| dc.description.abstract | The most serious of the postulated accidents considered in
the design of the Liquid Metal Cooled Fast Breeder Reactor (LMFBR)
is the Loss of Pipe Integrity (LOPI) accident. Analysis models
used to calculate the consequences of this accident assume that
once boiling is initiated film dryout occurs in the hot assembly
as a result of rapid vapor bubble growth and consequent flow
stoppage or reversal. However, this assumption has not been put
to any real test.
Once boiling is initiated in the hot assembly during an LMFBR
LOPI accident, a substantial gravity pressure difference would exist
between this assembly and other colder assemblies in the core. This
condition would give rise to natural circulation flow boiling
accompanied by pressure and flow oscillations. It is possible that
such oscillations could prevent or delay dryout and provide
substantial post-voiding heat removal. The tests described in this
report were conceived with the objective of obtaining basic information and data relating to this possibility.
To accomplish this objective a natural circulation test loop
was designed to simulate LMFBR geometry and flow conditions predicted
to exist at the time boiling is initiated in a LOPI accident. The
test loop included: (l) a vertical tube test section, (2) upper and
lower plenum tanks, (3) an external down-commer, (4) sight flow
indicators and (5) instrumentation. The test section was an
electrically heated tube designed with a hydraulic diameter and
length similar to current LMFBR (FTR) design. The upper and lower
plenum tanks were provided with means for controlling liquid
subcooling above and below the test section. The down-commer was
large enough to eliminate down-commer hydraulics. Water at a
pressure of 1 atmosphere was used to simulate sodium. Sight flow
indicators were provided to observe flow conditions at the test
section inlet and exit. Instrumentation was provided to measure
test section pressures, inlet and exit temperatures, tube wall
temperatures, heat flux and oscillation frequencies.
Steady state tests were conducted for subcooled flow boiling,
saturated flow boiling, CHF and post CHF conditions. Subcooled
flow boiling was observed for heat fluxes below 1 x 104 BTU/hr ft2.
For this condition, both pressure oscillations and temperature
oscillations at the heated surface were observed; but the pressure
oscillations were not observed continuously. Saturated flow
boiling was observed for heat fluxes between 3 x 104 BTU/hr ft2
and CHF. For this condition, pressure oscillations were observed
continuously. As the CHF condition was approached, a periodic
downward expansion of vapor from the heated section was observed at
the bottom sight flow indicator and the flow regime appeared to be
annular at the top sight flow indicator. CHF was observed at the
top of the heated section when the heat flux reached 6.4 x 104
BTU/hr ft2, but rewetting occurred after a few seconds. As the
heat flux was increased further, the maximum surface temperature
reached before rewetting increased; until, at a heat flux of 7.15 x
104 BTU/hr ft2, the maximum temperature exceeded 9000 F and rewetting
no longer occurred. A transient test was conducted for a post CHF condition. The
heat flux was 7.3 x 104 BTU/hr ft2. The oscillations observed under
steady state conditions developed within a few seconds after the
power was turned on. The equilibrium tube wall temperature upstream
of the CHF location was reached in 10 seconds. The equilibrium tube
wall temperature at the CHF location was reached in about 135 seconds.
A similarity analysis was done in order to scale the test results
to LMFBR LOPI conditions. The results of this analysis indicate that
the CHF for the LMFBR (FTR) would be at least 6 x 104 BTU/hr ft2.
This corresponds to a critical average linear power for the hot
assembly of 1.06 kw/ft compared to an estimated 2.55 to 5.1 kw/ft being
transferred to the coolant at the time boiling begins during a LOPI
accident. On the basis of this analysis, the results of the water
tests indicate that CHF would occur. But, this conclusion is
conservative for a number of reasons and further experimental work
on a more prototypical system is suggested. | en |
