T. Gogniat and D. Hunter, " Response of the Yankee Reactor to Startup Accidents", January 1959
401: 3 5/8/61 q
8.
A safety rod group vill be cocked and ready for innertion, in case V
of an excursion, before dilution and heatup.
The physical mechanicm of the startup accident is as follows:
Operator is withdrawing rods as in a normal approach to criticality, a.
b.
First Failure - Operator error ignores 1 decade per minute born and light in the source range and continues withdrawing rods.
Second Failure - Equipment failure of startup rate scram.
c.
d.
Rod stop signal frm inte rmediate range, source or intermediate range startup rate scram, or scram at 35% power, with emplete rod insertion in less than 2 sec, would reduce power after it had peaked at less than 0.1% power.
The results of interest in the subpower range are expressed in terms of the time frcus start of rod withdrawal:
Time to rod stop at 1.5 decade / min 295 see Time to critical 325 see Time to 1% power 355 see k at 1% power 0.47%
The reactor does not go prompt critical.
m f
/
In the power range transient fuel temperature and heat flux behavior
' were estimated on a steady state basis assuming total reactivity is zero at any time after the 1% power level is reached. This method was checked against the previous Yankee startup analysis, performed with an analog computer, and found to be very good for estimating the first minute of the temperature transient. This method utilizes reactivity coefficients for power, moderator temperature, and pressure, as well as the reactor loop thermal characteristics.
On these bases the apparent core power, as indicated by fuel tem-perature and heat transfer rate, reaches 485 mvt in 10 seconds after reaching 1% power, and increases at the rate of 12 mut/sec. Center of fuel melting, based on hot ~ spot fuel condu:tivity~of 1.15 Btu /hr-ft-F, is reached 26 seconds after the 1% power-point. Departure from nucleate boiling, based on Gunther's correlation, is reached about 30 seconds after the 1% power point.-
Even if the startup accident considered above were allowed to pro-ceed to empletion, assuming.no corrective action by the operator and without the operation of any mechanical protective device, the resultant condition would be considerably less serious than that considered in the Hypothetical Accident.. (SeeSection403). Any reactor subjected to uncontrolled continu-ous reactivity insertion vill eventually reach a damaging condition (because of the fact that temperature effects are balancing ever increasing core re-activity). The safety of a reactor, therefore, depends on the number of fm '
~ protective devices, the administrative controls, and the time available for these.to take effect.
41:L 3/ 31/61 Hot Startup - The hot startup accident is hypothesized to occur
[_}
when the main coolant system temperature is 516 F, and the pressure is 1200 V
psia.
This accident is guarded against by the same equi;nent and adminir-trative controls previously listed for the cold startup.
The reactivity coefficients used are:
- 1 5 x 10 8k/F "g
=
~
+ 1 5 x 10 6k/ psia oc.
=
P
~c
- 2 5 x 10 ' 6k/=vt oc
=
q Eecause of the higher power coefficient, power ls less than Lo5 =vt for =cre than two minutes after reaching 1% pcVer.
Pressure is the first prcblem en-countered, with 2500 psia being reached in 25 seconds if no spray or relief is actuated. The time to fill the pressurizer is Greater than LO seccnds.
DNB does not occur until after the entire primary system reaches saturaticn temperature.
Continuous Rod Withdrawal at power Another type of reactivity accident is continuous rod withdrawal at power. In this case, the reactor is operating initially at or near 100%
power, and a continuous withdrawal of control rods at the maxi =u= conceivable
(~ ')
speed occurs. Thus, reactivity is introduced at 15.A x 10-5 o k/sec. Pro-V tection is afforded by the folleving:
1.
Level scra= - set at 120% for 4-loop operation and 110% for 3-lcop operation.
2.
Autcr::atic rod insertion (overriding any " rods out" signal) when T"Y8 goes above dead band.
3 Light when rods are going out gives indication to operator.
4.
High Th * ""
- 5 High~ pressure alar =.
6.
High pressurizer level alarm.
7 Administrative control exercised through nor=al operator cognizance of the situation.
The physical mechanism of the rod withdrawal at power accident is as follows:
Reactor is at or near 100% power while on autcnatic control.
a.
b.
T drif ts down to the lover end of the " dead band", initiating a b out" signal.
401: 5 5/8/61
[]
c.
First Failure - Equipnent failure - As T reaches t..e upper end O
of the dead band, the " rods in" signal Es not occur and rods continue to move out.
d.
Second Fa!1ure - Operator error ignores indication of " rods out" signal (light for rods out), high T alarm and high pressure alarm.
h High power level scram at 120% (4 loops) or 110% (3 loops) wi',hout e.
core d6 sage. 'Ibece levels have been conservatively set to prevent core damage.
This case was analyzed with an analog emputer. The rod rate and reactivity coefficients ascumed were the esme as for the hot startup accident.
Initial conditions were 2000 psia, 516 F core average temperature, and 485 mvt. From the start of rod withdrawal, the 120% power level scram would be reached in 16 seconds. At this time bc 9 seram and automatic control rod insertion are initiated by different elements of the control and safety system.
If both safety actions fail, and the operator does not respond to alarms, center of fuel molting, based on fuel conductivity of 1.15 Btu /hr-ft-F and gap con-2 ductance of 1000 Btu /hr-ft -F (the minimum expected values at the hot spot),
occurs in 27 seconds; pressurizer relief valves open at 2400 pain in 38 seconds; DNB occurs in 82 seconds; design pressure of 2500 psia is reached in 85 seconds, and the pressurizer is filled in 100 seconds, page 401:6.
The results of this accident may be compared with the similar e
accident analyzed in YAEC-109, in which 120% (of 392 ut) power was reached (n) in about 25 seconds, and power approached 137% asymptotically. The principal reason why the power rises more rapidly in the p/sec vs 21 x 10-5 6 k/sec) is resent study, despite the lower reactivity insertion rate (15 4 x 10-5 4 k that the power coefficient due to Doppler effect, 2 5 x 10-5 8k/mvt is less than half the value of 5 7 x 10-5 6 h/mwt indicated by the ncninal Doppler coefficient of 4 x 10-5 4 h/F used in YAEC-109 Other factors contributing to a greater power increase in the present study are a smaller ne6ative moderator coefficient and a finite positive pressure coefficient, with pressure being calculated frcn the coolant volume increase.
Cold Water Accidents Addition of an Isolated Loop - Since the Yankee reactor has a rela-tively large negative temperature coefficient of reactivity, a lowering of temperature represents an addition of reactivity. Such a downward temperature change might come about through opening valves which previously had isolated a coolant loop containing water at a temperature below that of the water in the reactor core. Protection against such an occurrence is afforded by the fol-lowing:
1.
Temperaturo interlock set at 30 F or less prevents valve opening unless the differential between the cold les loop temperature and the highest cold leg temperature in the other loops is within the interlock setting.
401:6 3/31/61 O
16 0 5
\\
a ONS OCCURS
$w g
~ LOCALIZED FUEL MELTING BEGINS 12 0 T
AurOMATic ROD INSERTION ayo 120 % POWER LEVEL SCRAM i
I I
I I
,oo O
20 40 60 80 100 TIME FROM INITIATION OF ROD MOVEMENT-SECONDS O
2600 SAFETY VALVES OPEN PRESSURIZER RELIEF 4
VALVES OPEN
/
p 2400 PRESSURIZER FILLED WITH WATER E
2200 l
1 l
l
- zow,
,o
,o
,o TIME FROM INITI ATION OF ROD MOVEMENT-SECONDS TRANSIENT RESPONSE TO UNCONTROLLED INSERTION OF O
seaCrivirv raOm 4es awT eowen
401:7
$/8/61 2.
Slov opening valve permits mixin6 of isointed loop water with water in the active loops. Loop transit time is short emipared to valve opening time which increases mixing, as does the configuration of reactor piping and internals.
3 Cold leg lov temperature alarm.
4.
Startup rate protection and 35% power level scram, if at low power levels, and 120% power level scram, if at povar.
5 Administrative control dictates that the isolated loop will be borated when it is more than 30 F colder than the highest core inlet water temperature.
6.
Administrat!ve control dictates that temperature and boron concentrn-
. tion dil. be matched before a loop is cut in.
A combination of interlock failure with failure to borate or failure to match temperatures is analyzed in the following manner:
a.
First Failure - Equipnent failure - 30 F interlock does not function.
b.
Second Failure - Operator error - does not match the loop tampera-tures before cutting in isolated loop.
Cold.(VO F) borated (950 ppm) loop is cut into a hot (516 F) critical I
c.
O system.
d.
The initial. positive reactivity transient that might be caused if the cold loop " hit" the core as a slug, which is highly unlikely because of mixing, is reduced to a negative reactivity transient by the-boron added to the system, i.e., 950 ppm of boron relative to zero ppm is.vorth more in ne6ative reactivity than 70 F relative to 516 T-is worth in positive reactivity.
e.
The long_ term effects caused by the mixin6 of the isolated loop with the active -loops is. complicated by the fact that the isolated loop
. steam generator acts as a heat sink, effectively increasing the total cold water available. Conservative calculations indicate that reactivity would be added at a rate less than 1 5 x 10-4 A k/sei:..
t 1
This rate can be handled by the control rods; moreover, this rate is
. slower than the reactivity insertion rate associaf ed with rod with-
- drawal at power.. Thus,-the transients are similar, but less serious.
Even it.this accident is allowed to proceed to=ccmpletion, sufficient control' remains- (1% shutdown in boron + 1% shutdown in control rods) to render the core suberitical in the final equilibrium condition.
The presence.of fission products would increase the available shut-
'down control. The-insertion.of control rods vould be pranpted by either-operator response to the cold' leg Icv temperature alarm or nutomatic control system respor.se to period or power level signals
.g..
before core damage occurs.
A
+
n
+
)
t f
M
/
t
-1 4-P-n*v' r
401:8 3/ 31/#1 If the boration procedure should also fail and the accident is allovei to proceed to ecxnpletion, core damage might occur; however, the consequences would in any case be less than those of the Hypothetical Accident. The ex-tent and the probability of core damage occurring depends on the manner in which the isolated loop " hits" the core; for a fully mixed cold loop, no damage occurs, while for the other extreme (slug flow), core destruction r
would occur. Although mixing data gathered as part of the thermal and hy-draulic design studies indicates gross mixing would probably occur, it is impossible to say concretely that no point or pointa in the core vill be subjected to a sudden decrease in temperature before complete mixing occurs.
Because of these uncertainties, administrative control and mechanical inter-locks have been incorporated in the design in order to prevent cold water accidents.
Essentially the same argwnents and conclusions apply to either the 392 mv or 485 av core.
Starting of Boiler Feedvater Pumps - An additional cold water acci-dent results from the possibility of suddenly turning on the boiler feedvater pumps while operating at low power. This could result from operation of the boiler level control manually during which time the level is allowed to fall below the rormal level. Then the feedvater flow would be increased to re-store level, either manually or by switching into automatic level control.
The result would be a sudden cooling of the secondary and primary water. An analogue ccanputer simulation of this accident has shown that the resulting rate of reactivity addition is approximately the same as the rate assumed for f}
the study of the continuous control rod withdrawal accident. Thus the conse-quences are less severe than the consequences nf those accidents, since the s-primary temperature and pressure incresces accompucying those accidents do not occur because the reactivity is being insert *
- by cooling of the water, rather than by control rod motion.
The severity of this accident could be increased, however, by simultaneous removal of control rods by the automatic control system in re-sponse to the cold water signal. For this reason the autamatic control system is not used until the steam generator level is at steady state under automatic level control.
In order to prevent this conditicn an administrative control of the boiler feedvater pump operation hris been adopted which requires the establishment of correct levels on th6 -team generators before the turbine is loaded to a level which would cause operation of the pennissive relay and thac allow autcznatic control rod withdrawal.
Boron Concentration Accidents Another type of reactivity accident is an unscheduled decrease in the boron concentration. Two general types of boron changes are possible; one type is a slow rate of change in boron concentration, and the other $n-volves a rapid rate. Representative cores have been analyzed for both and are presented as follows:
g s
401:9 5/8/61 I.
Slow Changen in Boron Concentration A.
Failure to Borate Before a Reactor Cool _down Protection is afforded by the following:
1.
Startup rate protection.
2.
3$% power level scram.
3 Pressurizer pressure and loop temperature alarms.
4.
Administrat,1ve control requires boration to cold shut-down concentration before cooldown.
5 Administrative control requires cocked safety rods during cooldown.
The physical mechanism of the credible accident is a com-bination of a, b ar.d c below:
a.
Operator error - failure to borate before cooldown from 514 F with the reacter suberitical on control rods.
b.
Operator error - failure to leave safety rod group fq cocked before cooldown.
O c.
Operator error - failure to observe pressure and tem-perature alarms.
- d.. Combination of a'+ c leads to startup rate alam and scram or high pressure alarm and level scram before core damage, e.
Cambination of a + b leads to operator cognizance of
.the problem and boration before core damage.
In either d or e above,/sec; which corresponds to the reactivity is being added at f.-
less than 6 x 10-6 A h maximm design cooldown rate. Neither case'is con-sidered to represent a serioua operational hazard.
'B.
Dilution of Boron While in the Cold Shutdown Condition In most cases a safety group of rods is available. Although it is assumed that no cocked safety group of rods is avail-able protection is afforded by the following:
'l.
12 wt % boric acid available from the chemical shutdown system ready to insert negative reactivity at -1 x 10-g.
75
.Ak/sec.
I 9
1
-l c-
.__=_.
401:10 3/31/61 2.
Administrative control does not permit baron dilution
'1 before 250 F is reached.
l 3
Administrative control calls for continuous operator attention and check of the systems while cold.
4.
Relief valve on shutdown cooling system serves as l
partial pressure relief.
5 Administrative control calls for periodic chemical analysis of the system.
The physical mechanism of the credible accident is as follows:
1 The reactor is 5% shutdown, head on and control rods a.
ins?rted.
b.
First Error - Operator dilutes boron.
c.
Second Error - Operator. fails to observe increasing count rate.
-d.
Power level rises'due to reactivity addition at 1.6 x 10-5 ak/sec, based on 100 gpm bleed and feed. The temperature _ and pressure follows the power closely.
- e.
The operator notes rise in temperature and pressure,
't
. \\.'
stops dilution and initiates injection of 12 vt %
~ boric acid solution.
. Even when carried to ecznpletion with sequential operator errors, none of the slow reactivity addition cases can
~
result in consequences as severe as the Hypothetical Acci-
. dent.- Moreover, the rates of addition are so slow that core damage can be envisioned only under the most unlikely r
ccabination'of. circumstances.
~
II.. Fast Changes'in Boron Concentration *
!I Isolated Loop Accident Protection against cutting an unborated loop.into operation while. alli other loops are borated is provided by the follow-Ling:
~
i l.' Slow opening valves-permit' mixing of the-isolated loop water
- with the borated. vater in'the active loops.
- E As pointed out-in Section 106,i chemistry ' tests have shown that " boron M
1 hideout" for the Yankee water:. chemistry is impossible. -Therefore, this accident bas not been evaluated.
(1.,
~
'N..
-i.
401:11
$/8/61
\\
2.
Startup rate protection and 35% flux level scram.
3 The reactor is always at least 3% subcritical whenever the nonisolated loops are fully borated.
h.
Administrative control dictates that the isolated loop be borated as soon as it becomes 30 F cooler than the other loops.
5 Administrative. control dictates that boron concentration be checked and matched before a loop is cut in.
In any event, the condition caused by cutting an unborated cold loop into three borated cold loops has been analyzed and found to present no problem since, under equilibrium conditions, the cold clean core with all rods inserted would remain about 4% suberitical.
1 The initial local positive reactivity effect caused by a "slu6" of unborated water is reduced by the slow opening valves and the inherent mixing effects.
m)
?
h h01:12 6/1/62 Reactivity Accidents with Core II at $h0 Mwt Startup Accidents Startup accidents are not significantly affected by a change in power rating. The measures to protect against these accidents outlined on
.pages h01:2 and h01:3 are still in effect. Minimum measured values of negative reactivity coefficients are also essentially the same as those used in tr.e original analysis of startup accidents. Therefore, the conclusions reached in these analyses still apply and the accidents have not been reanalyzed.
Rod Withdrawal at Power Analog computer studies of rod withdrawal accidents were performed covering the full range of incremental rod group worths, power and temperature reactivity coefficients, and pressure and temperature conditions. A scram set point of 108% of power was assumed, with a maximum overpower of 116% including the worst possible combination of instrument uncertanties (see page 102:15,17).
Tnree additional safeguards against excessive rod withdrawal were also assumed to be in effect above a power level of h85 MWT:
1.
Elimination of automatic rod withdrawal.
4 2.
A rod stop has been installed which must be manually reset after each rod step.
3.
Maximum rod withdrawal is administrative 1y limited to 3" in any e-hour period.
As might be expected, removal of the highest worth rod group (6 x 10 h f k/ inch) results in the most severe conditions. A plot of core c
. response to withdrawal of the maximum worth group is shown on page h01:13.
The coordinates are outlet temperature, and nuclear or thermal power. The jagged line represents nuclear power, which responds instantaneously te step
- motion of control rods, then decreases as temperature rises. The scalloped curve represents the actual heat transfer and temperature conditions for the he' channel. The solid lines represent the expected variation of nuclear and thermal power with scram at 116%. The dashed extensions of these lines show what would happen without a scram. The slanted straight line indicates the conditions under:which a DNB ration of 1.25 (W-2 correlation) would be reached in the hot channel with pressure at its worst value possible during the tran-sient. The W-1 correlation is not shown because it is never limiting in this accident.
-It can be seen that DNIR=1.25 is not reached in the transient; the minimum DNBR at 116% of power is 1.3.
Without scram, the DNBR would drop to 1.23. - By comparison,the W-1 correlation yields a DNBR of 1.h8 without scram.
-p t/
)
O Obli 560 H-DN BR= l. 25 m
E 555
[ MAXIMUM SCRAM 3,,N s
E EQUILIBRIUM N
g 550 - WITHOUT SCRAM Ns
/\\
g SCRAM ON
,/ - -
- f 545 HIGH POWER
._s g
O a
x LEAR POWER -
(
540 l-THERMAL POWER I
I I
I 535
!00 105 110 ll5 120 125 -
PERCENT OF FULL POWER (540 MWT)
RESPONSE TO 3-INCH R0D WITHDRAWAL WITH MAXIMUM R0D GROUP INCREMENTAL WORTH f
E A
h01:lh c/1/62 Cold Water Accidents The safeguards against cold water accidents are described on pages LOl:5 - LOl:8. All of these safegnanis are still in effect, a.d the sane arguents and cenelusions apply regardless of reactor power rating.
However, accidents involving eennecticn of a cold, borated loop, and a hot, unborated loop have new been studied quantitatively at the mdm 3-loop operating pcwer 11=it of 378 W. These studies indicate that the resulting transients would be temMted by high pcwer scran, with a
- 4 m DNB ratio always greater than 1.25.
The transient which would be caused by rnd-un boiler feedwater flow at $hO Nt (see page LOl:8) has also been analysed. A mininux Iri3 ratio of 1.3 is calculated for this condition using the W-2 correlation. The W-1 corre-lation would predict a DN3 ratio of 1.62.
Boron Concentration Accidents Boron concentration accidents are discussed in detail en pages
.h0128 - h01:11. Since these accidents are not affected by reactor power rating they have not been re-analyzed.
Bod Drop Accident
/D i )
The effect of a dropped centrol rod on core power distributions has been studied for an initial operating pcwer of $h0 St.
The inxediate result of the negative reactivity insertion would be a rapid decrease in core average flux to some minimum, ubich is calculated as a functicn of the reactivity loss.
Without 'ecrrective action, power would then rise to a new equilibrium value below the initial power, at reduced coolant teeperature.
The potential problem associated with the dropped rod lies in the distorted power distribution which could result frca an asy= metrical red pattern. Such distortion could increase hot channel enthalpy above its initial steady state value and, under certain conditiens, it would reduce the DK3 ratio.
The analysis of this accident was based on experimental results re-lating the effects of rod drop to power distribution. This information was adjusted to the conditions of the accident by assening the increase in Fgi
- to be a linear function of-the reactivity worth of the dropped rod, and -
proportional to the 30% increase observed for a -0.25% reactivity change. 1
-siy41av increase in Fq would result fzrm the asartetrical rod cordguration.
However, it was determined that enthalpy rather than heat flux is Hwiting in this accident.
- The mar *xcm power and te=perature reactivity ecefficients were cen-servatively assumed. Scram was assumed to be triggered by the initial drop in average core power as measured by any single power range channel.
h The results of this a alysis indicate that the -4Mm DN3 ratio
'V will not drop below 1.25 (W-2 correlation), if the scran set point'is 85% of power 'or above. The W-1 correlation would predict a DNER of 1.h9 -for the
- same conditions. Therefore, 'a 1cv power scran will be initiated if a.y single power range detector inMeates a power reduction of 15%.
h01:15 7/19/63
. Reactivity Accidents - Core III Reactivity accidents have been analyced for Core III in the same manner that they were studied for Core II (pages h01:12--h01:lk). However, calculations were performed only for the end of life conditions at 600 MWt 4
and Tavg = 527 0F.
Reanalysis of the $h0 MWt power level was unnecessary because the beginning of life operating conditions at $h0 MWt and $lh F are less severe than those pnsviously studied for Core II.
(see page 102:2h).
Actually, the consequences of the reactivity accidents from 600 MWt and $27 0F are also less severe than mported for Core II due to substantially lower hot channel factors at the end of Core III life.
This is illustrated by the graphs on pages h01:13 and h01:16, which represent the response of the system to withdrawal of the highest worth con-trol rod group with the reactor initially at nav4=um conditions of power and temperature. For both Core II and Core III it can be seen that the reactor would scram on high power before the DNB ratio reached 1.25. However, in Core III the INB ratio would remain above 1.25 even without scram, which is not the case in Core II.
However,. essentially the same restrictions on rod withdrawal will
. be retained through Core III above 90% of rated power:
1.
Elimination of automatic rod withdrawal.
2.
A rod stop has been installed which must be manually reset
-(
after each rod step.
3.
A==vi=n= rod withdrawal is administratively limited to a reactivity insertion of 0.18% in any 8-hour period, (equiv-alent to 3" of rod motion at the highest incremental rod group worth).
c-w e
I f
hoi : 1.i-O H-DNBR 1.25 i
MAXIMUM 575
[ SCRAM 1
EQUILIBRIUM [N g
WITHOUT SCRAM
\\
r
\\
h565 l\\
j\\
M
\\
t; SCRAM ON g
O g560 J u/ "
h o
HIGH POWER O
s(
kNUCLEAR o
POWER y555 1
2 THERMAL POWER 550 I
I I
I I
100 105 l10 115 120 125 PERCENT OF FULL POWER, 600 MWt RESPONSE TO 3-INCH R0D WITHDRAWAL WITH MAXIMUM ROD GROUP INCREMENTAL WORTH O
9
.-y y
9
.,9e-
.g
--g g
4
---,,*9,w-,
,99.-*9 g-9,
,a.---.
m y
.,wy99-.
r--
1 I;O1:17 10/7/65
.Q Reactivity Accidents - Core V Reactivity accidents have been reviewed with respect to the two zirealey test assemblies included in Core 7.
The transient behavior cf the core is not affected by the presence of these test assemblies. As explained on page 103:56, the reactivity coefficients used in previous accident analysis are appropriate for Core V.
Moreover, themal conditions in Core 7 are no r. ore worse than those found in previous studies. (See pages 102:27 and 102:28).
Therefore, -it is not necessary to reanalyze any of the reactivity accidents in detail. In every case, the consequences of-such accidents will be no nore severe than previously reported.
LO nv
402:1 3/31/61 402 MECHANICAL ACCIDENTS (j
IDSS OF COOIANT FIDW ACCIDEhr The loss of coolant flow accidents described here assume complete and simultaneous loss of power to all pumps in the primary system for an extended period of time.
It is highly improbable that such an accident would occur since the four pump motors are divided, two being fed from a transformer connected to the main generator leads and each of the other two from transformers connected to separate, incoming 115 kV lines. Section 226, ELECTRICAL SYSTEM, contains a description of this systec. These may be con-sidered to be essentially independent sources of power. However, for the purposes of these studies, it has been assumed that all cources would be lost simultaneously. Such an accident is very unlikely but could conceivably occur as a result of a hurricane or tornado which caused an electrical failure of long duration between the switchyard and the electric generator.
he analysis of this accident shows that the temperature of the fuel clad might reach a maximum of about 1,430 F.
Since the temperature is considerably below the annealing temperature of the clad and the temperature at which the fuel assemblies were brazed, there is no danger of rupture of cladding. Because the fuel elements may become distorted if the clad tempera-ture exceeds 1,300 F, a new analysis based on actual conditions at the time of the accident, and/or inspection of the core, vill be required before returning to full power if complete loss of flov should occur.
'Ihe physical mechanism of the accident and the related thermal and
(
)
neutron kinetics of the core are as follows:
As the coolant flow decreases, the core outlet coolant temperature increases and the film heat transfer coefficient decreases. 'Ihe increase in average coolant temperature causes a decrease in re-activity which, in turn, puts the reactor on a negative period and results in a reduction in power level. Tending to counter-balance the decrease in reactivity caused by coolant temperature increase is the lesser positive reactivity effect caused by a decrease in fuel temperature.
Uranium dioxide has, compared to metallic fuel materials, a lov thermal conductivity and high heat storage capacity. The specific power of the reactcr and the geometry of the fuel, that is, fuel rod diameter, are such that a high temperature gradient is estab-lished in the fuel during steady state operation at or near full power, Consequently, the average fuel temperature is well above the average coolant temperature. During a flow coastdown tran-
'sient, the large amount of stored energy in the fuel tends to maintain.a high surface heat flux while the burnout heat flux de-creases. The result of these two effects is to raise the clad temperature.
q
402:2 3/31/61
)~
In order to evaluate the temperature excursions following complete loss of coolant flow in the open lattice Yankee reactor, a detailed knowl-edge of transient boiling heat transfer and flow redistribution is required.
A significant effort has been expended on the Research and Development Pro-gram in order to analyze this problem. The evolution of methods of analysis is evident from the examination of the earlier efforts, YAEC-72* and YAEC-83**,
as compared to the analysis presented in YAEC-132+.
YAEC-132 t'orms the basis for the results and conclusions discussed herein.
The method of solution makes use of an iterative procedure employ-i i
ing analog.and digital computer techniques. Briefly, the steps followed were:
l.
Using a predefined flow coastdown curve, (figure on page 402:3)
'as obtained from pump characteristics in the Yankee system, and a given axial power distribution (figure on page 402:4), the normal channel was simulated on the analog computer and the clad to water heat flow for this normal channel was obtained.
2.
Using this heat flow, converted by hot channel factors to yield hot channel heat flow, the CAT ++ code was used to calculate the flow transient along the hot channel. Since there is no provision for the inclusion of the effect of a change in total heat transfer coefficient upon heat flux, this digital solution was.used to obtain parameter inputs to an analog hot channel J.
model.
3 Using the inputs of average coolant mass, average mass velocity, average heat transfer coefficient and average specific heat, an analog model was used to find the hot channel clad to water heat flow for an axially sectionalized core model.
4.
'Ihe analog output was then used as the input to the first digital iteration. 'Ihe solution of this iteration for enthalpy was then compared to the analog solution for convergence. 'Ihe time and position of departure from nucleate boiling were found in the
~
}
digital solution.
j i5 An analog model of the hot spot was used to find the transient temperature.at the point of. initial departure from nucleate boiling.-
- YAEC-79.. A. Bournia." Studies of Thermal Behavior Under Loss of Pump
~
~ Power Transient Conditions," August,~1958.
- YAEC-83, J. M. Gallagher, Jr. and D.~ Hunter, "A Study of Complete Ioss
~
'of Coolant Flow in the' Yankee Reactor," November, :1958.
f+YAEC-132,.'J. P. Cunningham,-J. M._Gallagher, Jr. and T. Gogniat,
'" Transient Boiling-Behavior of the Yankee Reactor Following Loss
~ f Coolant Flow."
(To be issued).
o 7
- ++YAEP-145, A. Bournia, J. P. Cunningham, M. Neuman and L. S. Tong,
' 'J' -
?" CAT -?An IBM-704'. Code for the Solution of Flow Transients in an
- Open Lattice Reactor Core,".1959
~
?
5
4 02:3 9/15/ 59 O
l.O O.8 i O.6 S
u.
\\
d 0.4 g
i
\\
e n
%)
0.2 x N N
00 1
2 3
4 5
6 7
8 9
10 TIME FROM LOSS OF ELECTRICAL POWER, SEC i
e I
NORMALIZED FLOW COASTDOWN CURVE AS A FUNCTION OF TIME AFTER LOSS OF PUMP POWER
.o-Q,J.
s o 3*
- ebR3 09 O
0 8
g}
07 S
E 0 H 6 C N
I E
R O
C 0
5FO P
O O
T M
0O 4R F
E C
N
/
A
/
0 ST 3ID 0
2 0
1 2
0 a
6 4
2 0
8 6,
4 2
I 1
1 1
2 2
zOEhN o 35>r mQm2hN O
>X > F T 2 Oyi25Ci5Z ZO2E>[ Nmo g iI 2m m
><m2 >k >X Dr mO$A. A n Y s^N-O ir O
1 3
402:5 3/31/61 In the solution of problems involving transient boiling, which gov-erns the transient response of the Yankee system to loss of coolant flow, certain assumptions are necessary in order to solve the problems. The major assumptions which vere made are:
1.
The Dittus-Boelter correlation is applicable in forced convec-tion subcooled heat transfer.
2.
The pellet-to-clad gap " film" heat transfer coefficient is con-stantwithavalueof1,200 Btu /hr-sqft-F.
3 The heat flux at departure from nucleate boiling, pDNB' is given I
-2 5 pag 3-032x10 g-0.0024L/D 3
10 f = 150 Btu /hr-sq ft-F.
and h where h
= film heat transfer coefficient from clad to coolant p
H
= enthalpy at point L in the channel L/D
= length.to diameter ratio for the DNB point 4.
The thermal conductivity of UO is 1.0 Btu /hr-ft-F.
2 I
- 5. ~ The moderator coefficient is -2 x 10 5K/F,theDopple coefficient
~
-4.6'x 10-5 J K/F and the pressure coefficient 2 x 10- J K/ psi.
- Ibe previous study of this accident was based on the assumptions
^
-listed. Since the performance of that analysis, new data or re-evaluation of data in several pertinent areas have' indicated that changes'in several assumptions employed-in that study-are in order. Probably the most significant change is in the expected core power distribution. As discussed in Section 103 under " Core Power Distribution," recent calculations, verified by experiments in the Yankee reactor, show that hot channel factors are significantly lower than the desi6n factors. In addition to hot channel factors, estimates of power distribution are available in mora detail than before. Valuesof1.15 Btu /hr-ft-F for..the thermal conductivity. of UO2 and-1,000 Btu /hr-sq ft-F for the gap conductance, have been shown by later work to be realistic assumptions, as discussed on.page 102:11. A recent correlation of film' boiling data has indicated that a film boiling coefficient of 130 (instead of 150) Btu /hr-sq ft-F
'is appropriate to the conditions encountered at DNB (mass velocity = 5 x 105 lb/
hr-sq ft, steam quality = 12%,in the worst case). Finally,.the coefficient-in-theLexpression for DNB flux has bee changed to 0.24 (instead of 0 32) to coincide with new experimental data
+W.-H. McAdams, " Heat Transmission" 2nd Edition, McGraw Hill New York, 1942.
- Technical report 62, Mine Safety Appliance-Research Corporation, 73 J..B. McDonough,..et,al, 1958.
~
4 E
. **ANL 6063, P. A. ~ Lottes, et al,1959
402:6 3/31/61 Because any of these factors might have a significant effect on the results of the loss-of-flow analysis, a new study was performed. Because the
-s( )
most important effects of these changes are manifest in the time and location of departure from nucleate boiling and on the transients following local boiling, rather than on the behavior of power and heat flux before DNB, the assumption was made that the time behavior of core average power, core average heat flux, and hot channel average heat flux is as indicated by the preliminary analysis.
A presentation of the new analysis, replacing steps 4 and 5 of the preliminary study, follows.
Hot Channel Factors - The loss of coolant flow accident was studied forthetwospatig'heatfluxdistributionsdescribedonpage103:27 For thefirstcase,Fq=388andforthesecondcase,FqN = 3 20.
For both, F
= 2.0.
T For both distributions the engineering hot channel factors studied were E
F
= 1.08 q
F E, 1,93 g
F
= 1 37 g
Four Pump Loss - Flow conditions for the case of complete loss of pcVer to all four pump 6 ere analyzed using the CAT code. Initial coolant flow e w rate was 0 9 x 41 x 10 lb/hr. The same flow coastdown curve was used as was used for the preliminary study. Heat flux time behavior was assumed to
^
be the some as that calculated in the preliminary study, with normal channel heat flux decreasing to 80%, and hot channel heat flux to 95%, of the original values in two seconds. Core average coolant temperature was assumed to be 518 F, at 2000 psia.
The result is that for 485 mvt initial power, departure from nucleate boiling first occurs 2.1 seconds after loss of pump power with the flux dis-N = 3.88), and in 1 5 seconds with the flux distribution tributionofgase1(F= 3 20)q For an initial power of 392 mvt, DUB does not occur of Case 2 (Fq within 3 seconds (the end of the calculated flow transient) for Case 1, but does occur in 2 3 seconds for Case 2.
In order to find the maximum clad temperature reached following DUB and scram, temperature transient data which were calculated for a previous study (in conjunction with the BR-3 project) were utilized. These data vere calculated using a 7-radial-region fuel pellet model and a 2-radial-region clad model. A step from nucleate boiling to constant coefficient film boiling, and a power transient determined by a -25% reactivity insertion in 0.6 seconds were assumed. A variety of film and gap heat transfer coefficients were studied. With properly adjusted time and temperature scales these data can be applied to any clad pellet design with roughly equivalent clad thickness-to-6
1:' 2:7 s
3/31/61 diameter ratio, with negligible theoretical error. This principle was tested by application to two loss of flov transients solved in the preliminaq (n) study; the error was 1 F in each case.
In all these calculations, credit was taken for the decrease of power before scram, as calculated in the preliminary loss-of-flow study.
For the maximum clad temperature estimates, the new hot-spot parameters assumed are:
fuel conductivity kf = 1.15 Btu /hr-ft-F gap coefficient hg = 1000 Btu /hr-sq ft-F film boiling coefficient hr = 130 Btu /hr-sq ft-F scre= time ts = 2.8 seconds after loss-of-flow Since DNB occurs first at a moderate-flux point in the hot channel and progresses with time toward the maximum flux point, maximum clad te=peratures were calculated for several points in the channel in each case in order to find the true maximum. The temperature transient was then found for the point at which the true maximum te=perature was reached.
The resulting maximum tempera-tures are here listed:
Initial Power Flux Distribution Maxi =um Temperatures 485 mvt Case 1 1270 F h85 mwt Case 2 1430 F 392 mwt Case 1
_)
392 mvt Case 2 1210 F
(
Cases 1 and 2 represent two selected points in core life when the axial power distribution is expected to be most unfavorable. These cases are discussed in detail on page 103:27 Case 1 occurs early in life when the peak is near the bottoci of the core and the axial peak-to-average ratio is 194.
Case 2 occurs later in. Life when the peak is near the top of the core and the axial peak-to-average ratio is 1.60.
Both cases were assumed'to occur when
-the radial peaking factor was 2.0.
Case 1 for Y pt operation was not calculated be.cause the maximum clad tempers ^,ure for a +t v.1dition was obviously not limiting. The curves on page 402 :8 show the m 1ation of cladding inner surface temperature with time for the point at which the maximum temperature occurs. Both cases are shown for initial power operation at 485 mvt.
Since the clad temperature does not approach the annealing temperature of 1875 F in any case, there is no danger of rupture of cladding.
Thus the loss of coolent flow is not a hazardous accident.
Y
402:8 3/31/61 O
l 1800 INITIAL POWER = 485 MWT 1600 FLUX DISTRIBUTION CASE 2 1400 W:)
D cr 1200 W
FLUX OlSTRIBUTION CASEI j
5 1000 u
800 600 0 2
4 6
8 10 12 TIME AFTER LOSS OF FLOW -SECONDS CLADDING TEMPERATURE TRANSlENT AT POINT OF MAXIMUM TEMPERATURE FOLLOWING FOUR PUMP LOSS OF FLOW ACCIDENT O
h02:9 3/31/61 If, however, the clad te=perature exceeds 1300 F during the loss of flow transient, the fuel elements involved may beccc:e distorted to the extent j
- that subsequent operation at full power would result in clad rupture, and the
~
release of fission products into the primary coolant. For this reason it is of interest to empare critically the results of this study (in particular the peak temperature in the vorst case, 1430 F) vith the 1300 F criterion. In areas in which uncertainties in properties and phenomena exist, it has been necessary to base this analysis on the most pessimistic credible assu=ptions.
'Ibe uncertainties in some cases, notably burnout correlations, UO2 therral conductivity, and transient r hnnges in the clad-to-gap heat transfer coefficient, 0
could each be translated into an error of 50 to 150 in the estimated peak temperature. Confident elimination of these possible errors depends on advances in the state of the arts involved.
In the areas of flux distribution and effective scram delay, however, it has been expedient to study the vorst of each condition in order to define an upper bound of clad temperature ex-cursions. For example, if the flux peak is near the top of the core, as it is in the vorst case, a significant negative reactivity may be inserted by a small rod insertion, very shortly following the one second serem instrumentation delay. A reduction of even one-half second in the assumed 2.8 seconds effective scram delay time is worth about 60 F in the maximu= clad temperature. Again, it is known that the flux distributions studied vill exist for only a short
, time in core life. A reduction of 10% in F and F is worth more than 80 F g
in the maximum clad temperature. WhileamOreextensiveanalysisisnot warranted at this time in view of the improbability of occurrence of the four-pump-loss accident, it is clear that a much more detailed study vould yield lower ' estimates, and for most times in core life significantly lover estimates of clad temperature than the worst case indicated here.
O
..In conclusion, since the loss of flow accident is not a direct hazard, its possibility does not constrain operation to less than h85 mvt.
- If a loss of-flov should occur, there vill be available sufficient information concerning flux-distribution and control rod worth to enable a less conserva-tive calculation of the actual maximum clad temperature, to determine whether the core should be returned to full power or examined for distortion.
Following the several seconds covered in this analysis, natural circulation inithe primary system together with operation of the safety and relief valves in the second system is more then sufficient to provide for dissipation of stored and decay heat. For a long time loss of pump power,
' lasting several hours, damage to the steam generators, which might be caused by' boiling the shell side dry, is prevented by supplying vater to the secondary system from a pump operated by power from an engine-driven generator. Since
.'one of the primary system charging pumps is used for this purpose, vater can Ealso be charged to the primary system in case of need.
Two-Pump loss -- To assess the vorst possible D:a conditions following loss of two pumps, CAT code calculations were made for steady state,' at h85 mvt, i
with 54% of full flow. With the Case 1 flux distribution, the minimum IIB ratio' was 177; and with the Case.2 distribution, the minimum IIB ratio-vas 1.13 1 Thus burnout is not expected during the. two-pump. loss-of-flow transient.
'u 4
+
y y
y
k2:10 3/31/61 LOSS OF LOAD ACCIDENT A primary plant loss of load accident is defined as an almost in-stantaneous load decrease, from any value above that corresponding to an electrical output of 15 mv to zero or station auxiliary load. This loss of load condition may or may not cause an automatic reactor scram. Experimental data taken during operation of the Yankee plant demonstrated the need for a significant revision of the mathematical description of the plant used in previous loss of load accident analyses.
In particular, it became evident that the constant-heat-capacity, variable-load representation of the steam generator secondary side was not adequate to describe the important effects of large feedwater temperature changes and steam volume changes on the over-all plant response. Following a loss of load, the feedvater temperature decreases rapidly, because flow of steam to the feedvater heaters decreases.
At the same time the steam generator water level d ureases, and as a result the level control system requires an increase ir. the mass of water.
The two effects ccubine to cool the steam generator vn:,er well below saturation.
The cool secondary water presents a sink for che primary system, actira as a load on the primary after the secondary loel has stopped.
Thus the magnitude of load reduction felt by the primary system is less than the load reduction on the secondary system. Because the constant heat capacity representation specified a load reduction at the primary system equal to the load reduction at the secondary system the results obtained with that representation were in general pessimistic for the + 10% step load changes and the 3%/ min loading and unloading rates. Theie load variations were used to determine the behavior of the controlled plant under normal operating conditions.
\\p U)
However for the condition of turbine trip followed by reactor scram or vice versa the constant heat capacity description predicted transients considerably more optimistic than those which actually were observed. The only exceptions are those cases where the feedvater pumps were tripped. Aside from this exception the Yankee test results showed an appreciable cooling of the primary system following scram and turbine trip which was not predicted by prior analytical results.
In order to study the problems introduced by steam generator cool-ing an improved steam generator simulation'was developed. Briefly the analysis was accomplished by describing the steam and water phases as separate regions which then allowed subcooled as well as saturated water states. A precalculated equilibrium program for the steam volume below normal water level was used in the model. The secondary side steam and feed-water conditions were obtained from the Stone & Webster heat balances.
The loss of load which causes an autcnatic scram and which most probably is caused by either failure of the electrical tra2smission lines connecting the plant to the interconnected system or by failure of secondary plant equipment will occur infrequently. The automatic tripout of the tur-bine steam valves results in an interlock arrangement functioning (with load about 15 we) to cause an automatic reactor scram.
,m
402:11 3/31/61 i
Results for simultaneous reactor scram and turbine trip from 485 mwt l
are shown on page 402:13 These curves show the transients caused by scram
.V '
_ followed by turbine trip in two seconds and the follcwing three conditions on j
feedwater:
1.
No feedwater trip; the feedwater pumps continue to operate and the steam generator water level is restored to the setpoint value. For i
this condition the reactor inlet temperature is decreased to 466 F in about three minutes after turbine trip.
In addition the pressur-izer was emptied about 70 seconds after reactor scram. Under this condition it is difficult to estimate the steady state increase in reactivity caused by the cooling of the primary system. Just prior to emptying the pressurizer, approximately 3 1/2% increase in S k 1
. resulted frm the decrease in both power generated and reactor core moderator temperature was approximately 11/2% greater than the in-crease attributable to the power coefficient (approximately 2% in 8k for a power coefficient of 4 x 10~5 8 k/mwt). The additional 3
increase in 8 k for a moderator coefficient of 4 x 10-8 k/ F f
would be on the order of 2%, which would result in a net increase in. 8k of 4%. However, when the pressurizer was emptied the primary system pressure vould decrease to approximately 500 psia which would cause a reduction in 8 k of 0.6%. Therefore for the transient in question the net' change in 6k would have been on the order of an addition of 3 4%. The minimum shutdown 8k available at normal moderator temperature is 3%; therefore, with the power coefficient there would be 5% available for scram from full power. If the reactor had been scrammed by a vorth of 5% there would be 1.6% 8k for shutdown after the calculated cooling of the system.
~2.
Manual feedvater trip; the feedvater pumps were tripped one minute I
after reactor scram.
For this case the reactor inlet temperature was decreased to 479 F.
The reactivity addition caused by the power decrease and.cooldown of the. moderator was approximately 3 4%
~,
which-in this case assumed that the primary system pressure did not increase.by a significant amount. ~ Again the net shutdown for a 5%
scram would be approximately 1.6%.
The resultant negative surge was of such a magnitude that the con-clusion as' to whether or not the pressurizer would be emptied is dependent upon the exact initial conditions for the pressurizer
.. volume and_is thus' indeterminate from the study as performed here.
- In order to assure that the pressurizer would not be emptied the feedvater pumps should be tripped well before a mir ate had elapsed.
- 3. L cAutmatic feeuwater trip; the feedwater pumps' vere tripped 5 seconds after reactor scram.: The reactor. inlet-temperature did not de-crease below the full power value-and the pressurizer surge was approximately 30 ft3 The increase.in reactivity following scram
_ vas-about 2 5%; 8k which would result in 2 5% shutdown for a 5%'
1--
. scram.
-Q.:D):
b j
l ep 1
m
- w v
4 rr, -
402:12 9 /15 /5 9 6
10 0 8
,N%L a.
O ge so U
9 2500 7
al
/
E 2000 a.
h 200 2
10 0 w d f
s m
E
/[/
580 I
560
-//L
/
/
/ A/
a u O
OUTLET /
/
F a ; 520
/
$g AVERAGE /
MEACTIVITY COEFFICIENTS IN DOPPL E R - 4.6 X 10 4k /
- F j
w y 500 INLET MODE R ATOR - 2.0 X 10~4 6k/*F
~
+2.0 X l0 *6 k / psi 480 460 E. '
y h
500 c
g 2 XIO' da 2
IX IO' 5-i i
s I
m O
O 10 20 30 40 50 60 70 80 90 10 0 SECONDS LOSS OF 485 MW LOAD-NO REACTOR SCRAM
k?2:13 3/31/61 O
5 2
NO EEO WATER TRIP y
4 U[,
3 N MANUAL FEEO WATER TRIP (60 SEC)
\\
2 - - - - -
AUTOM ATIC FEED WATER TRIP (5 SEC)
S OOPPLER EFFECT a"
1 I
I I
I I
O 40 00 120 16 0 200 240 SECONDS O
AUTOM ATIC FEEO WATER TRIP (5 SEC)
-20 N
g w
N E Wa -40 D ew 3
-60 MANU AL FEEO WATER TRIP (60 SEC) a p)
-80
(
PT NO FEED WATER TRIP
~
l I
l l
1 l
0 40 80 820 160 200 24 0 SECONDS AUTOM ATIC FEEO WATER TRIP (5 SEC)
E 500 j
M ANUAL FEEO WATER TRIP (60 SEC) sg x
e-SI 480 g
E NO FEEO WATER TPIP O
40 80 12 0 160 200 240 SECONOS TR ANSIENT RESPONSE TO TURBINE TRIP WITH.
O SIMULTANEOUS SCRAM AT 485 MWT POWER L)
h02:1h O
7/28/61 In view of the improved conditions resulting if the feedwater flow is terminated in the event of reactor scram, automatic control circuits to accomplish this are provided.
The aforementioned results were obtained for a plant descrip-tion which assumed that all four main coolant pumps continued to operate after turbine trip. In actual practice two pumps are operated from the plant generator and would be de-energized sometime af ter turbine trip. Normally this will occur between one to two minutes after turbine trip and may therefore result in a higher core inlet temperature than shown here. However, the pressurizer volume surge would be approximately the samu since two loops would cool down further than shown in these results.
The loss of load which does rot cause an automatic reactor scram is extremely improbable. If the loss of load occurs by the opening of both main circuit breakers through some operating error, or by some fault of the inter-connected system causing certain circuit breakers to open preventing the plant from feeding the interconnheted system, load will drop to auxiliary load level.
The turbine speed governors are capable of reducing steam flow to auxiliary power level so that the turbine would not trip on overspeed. As a result of the reactor not scramming automatically, the primary plant pressure and temper-ature would rise rapidly as shown in the figure on page h02:12, resulting in pressurizer safety valve dischargo and a condition which requires emergency action. A more extensive analysis of this transient has been performed which
,m
()
indicates slightly more severe results than shown in the figure on page h02:12.
However, these curves have been retained in order to show qualitatively the nature of the transient. The magnitude of the positive surge resulting from the present analysis is given below.
In all cases-the circumstances assumed in this analysis are such as to increase the severity of the transient, thus increasing the conservatism of the result while reducing still further the probability of such a condition.
The basis for the analysis is as follows:
1.
The reactor before the transient is operating at a steady state power level of h85 MW thermal.
. 2.
The incident occurs near the end of core life when the ratio of the moderator temperature coefficient of reactivity to the power coeffi-cient of reactivity is at a mini-m.
The values of reactivity coeffi-cients used.are based on extrapolation of values determinedexperiment-ally at the plant.
. Moderator temperature coefficient
-1.5 x 10-b6k/F Power coefficient
-h.3 x 10-I k/MW t
3 It is assumed:that no reactor scram takes place. An actual load drop from 60 MW electric, performed as a test, indicated that high level in the moisture separators would in all likelihood trip the turbine, n.
thereby scraming the reactor, for. load drops from levels above approximately 100 W electric.
h02:15
(~]
7/28/61 v
h.
It is assumed that the reactor is cn manual temperature control at the time of the transient. Normal operation is on automatic control, which would promptly insert control rods as the temperature began to increase.
5.
No operator action is assumed, although the operator would have ample time available to scram the reactor manually.
6.
It is assumed that the bleed and feed system removes no water from the primary system during the transient. Under actual conditions, as soon as pressuri::er level began to rise, 25 GPM would be automatically re-lieved to the low pressure surge tank.
Under the above conditions the analysis shows that the volume surge into the pressurizer would approach an equilibrium value of 213 cu ft, between 50 and 100 seconds after the loss of load. Since there is nominally 205 cu ft of steam in the pressurizer, this result indicates that the pressurizer would fill with water and that a small amount would be passed through the pressurizer relief valves.
If a further pessimistic r.ssumption is made - that the boiler feed pumps trip out during the transient, the magnitude of the volume surge would be in-creased, thereby increasing the volume of water passed through the relief valves.
Although a boiler feed pump trip is unlikely, it cannot be ruled out entirely a
since either low suction header pressure or a motor overload could initiate
(,)
this action.
As a result of the foregoing analysis, and in order to maintain the design ground rule of no water relief from the pressurizer, a high pressurizer water level scram circuit is provided. Assuming that this scram is set to operate when the surge exceeds 125 cu ft, at this point reactor power will have decreased to 73 per cent of full power through the effect of the moderator coefficient. Scram would follow after a maximum of 2 3 seconds. After scram, the energy stored in the fuel, worth 6.7 seconds at this power, is transferred into the coolant. Thus, after the scram level is reached the additional energy
' to be transferred into the primary system is equivalent to 9 seconds at 73 per cent of full power. This amount of energy would result in a maximum additional surge of 72 cu ft calculated on the basis of no heat transfer at all to the secondary system during this time.
The actual surge would he something less than this depending on whether or not and at what point the boiler feed pumps tripped out.
l Based on the foregoing analysis the pressurizer level scram set point is at a-level not more than 125 cu ft above normal, which provides 80 cu ft above this point.
_,ev-
Lo2:16 3/31/61 LOSS OF CoOIE T ACCIDEhf
(~)'
Mechanism of Blowdown V
General - Any rupture or break in the main coolant system results in a rapid expulsion of high pressure primary coolant water. The safety injection system is provided to guard against the consequences of this situation. In order to properly design this system, and to study the consequences of its nalfunction, a time history of conditions in the main coolant system following a break is required. The basic variable in calculating this history is the size of the break. Three ranges of rupture si: es were chosen to cover all possible loss of coolant accident conditions. The history of system conditions for each of these cases was determined by a step-vise calculation based on pressure intervals. Mass and energy balances were written for the system including the effects of core decay and stored heat. For each pressure inter-val, a trial-and-error calculation gave the time increment for this interval and the system vapor and liquid volumes at the final pressure. These condi-tions become the initial values for the next interval. The basic procedure followed is outlined in detail in WAPD-T-861.* All numerical co=putations were done on IBM-704 and IBM-7090 computers.
Small Breaks - This range includes all breaks up to 0.002 sq ft in area (equivalent to a maximum rupture diameter of 0.6 in.).
The first stage of the discharge is the expulsion of subcooled water until the pressurizer is emptied. For these breaks, this process vill take 120 see or more. The reactor is equipped with a high pressure charging system having a maxi =un capacity of 100 gpm, divided among three pu=ps.
If the rupture is less than N
0.00035 sq ft in area (0.25 in.in diameter), the automatically available (d
charging flow of over 50 gpm from the two operating charging pu=ps is suffi-cient to maintain pressurizer level and pressure. Reactor operation is un-affected. When the break area is between 0.00035 and o.002 sq ft, syste=
overpressure vill be lost und the low pressure signal from the pressuriner vill initiate a reactor scram.
(This assumes no corrective action is taken by the operator.) Daring the subcooled blowdown period, sufficient time is available to activate the third charging pu=p, bringing the charging flow up to 100 gpm. For any rupture in this range, 100 gpm is adequate to main-tain values (2830 cu ft and 756 psi gage). A small break thus presents no serious hazard beyond the ejection of slightly radioactive water and steam into the vapor container. These conclusions are applicable to breaks occur-ring at 485 as well as 392 MRt.
Medium Breaks - This range includes rupture areas from 0.002 to 0.25 sq ft (equivalent to diameters of 0.6 to 6.8 in.).
The discharge from this size break occurs in three fairly distinct stages:
- a. Expulsion of subcooled solid vater until the pressurizer has emptied.
System pressure falls to the saturation pressure corresponding to the average coolant temperature.
- b. Two-phase flow of water-steam mixture until the level of water in the reactor vessel reaches the inlet and outlet nozzles.
- WAPD-T-861, " Analysis of the Coolant E:cpansion Due to a Ioss of Coolant
)'
Accident in a Pressurized Water Nuclear Power Plant", T. A. Harris, Westinghouse Bettis Plant, Dece=ber 1958.
0Na
,3RS 000 1
O S
S D
D Y.
N N T O
E P C
M E
NE S
W S
OI E
D G
WR A
OE T
L Z S
BI R 0
N U
' 0 1
W D$
O ES D
L E OR W
OP O
L C
B B N UE S H D
.W E
E L
T O
O O
N C
B U
S 0
1 E
TE L
P t.
M O
T F
C O
5 S
O K
T 2
A 0
D E
R E
B R
I IU L
L Q
A
- I E
M R
S E
S M
M K
IU I
A T
p D E J
E R MB 2
T F
S E
2 K
G 0
A R
0 E
A 0
R L B I.
1O 1
1 0
0 0
0 0
O gNm o9 :s % aE 7 uoc $ m
,$4 O
OM# O" ooOF#z4 >OObm24 VT 2CU4C3N $NM CEoOOFQ ErOEOOE2 2>dO2 <4 D
D-(
402:18 3/31/61
- c. Flov of steam with small amounts of water entrainment, depen >".f ng on the break size.
Breaks of 0.1, 0.056 and 0.01 sq ft were chosen and analyzed as representative of this group.
O.1 sq ft corresponds to a complete break of a 5 in. Schedule 160 pipe.
0.056 sq ft corresponds to the maximum ductile rupture of a 20 or 24 in. main coolant loop.
0.01 sq ft corresponds to a complete break of a 1-1/2 in. Schedule 160 pipe The results of the calculation for the largest of these breaks, 0.1
. sq ft, occurring at a power level at 485 W t are summarized as follows.
The first stage of discharge is complete in 2 5 see, and it is assumed that the reactor scram has been completed during this time. During the first 60 see of the subsequent two-phase discharge, the heat added by the core is sufficient to produce a greater volume of steam than the volume of mixture being discharged. As a result, the pressure in the system rises slowly until the volume of new steam no longer balances the discharge volume. The pressure falls off from this point until the nozzles are uncovered, 200 see after the rupture. No reliable method is known for evaluating the_ amount of water carried away with the subsequent steam flow. Two cases were investigated,
-)
representing the upper and lower limits for this process. The first case t
assumes a flow of homogeneous fluid similar to the second stage of discharge.
The second case considers only pure steam flow, with no entrained water. The results for the case of homogeneous flov, representing the more adverse blow-
_down conditions, are presented in the curves on page 402:19 and sum =arized
- below.
-The top of the-core.is exposed approximately 220 see after rupture, the hot spot (30% of the core height below the top) is. exposed 13 see later, and the bottom of the core is uncovered 260 see after rupture.
- During this period, the pressure is between 820 psi gage and 690 psi gage.: The safety injection system.is capable of adding water to the reac-
-tor when system pressure falls below 245 psi gage. At most, 335 see elapse
- before pressure falls to this value..As a result, the hot spot may have been
~
. uncovered for as long as'.100 see before safety injection water can be added.
However,.as system pressure decreases, the charging rate of the safety injec-tion pumps increases and core water level is raised. Using pessimistic-
. assumptions, -calculations indicate that the core vill be covered 115 see or M( ;
f
t
-fOF.O.HD N
wNWHN%H O<
E>E nooF$w my*2 ymmy=m(a
- 0 0
0 g
0 0
0 0
o 0
0 8
6 4
2 0
1
/
N9g1 YeT
=
T C E E F
J A N SI gl 0 0
N 3
_s
_\\
_s s_
_ \\_
N
_N S
D A
N O
, \\.
C E
gs_
0 S 0
l g_
2 ER U
g s
P T
O UR N
t s
E f
T F
\\
D A
\\
E S
E
\\
O M
I P
\\
T X
E p\\
S 0+N N
E t
L D
0 pu k
0 l
Z E
AN Z
D R
1 S
\\
0 E
E S
S V
E N
R
\\
O O
L P
C P
\\
E X
N N
S E
U S
\\
E E
Y N
V R
L O
L A
R C
U F
O F
N T
O E
g C
R A
P O
N E
O C
s R
T
~ ~
_~~
0
~ - - - ~ -
0 0
0 0
0 0
0 0
0 0
3 2
1 TDI$4*$4 g,rGO<oEgae2=5 Am nimg C.0C6 <OtCEm p2O T2$mC2m E>g p2t-(
,OF o$ ZO p FO(nM 0M oOOrpZ4 > o OMZ o<y, m$ E TNa9 g Q' f *rioE9 $*EEH r
h02:20 0
7/2e/61 less after safety injection flow begins (h50 see after the rupture). Mel+-
down calculations for this break size show that at this time, the tenperature of the cladding at the hot spot is well belew the melting point.
Prcper operation of the safety injection syste= will, therefore, prevent any melting of cladding material.
Conditions following the 0.056 and 0.01 sq ft breaks (page h02:17) are correspondingly less severe than these described for the 0.1 sq ft break.
Wese results are s~una*1 zed belew:
Break size, sq ft 0.056 0.01 Time after mpture, see Subcooled blowdcwn ec=plete h.5 25.5 Nostles uncovered 3hD lLEO Top of core uncovered 37h Hot spot uncovered 392 Bottom of core uncovered h33 Hot spot recovered 6SO Iarge Breaks - This range includes breaks in areas up to 2.86 sq ft, which represents a complete rupture of ore of the 20 in, rain coolant pipes, with both brcken ends discharging fluid. This size break is included to p
present the most pessimistic assu=ption possible. A :7 break involving a pipe d
of this size is exceedingly unlikely, but if a break were to occur, the ductile longitudinal rupture ending with a bmak area of 0.056 square feet is cen-sidered to be much more realistic, since the hoop stress in a pipe is always much greater than the nrini stress. The 0.056 sq ft break is analyzed under
" Medium breaks" and is mich less severe.
The subcooled blowdevn is virtually instantaneous for the 2.66 sq ft break. Because of the high flow rates, the entin centents of the pri:mry system are assumed to flash and form a hemogenecus froth of water and steam.
This adxture ficws thrcugh the breaks until the reacter vessel is emptied.
Calculations abow that automatic and comple+4 scra= cf the control rods will occur despite the flow of water and steam upward through the reactor core.
A passure differential greater that 77 psi thrtngh the core is req 2 ired to exceed the gravitational force o-the centrol rods. This pressure drop is not approached even under thin extreme conditien.
The top of the core is exposed apprMmately 13.h see after rupture, the hot spot 0.6 see later, and the bottc= of the co e 15.6 see after the rupture. In 23 seconds the main coolant system is essentially empty. System pressure reaches the shutoff head of the safety injection pumps in about lh seconds; however, safety injection startup requires 20 see af+4r the initiating signal is actuated. To ensure that safety injection is available at the earliest possible time, the initiating pressure for start of valve opening is set at 800 psig or above and the initiating presm:re for pump starting at not less than 270 psig.
fm 1
Womogenecus flow model no longer valid. Steam flew model does not result in core uncovering.
402:21 3/31/61 In analyzing this accident, the flow reaching the core was determined
(
by subtracting the portion escaping through the rupture fr" Se total flow delivered by the safety injection system. Under these cond.
.ns, the hot spot is recovered 90 see after the rupture. 'Ihe main coolant pressure and liquid volume history is shown on page 402:22.
Hot spot cladding temperatures were calculated for this accident, and the results, plotted an page 402:23, show that melting of the cladding vill not occur. 'Ihe following assumptions were made in this analysis.
N 1.
Fq is 3 88 for steady state conditions and 3 23 for decay heat calculations.
2.
Steady state thennal conditions exist in the fuel rod for 15 seconds following rupture.
3 Prior to uncovering the core, the convective heat transfer coefficientforthewaterfilmis500 Btu /hr-sqft-Fandthe conductance of the gas gap is 1000 Btu /hr-sq ft-F.
4.
Followinguncoveringofthecore,thegapconductanceis250 Btu /
hr-sq ft-F and there is no heat removal from the cladding surface.
5 After recovering the bottom of the core, there is convective heat transfer to steam with a coefficient of 13 Btt/hr-sq ft-F.
As shown by the curve, page 402:23, steam formation caused by restoration of water to a level above the bottom of the core is effective in limiting the maximum hot spot cladding temperature to about 1980 F.
The number of fuel rods reaching this temperature is, of course, very small in proportion to the total; probably of the order of 2 per cent. At this temperature some local yielding of the cladding might occur due.to internal gas pressure. Although the possibility exists of moderate release of fission product gases through defects caused by such a transient, the probability and magnitude of the release are felt to be small. No gross cladding failures are expected at these temperatures.
-Criticality of the Core During Blowdown General - Reactivity increases during a blowdown transient because of the decrease in main coolant temperature. Several factors tend to counter-balance this reactivity increase; namely, void production, uncovering of the core, system pressure decrease, control rod insertion, and boron injection.
Using the data developed in Section 103, CORE NUCIEAR DESIGN, for temperature (page 103:4), void (Page 103:21) and pressure effects on reactivity, kerf was evaluated as a function of time for all sizes of breaks. All cases have been analyzed for the beginning of life conditions where the control requirenents are most stringent.
Small Breaks _ -
2 (a) up to.00035 ft Since the charging system can maintain pressurizer level and pressure,' criticality and power operation vill be maintained for.
this size break. Movement of control rods to balance slight pressure i
- changes vill be necessary but will not present an operational problem.
".oO rT w WtmH G)
Ee2 n8c&4 N$a v$$$, ' a 0
0 0
0 0
o 0
0 0
0 w
0 2
0 4
2 Oo 3
l 8
s
~
~
i
~
I 04 i
E P
I P
- 0 G 2
N I
F T
0
~
OTA 2
WR E
K MP A
O E5 R8 N
B4 O
S T
D D
I D
E O N R C E EE R
I O O DN WJ E
i C EON V
E PI O
S C
ELY L A T T
E
~
BI E O
R UTF P
U OI A S
0 T N
DI S F
l 8 P T*
\\I OO U
H O R
" G O
[
E2 R
R=
E
~
T Og F
CT S
A N
I 0
G 6 E E
M B
TI N
Eb R i O
T C
C E
E R
F J
0 O
O N
C I
4 M
Y F
O T
O T E E
P T F R
O OA U
T B
S SS I
E
'\\
0 R
g\\g 2
1 g
E P
I M
\\g\\
I U
g LO V
\\
s O
0 0
0 0
0 0
0 0
0 0
3 2
1
~ ogo $c g,,.z z nOOQN E3E mHJ t
EZ hOrDz4 $y m CoC O orc m DZO gwmuUCg s)
,4 7OFrOg.zO IO Q Oh;Z4 >oo O zn 1
3, ar>2 N mxmpX n$ m~
~/Q
402:23 3/31/61 O
2600 CLADDING MELTING TEMPERATURE,7 --
2400 1
I 2200 l
BOTTOM OF HOT SPOT
,k.
CORE RE-COVERED RE-COVERED o 2000 m
8 x
w 1800 N
a y 1600 DOUBLE ENDED BREAK OF 3
20 INCH PIPE E
INITIAL POWER -485 MWT (O
SAFETY INJECTION V;
g 1400 OP ER ATING w
F 5 1200 d
~
l000 800 BOTTOM OF CORE UNCOVERED I
I I
I I
I I
I I
I I
I I
600O 20 40 60 80 10 0 12 0 14 0 TIME AFTER RUPTURE-SECONDS CLADDING TEMPER ATURE AT HOT SPOT FOLLOWING A LOSS OF COOL ANT ACCIDENT i LARGE BRE AK CASE )
<'y.,',
402:23A 3/31/61 2
(b)
.00035
.002 ft Since the charging system is capable of maintaining saturation conditions corresponding to 514 F, the only resetivity changes vill be the addition of negative reactivity by voids, control rod insertion and pressure decrease. Hence, even at the beginning i
of life the core vill be brought at least to its hot zero power keff value of 0 955 on rods alone. credit for pressure decrease, 7
voids and any fission product buildup would lower keff even further, depending upon the extent of burnup.
Medium Breaks -
2 (a) 0.1 ft he heff of the core vill drop to a value of 0 995 as pressure de-creases. Reactor scram (completed by 2 5 see) initiated as the pressure drops vill render the core at least 5% suberitical after the scram until 100 see of the transient have elapsed. The decrease in reactivity associated with the density decrease after 100 see is more than sufficient to balance the increase in reactivity caused by temperature decrease. In fact, keff vill be less than 0 9 when safety injection occurs. The injection of the borated solution, coupled with the rod scram vill maintain the core at least 5% sub-critical even if the system is cooled to 70 F.
Any credit for burnup and fission product buildup would substantially decrease keff; i.e.,
at any point during core life the shutdown is greater than 5%.
2 (b) ;0.01 ft The reactivity effects are similar to those described above except with a slightly different time scale. During the 2 5 see before scram is completed, reactivity will d*op to a value slightly less than 1.00.
During the subcooled blowdown (24 sec), the reactor will be held 5% subcritical by rods. As the system pressure builds up following the subcooled blovdown, reactivity vill decrease as T avf increases, the system again being held at least 5% suberitical.
B using feed for the charging pumps from the chemical shutdown system, the core reactivity vill be decreased even more than this 5%. As the
. system pressure and t 7perature decrease after 200 sec, the presence
' of_ the borated solution plus the control rods give adequate protection against a possible return to criticality. Since the exact hydrody-
'namics'of the accident are not known, i.e., void formation in the core, it is' impossible to estimate the exact value of keff throughout the transient. However, one can say with certainty that with the use of the borated feed plus a reactor scram that the core vill be held
. greater'than 5% shutdown even-if the core. vere cooled to 70 F.
Large' Breaks - Reactivity drops initially as pressure drops and continues to drop as voids form. Even without reactor scram the core vill be 25% sub-critical
- after 5 sec. ' Reactor scram would be sufficient. to reduce heff at least I:
another 5%. credit for fission products vould tend to make the system even more suberitical.-
g'() '
3
- *The density of the ' mixture in the core has dropped to 25 lb/ft,
1 r
.-~ -.
kO2:23B 3/31/61 Core Meltdown f}
General - Partial or complete melting of the reactor core following rupture and the loss of all water from the main coolant system would result in release of fission products into the vapor container. To prevent such an occurrence, an automatically initiated safety injection system with ample storage of borated water, a highly reliable means of introducing this borated water into the reactor vessel, and a dependable power supply, are provided.
The safety injection system pumps have been sized such than even for the vorst possible break conditions full core coverage vill be attained before possible core damage could occur and thus, meltdown problems are guarded against.
Nevertheless, a core meltdown event, in which it is assumed that the safety injection system does not operate, has been carried to its conclusion to determine the rate of melting, fission production release and possible criticality of the pellets in the bottom of the vessel.
Mechanism of Core Meltdown - At 392 mvt, the maximum normal operating fuel temperature at the center of the hottest pellet is 4,330 F vhile the corresponding maximum cladding temperature is 663 F.*
The steep gradient in temperature between the center of the pellet and the fuel cladding surface is the result of a high heat flux which prevails and the lov thermal conductivity of sintered UO. A significant decrease in the rate of heat removal from the 2
surface of the cladding vill cause the te=perature gradient from the pellet to clad to decrease and the cladding temperature to approach the fuel temperature.
The melting point of type 348 stainless steel is 2,550 F, while that of p
uranium dioxide is approximately 5,000 F.
Therefore, stainless steel cladding y,/
vill melt before the fuel melta if heat generation continues within the fuel t
while the rate of heat removal from the cladding surface decreases, as is the case if the fuel cladding tube is surrounded by steam. Assuredly, as soon as any point on a fuel rod reaches the melting point of 2,550 F, the cladding vill rupture, allowing some of the gaseous and volatile fission products to escape.
During the transient following loss of coolant, the high temperature of the fuel assemblies as they are heated to melting coupled with the nonabsorbing medium present will lead to radiant heat losses to the control rods. Conservative calculations of the radiation losses indicate that the Ag-In-Cd control rods vill melt due to the radiant heat transfer.
This is due in part to the fact that the Ag-In-Cd melting point is 1,500 F campared to the relative high melting point of the stainless steel structure and of the UO2 #"*1*
Criticality Considerations Following Melting of the Core - Calcula-tions have been made to determine whether a criticality problem exists when the molten clad, control rods and the pellets fall to the bottom of the vessel.
- The hot channel factors which allow operation at 485 mv vould produce approximately the same tenperatures at the hottest pellet.
p$
\\%/
I
LO2:23C 3/31/61 Assuming the safety injection system does not function following the large break and complete core meltdown occurs, the density of the water-steam mixture vill be so low in the bottom of the vessel that a criticality problem cannot exist. As shown in the figure on page 402:18, the primary system is essentially empty by 15 see; thus, by the time the first clad melts at 155 sec, the system would be essentially void of moderator.
Calculation of this sytem indicates that the kggf would be less than 0 5 A more serious problem might exist if following complete meltdown, the safety injection system then functions. This vould cause water containing 950 Ppn boron to be pumped into the " melted mass" at the bottom of the vessel.
An analysis of this system was carried out with the folleving assumptions:
(1) Ccxnplete core meltdown as noted above.
(2) Moderation by water (containing 950 pga boron) at a density corresponding to saturated liquid at one atmosphere.
(3) spherical geometry which leads to minimum leakage.
(4) Equilibrium Samarium, but no Xe and no other fission p
products.
J (5) Random packing of chipped cylindrical pellets which gives an experimental water-to-UO2 ratio of 0.66.
On the basis of these conservative assumptions, keff was found to be 0 91. If credit were taken for fission product buildup corresponding to the end of life, keff would be 0.85 Credit for the Doppler addition, con-centration increase of the boron solution due to boiling of the water and void content would further reduce keff*
Fission Product Activity at the End of Core Cycle - At the end of the core cycle, a variety of fission products is present in the fuel material.
An analysis of the gross gamma activities of the gaseous and volatile fission products has been made, since fission products of this type can be released if core melting is not prevented. It is assumed that the core has been oper-ated at full power of M5 mv* for an infinite length of time. This is a
- This power level is used rather than the core design level of 392 mv, since this is the more serious condition and since this condition vill exist in the future when reactor power is increased.
%)
!+02: 21+
3/31/61 e
O INTEGR ATED HEAT RELE ASE, EQUlVALENT SECONDS AT A GIVEN POWER N
b N
O O
9 o
o=E I
I I
i i
E e
2 i
d
'9 U
hb
~ ;
E 8!
O 2 vs N
tr U
Wg
.L O v;e DECAY HEAT POW E R, PERCENT OF INITI AL VALUE TOTAL DECAY HEAT GENERATION (BETA, GAMMA, U-238 CAPTURE)
AFTER 500 HR OPERATION AT A GIVEN POWER
402:25 9/15/59 conservative assumption which yields somewhat higher activities than those
-)
which might actually be present.
v The noble gases and the halogens are considered to constitute the gaseous and volatile fission products. Altogether,12 isotopes of bromine, krypton, iodine and xenon are adjudged to posses significant gamma activity to be considered. Five elements are volatile at, or below, 450 C in combi-nation with oxygen and/or one of the halogens. These elements are arsenic, molybdenum, antimony, tin, and tellurium. Of these elements, the isotopes of arsenic and tin have a low yield in the fission process. Fifteen isotopes of the other three elements have been examined and significant activities have been included in the totals. Gross fission product gamma activities are given in the table below at three different times after reactor shutdown from 485 mv.
Fission Product Gamra Activity Following Shutdown (In units of 1018 mev/sec)
Time after shutdown 0
5 min 1 hr Gases 9 36 5.28 3.26 Elements volatile as compounds
.99 89
_,54 Total 10.35 6.17 3.80 Fission product gamma activity is plotted graphically as a function of time after shutdown in the figure on page 402:26. The lower curve scale of 1019 mev/sec gives the gamma activity of all the fission products follow-(~/
18 mev/seeincludesonlythegaseous
\\
ing shutdown; the upper curve scale of 10 and volatile fission products and may be compared with the data presented in the table above.
Fission Product Release to Vapor Container - The fission product gamma activity which is present in the core after infinite time of operation at 485 mv is shown in the figure on page 402:26. Following an accident in which no use is made of the safoty injection system and complete core melting results, the activity attributable to gaseous and volatile fission products could be released into the vapor container. If the safety injection system is used but is not fully effective, partial melting occurs and the activities released to the vapor container would be some fraction of the values shown.
Chemical Accident Even in the event of core meltdown, no significant release of chem-ical energy due to the reaction of water with the stainless steel fuel cladding is expected. While a reaction of this type is thermodynamically possible, as evidence by Aerojet Generals 4 tests with a molten stainless steel spray at 2,270 C, an explosive reaction is highly improbable, if not impossible. Two
+IDO-28000, H. M. Higgins, and R. D. Schultz, "The Reaction of Metals with Water and Oxidizing Gaseous at High Temperature", April 1957.
.._l L
1
402:26 9/15/59 GASEOUS & VOLATILE FISSION PRODUCT ACTIV!TY, mov/sec.
~o xn N
o m
e n
e a
w n
n l
t i
i 9
t l
aI I
W e d d
l 7
E Eo i
8>
o
_ _ g ',
o a
s 5
S I
us oMr gw o-3 F
/
7 X
-o e
Xn N
o e
e n
e e
g n
N TOT AL FISSION PRODUCT ACTIVITY, mov/sec.
l l
l l
FISSION PRODUCT GAMM A ACTIVITY AFTER INFINITE
- O TIME OF OPERATION AT 485 mw. POWER l
l l
l
402:27 9/15/59 types of pressure build-up ard energy release are possible for a stainless
(
steel vater reaction; one, a relatively slow increase in pressure and gradual t
energy release such as that encountered in the production of fine mesh steel powder by the water granulatien process' and the other in which pressure builds up rapidly and energy release takes the form of an explosion as in the Aerojet General tests. Since meltirg of the clad will not produce a finely divided spray, but rather a molten stream, the reaction vill correspond more nearly to the process in which no violent chenical reactien is evidenced.
Although zirconium is present in the core in the form of control rod followers, no hazards are contemplated due to its presence. Since the zirconium vill be in massive form rather than the fine particles necessary to maintain the reaction and vill not be in direct contact with the fuel rods where most heat generation occurs during an excursion, no chain of events is envisioned that would catalyse the latent zirconium-water reaction.
Vmigr, Contain=ent The vapor container is designed to retain all vapors, gases, liquids and solid materials released as a result of a loss of coolant accident. The maximum loss of coolant accident employed in the vapor container design con-sists oft Complete severance of one 20 in. main coolant line, with two open pipe ends.
..O-_
Simultaneous rupture of one secondary main steam line inside the vapor container. The placement of each main coolant loop in a separate concrete shielded compartment and the installa-tion of a nenreturn valve in the main steam line from each steam generator limit this part of the accident to the nip-turing of a single secondary main steam line.
Detachment of an object or metal fragment from the pressurized system in such a way that it acquires kinetic energy, which, unless restrained or stopped by a barrier, might perforate the steel shell of the containment vessel, thus releasing contaminated vapor following the loss of water accident.
The figure on page 402:29 shows the initial pressure transient fol-lovig the release of.178,600 lb of fluid from the main coolant system and one secondary coolant circuit into the net volume of the vapor container of 860,000 cu ft. The maximum differential pressure between the concrete con-partment and the vapor container is 6 psi,- and this pressure is reached in 0.2.sec.
A port area of 400 sq ft in any one loop shield co. partment is required to' limit the pressure differential across the concrete valls to this
- In the production of fine mesh stainless steel powder by the water
- granulation method, molten stainless is poured in'a stream about
.l'l/2 in. in diameter. This stream of stainless is hit with a high -
velocity water. stream. ' The resultant particles, varying from p_
400 mesh powder to 11/2 in. diam globales, are so slightly oxidized Q
that most of the powder is sold without further proce.1 sing. No ener-
.getic chenical reaction between the steel and water has been observed.
402:28 1/10/60 value. To provide this vent area and to prevent the maximum differential pressure from exceeding 6 pai, openings are located at the bottens of the partition valls between the compartments, openings are located over the stee,m generators and openings vill be available when some of the removable slabs lift at about 3 psi gage differential pressure. The concrete battered valls adjacent to the equipment access opening are designed for a maximum differential pressure of 6 psi. All coolant is released from the main cool-ant system within approximately 18 see and equilibrium is attained inside the vapor container at a final pressure " called" 34 5 psi gase, or 49 2 psia.
The corresynding vapor temperature is 249 F and the energy released is 92.2 x 100 Btu.
The figure on page 402: 30 shows the long-time effect after the re-lease of vapor and initial pressure rise to 34 5 psi gage. During the first 2 hr, there is a marked decrease in pressure due to thermal radiation and convection from the uninsulated vapor container shell and due to the diffu-sion of heat into the inner concrete structure. Subsequently, there is a gradual decrease in pressure with a small secondary rise, peaking in 4 hr at 15 psi gage, due to the continued release of decay heat from the reactor
-core.
The air-vapor mixture pressure within the vapor container after the maximum loss of coolant accident is based on the assumption that the total internal energy of the fluid remains the same before and after the rupture.
':'his is based on the conservation of energy relation:
4 = AW + E Q = Net heat release, Btu Where A = Reciprocal of mechanical equivalent of heat W = Mechanical work performed, ft-lb M = Change of internal energy, Btu During the brief interval after the initial burst, it is assumed that there is not heat loss, or Q = 0.
There is no work done, since the fluid begins
. and ends in a state of rest, or AW = 0.
Therefore, the internal energy be-fore the accident-is the same as that after the accident, or M = 0.
The summary of the. principal data for the major loss of water accident is as follows:
Main coolant pressure, nominal upper operating limit, psi gage 2,150 Average temperature main coolant, upper
._ operating limit, F 518 Total volume of water in main coolant
- system, cu ft Reactor 1,460 Pressurizer 75
~
Steam generators 832 Piping 404
' A)-
Pumps (4;at 21 each)'
84
(
Gate and check valves 49 Total 2,904
402:29 9/IS/59 O
i j
i CALCULATED MAXIMUM PRESSURE 34.5 PSI GAGE IN VAPOR CONTAINER 35 1
35
(
1 30 3o g
I.
I 25 25 g
g E
I o
e E
I E
I^
g 20 20 g If 5
5 e
e f
' 15 15 *
/
PREDICTED RISE IN VAPOR CONTAINER to 10 PRESSURE RISE IN STEAM GENERATOR CC"DARTMENT FOLLOWING A RUPTURE IN Y. E MAIN COOLANT LOOP BASED ON AN OUTFLOW AREA OF 400 SQ FT I
/,#
5 j,
5
/
MAXIMUM PRESSURE
/
DIFFERENTI AL BETWEEN
./
COMPARTMENT AND VAPOR v
CONTAINER APPROX 6 PSI n
e i
g 0.01 0.05 0.1 0.5 1
5 10 50 10 0 EL APSED TIME AFTER RUPTURE, SEC PRESSURE RISE IN VAPOR CONTAINER vs ELAPSED TIME AFTER RUPTURE C'1 V
REY. 7 15-57
402:30 9/:5/59 40 40 j
i PEAK REACHED APPROX.18 SECONDS AFTER RUPTURE (34.5 PSI GAGE,249 F) 35 35 30 30 25 25 g
g E
E o
o g 20 20 g 5
5 E
E E
- 15 15 '
([
\\
~_
10 10 5
5 0
2 4
6 8
to 12 14 16 18 20 22 24 26 28 30 ELAPSED TIME, HOURS PRESSURE IN VAPOR CONTAINER FOLLOWING A NUCLEAR ACCIDENT NO INSULATION ON VAPOR CONTAINER SHELL AU REV 7-IS S7 J
402:31 9/15/59 Total volume of steam in main coolant system,
/
cu ft Pressurizer (t = 645 F P = 2,150 psia) 251 Total volume of water in one secondary loop, cu ft at P = 770 psia t = 512 F 641 Total volume of steam in one secondary loop, eu ft at P = 770 psia t = 512 F 448 Gross volume of vapor container, eu ft 1,020,000 Net effective volume of vapor container, cu ft 860,000 Weight of fluids in main coolant system and one secondary circuit, lb 178,600 Internal energy of released fluids, Btu 92,205,000 Final pressure, psia Vapor 28.75 Air 19.8 Total, psia 48.55 Total, psi gage 33.85
.(m)
Final tem:perature, F.
248 For design purposes, the final pressure is called 34.5 psi gage and the temperature 249 F.
Subsequent to the initial pressure release, the following heat transfer effects proceed simultaneously, with the net integrated effect of these on the vapor container pressure shown in figure on page 402:30. These effects are as follows:
Decay heat is released from the reactor core in accordance with the following relation:
f=,.076 G*
o Where P = rate of heat release after time G, av
~
3 = initial rate of heat release, 485 mv
.P
=
time after reactor shutdown, see The rate of release of decay heat is dependent upon the number of hours which the core has operated; the longer the operating
. period, the greater the rate of release of. decay heat.
jr3 t>
J
=.
d 402:32 9/15/59 This relationship is based on an infinite time of operation, f"s which corresponds substantially to the rate of heat release after many hours of operation, and is thus conservative.
Heat is lost be radiation and convection through the spherical shell. This rate of heat release depends on the ambient tem-perature within the vapor container, the outside ambient tem-perature and a radiation and convection coefficient which available data indicate for large spheres is 2.2 Btu per sq ft, per hr, per degree F.
The outside ambient temperature is taken as 70 F, the average of a summer day. The initial am-bient temperature within the vapor container prior to the accident is 120 F.
Heat is absorbed by the vapor container metal. This rate of absorption is proportional to the ambient temperature within the vapor container. The weight of the containment vessel is approximately 2,500,000 lb and the specific heat is 0.12 Btu per lb, per degree F.
Heat is slowly released from metal parts which have been operat-ing at normal temperature. This rate of heat release is pro-portional to the ambient temperature within the vapor container.
The insulated metal parts weigh approximately 1,500,000 lb, and have a specific heat of 0.12 Btu per lb per degree F.
Insulation is provided by 4 in. of calcium silicate with an average thermal conductivity of 0 55 Btu per sq ft, per degree F, per hr, per in.
hm of thickness.
Heat is absorbed by the internal concrete structures. The rate of diffusion of heat into the concrete with time is dependent on the ambient temperature within the vapor container. The temperature-time relationship for the concrete was determined by use of the Schmidt method, using a specific heat of 0.22 Btu
.per lb, per degree F, a thermal conductivity of 0.5 Btu per sq ft, per degree F,-per hr, per ft of thickness, and a density of 150_1b per eu ft.
.The ambient temperature is an independent variable in all of these
. factors, except decay heat, contributing to the redistribution of heat.
This permits the determination of the new vapor temperature and total pres-sure of.the air-vapor mixture after any elapsed time. Only the first two and last items in this list contribute importantly to the redistribution of heat. The results of the calculation are shown in figure on page 402:30.
The conservative assumption has been made that there is no condensation dur-
-ing the first few minutes after the initial rupture to reduce the calculated initial pressure..
~
Missile Protection L
Although it is believed that no plausible missile could be released
. by the main coolant < system, protection-is provided by the inner concrete
~n-structure. 'It'is considered that the ductile austenitic materials of con-rM
- struction.of the main coolant system piping vill-not fail in a brittle manner
~
p m
e
402s33 9/15/59 when in contact with the hot compressed coolant. It is likewise con:sidered that the stainless clad, carbon steel reactor vessel, fabricated and tested acconling to the best techniques and under the proper codes and in contact with the hot compressed fluid, vill not fail, and is not a feasible source of missiles.
9' The internal reinforced concrete structure serves '.s a secondary biological shield, as structural support for equipment and, in addition, pro-vides a misaile barrier. For biological radiation dose considerations, the valls and bottom of this structure consist cf 4 5 to 6 ft of reinforced con-crete, and the top or upper floor level consists of a minimum of 3 ft.
Since the control rod drive mechanisms are located in the cavity above the reactor vessel and are not covered by the reinforced concrete floor, a special missile shield is provided just below the floor grade, fabricated from 1 and 2 in, thick steel plate backed'up by 18 in. of concrete.
O
- (/
h02:3h 6/1/62 Ioss of Coolant Flow Accident with Core II at $h0 Wt The loss of coolant flow accident was re-evaluated for loss of one, two, and four pumps while operating with four loops at 5h0 Wt, and loss of three pumps while at the maximum 3-loop operating power of 378 Wt. In each case, complete and simultaneous loss of power t,o the affected pumps was assumed.
The worst measured power distributions (pages 103:39 and 103:h0) were used in the analysis. The consequences of each accident were evaluated in terms of both the W-1 and the W-2 INB correlations. In calculating fuel temperature, a W-2 DNB ratio of 1.25 was made equivalent to DNB. Actually, DiBR=1.25 corres-ponds to a 95% confidence that DNB will not occur. The results of the study are sumarized in the following tables DNB Case Correlation Maximum Cladding Temperature Two pump loss at 5h0 Wt W-1 1200 F W-2 1150 F*
Four pump loss at 5hD Nt W-1 1hh0 F W-2 1330 F"
Three pump loss at 378 E t W-2 1120 F One pump loss at 5hD Wt W-1 No DNB W-2 DNBR >1.25 O
- 1
- a o 175 er rea ith o"sa<125
- less than 0.17% of rods within 100 0F of maximum The analysis shows that in no case would cladding temperatures exceed the annealing point, where the likelihood of collapse is. significant.
The highly improbable loss of all four pumps could, under the most adverse conditions of power distribution, result in permanent distortion of a small number of fuel rods. Should this accident occur, power distribution data applicable at the time of the accident would be used to ascertain whether an examination of the core would be required before returning to power.
In the case of a two-pump loss, the transient is less severe than in the four-pump loss, and even under the most unfavorable circumstances no pemanent deformation would occur. Similarly, no damage would result from the loss of a single pump, or from complete loss of flow from a condition of three-pump operation. In all cases, the accidents analyzed reflect the most conservative assumptions relative to power distribution, instrument errors and control band limits. The individual cases are discussed in somewhat more detail in the following paragraphs.
Two Pump Ioss The flow coastdown curve for a two pump loss of flow with scram is 6
shown en page h02:36. The initial flow rate is 36.7 x 10 lb/hr. The acci-dent is limited by high coolant enthalpy in the W-2 correlation. Since DNB
'[Vl first occurs at a moderate f]ux point in the hot channel and progresses with time towarti the maximum flux point, the clad temperature was calculated for several points in the channel in each case in order to find the true mari=n=.
-=.
h02s35 6/1/62 With the W-2 correlation, 3IB first occurs in 1.3 seconds following O
loss of flow, and the maximum clad temperature reached is 1150 F.
Page h02:37 shows the variation in average clad temperature with time for the point at which the maximum temperature occurs. A calculation was made of the percent af rods actually seeing these conditions by statistically combining the nuclear snd engineering hot channel factors. This indicates that only 0.17% of the rods will have hot channel factors greater than 93% of the maximum values.
These rods will range in temperature from 700 F - 1150 F.
For hot channel factors less than 93% of the maximum, DNER would never be less than 1.25 following a two pump loss of flow accident with scram.
With the W-1 correlation and maximum hot channel factors, DNB first 0
occurs in 1.9 seconds and the maximum clad temperature is 1200 F, Four Pump Loss The flow coastdown curve for the four pump loss is also shown on page h02:36 With the W-2 correlation, the analysis shows that DNBR=1.25
. first occurs in 0.7 seconds and the maximum clad temperature reached is 1330 F.
' The variation in clad temperature with time at the point at which the maximum occurs is shown on page h02:37. An analysis made with the W-1 correlation 0
predicts DNB in 1.3 seconds and a maximum clad temperature of 1hh0 F.
Three Pump loss An analysis was made for a three pump loss of flow accident from O
three loop operation at a power level of 378 MWt. The maximum cladding O
temperature following DNB was found to be 1120 F.
The accident is therefore less severe than the two pump loss of flow at 5h0 MWt.
One Pump Ioss A transient analysis was performed for the loss of a single pump while operating at a power level of $h0 MWt. The analysis showed that DNB would be prevented by an automatic low flow scram for both W-1 and W-2 cor-relations.
The case of one pump loss without scram was approximated by a steady state calculation for three pump operation at 5hD MWt using==vi==
hot channel factors. The W-2 correlation predicts a DNB ratio of 1.23 for this case, whereas the W-1 correlation gave a mini== INB ratio of 1.8.
~.
J ho2:36 O
6/1/62 1.0 1
- 0. 8 LOSS OF TWO PUMPS
)
y
' O. 6 3::
S u.
S
$ 0. 4 O
LOSS OF F0uR euMeS E
8
- 0. 2 I
I I
I O
0 1
2 3
4 5
6 TIME FROM LOSS OF ELECTRICAL POWER - SEC NORMAllZED FLOW C0ASTDOWN AS A FUNCTION OF TIME AFTER LOSS OF PUMP POWER
" O 1
.., -.., ~..,. -.,.., --,.
., _ _ -......... ~..
l h02:37 6/1/62 O
1600 IU LOSS OF FOUR PUMPS u.
i No 3
1200 u.
b s
8 1000 O
2 LOSS OF TWO euMeS l
E l
E FOR WORST POWER l
D DISTRIBUTION CASE:
800 3
l PEAK IN UPPER CORE i
I l
g 0
2 4
6 8
TIME AFTER LOSS OF FLOW-SECONDS l
l MAXIMUM CLADDING TEMPERATURE FOLLOWING TWO-AND FOUR-PUMP LOSS OF COOLANT FLOW ACCIDENTS AT 540 MWT
!O i
i l
l Y
l h02:38 6/1/62 Loss of load Accident with Core II at 5h0 MWt The loss of load accident is discussed in detail on pages h02:10 to h02:15. The conditions and precautions outlined in these pages also apply at
$h0 MWt. In addition, the high pressurizer water level scram has been re-established at 200" to provide an additional margin for expansion. No opera-ting problems arise with this set point as the water level remains very close to 120" under normal conditions.
Loss of load with automatic reactor scram is not significantly dif-femnt at 5h0 Wt than at h85 MWt. Therefore, this accident has not been reanalysed.
Without reactor scram, pressure would reach the relief valve setting of 2h85 pai and steam would be vented from the pressurizer. The liquid level would rise to 200r where the reactor would be automatically scrammed. At the 3
time of scram, the steam volume in the pressurizer would be 130 ft. The post-scram water surge would be less than 110 ft), leaving a margin of at least 20 ft3 before liquid discharge from the pressurizer would occur.
(v3
.q
(.s
h02:38A 7/19/63 Loss of Coolant Flow - Core III p
\\ )
~
The lost, of flow a:cident was analyzed for Cors III at 600 K4t and Tavg = $27 F.
The power distribution used is shovn on page 103:$2. It is not necessary to analyza the $ho K4t case beenuse hot channel conditions are less severe than previously reported for Core II.
The method of analysis is the same as that used on Com II (page h02:3h) except that the maximum, three-loop operating power was taken as h50 K4t, in keeping with a four-loop power of 600 K4t. The results are as follows:
Maximum Cladding Temperature One pump loss
- 600 K4t DNBR >1.25 Two pump loss - 600 K4t 1120 F
Three pump loss - h50 K4t 1180 0F Four pump loss - 600 K4t lh80 0F These results are essentially the same as reported for Com II on 0
page h02:3h. In no case would cladding temperatures ext:eed 1600 F, where the likelihood of collapse is significant. The highly improbable loss of all four pumps could, under the most adverse conditions of power distribu-tion, result in permanent distortion of a small number of fuel rods. Should this accident occur, power distribution data applicable at the time of the I
accident would be used to ascertain whether an examination of the core would
~
be required before returning to power.
In the case of a two-pump loss, the transient is less severe than in the four-pt.mp loss, and even under the most unfavorable circumstances no permanent deformation would occur. Similarly, no damage would result from the loss of a single pump, or from complete loss of flow from a condi-tion of three-pump operation. In all cases, the accidents analyzed reflect the most conservative assumptions relative to power distribution, instrument errors and control band limits.
~
Loss of Load - Core III The loss of load accident is discussed in detail on pages h02:10 to h02:15 and also on page h02:38. The conditions and precautions outlined on these pages also apply to Core III.
Loss of load with the associated reactor scram is not significantly affected by power level. Therefore, this accident has not been re-analyzed.
Without automatic scram, pressure would reach the relief valve setting of 2h85 pai and steam would be vented from the pressurizer. The At the time of scram, the steam volume in the pressurizer would be 130 ftg.
liquid level would rise to 200 in.where the reactor would then be scramme 3
The post-scram water surge would be 63 ft, leaving a margin of 67 ft3 before liquid discharge from the pressurizer would occur.
h02:39 7/19/63 Loss of Coolant Accident at 600 W t An analysis of the loss of coolant accident for the case of h85 W t operation with T 51h0F was presented on pages h02:16-32. A com-parable analysis is pr$sented here for the case of 600 !"dt operation with av T
527 F.
ayg The general description of the mechanism of main coolant system rupture and coolant blowdown on pages h02:16-18, as well as the reactivity analysis on pages h02:21-h02:23c, are not affected by the change of initial power level and are not repeated here in detail.
Calculation of the gatem pressure-voluma history following various postulated breaks was perfomed with the digital computer code IDCO*.
This method makes use of step-by-step mass and energy balance equations taking into account the changes in stored heat and residual core power, as well as the delivery characteristics of the safety injection system. The method is the same as that,employad in the previous analysis at h85 W t.
However, two new assumptions were made in accordance with plant changes which affect the capability for injecting water into the main cool-ant system. It was assumed that two of the three charging pumps will be available for service in the event of a loss of coolant accident, and will be actuated by a low pressurizer level signal.
(The effect of charging flow was neglected in the previous loss of coolant study.) Further, a third safety injection pump has been installed in parallel with the two existing pumps.
Q,_,
The shutoff head of the new pump is approximately 770 psi, and its maximum flow is 1800 gpm. This pump will initiate safety injection while main cool-ant pressure is substantially above the shutoff head of the two original pumps. Other assumptions regarding the blowdown are the same as in the pre-vious analysis.
Favi== clad and control rod temperatures were calculated for each loss-of-coolant accident case which results in exposure of the core hot spot to a steam atmosphePe. The liquid level curve obtained from the IDC0 code was used as input for the transient heat transfer calculations. The method for determining clad temperature was the same as that used in previous studies
- and is described in YAEC-188**. An improved decay heat relationship, giving a slightly higher heat input for decay times less that 1000 seconds, was used in the analysis of this part of the accident.
WAPD-T-661 " Analysis of the Coolant Expansion Due to a loss of Coolant Accident in a Pressurized Water Nuclear Power Plant" by T. A. Harris, December 1958.
- ~ YAEco188
" Semi-Annual Progress Report for the Period July 1 to December 31, 1960",~ Februar/- 15, 1961.
~p J
h02th0 7/19/63
(
The loss of coolant calculations were performed for the following b) cases at the 600 MWt power levels 2
1.
1.h2 ft break
-- complete severance of the 20 inch main coolant pipe, with unhindered flow from both ends.
2.
0.1 ft2 break
-- complete ceverance of the largest connect-ing line, the $ inch bypass line, with flow from one end.
3.
0.056 ft2 break
-- representing a ductile rupture of the main coolant pipe, h.
0.0375 ft2 break
-- complete severance of a 3 inch safety injection line.
- 5. 0.01 ft2 break
-- complete severance of a 1-1/2 inch connect-ing line.
Additional breaks of intermediate and smaller size were also checked to insure that the most severe case is actually the 20 inch pipe break.
Severance of 20-inch Main Coolant Pipe Although this accident is not considered credible, it is analyzed to represent the most pessimistic assumptions possible. Actually, failure in a large pipe, fabricated of a ductile material, characteristically occurs as
(' 1 a longitudinal split, originating at a point of metal defect. This type of 1?
rupture is represented by the.056 ft2 break discussed on page h02th3.
Results for the case of the 20-inch break are summarized as follows:
2 Break size (each pipe end), ft 1.h2 Top of core uncovered, sec af ter rupture 10.8 Hot spot uncovered, see after rupture 11.3 Bottom of core uncovered, see after rupture 12.$
Bottom of core re-covered, see af ter rupture 81.0 0
Maximum hot spot clad temperature F
1920 Maximum control rod temperature, gF 1030 Curves of main coolant system pressure and liquid volume versus time are shown on page h02thl. The ma*= clad temperature transient is shown on page h02:h2.
This accident is characterized by a very rapid depressurization (negligible period of sub-cooled blowdown), and homogeneous discharge of a steam-water mixture which persists until the reactor coolant system is empty.
To prevent gross core damage, the safety injection signals are set' high enough to assure full speed operation of all safety injection pumps as soon as system pressure drops below the respective pup shutoff heads.
A conservative estimate of the internal gas pressure in the fuel rods at the end of core life indicates that no rods will undergo stresses in excess of yield strength at the temperatures attained in this accident. Thus no release of fission gas is anticipated.
]
O
!';.0.;
l 1
SAFETY INJECTION OPERATING
,g 2500 1000 Di n.
W
- t PRESSURE 2m f2000 800y o
5 o
[cyl500 VOLUME g{
a O
g 8
8 400g v1000 z
E TOP OF CORE y
- E 8o
_ ___.lToM 0F10RE_ _ _ _ _ _ _ -
200 Sm =____
BOTTOM RECOVERS AT 81 SECOND I
I I
I i
0 0
0 10 20 30 40 50 60 TIME AFTER RUPTURE - SECONDS PRESSURE AND VOLUME TRANSIENTS FOLLOWING 20-INCH PlPE BREAK AT 600 MWt CONDITIONS v
a:...,
w~
CLADDING MELTING TEMPERATURE 2200 2l00 O
BOTTOM 0F CORE g
REC 0VERED w
1800 E
1 e
l I
Is!
1600 I
a<
l O
e l
s r
i E
I.42 FT BREAK I
l g
1200 INITIAL POWER-600 MWt 5
SAFETY INJECTION l
U OPERATING l
1000 i
l BOTTOM 0F CORE UNCOVERED l
800 I
I I
I 600 I
I I
I I
I II 0
10, 20 30 40 50 60 70 80 90 TIME AFTER RUPTURE (SEC.)
CLADDING TEMPERATURE AT HOT SPOT FOLLOWING A LOSS OF COOLANT ACCIDENT q
2 1.42 FT BREAK kJ i
h02 h3 7/19/63 e.
The W== control rod temperature would remain well below the
(
melting point of Ag-In-Cd (1500 F), hence no reactivity gain due to loss of absorbing material would occur. With hafnium control rods, no melting problem could exist because the melting point of this material is higher than that of the stainless steel fuel cladding.
Severance of 5-inch Bypass Line Significant data for the 5-inch line rupture are:
2 Break size, ft 0.1 Top of core uncovered, see after rupture 192 Hot spot uncovered, see after rupture 200 Bottom of core uncovered, see after rupture 226 Bottom of core re-covered, see after rupture 3h8 Maximum hot spot clad temperature, OF 1819 MaH== hot spot control rod temperature, F 1190 Curves of system pressure and volume and maximum clad temperature are shown on pages h02:hh and h02:h5, respectively. This accident differs from the 20-inch double ended break in that the main coolant system contents are retained at a higher pressure during the early stages of blowdown. Thus the homogeneous discharge phase is prolonged, and the systea pressure does not fall below safety injection delivery pressure until 230 seconds after the rupture. Prior to this time the only water addition is by means of the two charging pumps.
2 Comparison of the pressure-time curve for the 0.1 ft break at 600 MWt with the corresponding curve for h85 MWt (pagt h02:19) shows that the pressure, and therefore the temperature during the saturated blowdown phase, differs but slightly in the two cases until after an effective safety injection flow rate is attai~ned.
The madem fuel temperature reached is slightly lower than for the 20-inch pipe break. However, the control rods reach a somewhat higher
-temperature..
2 2
2 0.056 ft, o,o37g fg, o,oy fg Breaks 2
The three ramining cases, the 0.056 ft ducplerupture,the 2
. 0.0375 ft. safety injection line break and the 0.01 ft "small line" break, are analyzed together, since their pressure-volume histories are similar.
The-sub-cooled blowdown phase is longer, and the pressure peak at then beginning of the saturated blowdown phase is higher, for smaller break sizes.. In the accidents anal;7 zed, the mode of discharge changes from homo-geneous flow to steam flow with less than 1% enstrainment when the water level falls below the vessel nozzles. This change results in a more rapid pressure decrease and a lower rate of volume decrease. In each case, safety injection prevents.the hot spot from being uncovered, averting any significant
.( }
temperature increase in'either the clad or control rod material.
U'
p h02:Lh V
7/19/63 SAFETY INJECTION OPERATING PRESSURE 2500 -
T 1000 G
,n C
si a
M tM y2000
~
20 y VOLUME O-O 5
o E5 G
S1500 600 $
n
!;E
!iE v
5 5
8 8
" 1000 400 "a zE TOP OF CORE
?#E
- E 500 __ __B0_T_T_0M_OF C_0 RE_ _ _ _
200 1
0 100 200 300 400 TIME AFTER RUPTURE, SECONDS PRESSURE AND VOLUME TRANSIENTS FOLLOWING 5-INCH PIPE BREAK AT 600 MWt CONDITIONS 4
i
O
,)
?:
2500 CLADDING MELTING TEMPERATURE 2
.I FT BREAK 2300 INITIAL POWER-600 MWT SAFETY INJECTION OPERATING 2l00 m
o E5 o-BOTTOM 0F CORE 1900 RECOVERED m
E 1700 l
l O
e I
E 1500 I
f l
g 1300 l
E I
3 I
g l100 i
I 900 l
I HOT SPOT l
700 -I UNCOVERED I
I I
i 500 I
I I
I I
I I
I I 200 220 240 260 280 300 320 340 360 TIME AFTER RUPTURE (SEC.)
CLADDING TEMPERATURE AT HOT SPOT FOLLOWING A LOSS OF COOLANT ACCIDENT 2
.I FT BREAK
h02:h6 7/19/63
~
Significant data for the two cases are as follows:
2
~
Break size, ft 0.0$6 0.0375 0.010 Top of core uncovered, sec 363
$31 Top of core re-covered, sec 379
$86
- Top of core remains covered.
2 The pressure - volume histories for the 0.056 ft and 0.0375 ft breaks, are shown on pages h02:h7 and h02:h8, respectively.
Vapor Containment The design pressure of the vapor container was established as that resulting from the 20-inch pipe break, with unhindered flow from both ends, accompanied by simultaneous rupture of one main steam line inside the vapor container. The calculations performed to verify the container design and to define the pressure transient for a power level of h85 W t are described in detail on pages h02:27 - h02:33. These calculations have also been performed for 600 N t with T
= $27oF and the results are described herein.
avg The Mmm pressure surge occurs immediately after the rupture and is a function of the stored energy of the systems discharging to the containment. While the volume of the main coolant system is the same for the 600 Wt condition as for the previously licensed power level of h85 Wt, the average temperature and stored heat in the system are higher. There-t fore, the pressure transient in the vaper container was completely re-analyzed. It was found that the additional stored heat has no significant effect on the pressure peak. In fact, the initial surge of 32.2 psi is lower than the design value of 3h.$ psi. This is due to the fact that the original calculations were performed before precise numbers were available on the water volumes in certain components. It was the:efore necessary to make conservative assumptions for these values. The calculations at 600 W t are based on actual water volumes in the components as installed.
'"he pressure-time characteristics of the vapor-air mixture in the container during the period following the initial peak were also examined because of the effect of increased power level on the decay heat generation.
As previously noted, there are three significant heat transfer processes which largely deteImine these pressure-time characteristics: the evolution of decay heat by the core, the absorption of heat by internal concrete structures, and the loss of heat by radiation and convection through the spherical shell. The relationship among these factors is' evident from the curves on page h02:h9. The rapid drop in pressure during the first 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> following the accident is due to the predeninant effect of heat absorption by internal concrete structures. After the cuter layers of concrete essen-tially reach thermal equilibrium with the container atmosphere, the pressure increases again to a secondary peak as the sphere approaches a temperature sufficient to dissipate heat at a rate equal to the decay heat generation.
Thereafter, tha pressure decreases as the decay heat generation rate dimi-nishes. Further redistribution of heat between the concrete structure and 4
the container atmosphere occuza slowly during the latter stages of the transient, and tends to damp any subsequent pressure variations.
e O
9 3000 l200 E
\\
E E
\\
m m
n 4
N o 8 y 2500
\\
f,o 1000,
$55 >
\\
5 Msg Q i
N 9,
E 054 4
\\
8 E > a $ 2m0
\\
8m @
962 E 2$C 5 w
A Q@ E
\\
o gg g 1500 g
g 600 2
s$8 E N
2 r z EE M
\\
Q u+m s
m 5o>5 5 1000 REACTOR VESSEL N0ZZLES EXPOSED
'y
^
7'~
400 E
WG m
\\
~
aRg 5 TOP OF CORE EXPOSED o
s j
5r" m
$.$ A 3
g 500 CORE FULLY UNCOVERED 200
!;, 8 I
I I
I I
I I
0 O
1 0
100 200 300 400 500 600 700 TIME AFTER RUPTURE, SECONDS
4 0
9
~
3000 l200 5::>
\\
E Eo f8$2500 +/
1000 2::
,,o 909 %
E M-5 4 fn n
5 E
\\
- g 8
,a > 4 i
E E @ m 20m m0 s
m
=
2 5
3 $ r- ?
m
" o 25 55
\\
d
-mc a
~
58o < 1500
\\
600 2
E 5$
\\
s M
\\
REACTOR VESSEL N0ZZLES
/
h o
5gj ji0m EXPOSED
$ rn m o
TOP 05 CORE N
u 3:: 5g n
e 3 $0 c
500 BOTTOM 0F CORE 200 g
5g I
I E
0 0
100 200 300 400 500 TIME AFTER RUPTURE, SECONDS
PEAK REACHED APPROX.18 SEC.
35 AFTER RUPTURE
)
30 5
? 25 u
h DECAY PRESSURE FOR e
600MWT CONDITIONS a.
g 20 0
$8 DECAY PRESSURE FOR
!15 485 MWf CONDITIONS s
v s\\
'~
5 I
0 0
4 8
12 16 20 TIME AFTER RUPTURE - HR PRESSURE IN VAPOR CONTAINER FOLLOWING 9
HYPOTHETICAL ACCIDENT (N0 INSULATION ON VAPOR CONTAINER SHELU
h02:$0 7/19/63
(',
The pressure curves for both h85 Mdt and 600 Mdt are shown on page h02:h9. The effect of increased decay heat is seen as an upward displacement of the secondary peak, which does not approach the limitirg value for design established by the initial peak. It can be concluded that the vessel design pressure of 3h.$ psig is more than adequate for a plant rating of 600 Mdt.
Based on the results of the pressure transient calculation, the assumed drivirg force for long-term post accident leakage from the vapor container was increased from 15 psi (h85 Mdt) to 21 psi (600 Mdt).
h02:51 10/7/65
^ hanical Accidents - Core V c
The only mechanical accidents that can conceivably be affected by the pre:ence of the two zircaloy test assemblies in Core V are Loss of Flow and Losa of Coolant. These accidents have been reanalyzed and the results are reported as follows.
Loss of Coolant Flow The consequences of a loss of flow accident have previously been evaluated with respect to the basic, stainless steel fuel design, and the results of these ctudies are reported on pages h02:3h - h02:38A. Because thermal conditions in C;ra V are less severe than in previous cores, it is not necessary to reanalyze the loss of flow accident for the stainless fuel.
However, the effect of loss of flow on the two zircaloy assemblies has been evaluated. For a one-pump or two-pump loss at 600 Et, DNB will not occur in the zirealoy fuel and clad temperature will not rise above the normal operating 0
1; val of 6h20F. For a four-pump loss, the clad would reach 1350 F.
These temperatures are much lower than reported for the stainless steel fuel on page h02:38A.
No zircaloy clad damage would occur in either the one-pump or two-pump loss cf flow. A few rods could collapse if a four-pump loss occurred. However, it ould be noted that simultaneous loss of four main coolant pumps is virtually credible, because these pumps are fed from three independent power supplies.
Loss of Coolant The loss of coolant accident has also been evaluated for the two zircaloy as;emblies. For medium and small pipe breaks (5 5"I.D.), the temperatures reached by the zirealey are much lower than reported for the stainless steel furl on pages h02:h0 - h02:h6. The naximm temperature reached by the zircaloy aft:r at 5" break is 15350F, compared to 1819 F in the stainlcss steel.
However, following a very large break, a zirconium-water reaction could occur in the test assemblies. For a double-ended, 20" pipe break, the hottest zircaloy rod would reach 1800 F approximately 20 seconds after the accident.
Bared on the parabolic rate law (ANL-65h8), a zirconium-water reaction would proceed at a rapid rate and temperatures in the cladding could exceed the molting point of zirconium (3300 F) before the reaction is terminated by safety 0
injection in about 110 seconds.
Once the cladding has reached the melting point, its behavior is difficult to predict. Part or all of the molten zircaloy could drop into the water below
- and the reaction could be quenched. On the other hand, the ZrOp formed by the rea* tion could contain most of the molten zircaloy, permitting the reaction to proceed.
In this analysis it is simply assumed that 100% of the zirconium reacts.
'O is done on3y to define an upper limit for energy addition to the primary
. tem and to the containment, and is not intended to represent the physical onditions which actually exist in the core.
h02:52 10/7/65 O
Thp energy liberated by 100% reaction of the two test assemblies is 5.1 x 107 BN. Recombinatjonofallthehydrogengenergtedbythereaction could add another h.3 x 10' BTU. Thg total of 9.h x 10' BTU is orders of magnitude smaller than the 92.2 x 10 BM released to the vapor container by the primary coolant (see page h02:31).
If the Zr-H O and hydrogen energy 2
were added instantaneously to that from the coolant, tne pressure increment would be less than 0.5 psig. This is insignificant to the design of the vapor container,(because the maximum container pressure is well below the design pressure see page h02 h6).
It is therefore concluded that there is no safety problem associated with the theoretical possibility of a zircaloy-water reaction in the two test assemblies.
(3 x) n
~j'
403:1 9/15 /5 9 403 HAZARDS FROM REACTOR ACCIDENTS Maximum Credible Accident In foregoing sections, analyses have been made of a number of accidents, the most serious of which is the large loss-of-water accident.
Such an accident might occur through rupture or severance of a 20 in main
- coolant line, resulting in depressurization and virtually complete loss'of water from the primary system. The danger, of course, is the possibility of' core meltdown an1 release of fission products to the vapor container, thereby causing a radiation hazard to the public. To prevent such an occurrence, a safety injection system is provided. It is believed that this -
system and its controls assure sufficient cooling water in time to protect against core meltdown and can not in any credible way be disabled or rendered inoperative by the primary effects of the accident. In the opinion of Yankee Atomic Electric Company and of the reactor designers, the maximum credible accident can be defined as this large loss-of-water accident, with core melting prevented by the safety injection system, no appreciable release of fission products from the core, and therefore no hazard to the public.
Hypothetical-Accident The unique danger from a nuclear reactor installation is the accidental release of fission products from the plant and the creation, thereby, of external radiation hazards. The present state of reactor technology demands that all reasonable measures be taken to guard against W
- even the most unlikely event by incorporating effective safety features in
' ();
the plant design. Accordingly, even though the maximum credible accident does not. result.in the release of fission products and does not cause ex-ternal radiation hazards, a hypothetical accident in which such a release
-does occur has been postulated and analyzed in order to evaluate the effectiveness of containment and other safety features which are incor-
~
porated in the plant design.
The hypothetical accident is based on the following assumed condi-
'tions and. sequence of events:
'A 20 in. pipe severance occurs:in the primary system, resulting fin depressurization and virtually complete loss of water from the primary system.
Blowdown of. the. primary system results in an initial pressure rise'in the vapor container to 33 85 psi gage, decreasing to approximately 15 psi gage after 2 hr.
' It:is ' assumed that,- for unexplained reasons, the safety in-jection system does not-function and that partial core meltdown occurs.-
~~
No criticality of the fue1 pellets released by clad melting occurs.
i U
y
.m
.2.
h03:2
- V]
7/28/61
(
~
20 per cent of the gaseous and volatile fission products present in the core after 10,000 hr of reactor operation at h85 mwt,are released to the vapor container and dispersed homogeneously therein.
Gaseous and volatile fission products are released instant-aneously from the fuel into the vapor container, although the release actually would occur after a finite time interval which would begin several minutes af ter the start of the accident.
The vapor container has a leak rate of 70 cu ft per hr (STP) with the vapor container internal pressure at 15 psi above atmosphere, and this leak rate and driving head are assumed to continue indefinitely.
These assumptions are believed to be conservative for the following reasons:
Complete severance of a 20 in. pipe is, as far as is known, an unheard of event in the operation of steam-electric power stations.
The safety injection system would normally be actuated auto-matically by the drop in primary system pressure. Failing this, it' would be actuated, within seconds, by the operator acting on the numerous alarms and signals indicating the situation.
20% release of the gaseous and volatile fission products to the vapor container is an arbitrarily conservative number. chosen solely because little or no information is available on such a release from uranium dioxide fuel. Experiments
- have been per-
-formed on the actual meltdown of metal fuel which showed a maH=im iodine release of 25.8%, most of which remained in the meltdown chamber, which would correspond to the reactor vessel.
Only about 2% of the available iodine escaped into what would correspond to the vapor container. There would be no actual meltinE of the uranium dioxide fuel. Only the stainless steel clad is subject to melting under the conditions involved. This would further act to reduce the actual release of radioactivity.
The actual' leak rate from the vapor container at 15 psi gage has been experimentally determined to be less than 21 cu ft per hr(STP).
The driving head would actually continue to drop off with time.
- NP-7071, Technical Report 63, Contract N0bs-65h26, Index No. NS-200-021 " Fission Product Release During a
. Simulated Meltdown of a PWR Type Core"
{}
October 20,.1958. S.J.Rodgers, G.E. Kennedy.
"n
,m n
403:3 9/15 /5 9 Direct Radiation On-site V
In the hypothetical accident, fission products are released into the vapor container, and c radiation source exists at the site. Calculated radiation levals at a point external to the vapor container are based only on that activity v1ich is not attenuated by the internal secondary shield.
To obtain radiation levels, no credit is taken for any delay time due to slow melting of the core before release of the fission products begins. The figures on pages 403:4 and 403:5 show gamma dose rates and integrated dose l
~ as a function of distance from the sphere. The control room, which has additional shielding, provides a place where essential plant operating per-
.sonnel can gather for protection from radiation. The total direct dose re-ceived in the control room under these conditions is less than 240 mr in the i
first hour and less than 2.4 r in the first 24 hr following the release. All plant personnel 'other than essential personnel vill proceed to an assembly 4
point 2,500 ft from the plant on the sounding of a clearly audible signal.
Since a period of several minutes is available before a significant portion of the core melts, plant personnel can take stations or evacuate the plant without receiving harmful direct radiation doses.
Direct Radiation Off-site 4
Dose rate due to direct radiation emanating from the vapor container in the case of a large primary system rupture followed by partial core melt-down is shown on page 403:4 as a function of distance from the source. Inte-
' grated doses for 5 min and for 1 hr after the instantaneous release as a
-A function of distance are shown on page 403:5 V
The dose at the public road across from Sherman Pond,1,300 ft from the vapor container, is 6 r during the first hour after release. Hence, several hours are available to remove persons and vehicles that might be on
.the road at the time to a safe distance from the site. This is based on the once-in-a-lifetime direct radiation dose limit of 25 r indicated in National Bureau of Standards Handbook 59 Because the power plant is located at the bottom of a deep narrow valley, direct radiation from the vapor container does not reach inhabited buildings, except for one dwelling approximately 4,000 ft_ distant in the down-river direction.- Since the integrated dose for the first hour is 24 mr, the occupants of this house could remain there for several weeks without-serious exposure. -In the up-river direction, there are no buildings within 9,000 ft, and, beyond that point, all buildings are shielded by hills.
Except for the dwelling mentioned above, all buildings within one
' mile,of the site are owned.in fee by Yankee Atomic Electric Company or New England _ Power Company and are considered to be under administrative control 1 of these two : companies..
4
/%
h aa.; -
t-Y
403:4 sO/2/59 O
l 1
l l
GAMMA DOSE R ATE DUE TO RELEASE OF 20% OF THE VOL ATILE FISSION PRODUCTS INTO VAPOR CONTAINER 3
10 125 FT DI AM SPHERICAL VAPO R CONTAINER 485 mv FULL POWER CORE 2
80 k
s
/4 w l0 y
or 8
(
i h
k 2
O
=
o 4
10*'
/
f t
+
.N 10 2 0
1000 2000 3000 4000 5000 6000 OlSTANCE FROM SPHERE, FT GAMMA DOSE RATE FOR OUTSIDE EXPOSURE O
.=
1
403:5 10/2 /59 O
10 4 i
g i
i g
i INTEGRATED GAMMA DOSE DUE TO RELEASE OF 20% OF THE VOLATILE FISSION PRODUCTS INTO VAPOR CONTAINE R l
10 3 125 F T DIAM SPHERIC ALVAPOR CONTAINER 485 mv FULL POWER CORE 2
, 10 ui 8 25 O
4 10 3
s
- p O
o e
g i
P e
O
'e o
+4 10 -8 e,
P9 10-2 f
i O
1000 2000 3000 4000 5000 6000 l
DISTANCE FROM SPHERE,FT l
l INTEGRATED GAMMA DOSE FOR OUTSIDE EXPOSURE i
o
403:6 1/10/ 00 Vapor Container Leakage and Air-borne Radiation (d
In the hypothetical accident, 20% of the gaseous and volatile fission products are assumed to be ho:noSeneously aispersed in the vapor con-tainer. Leakage from the vapor container at the assumed leak rate vill release these fission products to the atmosphere and, under certain meteor-ological conditions, they can be carried to populated areas where they may be inhaled or ingested.
Of the volatile and nonvolatile fission products in the core, radio-iodine and radio-strontium provide the important activities with respect to the inhalation dose, with iodine being selectively absorbed by the thyroid and strontium by the bone. For the purpose of this report, it has been conserva-tively assumed that 20% of all the iodine and strontium have been released frcxn the core. However, the release of strontium unde-- laboratory conditions has been reported more nearly as 1 to 5%. In addition, the distribution of iodine betvcen the liquid and vapor phases of water which would be present in the vapor container as the result of a major loss of coolant accident might reduce the amount of iodine available for leaka6e as a gas by a factor of several hundred.
The total activity of iodine and strontium assu=ed to be present is as follows:
Activity - Curies Vapor Container Concentration
- Isotope h85 mv Core Microcuries/ Milliliter mb I-131 1.2 x 107 2
0 99 x 10 I-132 1.8 x 107 15 x 102 I-133 2.6 x 107
'2.1 x 102 I-134 31x107 2.6 x 102 I-135 2.4 x 107 2.0 x 102 Sr-89 2.0 x 107 1.6 x 102 Sr-90 6 7 x 105 0.055 x 102
- 20% of core activity divided by net volume of vapor container (o60,000 cu ft or 2.43 x 1010 ml)
Kuper and Cowan have determined the effects of an accident where all volatile fission products were released as well as 1% of the contained strontium (1). Under these conditions, significant exposure results only from direct radiation and iodine isotopes. The exposure caused by strontium was found to constitute a negligible risk. Following this reasoning, the strontium hazard is not further evaluated here.
The biological hazard of the several iodine isotopes may be evaluated in terms of I-131 by consideration of the product of their effective energy delivered to the thyroid tissue and their effective half life in the body. By (1) Exposure criteria for estimating the consequences of a catastrophe in a nuclear plant. J.B.H. Kuper and F.P. Cowan, presented at Geneva
[_
Conference, September 1958. Conference paper number 430.
403: 7 1/10/60 this method, the equivalent concentration of all radio-iodine in the vapor container was estimated to 170 microcuries per milliliter as I-131.
The concentration of air-borne activity at points off the plant site is a function of the assumed vapor container concentration, the assu=ed vapor container leak rate of 70 cu ft per hr, and the dilution obtained by meteorological processes.
The poorest dilution condztions occur when valley winds are light and in a down-valley direction, since the width of the vall.ey restricts lateral diffusion and the down-slope vinds from the sides of the valley tend to concentrate air-borne materials within the valley. A temperature inver-
,sion may also exist at the same time.
Smoke tests have shown that, under these meteorological conditions, vapors released from the vicinity of the vapor container can be expected to flow in a layer on the ground approximately 300 ft deep at the nearest in-habited area 4,000 ft away. At this point, the width of the layer is about
'1,000 ft.
Using this layer cross-section and an average vind speed of 4 fps gives a down-valley air. stream available for dilution of 4.2 x 109 cu ft per hr, neglecting down-slope vinds. This dilution air, combined with the assumed
. radio-iodine leakage from the vapor container, results in an estimated con-centration of 2.8 x 10-6p/ml, as I-131, at the nearest inhabited area.
In ord6r to estimate the off-site concentrations of radio-iodine at the nearest inhabited area under moderate lapse or unstable atmospheric O
conditions, the equations of Sutton and Cramer may be used with the diffusion V
parameters presented in Section 301, METEOROLOGY. The concentrations were calculated for " Center of Cloud" position on the assumption that, in the worst case, the cloud-could descend to ground level. Wind speeds were selected on the basis of records tabulated in Section 301.
. The resultant air-borne concentrations for each of the above con-ditions are'shown as follows:
Wind Iodine - 131 Meteorological
- Velocity, Concentration. ac/ml Condition fos Sutton Cramer Austin
. Light down-valley 4
2.8 x 10-6 vinds(inversion)
Moderate lapse 13 1.2 x 10-6 1.2 x 10-6 Unstable 13 0 3 x 10-6 1 5 x 10-8 Several authorities have attempted to establish exposure tolerances for catastrophic occurrences. These may be set on the basis of microcuries of activity received-by a critical organ or the rem of absorbed dose-in such
-organs. Here again, the limiting condition appears to be either direct radiation or_ radio-iodine concentrated in the thyroid. A value of 25 rem appears acceptable as mentioned on page 403:3 under " Direct Radiation Off-site".
. p).
w/
403:8 I
9/15 /5 9 K. Z. Morgan et al(2) have published figures which relate activity O
te ao e 1 the cr1*1c 1 orse - rer ea1 e-131, he 88estea 4--
ver-missible' inhalation of 17 microcuries which would result, at 15% uptake, in 2 5 microcuries in the gland and a resultant dose of 15 7 rem in one year.
In order to inhale 17 microcuries in 8 hr at the standard respiratory rate of 13 liters per minute, the air-borne concentration must be 2 7 x 10-Djac/ml.
Kuper and Cowan(l) suggested a value of 2,000 red or 400 microcuries in the thyroid at 24 hr after the incident as a value below which no dr:me-diate injury would be expected. in adults, although damage to children or de-layed effects in adults is a possibility. This corresponds to 2,660 micro-curies inhal d which for an 8 hr exposure calls for an air-borne concentration of 4.25 x 10 pc/ml. This is clearly an emergency level and is substantially higher than predicted under the most unfavorable meteorological conditions.
?.
'(2) MaxNun permissible concentration of radioisotopes in air and
-l water for short period exposure - Karl Z. Morgan, Walter S. Snyder
.:y
. and Mary R. Ford - Geneva Conference
-1956 - Conference Paper (j
iLNo. 79,;
1
h03:9 7/19/63 Hypothetical Accident at 600 K4t S
The hypothetical accident at h83 K4t is discussed in detail on pages h03:1 to h03:8. Internal and external dose rates have been recalculated for 600 K4t operation, using the came techniques emMoyed for h83 K4t. The effects of higher fission product inventory and greater vapor container leakage due to increased internal pressure (see page h02:h9) were considered.
Dose rates from direct radiation are shown as a function of distance from the vapor container on page h03:10. Integrated dose is plotted on page h03:11.
The direct dose in the shielded control room, where essential plant perronnel would gather following such an accident, is less than 3 rem in the first 2h hours after the release. Other plant personnel are instructed to pro-coed to an assembly area more than 2,500 f t from ',he container on the sounding of a clearly audible alam. This area is shielded from the container by the topography of the site, and dose rates frem direct radiation would be negligible.
The dose at the public road across Sherman Pond, and 1300 feet from the plant, will be less than 7 rem in the first hour following the release.
Therefore, several hours are available to close the road before anyone could receive a dose as high as 25 rem. which has been suggested as a permissible, emergency, whole-body exposure.
. As discussed on page h03:3, only one dwlling not controlled by the New England Power Company or the Yankee Atomic Electric Company will be exposed to direct radiation from the vapor container, and that dwelling is approximately 7
h,000 f t from the plant. The integrated dose at h,000 f t would be less than 30 mrem in the first hour following the release.
The calo ted iodine.
antration at h,000 f t under inversion con-ditionsish.1x1g0 pc/ml, in tm.
3 of equivalent I-121. This corresponds to a total thyroid dose of 36 rer., for an 8-hour exposure to the radioactive cloud.
It is clear that the occupants of the above-rrentioned dwelling could remain there for some time without receiving exposure approaching the emergency levels of 25 rem and 300 rem that have been suggested for the whole-body and the thyroid, respectively.
The hypothetical accident has also been evaluated using the methods described in TID-lh8hh*, although the metearological asewrptions used in this document are not strictly applicable to the deep river valley in which Yankee is located. The resulting 2 hr. desmat the minimum exclusion radius of 3100 ft are 1.2 rem to the whole body and 250 rem to the thyroid. These numbers are within the values set forth in 10CFR100 (25 rem and 300 rem, respectively). The minimum radius calculated for the low population zone le 7.5 miles and the popu-lation center distance 10 miles.
ActW1y the nearest population center to Yankee is Pittsfield, with
$8,000 persons at a distance of 21 miles, and the area betwen Pittsfield and the reactor is very sparsely populated. Within the low population zone there are approximately h,200 residents, all of whom could be evacuated if it ever should become necessary.
TID-lh8hh, " Calculation of Distance Factors for Poster and Test Reactor Sites", J.DiNunno, F. Anderson, R. Baker, R.Waterfield; March,1962.
1;> uo 1.'19163 GAMMA DOSE RATE DUE TO RELEASE OF VOLATILE FISSION PRODUCTS INTO VAPOR CONTAINER 125 FT DIAM SPHERICAL VAPOR CONTAINER 600 mw FULL POWER CORE 100 l
10 E
l l
4fe i
L b
- p?'4
~'
o 4
V M
l t g
0 I'
O.I
-/
I J
4,4 q 1
C i
i f
I l
I
\\
0.01 l
l i
l l
30b0 4000 5000 6000 0
1000 2000 DISTANCE FROM SPHERE, FT GAMMA DOSE RATE FOR OUTSIDE EXPOSURE f
r h
bo3.u 7/19/63 INTEGRATED GAMMA DOSE DUE TO RELEASE OF VOLATILE FISSION PRODUCTS INTO VAPOR CONTAINER 125 FT DIAM SPHERICAL VAPOR CONTAINER 600 mw FULL POWER CORE 10
- ?
g ur 4
o f0.i
'O 9++
s
\\
8
$\\
0.01 N
'O 1000 2000 3000 4000 5000 DISTANCE FROM SPHERE,FT.
INTEGRATED GAMMA DOSE FOR OUTSIDE EXPOSURE
!O
-,, - +
-r, '. _,,
,,--h
f 404:1 9/15/59 40/. CONCLUSIONS This pressurized water reactor possesses inherent stability because of its negative temperature and Doppler coefficients. Since it is normally operated with the coolant-moderator near its saturation temperature, further stability results from the formation of steam voids in any extensive rising power excursion. In addition to this inherent stability, mechanica control rods, capable of making the hot reactor suberitical, are provided to regulate power level and to control reactivity throughout the power production life-time of the core. A supplementary chemical control system is provided to bring the reactor to cold shutdown.
The plant design and the selection of n terials provide four sequential barriers to the escape of fission products to the environment.
These barriers, in order, are 0xide Fuel - The noncorrosive UO2 fuel acts as a first barrier to contain large percentages of the fission products within its matrix.
Stainless Steel Cladding - The fuel red cladding with only two end welds per full length tube acts as a second barrier to escape.
Main Coolant System - The third barrier to escape is the high integrity main coolant system.
O Vapor Container - This barrier acts as a fourth line of defense in the event of fission product release from the main coolant system.
An additional geographical barrier is inherent in the site selected for the plant. The nearest privately owned and occupied dwelling is approxi-mately 4,000 ft away, and the population density within a five-mile radius of the plant is 25 people per square mile.
Several reactor plant accidents have been investigated and analyzed.
Accidents involving reactivity additions during startup and at full power re-sult in transients bat give every indication of leveling off at power levels that are neither harmful nor dangerous. Accidents involving release of energy through chemical reaction between the water and the metallic constit-uents of the core are considered impossible because of the materials employed.
Mechanical accidents in the form of pump failures with ensuing decrease or i
loss of coolant flow can be handled without dangerous power or reactivity excursions. In failures of a single pump, stability is regained with prac-tically no increase in temperature level, even without scram. In failures I
of two or more pumps, however, a low flow scram occurs and the reactor is kept under control by this means.
Among the mechanical accidents that have been analyzed is one caused by a break in a 20 in, main coolant line at the worst possible location and involving loss of all water from the main coolant system but with proper functioning of the safety injection system. This is considered to be the
^
m h = credible accident.
1
404:2 9/15/59 "3
(V In none of these accidents is there any melting of the core, any appreciabi release of gaseous and volatile fission products to the vapor container, nor any hazard to the public.
However, an analysis has been made of a hypothetical accident in which core melting and fission product release are assumed. An accident has been examined in which it is assumed that a large break occurs in the main coolant system; virtually all water is lost from the system; the safety in-jection system does not function properly; partial core meltdown occurs; and 20 per cent of the gaseous and volatile fission products are released to the vapor container. The analysis shows that there would be no hazard to the general.public because of direct radiation from the vapor container. Since l
the vapor container has a finite leak rate, however, some of the fission l
products may escape to the atmosphere and, under certain meteorological con-1 ditions, the escaping fission products may be carried to nearby inhabited i
areas. At the nearest community, however, an 8 hr exposure to the indicated concentration ;f radioactivity, under the most unfavorable meteorological conditions, would result in less than tolerable once-in-a-lifetime inhalation and ingestion doses. Nevertheless, plans are being formulated in cooperation with the Massachusetts State Police and the AEC New York Cperations Office to handle any situation which might result in unacceptable hazards to the public.
j Yankee Atomic Electric Company, therefore, concludes that this reactor can-be operated without undue hazard to the public health and safety.
V E.
N,'
L.