ITIGATING
AEROSOL FIRE AND EXPLOSION HAZARDS
ASSOCIATED WITH LOW VAPOR PRESSURE LIQUIDS

Heather D. Willauer; Ramagopal Ananth; John B. Hoover; George W. Mushrush; and Frederick
W. Williams; Naval Research Laboratory, Code 6180, Navy Technology Center for Safety and
Survivability, 4555 Overlook Avenue, SW, Washington DC 20375


INTRODUCTION

This paper addresses the hazards associated with aerosolized hydraulic fluids and other liquids
designed to be relatively non-flammable with respect to flashpoint and vapor pressure. The
studies presented seek to understand and quantify conditions and circumstances that can produce
such hazardous conditions. US Navy personnel are particularly susceptible to fuel aerosol
hazards due to the widespread use and storage of these materials on both ships and airbases [1-
3]. As a result a great deal of research effort has been directed toward designing safer jet fuels,
lube oils, and hydraulic fluids. Most of these approaches have focused on improving the
thermodynamic properties of the liquids by increasing their flashpoint and lowering their liquid
vapor pressure. This approach was used by the Department of Defense (DOD) to reduce aviation
fires. The design and implementation of lower volatility aviation fuels such as JP5 (Navy) and
JP8 (Air Force) helped to reduce fuel ignition caused by gunfire in combat and help increase fire
safety aboard aircraft carriers [4].

The DOD has also put a great deal of effort in the development of fire-resistant hydraulic fluids
as a replacement for petroleum-based hydrocarbon fluids still in use today [1,2,5,6]. Though
these fluids looked promising, flammability tests suggested aerosols of these fluids had the
potential to ignite and burn [1,2,6]. The safety advantages of these fluids were not enough to
warrant the cost of switching to these fluids or the potential unknown operational risks of using
these fluids, thus they were never implemented [5,6].

Though changing and designing fluids with better thermodynamic properties increases fuel
safety in bulk, little work has been done to quantify, predict, and mitigate hazards associated
with these liquids when they become aerosolized. These systems are complex because the vapor
concentration is dependent on several parameters which include; droplet size, number density,
fuel flow rate, and droplet linear velocity. In these studies, large scale tests demonstrate the
potential consequences of the atomization of 2190 TEP and the existing technology used to
mitigate these hazards. In addition, current methods and techniques used to quantify aerosol
composition and the correlations established to identify and predict hazardous aerosol
compositions will be demonstrated.

EXPERIMENTAL

The large-scale simulations of hydraulic system explosions were conducted in an compartment in
the Naval Research Laboratorys (NRL) ex-USS Shadwell full-scale fire test facility, located in
Mobile, AL [7-9]. The compartment mockup was built in the port wing wall on the ex-USS

1 Shadwell with lateral dimensions of 8.5 meters by 4 meters. The properties of the hydraulic fluid
used in the test series are shown below in Table 1.

Table 1. 2190-TEP Fluid Properties
2190 TEP Property
Property Value
Manufacturer
Chevron Turbine Oil Symbol 2190 TEP
Composition
>99% Heavy paraffinic distillates
Absolute Viscosity (cP)
69.0 @ 40
o
C; 8.4 @
o
C
Flash Point (
o
C) 246
Boiling Point (
o
C) >315
Specific Gravity
0.86 0.87
Heat of Combustion (MJ/Kg)
42.7

The hydraulic fluid was pressurized to 10 MPa (1450 psi) in a 190 liter (50 gal.) pressure vessel
using a 12-cylinder nitrogen manifold. The cylinder was connected to a nozzle array using 1.3
cm (0.5 inch) diameter welded stainless steel pipe. The array consisted of five in-line positions,
which were plugged when not being used. The nozzles were Bete Fog Nozzles, Inc. model P24,
90
o
solid cone spray nozzles. The fluid temperature at the nozzles was approximately 45 50
o
C.

Ignition of the aerosols was achieved using an electric arc, energized by an Allson type 4258
transformer (120 VAC, 450 VA input; 15 kv, 30 mA output). The 3 cm spark gap was located
above the nozzle array. The pressure transducers (1-2 psi) (Omegadyne model PX02C1-
002G5T) were installed outside the test compartment and were connected to short, large bore
stainless steel tubing that penetrated the bulkhead. This allowed fast pressure transient
measurements to be made during the tests.

The fire extinguishing agents tested for mitigation were standard PKP, CO
2
, and AFFF
extinguishers and hand held water mist systems. In each test, the agents were completely
discharged into the mist cloud before the area was secured and the explosion was initiated. The
water mist systems had no characteristic discharge time, so water was applied to the mist cloud
the longest. Further experimental details are reported elsewhere [8,9].

Small scale aerosol studies were conducted using a rotary atomizer and JP5 kerosene. The
kerosene was purchased from Putuxent River Naval Air Station. Figure 1 shows a diagram of
the test apparatus used to generate the aerosols. The aluminum test stand has a trough and splash
guards to contain aerosol generation. An aluminum rotary disk, 3.0 inches (7.6 cm) in diameter
and 0.50 inches (1.3 cm) thick, is located at the center of the test stand. The disk has a cavity 1.0
inch (2.5 cm) in diameter and approximately 0.25 inches (0.63 cm) deep in the center with four
radial holes. The radial holes are spaced 90 degrees apart and were drilled from the outer rim of
the disk to the center of the cavity. The Figure shows fluid is delivered to the center of the disk
and a motor creates centrifugal forces that push the fluid through the radial holes, producing an
aerosol. A propane fed Bunsen burner, located 5.2 inches (13 cm) from the edge of the disk, is
the aerosol ignition source. Each aerosol system is studied over a range of rotational speeds
(1000 16000 rpm) and liquid flow rates to the disk (50 200 mL/min).

2 Data Acquisition
Tachometer
Pump
Disk
Motor
Motor
Splash
Guard
Trough
Burner
Igniter
Video


Figure 1. Diagram of rotary atomizer apparatus

A stainless steel thermocouple assembly was designed to arrange an ensemble of K type
thermocouples to detect and measure aerosol ignition and flame propagation. The horseshoe
shaped (260

degree) thermocouple assembly is 0.25 inch (0.63 cm) stainless steel and sits in the
trough of the pan. Ten K type thermocouples have been situated approximately 5.0 inches (13
cm) apart, with the first thermocouple 3.5 inches (8.9 cm) from the burner. A video system has
also been implemented to visually monitor both the ignition of the fuel mist and propagation of
the ignited flame. The video system consists of three cameras (color, IR, and black and white).

A similar apparatus was coupled to a droplet size analyzer (Malvern Spraytec Malvern
Instruments Inc., Southborough, MA) to provide droplet size distribution characteristics of the
mist generated. When the disk reached a steady rotational speed, liquid was pumped to the
center of the disk at a constant flow rate for approximately 3 to 5 seconds. The Spraytech,
configured for continuous mode, took size distribution measurements during this period. Further
apparatus and experimental details can be found elsewhere [10,11].


RESULTS

Aboard US Navy platforms, 2190 TEP functions as both a hydraulic fluid and lube oil. In many
instances 2190 TEP is operating at elevated temperatures and pressures that can increase its risk
of being aerosolized and becoming flammable. In the large-scale explosion tests aboard the ex-
USS Shadwell, 2190 TEP was heated to 50
o
C and pressurized to 1450 psi. Figure 2 shows the
explosion overpressures measured for a 65 second continuous spray of 2190 TEP. The spray
was generated using a single nozzle in the nozzle array and it was determined the mist flow rate
was approximately 2.45 liters/min. The results of the three baseline explosions show the average
overpressures generated were about 5.5 kPa.


3 Figure 3 shows the results of the explosion mitigation experiments using fire extinguishers. The
objective of testing these extinguishing agents was to determine if existing agents could prevent
or mitigate an explosion after the aerosol cloud was detected. The Figure shows that PKP was
the most effective by reducing the overpressures to less than a third of the baseline explosions.
In some instances it prevented the ignition of the fuel mist entirely. It is believed ignition was
prevented by the increase in effective breakdown voltage of the air. The CO
2
extinguisher
showed a 22% reduction in the explosion overpressures. The cause of mitigation may be
attributed to the dilution of the available oxygen needed for combustion. The experiments
conducted with water and AFFF are not shown because they were essentially ineffective in
reducing explosion overpressures.
kPa


Elapsed Time (sec)
Figure 2. Baseline 2190 TEP Aerosol Explosions


Figure 3. 2190 TEP Aerosol Mitigation Explosions
Elapsed Time (sec)
kPa
Baseline
AFFF
CO
2
PKP


4
erosol/air mixtures are complex systems because the vapor concentration is dependent on

Figure 4. Steady Combustion Temperature of JP5 measured at the first thermocouple of the
a