ML19220C629
| ML19220C629 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 04/25/1979 |
| From: | Marino G NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| Shared Package | |
| ML19220C630 | List: |
| References | |
| NUDOCS 7905110343 | |
| Download: ML19220C629 (22) | |
Text
.
- 8
,8 UNITE D STATES NUCLEAR REGULATORY COMMISSION y's O
WASHINGTON, D. C. 20555
% Nh/e'a ApR 2 5 A MEMORANDUM FOR:
Files FROM-G. P. Marino Fuel Behavior Research Branch Division of Reactor Safety Research
SUBJECT:
PRELIMINARY ASSESSMENT OF CORE DAMAGE FOR THREE MILE ISLAND INCIDENT In order to assess the extent of the core damage received by the Three Mile Island Plant #2 during the incident of March 28, 1979, many calcu-lations have been made by members of the RES Fuel Behavior Research Branch and other NRC personnel. Taken as a whole, these calculations can help us narrow down the current condition of the core. The major calculations performed and their results are given below. The core is believed to have suffered its. major damage between 1-1/2 and 2-1/2 hours after the accident with additional minor damage out to 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />. This is the period for which the calculations are made.
A.
Total Hydrogen Generated During the Incident.
These calculations can yield an idea of the extent of oxidation of the zircaloy cladding that took place, since each mole of Zr oxidized con-tributes two moles of hydrogen as a reaction product.
Calculations were made by G. Marino, R. Van Houten and D. Hoatson. Marino's calculations are given in attachment 1.
1.
Results For full oxidation of the active core, the total amount of hydrogen releasable is s 330,000 cubic feet at standard condi-tions (i.e., P = 1 atmosphere and T = 0 C).
The minimum hydrogen produced by the TMI core can be estimated from known hydrogen contents that were observed in the TMI containment, i.e.,
Hydrogen in containment after burn-off = 2% = 42000 cubic feet Hydrogen burned-off 9.9 hours1.041667e-4 days <br />0.0025 hours <br />1.488095e-5 weeks <br />3.4245e-6 months <br /> after scram
= 75000 cubic feet Total
=117000 cubic feet Thus, the minimum amount of core fully oxidized or,, the minimum percent of oxidation of the whole core is s 35%.
79 05110 3 N h
98 033
Files APR 2 51979 In the above, we have neglected any hydrogen contribution from radiolysis and any hydrogen dissolved in the core. We believe the radiolysis contribution to be negligible and the amount remaining in solution difficult to ascertain exactly. Note
~
that radiolytically produced H2 will decrease the estimated in solution would increase the oxidation, whereas the H2 estimated oxidation. Thus, the error in neglecting them will tend to be minimized. However, the above number (35%) may possibly be as high as (50%).
B.
Boil-off Rate of the Core Water After RCS Pump Trio These calculations were intended to determine if the core water could have uncovered (boiled away) during the initial time the reactor coolant pressure was below the saturation temperature (s 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />).
Calculations were made by D. Hoatson and M. L. Picklesimer and are shown in detail in attachments 2 and 3.
A brief summary of results is given below.
For full core power going into vaporizing of the water:
(decay heat-l%)
initial rate = 772 in/hr.
However, as the water boils down, less and less of the core energy is transmitted to the water.
Thus, the rate should decrease and go to zero at zero core level. A rough estimate would therefore be one-half of the initial rate. Thus, Average boil-off rate s 386 in/hr s 32 ft/hr.
Further calculations by Hoatson based on release of steam or water through the stuck open relief valve will diminish the above number, but not to the extent that the core doesn't uncover.
Further speculation by Hoatson and Beckner on he sequence of events at TMI appear in Attachment 11.
C.
Fission Gas Release.
Analysis of radiation levels at TMI and other information received from W. V. Johnston of FBRB indicate that 30 - 40% of the radioactive Xe gas produced in the core was released. Using these numbers, several investi-gators have deduced the temperature that must have been reached by the fuel during the transient.
R. O. Meyer of the Core Performance Branch of DSS related to W. V. Johnston that the temperature must have been s 4000 F.
Malinauskis of ORNL reported to R. Sherry that T t 2400 F.
J. Rest of ANL did some GRASS calculgtions and concluded via a telecon to G. Marino on April 11, 1979 that:
1.
Subsequent to April 11, 1979, J. Rest sumarized his results in a letter to G. P. Marino (see attachment 4) in which he details the assump-tions he used, and gives more accurate temperature ranges.
U 089
Files APR 2 5 573 1.
For a pellet at 12 kw/ft, and a heat-up rate of 1 F/sec, temperatures in excess of 4500 F would have to be reached for 20 - 40% Xe release.
S 2.
For a power of 6 kw/f t only 5 - 10% Xe could be released at 4800 F.
3.
However, he concluded that if U02 grain fragmentation occurred, 50% Xe release could be achieved for the 6 kw/ft rod at 2400 -
2700 F.
4.
R. Van Houten stated that fuel fragmentation (into individual grains) nas occurred in PBF tests under pressure ramps. The TMI core saw increasing pressure ramps of s 10 psi / minute.
Thus, it is possible to conclude that the measured gas release gives minimum fuel temperatures of at least 2400 F and as high as 4800 F.
D.
Calculations c' Maximum Steadv-State Temperature Reachable in Flowing Steam at a Given Power, Heat Transfer Coefficient, and Steam Coolant Temperature.
These calculations were performed by G. P. Marino to assess the kind of clad (fuel rod) temperatures attainable at steady-state for fuel rods in flowing steam. Attachment 5 gives the details of the calculations.
The results show that if the steam temperature stays constant at the level it was at the start of the transient, heat transfer coefficients 2
of less than 2 BTU /hr-ft - F are required to raise the temperature of a 0.1 kw/ft section of a fuel rod to above 2000 F.
Using an in-house computer code (CRAC) it was determined that the heat-up rates for HTC's less than 5 was about 2 - 2.5 F/sec.
In section E it is shown that such rates are essentially adiabatic rates.
Thus, due to the low heat capacity of the steam, it is highly probable that as the steam rises past the fuel rods it achieves essentially the same temperature as the rods do.
Therefore, the calculations in attachment (5) would be on the low side.
Even so, W. Beckner of the Separate Effects Branch used the results and bounded the minimum damage to be full oxidation of 25% to 35% of the core.
He used an HTC of 3 given to him by Y. Y. Hsu on April 6,1979.
See attachment 6 for details.
E.
Adiabatic Heat-up of the Fuel Rods In view of the results in section D above, the calculation of rod / steam heat-up under adiabatic conditions gives an upper bound on the temperatures reached by the uncovered fuel rods. G. Marino and M. L. Picklesimer independently calculated the heat-up rates (see attachments 7 and 8).
98 090
Files ppy,,,g7g The general expression is:
adiabatic heat-up rate = T = 22.3P 'F/sec where P is the power of the rod.
For an average rod at 1% decay heat P = 6 x.01 =.06 kw/ft and T = 1.34 *F/sec.
An 11 kw/ft rod at 1.li, decay heat yields T = 2.7 F/sec.
Thus, even an average rod could have a-hieved a maximum temperature of 4000 F in s 0.7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br /> under adiabatic heat-up conditions.
Additional corroboration of essentially adiabatic heat-up conditions being present in the TMI core are provided in attachments 9 and 10 by Ma ri no.
In attachment 9 it is shown that to heat the steam at the same rate as the fuel rod, the steam requires only s 0.03 fraction of the total decay heat available, and that for an H.T.C. from rod to steam of 5 BTU /hr-ft - F, a temperature difference of only 18 F between rod and 2
steam is required to heat the steam at the same rate as the rod. 0 computes the steady-state steam temperature profile required to cool a reactor with steam at the flow rates calculated by Hoatson (7 ft/ min).
It was assumed that the inlet steam temperature would be s 600 F.
The results showed that for a 10 foot rise the steam would increase in temperature to 4337 F, and for a 5 foot rise 2465 F.
These calculations - since they assume steady-state - are independent of the heat transfer coefficient between the rod and the steam.
F.
Conclusions In view of the preceding calculations, members of the Fuel Behavior Resea-ch Branch met on the afternoon of April 11, 1979 to assess the damage to the TMI core.
In attendance were W. V. Johnston, Chief, G. P. Marino, M. L. Picklesimer, R. S. Sherry, D. A. Hoatson and R. Van Houten.
The following conclusions were tentatively agreed upon.
1.
From the measured H2 values and calculations, at least 30% and as much as 50% of the core was fully oxidized.
2.
Fuel temperatures must have exceeded 4000 F to predict the measured gas release numbers if fuel fragmentation did not take place.
98 091
A Files APP 2 5 $79 3.
In view of 1 and 2 above and calculational results, some adiabatic heating of the rods and surrounding steam must have occurred, allowing temperatures in excess of at least 3500 F to be reached at some parts of the core.
4.
30 - 50% of the core must have been so oxidized that it fell apart into the coolant and remains as a layer of rubble in the core possibly causing flow blockage and core coolability problems.
G.
Current and Future Work.
1.
Picklesimer and Marino will run the computer code BUILDS to assess the oxidation of the core.
2.
Calculations by branch members to pin down the steam heat-up rate more accurately will be attempted.
3.
J. Rest will analyze the iMI gas release using the GRASS code in conjunction with the LIFE code to more accurately predict fuel temperatures prior to the transient.
4.
D. Hoatson will refine his calculations of core boil-off.
Y 'Jm G. P. Marino Fuel Behavior Research Branch Division of Reactor Safety Research
Enclosures:
- 1. Calculations of possible hydrogen production (Marino)
- 2. Boil-off calculations (Hoatson)
- 3. Boil-off calculations (Picklesimer)
- 4. Memo J. Rest to G. P. Marino, dtd. 4/16/79
- 5. Memo G. P. Marino to W. V. Johnston, dtd. 4/4/79
- 6. Memo W. D. Beckner to File, dtd. 4/6/79
- 7. Adiabatic heat-up rate of fuel rod (Marino)
- 8. Adiabatic heat-up rate of fuel rod (Picklesimer)
- 9. Calculations on decay heat (Marino)
- 10. Calculations on steady-state steam (Marino)
- 11. Additional speculations (Hoatson and Beckner) 98 0n, n
ATTACHMENT 1 April 17, 1979 Revised Calculations of Possib_leJdrogen Production G. P. Marino in TMI Core To determine the amount of hydrogen that can be produced by fully oxidizing, the Zr in the cladding, it is necessary to compute the total amount of
~
r Zr present. We will divide this into the two categories of (1) active-length fuel rods (that amount of Zr containing fuel), and (2) Zr from the plenm regions of the rods plus Zr from guide tubes and poison rods added to active length.
I.
Active-length fuel rods There are 36816 fuel rods in the TMI core, each having active dimensions of:
- 0. D. =.4 30", I.D.
.430"-(. 0265)x 2, length = 144" claidingthickness Total weight cf Zr in active length of one rod
OF - (
7 P] w x 144 x (2.54)3
[
3 cm x 6.55 gms = 519.211 gms iii3 cc
= 1.447 lbs Total weight of active length of core = l.911 x 107 gms = 42143 lbs.
Two moles of H are produced per mole of Zr oxidized.
2 Total moles H2 produced from active length =
= 4.19 x 105 x.082 1-atmos / mole x 273 = 331,195 cu ft H
2 28 321/cu ft H for active 2
length II.
Total Zr Present On August 23, 1978, D. Powers of DSS determined from Machi of B&W that the total length of fuel rods was 153.13" including upper and lower plenums.
- Thus, 98 093
ATTACHMENT 1 (Cont.)
Zr from total length of fuel rods = 42143 x 1E3.13 = 44815 lbs 144 7
2.033 x 10 gms Zr from 976 guide tubes
[(.530)
- (.408) ] [ x 153.13 x 6.55 x (2.54)3 976 = 1.442 x 106 gas = 3178.6 lbs Zr from 1085 poison rods =
s
[(.430)
- (.360):] [ x 153.13 x (2.54)3 x 6.55 x 1088 = 7.767 x lo gms 1712 lbs
=
7 Total gms Ze = 2.033 x 10 + 1.442 x 106 + 7.767 x 105 2.255 x 10 gms 7
=
= 49711 lbs Total volume of H possible at STP is:
2 10' x 2 x.082 x 273 = 390,814 cubic feet V
= 2.255 x H 2 9T.22'x 28.T2
~
98 094
ATTACHMENT 2 Boil-off Calculations (Hoatson)
Assume TMI core at 1-3/4 hours after primary pump are tripped off is at an average saturation pressure of 900 psig.
If decay heat available simply went into boiling off the water in the reactor vessel which would,then vent through the pressurizer relief valve - at what rate would the core boil dry?
For average rod power of 6.1 Kw/ft, decay power of 1% at 1-3/4 hr, 36,816 pins in core:
(6.1 Kw/ft) (.01) (12 ft.) (36,816 pins) = 26,950 Kw (26,950 Kw) (1000 W/Kw) (1 BTU /.293 w-br) = 92 mill. BTU /hr.
for steam at 900 psig (915 psia) hv - h1 = 1195 - 528 = 667 BTU /lb 6 BTU /hr)/(667 BTU /lb) = 138,000 lb/hr.
steaming rate = (92 x 10 On a per pin basis this is (138,000 lb/hr)/36,816 pins = 3.75 lb/hr pin 3
for saturated liquid at 900 psig, the specific volume is.0212 ft /lb 2
and using the TMI flow channel area of 0.178 in ) = 772 in/hr.
The initial boil-cff rate is certainly adequate to boil the core dry in a little over 10 minutes.
When the level reaches the midpoint of the core, only the decay heat in the lower half of the core is contributing to boil-off and the rate will reduce to 386 in/hr. As the boil-off proceeds, the rate will reduce to 0 when the bottom of the core is reached.
These initial calculations indicate that boiling the core dry using decay heat in a stagnant pool of water is a credible scenario. More refined calculations should account for:
additional water that may flow into the core from the loops; boil-off rates should be adjusted for the axial cosine power shape; and additional water from HPI pumps, if any.
98 095
ATTACHMENT 3 Calculations Regarding Boil-off in Core (Pickiesimer)
Area of subchannel = A = (0.568)2 -h0.430)2 o,177404 jn2 3
~
A = 1.22944 x 10 3 ft2 5
Vol
= 1.22944 x 10 3 ft: x 12 ft = 0.1475 ft3 s
V = 25.494 in3 s
Assume 12 f t core uncovered in 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> al = 12 f t/hr = 2.4 in/ min = 0.04 in/sec 3
Boiling rate = 25.494 in3 + 3600 sec = 7.08 x 10 3 in /sec
= 4.098 x 106 3
ft /sec 3
~
at 0.0216 ft /lb,
= 1.897 x 10 " lb/sec Per channel Boiling Latent Btu /sec Vg Vol/sec Velocity 3
3 ft /sec of steam ft /sec psi rate, lb/ set heat Btu /lb required in chennt; 1000 1.897 x 10
- 649.4 0.123 0.4456 8.45x10~5 4.0 ft/mir 700 1.897 x 10 "
709.7 0.135 0.6554 1.24x10 4 6.06 ft/mi
~
2770 MW thernal power at start of accident - 1% decay heat at 1-1/2 - 2 hrs.
27.70 x 106 + 36,800 = 752.7 w/ rod
= 0.721 Btu /sec/ rod
. 0.721 Btu /sec/ rod available = 5.86 times requirerient at 1000 ps1'0.123 Btu /sec/ rod required at 200 psi,0.721 Btu /sec/ rod available = 5.34 times requirement 0.135'3tu/sec/ rod requ ired 27.70 MN = 95.5 x 106 Btu /hr/ core at 1000 psi, boil-off whole core = 14?,059 lb/hr at 2000 psi, boil-off whole core = 134,564 lb/hr Core barrel = 141" 10 = 11.75 f t core vol
= 1301.2 f t3
-fuel rods = -445 ft3 147,059 = 2.94 1000 psi
-structures = - 55 ft3 Boil-off ratio = 49,920 = 2.69 200 psi water vol 800 ft3 = 49,900 lb.
=
98 096
ATTACHMENT 5 6m ph 0 t 99 s
MEMORANDUM FOR:
W. V. Johnston, Chief Fuel, Behavior Research Branch FROM:
G. P. Marino Fuel Behavior Research Branch
SUBJECT:
MAXIMUM STEADY-STATE Ct.AD TEMPERATURES REACHED IN TMI PLANT AFTER VOIDING According to current " sequence of events" reports received by NRC personnel, the THI reactor core may have been initially unccvered for approximately one hour at ih hours into the transient. At this time no stored energy shculd be left in the rods and the clad and fuel tempera-tures at the decay power level (apprgximately 1 percent) should be essentially that of the coolant (556 F). Under fuch conditions, it is a simple matter to calculate the r.aximum obtainable toperatures of the cladding given the power and heat transfer coefficient (HTC) to the surroundings.
When the cora is in static steam, the fuel rods will begin to heat up since the heat removal rate will be less than the heat generated by decay heat. Using tne CRAC code, developed by the author for in-house use, ft was found that a given rod will heat up at a rate of approximately 2.5-3 f/sec. for a power level of approxir.ately 0.11kw/ft (degaf eat of h
1% for initial rod pcwer of 11kw/ft) and a HTC of 5 BTU /Hk-ft - F.
Such a rate would allow the steady-state terperature of cladding to be reached in approximately 5-10 minutes. The steady-state temperature reached is easily calculated by equating the heat generated to the heat flux out of the rod @ssuming the decay heat power changes very little over 5-10 minutes).
i.e., Heat generated (in flux units) = H (Tclad - Tenoj)'
The resulting calculation yielded the following equation far the raximum attainable cladding temperature of the THI core for voiding at approxi-mately 14 hours1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br /> into the transient.
T
= 5% +
clad
- m >
..... ?
ems = = >
DefS M
. - = <.. * *
- Nac roam sie o.m mcw sus
- -.-----=="a-a...
98 0ny
/,
,Dr. C. Marino April 16, 1979
- 3. Assuming that extensive grain-boundary separation does occur, GRASS-SST results indicate that for the 6 KW/ft fuel 50% or more total gas release could occur at fuel temperatures on the order
~
of 2400-2700*F. Results for the 12 KW/ft fuel indicate that 50%
gas release could occur under these conditions at fuel te=peratures
~
on the order of 4500*F.
The calculations indicated that the substan-tially greater fraction.1 release of fission gas predicted to be released from the lower rating fuel than fron the higher rating fuel at temperatures of about 2400-2700*F in the event of extensive grain-boundary separation was due to the fact that the lower opera-ting te=peratures of the lower rating fuel resultef in much smaller (and hence more mobile) bubbles being generated vi nin its grains.
- 4. Experience with fuel f rom the H. B. Robinson Reactor (30,000 mwd / ton burnup ce=parec to the N1000 mwd / ton burnup of the Three Mile Island fuel) during Direct Electrical Heating (DEH) transient tests at heating rates substantially higher tl'n those calculated to occur at Three Mile Island indicates that s20% a-o 0% release occurs in fuel regions where the te=peratures rer 2750 and 3050*F, respectively. However, it is expected that the
_. higher concentration of gas in the H. B. Robinson fuel (very little gas was released during the irra-diation) facilitated the for=ation of the observed grain-surface and grain-edge channels, and this enabled the gas released from the grains to escape te the exterior of the fuel.
In addition, fairly extencive grain-boundary separation was observed to occur in this fuel as a result of the transient heating.
The above results are based on assumptions about the properties of the fuel (e.g., grain-size), assumptions on the irradiation history of the fuel rods, and on the suggested scenarios of the accident.
For exa=ple, from the above results it is clear that the irradiation history of the fuel (e.g., 6 KW/f t vs. 12 KW/ft) significantly affects the predicted gas release during the accident.
Sincerely,
/
J. Rest 3R/ez cc:
Dr. W. V. Johnston, NRC/RSR R. W. Weeks 98 090 t