Key Findings of Experimental and Theoretical Studies on

Forced Mitigation System for an LNG Spill Emergency

Byung Kyu Kim, Ray A. Mentzer, M. Sam Mannan*

Mary Kay O’Connor Process Safety Center, Artie McFerrin Department of Chemical Engineering,
Texas A&M University System, College Station, Texas 77843-3122, USA
*Corresponding author. Tel.: +1 979 862 3985. E-mail address: [email protected] (M.S. Mannan)

Abstract

Mitigation measures have been categorized as one of the essential elements that require further
investigation in determining the potential impact on public safety from an LNG spill. Several outdoor
LNG spill experiments were conducted at the Brayton Fire Training Field to verify the effectiveness of
the direct application of a water spray on LNG vapors as part of LNG Safety and Spill Response
research at the Mary Kay O’Connor Process Safety Center (MKOPSC) at Texas A&M University
(TAMU). Theoretical modeling was conducted of LNG forced mitigation using an upward-oriented
conical water spray application. The computational fluid dynamics (CFD) model had been applied to
investigate the dominant physical mechanisms involved in LNG forced mitigation and essential key
design parameters were evaluated. Findings from the experimental and theoretical analysis on the LNG
forced mitigation are summarized. Finally, recommendations on designing an effective forced mitigation
system and potential hazards of applying the water curtain system are discussed. This work aims to
provide a comprehensive understanding of the forced mitigation system with direct application to LNG
facilities.

1. Introduction

The liquefaction capacity of natural gas is expected to double as the demand for natural gas
expands substantially [1, 2]. The new emerging LNG supply from North America will also play a
significant role in reorganizing the world gas supply chain in the near future [3]. With the transformation
from an importer to exporter of LNG in North America, the new challenges rise from introduction of
additional hazards from the liquefaction process and regulatory compliance to ensure safe operation [4].
NFPA 59A revised in 2013 requires all the new LNG facilities and any facilities that go through major
modifications to conduct quantitative risk assessments to ensure that the facilities do not impose beyond
tolerable risk to the nearby communities [5]. It is recommended to apply additional safety measures to
meet the risk criteria. The effects of various mitigation techniques applicable to the LNG facilities still
require further investigation to minimize the uncertainty in determining the impact of an LNG spill on
the public safety and security [6]. This paper summarizes the research of water spray applications
conducted at the Mary Kay O’Connor Process Safety Center (MKOPSC) at Texas A&M University
(TAMU) as part of the LNG Safety Research Program.

2. Previous water spray research on LNG vapor clouds

The US Coast Guard conducted a small-scale LNG spill experiment to simulate a spill on an
LNG transport ship in 1976 [7, 8]. It was concluded from the test results that the water spray curtain was
effective in reducing the methane concentration and that the mechanical turbulence induced from the
water droplets improved the mixing of the LNG vapors, enhancing the vapor dilution.
 The Gas Research Institute conducted three-phase research, which consisted of theoretical
analysis, small-scale spill tests, and wind tunnel experiments [9]. The downward spray application
showed a reduction in concentration to 2–5% by volume in the downwind region. The upward sprays
had diluted vapor concentrations approximately to 1 to 2% by volume; however, the effectiveness of
vapor concentration reduction was not consistent with the theoretical estimation. It was concluded from
the wind tunnel experiments that the upward sprays and downward-inclined sprays were the only two
applications, which showed an effective dilution.
Meroney and Neff (1985) provided forced mitigation effects for both box and slab models
assuming that the local entrainments increase with the application of the water curtain [10]. The
enhanced air entrainment effects were incorporated into the model by applying a multiplicative or
additive factor to the regular entrainment rates. The integral-type modeling provided a quick prediction
compared to dense gas dispersion, and has shown reasonable predictions when compared to small-scale
experiments. However, the integral-models present the fluid system in a way that limits the
understanding of the complex interactions of water droplets with the gas flow. Also, there had been
some concerns over the under-prediction from the integral-models in certain scenarios, mainly because
of the simplified assumption [11].

3. Experimental work from MKOPSC

To investigate the complex physical mechanisms of LNG forced dispersion, the Mary Kay
O’Connor Process Safety Center (MKOPSC) conducted outdoor LNG spill experiments at the Brayton
Fire Training Field to verify the dominant mechanisms and evaluate the effectiveness in dilution the
LNG vapor concentration [12, 13].

Fig. 1. LNG dispersion test: (a) without water curtain, (b) with water curtain (full cone) [14]*

Rana, Guo & Mannan (2010) discuss one of the experimental setups that had been conducted in
2007 [14]. Fig. 1 shows the experimental work of the water spray application on LNG vapor clouds.
Two different types of water spray nozzles were mainly compared: 60° full-cone spray nozzle and 180°
flat-fan spray nozzle [15]. The full-cone type produces finer droplets, while the flat-fan type creates a
thin barrier in the vicinity of the nozzle and very course droplets as the bulk water travel and break into
smaller droplets. The full-cone spray type provides more effective mixing with the air through high air
entrainments (Fig. 2(a)). The full-cone type showed an effective dilution at all elevations. The flat-fan
type creates a physical barrier in the pathway of LNG vapors and provides high momentum from the
droplets created from the water pressure. The flat-fan type pushed the LNG vapors to higher elevations,
resulting in an increase in concentration behind the water spray region (Fig. 2(b)).

*
Reprinted figures with permission from Rana, M., Guo, Y., & Mannan, M. S. (2010). Use of water spray curtain to disperse
LNG vapor clouds. Journal of Loss Prevention in the Process Industries, 23(1), 77-78. Copyright 2010 Elsevier.
 Fig. 2. Downwind concentration at three different heights with and without
(a) full-cone water curtain and (b) flat fan type [14]*

The temperature changes collected from the experiments suggested that the water droplets
provided heat transfer to the LNG vapors. The overall heat transfer evaluated from the experimental
results showed that the droplets produced from the full-cone type nozzle provided more heat transfer to
the LNG vapors than the flat-fan application (Fig. 3).

Fig. 3. Calculated heat loss by water curtain (Modified from [14])

The full-cone type nozzle produces smaller droplets, which increases the surface area. It was
concluded that the mixing effects through the entrained air promotes the LNG vapors most effectively.
The heat transfer from the droplets ensures that the vapors are sufficiently warmed to become positively
buoyant.

4. Modeling approach from MKOPSC

The CFD tool was adopted to conduct a comprehensive assessment of the effects of the water
curtain applications on LNG vapor clouds. In the present work, ANSYS Fluent was used to simulate the
Eulerian–Lagrangian spray modeling coupled with the LNG vapor flows. The essential parameters for
simulating the detailed description of LNG flow and interactions of the droplet-vapor system were
calibrated against the experimental results obtained from the LNG spill work conducted at the Brayton

*
Reprinted figures with permission from Rana, M., Guo, Y., & Mannan, M. S. (2010). Use of water spray curtain to disperse
LNG vapor clouds. Journal of Loss Prevention in the Process Industries, 23(1), 77-78. Copyright 2010 Elsevier.
Fire Training Field. The physical parameters that are essential inputs in setting up the CFD simulation of
the water-spray LNG system are discussed elsewhere [16].

Fig. 4. (a) Volumetric concentration for natural/forced(RM=5.32)/forced(RM=12.76) dispersion and (b)

Fig. 4 shows the dilution effects of two different spray settings applied (ܴெ = 5.32, and 12.76)
and heat transfer involved for various droplet temperatures simulated using the LNG forced dispersion
model. ܴெ is the momentum ratio, which is the ratio of the momentum imparted from the droplets to the
momentum of the gas clouds [17]. Different levels of safety distance reduction were observed for
various momentum ratios applied from the water spray as shown in Fig. 4 (a). As the water nozzle
discharged more momentum to the LNG vapors, the dilution effects increased. Also, the heat transfer
rate increased as the droplet temperature increased as shown in Fig. 4 (b). The effectiveness of dilution
improved as well, as more heat transfer was involved at different elevations (z = 1.2 and 2.1 m). These
results indicate that the LNG forced dispersion modeling using the Eulerian-Lagrangian spray model is
capable of simulating the dilution effects and influences on LNG vapor behavior by taking into account
the detailed droplet characteristics.

Fig. 5. Dilution effects from different mass flow rates and droplet velocities

Using LNG forced dispersion modeling, the physical mechanisms of the LNG forced dispersion
were investigated. The effects of momentum imparting from the droplets to the air-vapor mixture,

**
Reprinted figures with permission from Kim, B. K., Ng, D., Mentzer, R. A., & Mannan, M. S. (2012). Modeling of water
spray application in the forced dispersion of LNG vapor cloud using a combined Eulerian-Lagrangian approach. Industrial &
Engineering Chemistry Research, 51, 13803-13814. Copyright 2012 American Chemical Society.
thermal effects between two phases (vapor/droplet), and different level of air entrainments on the LNG
vapor behavior were evaluated.
The dilution effects of LNG forced dispersion were evaluated with different ranges of droplet
velocity (3–30 m/s) at two different mass flow rates (1 and 5 kg/s) as shown in Fig. 5. The dilution
effects increased as the droplet velocity increased. However, the spray with mass flow rate of 1 kg/s
showed only limited dilution effects even with a high droplet velocity applied. This is because even with
the high droplet velocity to provide effective mixing of the air-vapor mixture, the insufficient water
introduced from the nozzle limited the effective dilution of LNG vapors. This result indicates that both
of the design elements, the droplet velocity and mass flow rate, must be set above a certain range to
ensure an effective mitigation effect. From the water spray setting applied in this research, it was above
3 kg/s of mass flow rate with 12–15 m/s droplet velocity.
The thermal effects from the water droplets to the vapor clouds, particularly for cold gas releases,
can enhance dissipation of the vapor cloud by warming to gain positive buoyancy [18]. The temperature
of the water droplet was varied from 283 to 313 K to evaluate the turbulence effects induced from
different levels of heat transfer effects of the droplets, as shown in Fig. 6.

Fig. 6. (a) Turbulence kinetic energy and (b) vapor concentration from
various droplet temperature applications at 8 m elevation

The turbulence kinetic energy showed the highest level being induced from the water droplet
with highest temperature (313 K). The 8 m elevation was chosen to show the vapor behavior at high
elevation. The concentration also increased significantly with the highest water droplet temperature,
indicating that the vapors are sufficiently warmed to travel upward to the atmosphere. This result shows
that the thermal effects from the droplet enhance the turbulent flow within the vapor clouds, resulting in
an effective dilution of LNG vapors.
The air entrainment rate is determined by the spray characteristics: droplet sizes, droplet velocity,
spray location, and configuration. Rana (2009) have verified that the entrained air rate into the spray
increases as the nozzle angle size increases, while the operating pressure does not affect the rate of the
air entrainment significantly [13]. A total of four different angle sizes (30°, 45°, 60°, and 75°) were
simulated, as shown in Fig. 7.
 Fig. 7. (a) Turbulence kinetic energy and (b) vapor concentration from
various nozzle angle size applications at 8 m elevation

The turbulence induced at 8 m elevation showed that the turbulence effects from various angle
sizes were similar. On the other hand, the vapor concentration showed an increasing trend for the largest
nozzle angle (75°) at 8 m elevation, while the concentration for the rest of other applications showed
similar results once passed through the water spray. This indicates that the air entrainment setting that
could induce a large amount of air is more effective in dispersing the LNG vapors through effective
mixing, with turbulence effects playing a less significant role.

Fig. 8. (a) Heat transfer rate from droplets to air-vapor mixture and (b) safety distance
evaluated from various installation distances [19] ***

The size of droplets is determined by the water pressure and nozzle design, which is an intrinsic
characteristic for a specific nozzle design. Different droplet sizes were simulated to evaluate the heat
transfer rates (Fig. 8 (a)). The total heat transfer clearly indicated that the total amount of heat
transferred from the droplet to LNG increased as the flow rate increased. However, the heat transfer
from the surface decreased significantly as the droplet size increased, because of the larger droplets
providing smaller surface area. The installation distance from the LNG source was investigated (Fig.
8(b)). The safe distance, defined as where the LFL is reduced by 50%, was reduced significantly as the
water curtain was installed closer to the LNG source. These results can be used to evaluate the impact of
the distance factor in calculating the dilution effects.

***
Reprinted figures with permission from Kim, Byung Kyu, Ng, Dedy, Mentzer, Ray A., & Mannan, M. Sam. (2013). Key
parametric analysis on designing an effective forced mitigation system for LNG spill emergency. Journal of Loss Prevention
in the Process Industries. doi: http://dx.doi.org/10.1016/j.jlp.2013.01.007. Copyright 2013 Elsevier.
 Fig. 9. Vapor concentration increase [%] at 1 m and 6 m with (a) modified nozzle installations (tilted at
30°, 45°, 60°, and 90°) and (b) different nozzle sizes (30°, 45°, 60°, and 75°) [19] ***

The tilted installation designs at various nozzle angles were evaluated (30°, 45°, 60°, and 90°).
Per Fig. 9 (a), the vertical installation showed the most effective dilution at lower heights (1 m), while at
higher elevation (6 m), the nozzles tilted at 60° showed the largest increase of vapor concentration. This
indicates that the nozzles tilted at 60° discharge the droplets that could effectively envelop the LNG
vapors and enhance dilution of the air-vapor mixture. Various levels of air entrainments were applied to
evaluate the concentration increase in the atmosphere are plotted in Fig. 9 (b). The angle sizes of the
nozzle were modified in 30°, 45°, 60°, and 75° to simulate different levels of entrainments. The smaller
angle nozzles were more effective in pushing the vapors to lower heights. On the other hand, the nozzle
with largest angle size (75°) showed the highest increase of vapor concentration at higher elevation (6
m). This indicates that the nozzle with the largest entrainment air setting provides enhanced mixing that
could effectively dilute the LNG vapors to travel upward to the atmosphere.

5. Conclusion

The engineering guidelines in designing an effective water spray system for an LNG facility are
currently lacking mainly due to the existing gaps in the experimental results and limited theoretical
modeling developed using integral-type models. The CFD code had been applied in evaluating the
forced dispersion of LNG vapors using a water curtain application. The proposed model uses a
Eulerian–Lagrangian approach, which takes into account the detailed characteristics of the water
droplets in evaluating the effects on the LNG vapor behavior. The model shows a promising solution in
investigating the complex interaction of the droplet-vapor system. The CFD code also can serve as a
rigorous tool in evaluating the design for a site-specified mitigation system, to predict any unforeseen
hazards that might rise from under-sized mitigation. The modeling approach applied in this work can
evaluate the risk reduction effects from applying the water spray application, which can directly be
implemented for quantitative risk assessments, as indicated by NFPA 59A.

References

1. International Energy Outlook 2011 (US Energy Information Administration), 2011, US Energy
Information Administration (EIA): Washington, DC.
***
Reprinted figures with permission from Kim, Byung Kyu, Ng, Dedy, Mentzer, Ray A., & Mannan, M. Sam. (2013). Key
parametric analysis on designing an effective forced mitigation system for LNG spill emergency. Journal of Loss Prevention
in the Process Industries. doi: http://dx.doi.org/10.1016/j.jlp.2013.01.007. Copyright 2013 Elsevier.
2. World Energy Outlook 2011: Are we entering a golden age of gas? 2011, International Energy
Agency: Paris, France.
3. Henderson, J., The potential impact of North American LNG exports, 2012, The Oxford Institute
for Energy Studies: Oxford, UK.
4. Chosnek, J. and V. Edwards. From LNG import to exports: Process safety and regulatory
challenges. Presented in 15th Annual Symposium Mary Kay O'Connor Process Safety Center.
2012. College Station, TX.
5. NFPA 59A: Standard for the Production, Storage, and Handling of Liquefied Natural Gas (LNG),
2013, National Fire Protection Agency: Quincy, MA.
6. Public Safety Consequences of a Terrorist Attack on a Tanker Carrying Liquefied Natural Gas
Need Clarification (Maritime Security GAO-07-316), United States Government Accountability
Office, 2007: Washington, D.C.
7. Brown, L., et al., Small scale tests on control methods for some liquefied natural gas hazards,
1976, US Coast Guard: Washington D.C.
8. Martinsen, W.E., S.P. Muhlenkamp, and J. Olson, Disperse LNG vapors with water.
Hydrocarbon Processing, 1977: p. 260-267.
9. Heskestad, G., et al. Effectiveness of water spray curtains in dispersing LNG vapor clouds. in In
proceedings of the American Gas Association Transmission Conference Paper, Paper No 83-T-
69, 169-183. 1983. Seattle, Washington.
10. Meroney, R. and D. Neff, Numerical modelling of water spray barriers for dispersing dense
gases. Boundary - Layer Meteorology, 1985. 31(3): p. 233-247.
11. Gavelli, F., E. Bullister, and H. Kytomaa, Application of CFD (Fluent) to LNG spills into
geometrically complex environments. Journal of Hazardous Materials, 2008. 159(1): p. 158-168.
12. MKOPSC, Data Report of MKOPSC LNG Spill Tests: 2005-2009, M.S. Mannan, Editor 2010,
Mary Kay O'Connor Process Safety Center, Texas A&M Univeristy: College Station, TX.
13. Rana, M., Forced dispersion of liquefied natural gas vapor clouds with water spray curtain
application, Chemical Engineering, Texas A&M University, PhD Dissertation, 2009. p. 232.
14. Rana, M., Y. Guo, and M.S. Mannan, Use of water spray curtain to disperse LNG vapor clouds.
Journal of Loss Prevention in the Process Industries, 2010. 23(1): p. 77-78.
15. Rana, M. and M.S. Mannan, Forced dispersion of LNG vapor with water curtain. Journal of Loss
Prevention in the Process Industries, 2010. 23(6): p. 768-772.
16. Kim, B.K., et al., Modeling of water spray application in the forced dispersion of LNG vapor
cloud using a combined Eulerian-Lagrangian approach. Industrial & Engineering Chemistry
Research, 2012. 51: p. 13803-13814.
17. Hald, K., et al., Heavy gas dispersion by water spray curtains: A research methodology. Journal
of Loss Prevention in the Process Industries, 2005. 18(4–6): p. 506-511.
18. CCPS, Guideline for postrelease mitigation technology in the chemical process industry. 1997,
New York: Center for Chemical Process Safety, American Institute of Chemical Engineers.
19. Kim, B.K., et al., Key parametric analysis on designing an effective forced mitigation system for
LNG spill emergency. Journal of Loss Prevention in the Process Industries, 2013(0).