5
An abrupt pressure rise occurred at thermal runaway
preceded by a drop in battery voltage (Figure 8). The volt-
age drop prior to thermal runaway indicates battery short-
ing. Figure 9 shows an expanded view of the pressure and
pressure time rate (dP/dt) during thermal runaway. An
apparent 60 Hz interfering signal (40 Hz perceived fre-
quency at 100 samples per second) became more promi-
nent as thermal runaway progressed. The interfering signal
is most prominent in the dP/dt time trace. It is thought that
the interfering signal originated at the disk heater power
leads and was conducted to the metallic enclosure wall in
electrical contact with the pressure sensor body. Taking the
interfering signal into consideration, the peak pressure and
pressure rising rate are estimated to be in excess of 700 psig
(48.3 barg) and 800 psi/s (55.2 bar/s), respectively.
The thermal runaway pressure distorted the enclosure
beyond allowable limits. Researchers checked joint clear-
ances with a feeler gauge after the test, and several gaps were
found in excess of 0.004 inch (0.1 mm) (Figure 10), which
does not conform with U.S. Code of Federal Regulations
[12]. Escaping gasses etched a path through soot that accu-
mulated in the cover joint as shown in Figure 11. Thermal
runaway deformed the XP enclosure bottom plate about
4 mm (0.16 inch) over the 18.7-cm (7.36-inch) dimen-
sion (Figure 12), exceeding the 0.040 inch per lineal foot
of permanent distortion permitted by U.S. Code of Federal
Regulations [12].
Figure 13 shows the effect of MH1 battery gas leakage
from the XP enclosure by comparison to sealed enclosures
in terms of thermal runaway pressures and dP/dt. The MH1
cell data for the sealed enclosures is from Dubaniewicz et
al. [10]. The sealed enclosure data was obtained under
near-adiabatic conditions that impeded heat loss through
the enclosure walls. The gas leakage from the XP enclosure
and heat sinking to the enclosure walls reduced the pres-
sure and pressure rise rate to a great extent, but not nearly
enough to prevent excessive distortions to the cover joint
and bottom wall.
The battery thermal runaway generated high levels of
carbon monoxide (1,440 ppm) and carbon dioxide (989
ppm) as sampled several meters downwind from the bat-
tery (Figure 14). Besides being toxic, carbon monoxide-air
mixtures are flammable between 12% and 75% by volume
Figure 8. Battery voltage dropped several minutes before
thermal runaway pressure spiked
Figure 9. Time traces of pressure and time rate of pressure
during thermal runaway
Figure 10. A 0.004-inch feeler gauge penetrated the cover
joint in several places after the thermal runaway test
An abrupt pressure rise occurred at thermal runaway
preceded by a drop in battery voltage (Figure 8). The volt-
age drop prior to thermal runaway indicates battery short-
ing. Figure 9 shows an expanded view of the pressure and
pressure time rate (dP/dt) during thermal runaway. An
apparent 60 Hz interfering signal (40 Hz perceived fre-
quency at 100 samples per second) became more promi-
nent as thermal runaway progressed. The interfering signal
is most prominent in the dP/dt time trace. It is thought that
the interfering signal originated at the disk heater power
leads and was conducted to the metallic enclosure wall in
electrical contact with the pressure sensor body. Taking the
interfering signal into consideration, the peak pressure and
pressure rising rate are estimated to be in excess of 700 psig
(48.3 barg) and 800 psi/s (55.2 bar/s), respectively.
The thermal runaway pressure distorted the enclosure
beyond allowable limits. Researchers checked joint clear-
ances with a feeler gauge after the test, and several gaps were
found in excess of 0.004 inch (0.1 mm) (Figure 10), which
does not conform with U.S. Code of Federal Regulations
[12]. Escaping gasses etched a path through soot that accu-
mulated in the cover joint as shown in Figure 11. Thermal
runaway deformed the XP enclosure bottom plate about
4 mm (0.16 inch) over the 18.7-cm (7.36-inch) dimen-
sion (Figure 12), exceeding the 0.040 inch per lineal foot
of permanent distortion permitted by U.S. Code of Federal
Regulations [12].
Figure 13 shows the effect of MH1 battery gas leakage
from the XP enclosure by comparison to sealed enclosures
in terms of thermal runaway pressures and dP/dt. The MH1
cell data for the sealed enclosures is from Dubaniewicz et
al. [10]. The sealed enclosure data was obtained under
near-adiabatic conditions that impeded heat loss through
the enclosure walls. The gas leakage from the XP enclosure
and heat sinking to the enclosure walls reduced the pres-
sure and pressure rise rate to a great extent, but not nearly
enough to prevent excessive distortions to the cover joint
and bottom wall.
The battery thermal runaway generated high levels of
carbon monoxide (1,440 ppm) and carbon dioxide (989
ppm) as sampled several meters downwind from the bat-
tery (Figure 14). Besides being toxic, carbon monoxide-air
mixtures are flammable between 12% and 75% by volume
Figure 8. Battery voltage dropped several minutes before
thermal runaway pressure spiked
Figure 9. Time traces of pressure and time rate of pressure
during thermal runaway
Figure 10. A 0.004-inch feeler gauge penetrated the cover
joint in several places after the thermal runaway test