Untitled Document

Page 1 of 10

Home | Contact us

PCTRAN/SFP Fukushima Daiichi Unit 4 Spent Fuel Pool Accident Analysis

M icro-Simulation Technology April 10, 2011 The incident initial condition and event chronicle are based on the following report:

http://www.microsimtech.com/downloads/Fuku4.html

10/15/2011

Untitled Document

Page 2 of 10

http://www.microsimtech.com/downloads/Fuku4.html

10/15/2011

Untitled Document

Page 3 of 10

We assumed that Unit 4 was in refueling with the full core discharged into the spent fuel pool. The fuels have burn-up to 64 MWD/MTU and 15 days after shutdown. There are 560 assemblies. So the decay heat level was 15.47 MBtu/hr, or 4.53 MW (0.20 % of rated 2311 MW power). The pool has room to accommodate fuels of previous cycles. Since their exact amount is unknown to us and its contribution to the total decay heat is minimal, it is ignored in this study. The fuel pool cooling system was in operation to maintain the pool at 97 (36.1C). The seawater for cooling was 50 (10 F F C).

http://www.microsimtech.com/downloads/Fuku4.html

10/15/2011

Untitled Document

Page 4 of 10

Right after the Scale 9 earthquake the AC power was lost. Without cooling to the pool the code immediately calculated the heat-up rate to be 6.4 F/hr (3.5 C/h). It would take about 17.8 hours to boiling (in the upper left Time Estimate panel of the simulator mimic). We also assumed the earthquake has caused a crack to leak out 100 gpm (370 liters/min) water at the bottom of the pool. The projected time for fuel top exposure is 35.4 hours. Note the projected time is somewhat arbitrary, because we know neither the exact decay heat nor the crack size. Rather they are roughly consistent with the event chronicle to occur in 2 to 4 days. Following is the PCTRAN/SFP mimic right after event initiation. Note all circulating pumps are tripped so there is no means for heat removal.

Fig. 1 PCTRAN/SFP at event initiation Note full-core offload in the pool, the time estimate to boiling and fuel exposure at top left panel, all pumps are off on loss AC power. Table 1 Date Time Sequence of events according to NI SA of Japan report and PCTRAN/SFP simulation Time from start hours (seconds) 0 Event PCTRAN/SFP Run

March 11 14: 46

Earthquake and station Loss of offsite power, blackout Unit 4 full-core offload plus previous cycle spent fuels in the pool.

http://www.microsimtech.com/downloads/Fuku4.html

10/15/2011

Untitled Document

Page 5 of 10

1. Loss of SFP cooling (trip all circulating pumps).

1. Earthquake caused a crack at the pool bottom to initiate a drain flow of 370 ltr/m. March 14 4: 08 61.5 (221,400 s) Water temperature in SFP 84 C Water started boiling (212 (100 at 16.7 hr F C) (60,000 sec ) 1. Top of fuel assemblies exposed to air at about 36 hr (130,000 s). 2. Hydrogen from Zr -water reaction and Zr fire total > 10% of claddings. 3. Fire in the reactor building became possible (not simulated) By turning on the spray nozzle on top of the simulator (simulating the helicopter spray) the fire could be extinguished; but not simulated in this run.

March 15 9: 38

90 (324,000 s)

Fire occurred

March 15 12: 25

93 (334,800)

Fire extinguished

http://www.microsimtech.com/downloads/Fuku4.html

10/15/2011

Untitled Document

Page 6 of 10

Fig. 2

Pool bottom drain flow rate.

Fig. 3 Pool water level. The fuel top is about 12.5 feet from bottom. The top is exposed in about 130,000 seconds (36 hr).

http://www.microsimtech.com/downloads/Fuku4.html

10/15/2011

Untitled Document

Page 7 of 10

Fig. 4

Fuel decay heat.

http://www.microsimtech.com/downloads/Fuku4.html

10/15/2011

Untitled Document

Page 8 of 10

Fig. 5

Temperaturesof Peak fuel, peak clad of off-loaded core and older fuels, and Pool water. The pool boiled in about 60,000 sec (16.7 hrs).

http://www.microsimtech.com/downloads/Fuku4.html

10/15/2011

Untitled Document

Page 9 of 10

Fig. 6

Total released iodine and noble gasrate.

Fig. 7

Dose rate around the Reactor Building

http://www.microsimtech.com/downloads/Fuku4.html

10/15/2011

Untitled Document

Page 10 of 10

Fig. 8 Condition at end of the run about 170,000 seconds (47.2 hrs). Over 10% of fuel clad was oxidized by steam interaction and direct oxidation (burn). Hydrogen content in the building was combustible. A burn could start anytime. The fire waslater extinguished by spraying water from top. They were not simulated in thisrun.

Conclusion
In our first try to simulate this event, we made a gross estimate of the pool fuel inventory and its decay heat level. We also assumed a bottom leak that would expose the fuel in a couple of days. However, no attempt was made to match the timings exactly as the event chronicle. PCTRAN/SFP has turned out that the fuel cladding started to heat up and reached burning condition in about 2 days. Subsequent spraying water from top should extinguish the fire and stop radiological release. The fission gas release rate and dose around the building are compatible with those estimated or monitored during the event.

http://www.microsimtech.com/downloads/Fuku4.html

10/15/2011