3
0.25 inches (0.635 cm) thick. The cover flange flame path
was approximately 1.125 inches (2.86 cm) in length. A
5,550-ml (339 cubic inches) internal volume was measured
by filling it with water. With the battery and disk heater
placed inside, the free space within the enclosure was esti-
mated to be 4,870 ml. The free space-to-battery volume
ratio was estimated to be 7.54. Researchers bolted the cover
in place per manufacturer’s instructions and probed the
cover joint flame gaps with feeler gauges. A 0.0015-inch
(0.038-mm) feeler gauge penetrated the cover joint at two
locations shown in Figure 3. A 0.002-inch (0.051-mm)
feeler gauge was unable to penetrate the cover joint any-
where, indicating that the plane joint clearances were within
allowable limits (0.004 inch (0.1 mm)) per U.S. Code of
Federal Regulations [12]. Researchers fed a 4-conductor
type P cable through a stuffing box cable entrance and used
it to monitor battery voltage and to power the disk heater.
The cable was secured in the cable entrance per manufac-
turer’s instructions. Researchers modified the enclosure by
drilling and tapping three 7/16-inch holes for national pipe
taper (NPT) fittings to route thermocouples into the enclo-
sure and to connect a 3,000-psi (207-bar) pressure sensor
to the enclosure’s side wall. These modifications invalidate
the MSHA approval. Researchers selected the pressure sen-
sor range based on pressures measured for a comparable free
space-to-battery volume ratio [10]. Pressures were sampled
at 100 samples per second.
Thermocouples (TCs) were placed at various locations
on the battery and external enclosure surfaces and joints
(Figures 1 and 3). The TCs placed within the enclosure
were sheathed type N (0.81-mm diameter). These sheathed
TCs were epoxied within a steel tube fitting through the
enclosure wall to maintain an air-tight seal. Unsheathed
type K TCs were used on external surfaces and joints.
Temperatures were sampled at 10 samples per second.
The thermal runaway test was conducted in a container
with dimensions of 12-m long, 2.4-m wide, and 2.85-m
high (40 ft × 8 ft × 9.5 ft) (Figure 4) [13]. A fan installed at
one end of the container provided 0.2 m/s of ventilation
velocity. Gas samples were drawn by sampling tubes to an
infrared gas analyzer to measure carbon monoxide and car-
bon dioxide concentrations. The sampling tubes were
located a few meters downwind from the XP enclosure. Gas
measurements were sampled at 10 samples per second.
Video and thermal imaging cameras recorded the thermal
runaway test.
RESULTS AND DISCUSSION
The disk heater initiated a cascading battery thermal run-
away approximately 49.5 minutes into the test. Jet flames
emanating from the cover joint at several locations could be
discerned through thick smoke (Figure 5). The jet flames
appeared to be blue near the joint. Discharge of flame is a
failure criterion per U.S. Code of Federal Regulations [12].
The thermal imaging camera revealed hot gasses extending
well beyond the cover joint through the smoke. Joint and
cable entrance peak temperatures were 1,097 and 138 °C,
Figure 4. Drawing of battery thermal runaway test setup
with fan, cameras, and instrumentation connections
Figure 3. XP enclosure with bolted cover. Locations of the
largest flame-arresting gaps and placement of thermocouples
and fittings shown. Arrows point towards thermocouple
locations out of view
Table 1. Battery electrical parameters
Voltage Capacity (Ah) Energy (Wh)
Rated 47.7 9.6 458
Measured 53.7 9 429
0.25 inches (0.635 cm) thick. The cover flange flame path
was approximately 1.125 inches (2.86 cm) in length. A
5,550-ml (339 cubic inches) internal volume was measured
by filling it with water. With the battery and disk heater
placed inside, the free space within the enclosure was esti-
mated to be 4,870 ml. The free space-to-battery volume
ratio was estimated to be 7.54. Researchers bolted the cover
in place per manufacturer’s instructions and probed the
cover joint flame gaps with feeler gauges. A 0.0015-inch
(0.038-mm) feeler gauge penetrated the cover joint at two
locations shown in Figure 3. A 0.002-inch (0.051-mm)
feeler gauge was unable to penetrate the cover joint any-
where, indicating that the plane joint clearances were within
allowable limits (0.004 inch (0.1 mm)) per U.S. Code of
Federal Regulations [12]. Researchers fed a 4-conductor
type P cable through a stuffing box cable entrance and used
it to monitor battery voltage and to power the disk heater.
The cable was secured in the cable entrance per manufac-
turer’s instructions. Researchers modified the enclosure by
drilling and tapping three 7/16-inch holes for national pipe
taper (NPT) fittings to route thermocouples into the enclo-
sure and to connect a 3,000-psi (207-bar) pressure sensor
to the enclosure’s side wall. These modifications invalidate
the MSHA approval. Researchers selected the pressure sen-
sor range based on pressures measured for a comparable free
space-to-battery volume ratio [10]. Pressures were sampled
at 100 samples per second.
Thermocouples (TCs) were placed at various locations
on the battery and external enclosure surfaces and joints
(Figures 1 and 3). The TCs placed within the enclosure
were sheathed type N (0.81-mm diameter). These sheathed
TCs were epoxied within a steel tube fitting through the
enclosure wall to maintain an air-tight seal. Unsheathed
type K TCs were used on external surfaces and joints.
Temperatures were sampled at 10 samples per second.
The thermal runaway test was conducted in a container
with dimensions of 12-m long, 2.4-m wide, and 2.85-m
high (40 ft × 8 ft × 9.5 ft) (Figure 4) [13]. A fan installed at
one end of the container provided 0.2 m/s of ventilation
velocity. Gas samples were drawn by sampling tubes to an
infrared gas analyzer to measure carbon monoxide and car-
bon dioxide concentrations. The sampling tubes were
located a few meters downwind from the XP enclosure. Gas
measurements were sampled at 10 samples per second.
Video and thermal imaging cameras recorded the thermal
runaway test.
RESULTS AND DISCUSSION
The disk heater initiated a cascading battery thermal run-
away approximately 49.5 minutes into the test. Jet flames
emanating from the cover joint at several locations could be
discerned through thick smoke (Figure 5). The jet flames
appeared to be blue near the joint. Discharge of flame is a
failure criterion per U.S. Code of Federal Regulations [12].
The thermal imaging camera revealed hot gasses extending
well beyond the cover joint through the smoke. Joint and
cable entrance peak temperatures were 1,097 and 138 °C,
Figure 4. Drawing of battery thermal runaway test setup
with fan, cameras, and instrumentation connections
Figure 3. XP enclosure with bolted cover. Locations of the
largest flame-arresting gaps and placement of thermocouples
and fittings shown. Arrows point towards thermocouple
locations out of view
Table 1. Battery electrical parameters
Voltage Capacity (Ah) Energy (Wh)
Rated 47.7 9.6 458
Measured 53.7 9 429