ML20028B481

Thermal Margin Beyond Design Basis (Tmbdb) Melting Scenario.

ML20028B481
Person / Time
Site:	Clinch River
Issue date:	11/23/1982
From:	ENERGY, DEPT. OF, CLINCH RIVER BREEDER REACTOR PLANT
To:
Shared Package
ML20028B482	List: ML20028B481
References
NUDOCS 8211300431
Download: ML20028B481 (18)

Text

_

TMBDB MELTING SCENARTO I. Introduction The CRBRP Thermal Margin Beyond the Design Base (TMBDB) scenario assumes that a coolable core debris bed forms on the reactor cavity floor following a melt-through of the reactor vessel and the reactor vessel guard vessel. To ensure that a coolable bed does form, the reactor vessel guard vessel support skirt is provided with flow ports to assure that sufficient core debris is swept under the support skirt to produce coolable bed depths on the RC floor both inside and outside the support skirt (this is discussed at length in Appendix G.1 of Reference 1) .

Following the establishment of a coolable core debris bed, the EC floor liner is assumed to fail. In order to maximize the sodium-concrete contact in the Base Case scenario (Reference 1), the RC floor liner is conservatively assumed to " vanish" at the time of failure. The core debris is then assumed to remain at the interf ace of the sodium-concrete reaction product layer and the sodium, such that the debris bed remains cooled by the sodium pool. This report presents the results of an analysis to examine the ef fects of 'a 8211300431 021123 PDR A

ADOCK 05000537 l PDR

1 hypothetical scenario where the core debris is assumed to sink beneath the sodium-concrete reaction products, whereupon the reaction product layer would insulate the core debris from the cooling effects of the sodium pool.

The purpose of this " melting scenario" analysis is to evaluate the consequences of a non-coolable debris bed early in a TMBDB scenario in terms of TMBDB margins (Reference 1). Thus, the melting scenario is compared with the TMBDB base case scenario in terms of (1) containment vent time (36 hours4.166667e-4 days <br />0.01 hours <br />5.952381e-5 weeks <br />1.3698e-5 months <br /> in the base case),

(2) containment atmospheric and steel shell temperatures, (3) containment pressures, and (4) RCB atmosphere hydrogen concentration. This scenario is presented as a bounding case on the effects of core debris on the TMBDB scenario and is not assumed to be a realistic scenario since:

1. The RC floor liner is not expected to " vanish" at time zero and this scenario ignores the capability of the unfailed portions of the floor liner to '

support the core debris.

2. The sodium-concrete reaction products are '

believed to be a viscous liquid which would

provide a measure of support for the core debris.

II. Conclusions In the event that core debris on the reactor cavity floor assumed a non-coolable geometry (melting scenario), the following conclusions are drawn from the melting acenario study discussed in detail below:

o containment conditions would be acceptable throughout the scenario, o containment venting could be delayed for at least 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, o penetration into the basemat prior to sodium boildry would be about 5 feet, o existing TMBDB systems would be capable of handling the TMBDB loads.

III. Scenario The melting scenario was developed from the TMBDB base

_4_

case scenario of Reference 1, Section 3.2.1, so as' to permit direct comparison with the base case, thereby evaluating the impact of early melting on the beyond design base margins. Thus, the only changes were to replace the sodium-concrete reactions in the reactor cavity floor (after a 20 minute initial rapid sodium-concrete reaction as described below) with an extreme heat input to represent progression of a melt front (molten core debris and melting concrete) into the floor. All other features of the scenario, such as sequential liner failures, the resulting sodium-concrete reactions in other structures, the 50-hour vent bypass through the reactor vessel head, etc. were retained. Therefore, only the scenario details involving the melting into the RC floor will be described here.

1. Upon failure of the reactor and guard vessels, it is assumed that sodium and core debris pour onto t

the reactor cavity floor and spread uniformally over it as in the base case.

2. For twenty minutes, sodium-concrete reactions' occur at a rapid rate of 18 cm/ hour for a total

penetration of 6 cm. This rate and duration are' the recommended initial upper bound from Reference 2, page 50.

3. The reaction products formed from the initial 20 minute sodium-concrete reaction form a layer 5 inches thick (approximately twice the volume of the original concrete reaction tests), and the core debris sinks to the bottom of the reaction product layer. At this point the core debris layer (3.4 inches thick per Reference 1, page G.1-7) is separated by the 5 inch layer of reaction products, which has a thermal conductivity much less than sodium (assumed value 1.0 BTU /hr-f t-oF vs. M30 for liquid sodium) .

4. Also at this point the sodium pool is mechanically isolated from the concrete so that sodium-solid concrete reactions cease. The steam and CO 2 driven out of the concrete, by the heat of the core debris, pass through the debris and reaction product layers and react with the sodium pool, to generate additional sodium reaction products,-

hydrogen and heat (i.e., the sodium does not

migrate down to the concrete since, if it did; the debris would be cooled by the sodium and there would be no melting of concrete as described below).

5. The insulating effect of the reaction product layer causes the core debris layer to heat up (due to core decay heat). When temperatures reach the assumed melting point of concrete (2200 F)*, _the core debris melts into the concrete.

6. As the concrete melts, the core debris layer sinks through the molten concrete, and remains in contact with the unmelted concrete.

7. The core debris melt front continues down into the concrete, driving water and CO 2 UP into the sodium pool. Sideways melting is not considered since the extension of the wall liner anchor for the i

Margin Assessment Case (see Reference 3) would preclude sideways melting.

! *This is the lowest eutectic temperature of the concrete-core debris mixture (see Section 2.2 of Reference 1).

i

8. Af ter sodium boildry, the scenario is essentia'lly the same as the base cast post-boildry, except the post-boildry melt front starts at a lower elevation by the amount of melt front penetration prior to boildry.

IV. Analysis Methods The analysis was conducted in two steps, a TRUMP analysis of the thermal effects of fuel debris melting into concrete to determine fraction of decay heat driven into the concrete followed by a CACECO analysis to determine containment conditions resulting from the RC floor thermal effects.

1. TRUMP Analysis The TRUMP code was utilized to study the situation where fuel debris on the reactor cavity floor is separated from an overlying pool of sodium by an insulating layer of reaction products. In this situation, heat transfer'from the fuel to the sodium is by conduction through the insulatin'g layers of molten concrete and reaction products.

Under these conditions the fuel would be expe.cted to melt into the concrete, releasing more CO and 2

water than that calculated for the base case, and the resulting effects on containment using the CACECO code are to be calculated. The analysis using TRUMP, however, is a prerequisite in order to develop the input for CACECO.

A one-dimensional TRUMP model was developed as shown in Figure 1. It consists of one node representing the reaction product layer which is connected to a constant temperature boundary (the sodium pool); a second heat generating node representing the core debris layer, and 52 nodes representing 26 feet of concrete.

The thermal properties for the fuel and concrete 1 are the same as has been used in previous analyses of Reference 1. The thermal properties for the reaction product (assumed to be sodium carbonate)

, are:

i p 3

= 158 lb/ft C

p = .27 Btu /lb 0 F l

. k = 1.0 Btu /hr ft OF ~

The density and heat capacity of sodium carbone.te were obtained from page 3.121 of Perry's Chemical Engineer's Handbook, Fourth Edition. The thermal conductivity value was selected on the basis that it would be similar to that of concrete (0.6 to 1.5 Btu /hr f t oF) as no data were found specifically for sodium carbonate.

The results of the calculation are shown in Table 1 as the fraction of the generated heat which is transferred to the overlying sodium, and the fraction that is transferred to the concrete beneath the fuel. It is these two quantities that are used in developing the CACECO input. Except for the early phases (ceveral hours) of the transient, approximately 60% of the heat generated is transferred into the underlying concrete.

2. CACECO Analysis In the TMBDB scenario, fuel in the reactor cdvity would be expected to be in particulate form in a

l coolable debris bed until all of the sodium would be boiled away (Reference 1). However, as described earlier, this analysis will consider the core debris bed spread over the entire cavity floor to be uncoolable af ter sinking beneath a five inch layer of sodium-concrete reaction products. Based on the scenario described in Section III, the CACECO code was used to model this effect.

The CACECO model used in the analysis is the same as used in the Reference 1 base case except for the following: (a) the cavity floor was divided into 5 heat structures made up of a series of 2 inch nodes; (b) the sodium-concrete reaction was redefined (described later); (c) the boundary conditions applied to the RC floor were changed to simulate a melt front progression; and (d) venting was initiated at 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. The RC floor thermal model involved 5 connected heat structures to represent the basemat. Each of the heat structures was made up of' nodes 2 inches thick so as the melt front progressed, the volumetric -

release of water and carbon dioxide reacting with

the sodium would occur in realistic increments ~.

For this scenario, sodium-concrete reaction rates of 18 cm/hr were utilized for a 20 minute duration. (See Reference 2, page 50) .* The reaction energy for this interaction was the base case value of 331 Btu /lb. of concrete. With a layer of reaction product formed from the initial rapid sodium-concrete reaction, the core debris was assumed to sink below this layer and become isolated from the sodium pool. At this time the core debris cooling would be greatly reduced, causing the core debris temperatures to increase and melt the concrete. Based on the TRUMP results shown in Table 1, the fission product decay power l

was proportioned between the sodium pool and the concrete in the cavity floor to simulate a melting l process in the CACECO code. The melting process into the RC floor is simulated by forcing a heat flux into the first node of concrete. As the melting point of the concrete (220oF) was reached

• For the Base Case scenario, the sodium-concrete reaction rate is assumed to be 1/2" per hour for 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />.

l - - --

for a particular node, the thermal conductivity of that node was subsequently changed from O

approximately 1 Btu /hr ft F to 100, thus allowing the melt front to progress. As the concrete heated up, the water and CO released passed through the 2

core debris, the reaction product layer and the melted concrete to react exothermally with the sodium pool until sodium boildry.

Vent and annulus cooling were initiated at 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> with purging starting at 27 hours3.125e-4 days <br />0.0075 hours <br />4.464286e-5 weeks <br />1.02735e-5 months <br />. All other aspects of the scenario were identical to the base case.

V. Results The containment conditions from this CACECO analysis are compared with the TMBDB base case in Table 2. As expected, containment pressures, temperatures, and hydrogen concentrations are more severe for the m'elting scenario than for the base case, but the conditions are still acceptable in terms of vent time (see Reference 3). The peak hydrogen during the.

. venting blowdown exceeded 6 y/o (maximum 6.4 y/o) "for a short period, but this occurred at a time when containment oxygen was below 5 y/o. The initial (autocatalytic) hydrogen burn occurred earlier in time with a greater accumulation of hydrogen and resulted in higher, but still acceptable, pressure and temperature spikes (conservatively assuming instant burning).

Beyond 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> the upper containment conditions are acceptable. The peak vent rat" to the cleanup system and the peak heat load to the steel shell were slightly higher than the base case, but within other extreme sensitivity studies for which TMBDB systems were found adequate. These results indicate that existing TMBDB systems could accommodate the melting scenario. Temperatures are, of course, much higher deeper into the concrete basemat in the melting scenario, but the shortened boildry time would make the structural temperatures in the RC walls and pipeway cells less severe since the structures would be exposed to sodium boiling' temperatures for a shorter time. The steel shell temperatures are h1gher than the base case, but could be reduced by starting

the annulus cooling sooner without venting containment (in this calculation, both were started at 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />) .

It is concluded that the TMBDB systems would mitigate the threats to containment in the unlikely event that the core debris on the reactor cavity floor resulted in a non-coolable geometry (melting scenario).

References

1. CRBRP-3, Vol. 2, " Assessment of Thermal Margin Beyond the Design Base," dated March 1980.

2. HEDL-TME 82-15, " Sodium Concrete Reaction Executive Summary Report: Limestone Concrete,"

dated June 1982.

3. Letter HQ:S:82:112, J. Longenecker to P. Check, "TMBDB Margin Assessment Document," dated October 20, 1982.

9

I

~

Table 1 Heat Split Between Concrete and Sodium As Determined by the TRUMP Model i

Heat Split Time, hr Fraction Fraction

', Into to Concrete Sodium l

l l-10 .79 .21 10-25 .63 .37 25-35 .62 .38 35-50 .63 .37 50-* .60 .40 t

,_.,y, - _

, . - - . . . ,-- - - -, . . -- , - - --- - - - ~ ' '~ ~~

. . m .

Table 2 4

Comparison of Containment Conditiors for Base Case and Helting Scenario

- All Parameters: Base Case / Melting Scenario Initial Hydrogen Burn Time (Hrs.) 10.0/8.4 Pressure After Burn (psig) 22.4/27.9 RCB Atmosphere Temp. After Burn ('F) 845/1034 RCB Hydrogen Conc. Before Burn (v/o) 4.5/5.7 .

I' 5' Initiation of Venting Time (Hrs.) 36/24 RCB Hydrogen Conc. Peak During Vent (v/o) 4.5/6.4 Remainder of Scenario RCB Atmos.

Temp. ("F) RCB Pressure (psig) Steel Temp. (*F) Hydrogen (w/o)

Timb (Hrs.)

24 450/830 11.1/18.3 270/460 0/2.4 .

36 620/875 13.1/0 400/445 0/4.0 50 515/790 0/0 315/410 4.0/4.0 71 670/1000 0/0 360/495 3.7/4.0

~

I - . .

Table 2 (continued) -

l l .

Base Case Melting Scenario Bolldry Time (Hours) 133 71 RCB Maximum Vent Rate (ACFM) 24000 29000 8

Peak Heat Load on RCB Steel Shell (Btu /Hr.) 1.1 x 10 1.2 x 108 sodium Vapor to RCB at Boildry (Lbs.) 8.6 x 10 5 6.71 x 10 5 Sodium Vapor to RCB Max. (Lb/ Hour) 19300 29000

, RC Floor Temperature Above 500'F at Boildry 2.3 5.5 (Ft. from Liner)

RC Submerged Wall Temperature Above 500*F at Boildry 1.2 0.6 *

(Ft. from Liner) t; RC Floor Penetration at. Boildry (Ft.) 0.33 5.0 i

n

TMBDB MELTING SCENARIO Sodium Overlay-simulated by constant temperature .

boundary Reaction Product Layer } h 0

Molten Concrete Increaising 0

Fuel Debris } 3.4 "

Solid Concrete 26 ft. (and decreasing) i l

l ,

Figure 1 Schematic of TRUMP Model

'