I.A.

Karimi and Rajagopalan Srinivasan (Editors), Proceedings of the 11th International

Symposium on Process Systems Engineering, 15-19 July 2012, Singapore.
© 2012 Elsevier B.V. All rights reserved.

Quantitative Risk Analysis of New Energy Stations

by CFD-Based Explosion Simulation
Seungkyu Dana, Hyunmin Kima, Dongil Shinb, En Sup Yoona
a
Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul 151-742, Korea
b
Myongji University, Nam-dong, Cheoin-gu, Gyeonggi-do, 449-728, Korea

Abstract
In this research, we compared the safety of mixture fuel which is composed of 20% new
energy and 80% existing gas against existing fuel in the perspective of explosion risk
that would be the biggest concern in the operation of new-energy stations. The new
mixture will be used in the existing station in Korea without any changing the structure.
The explosion risk was analyzed and compared by three representative models:
empirical model, phenomenological model, and a CFD-based model, in increased model
complexity and computational efforts. Overpressures of explosion of mixture and
existing fuel, respectively, showed similar results, in all three models. Thus, it seems
like there is no additional risk in using mixture at the existing refueling facilities. CFD-
based explosion simulation was useful in finding out the exact overpressure distribution
and installation of prevention equipment if necessary. The final results are being
adopted as part of new safety regulations for the new and existing mixture stations to be
operated in Korea.

Keywords: Explosion Simulation; Quantitative Risk Analysis; Gas Explosion Modeling;

HCNG Mixture Fuel; DME-LPG Mixture Fuel

1. Introduction
Since the environmental pollution problems and climate issues related to ever-
increasing energy use have been emerging all over the world, developing efficient
renewable and/or alternative energy sources are considered as top-priority research
agenda [BP, 2009]. Among them, DME (dimethyl ether) has emerged as an attractive,
promising, alternative clean fuel for transportation, power generation and household.
Korean Government, with the leading role of KOGAS, is promoting commercial
introduction of this new clean energy, DME. This research has been conducted as part
of the KOGAS's core technology development activities covering from DME synthesis
from feedstock, such as natural gas, through utilization of DME as fuel by end users.
Also CNG, one of alternative energy, is used by fuel of bus in Korea. And it is going to
be used with hydrogen to prevent the exhaustion of natural gas and air pollution after
combustion.
Using the new mixtures still deal with inflammable gas. Therefore, when dealing with
new mixtures at the conventional filling stations, safety managements do not become an
automatically solved issue. Since the severe accident at Bucheon LPG filling station
[Park et al., 2006], operation, expansion and construction of LPG stations are heavily
regulated and controlled in Korea. Thus, converting and operating conventional filling
stations into new energy stations must be thoroughly analyzed and studied to get the
acceptance and approval by local communities and government authorities. The
development of a strategy for improving gas explosion models is also highly desirable
306 Seungkyu Dan, Hyunmin Kim, Dongil Shin, En Sup Yoon

[Jiang et al., 2001] as the use of gas as energy sources is expected to grow in near
future.

2. Comparison of explosion simulation models

We need an accurate assessment about possible accidents to reduce the loss caused by
overpressure shaped by gas explosion. There are many models ranging from
calculations using a simple expression to complex numerical formulas using CFD. The
more accurate the simulation results, the better confidence we may have in expecting
the real situations and the real risk and enforcing the results out of the analysis.
However, selecting and using appropriate models for the problem is important as the
information requirements and efforts to run the models increase with its model accuracy.
The estimation of the explosion risk have been conducted by following three
approaches: Empirical models are based on correlations obtained from analysis of
experimental data; and phenomenological models are simplified physical models, which
is using simple computer software based on emprical correlations, which seek to
represent only the essential physics of explosions; and using computer simulation
program based on Computational Fluid Dynamics (CFD) models find numerical
solutions to the partial differential equations governing the explosion process [Lea,
2002].
2.1. Empirical Model
Empirical models are based on correlations obtained from analysis of experimental data.
Among these models, the TNT equivalency model is a representative empirical model
and it is very simple and easy to use. TNT equivalency model is based on the
assumption that gas explosion is similarly to an explosive of high efficiency. There are
many differences between gas explosion and solid explosion, but using the utility factor
helps correcting the gap. Eqs. 1 and 2 below calculate overpressure from the explosion
effect of TNT. The calculation of scaled distance is necessary to obtain the overpressure
out of the calculation from Eq. 1.
K ME c
W (1)
E TNT
RG
ZG 1/3
(2)
W
where W is the equivalent mass of TNT (kg or lb), η is an empirical explosion
efficiency (unitless), M is the mass of hydrocarbon (kg or lb), Ec is the heat of
combustion of flammable gas (kJ/kg or Btu/lb), and ETNT is the heat of combusition of
TNT (4437-4765 kJ/kg or 1943-2049 Btu/lb) [CCPS, 1999].
2.2. Phenomenological Model
Phenomenological models are simplified physical models, which seek to represent only
the essential physics of explosions. It is made by simplication of the geometry, and it
considers a simplified condition for the simulation: wind, temperature, atmospheric
pressure, etc. PHAST, which has been selected as the phenomenological model for our
research, is a general tool for consequency analysis examining the progress of a
chemical process incident from initial release through formation of a cloud or pool to
final dispersion; calculating concentration, fire radiation, toxicity and explosion
overpressure. PHAST is a comprehensive hazard analysis package, applicable to all
stages of design and operation across a range of process and chemical industry sectors.
Quantitative Risk Analysis of New Energy Stations by CFD-Based Explosion Simulation 307

The VCE in the PHAST is calculating the overpressure based on TNT Equivalency
model and is calculated by below Eqs. 3. [DNV, 2007].
0 . 2518 (log z )  2 . 0225 (log z )  5 . 8095
2
log P0 (3)
2.3. Simulation Based on the CFD Model
Computer simulation programs based on CFD (Computational Fluid Dynamics) models
solve patial differential equations, based on Navier-Stokes equations that govern the
fliud flow, to calculate the phenomena controlling the explosion process. CFD
simulations can offer insight into the flow behavior in situations where it is impractical
or impossible to carry out experiments. Its use is being widely accepted for risk analysis
of explosion or dispersion.
The using equations in each model are based on Navier-Stokes equation but there are
some different ways for calculating results.
In this research, we utilized FLACS (FLame ACceleration Simulator) made by Gexcon
in Norway, which is a tool for dispersion and explosion simulations based on CFD
[Gexcon, 2010]. CFD-based complicated models are widely acknowledged for enabling
to show the necessary accuracy required for the adoption of simulation results for
detailed safety reviews and corresponding upgraded design of refilling stations.
For the comparion of aforementioned three models and the proper risk control, we need
to predict and estimate exact damages of existing and new energy refueling facilities’
explosion by running the simulation for a 3D structure of the refilling station.

Figure 1: Comparison of Overpressure by Propane Explosion for FLACS and Experiment

[Gexcon, 2010]

3. Explosion scenario
DME-LPG refilling stations will begin operating by setting two types of tank. One type
of tank is added as new tank which can store 5 ton of DME-LPG mixture gas, and it is
located nearby an existing LPG tank. Another way is using an existing tank which can
store 40 ton of DME-LPG. Installing tanks is separated by three types: one is
pressurized on the ground, another is a refrigeration method on the ground, and the
other is pressurized under the ground. Among them, the pressurized tank on the ground
was selected as the representative case in this research.
Once a release occurs caused by any accidents, flammable gases are shaped around the
refuelling facility. After the release, fire or explosion could happen if an ignition is
follwed by that accident, too. Between fire and explosion, we focused on the
overpressure hazard out of the explosion.
In the case of CNG or HCNG, because the state of stored fuel is gas, the possible
explosion by flammable gas is vapor cloud explosion, and BLEVE is not occurred.
Instead of BLEVE, it is possible that the tank is ruptured physically.
308 Seungkyu Dan, Hyunmin Kim, Dongil Shin, En Sup Yoon

4. Results

Figure 2: The results of explosion simulation by FLACS

Table 1: Summary of Results (DME-LPG)G
PHAST
Overpressure TNT Equivalency Model FLACS (VCE)
VCE BLEVE
5.8 bar at 10 m 1 bar at 10 m 18 bar at 10 m 0.026 bar
5 ton LPG at 0.43sec after the
0.13 bar at 50 m 0.4 bar at 50 m 0.1 bar at 50 m ignition

5.3 bar at 10 m 1 bar at 10 m 18 bar at 10 m

5 ton DME-LPG
0.11 bar at 50 m 0.39 bar at 50 m 0.09 bar at 50 m
Over 10 bar at 10 m 1 bar at 10 m 18 bar at 10 m
40 ton LPG
0.76 bar at 50 m 1 bar at 50 m 0.4 bar at 50 m
Over 10 bar at 10 m 1 bar at 10 m 18 bar at 10 m
40 ton DME-LPG
0.70 bar at 50 m 1 bar at 50 m 0.38 bar at 50 m

Table 2: Summary of Results (HCNG)G

PHAST
Overpressure TNT Equivalency Model FLACS (VCE)
(VCE)
Over 10 bar at 10 m 1 bar at 10 m 0.019
5 ton CNG at 17sec after the ignition
0.4 bar at 50 m 0.87 bar at 50 m
Over 10 bar at 10 m 1 bar at 10 m 0.016
5 ton HCNG at 0.34sec after the ignition
0.42 bar at 50 m 0.97 bar at 50 m
Over 10 bar at 10 m 1 bar at 10 m
40 ton CNG
1.79 bar at 50 m 0.87 bar at 100 m
40 ton HCNG Over 10 bar at 10 m 1 bar at 10 m
Quantitative Risk Analysis of New Energy Stations by CFD-Based Explosion Simulation 309

1.9 bar at 50 m 0.97 bar at 100 m

Results of simulations by three models were summarized as Table 1 and Table 2. In the
case of LPG and DME-LPG, compared against the TNT equivalency model, PHAST
results show a little lower value near to the ignition, but it shows a little bigger value far
from the ignition. The FLACS was simulated only for VCE, and comparison of its
BLEVE results is still being prepared. In the results of CNG and HCNG, although the
results of simulation about hcng mixture is a little higher than the cng, its values are
almost same.

5. Conclusions
This study has analyzed and compared the results of explosion risk prediction by three
representative models for estimation of explosion consequence, such as the TNT
equivalency model, PHAST, and FLACS. First, the three models were used and
compared in the explosion risk prediction for the conventional LPG station. When the
overpressures of gas explosion for both LPG and DME-LPG mixtures are calculated,
there was less difference between the simple TNT equivalency model and PHAST
simulation requiring more parameters. Overpressure plots of LPG and DME-LPG
mixtures relative to distance were similar to each other. Likewise, in the case of the
CNG and HCNG mixtures, the calculated results show almost same overpressure of
explosion. So when the HCNG mixtures using in the CNG filling station, it is expected
that almost no additional risk.
CFD-based complicated models only showed the necessary accuracy required for the
adoption of the simulation results for safety reviews and corresponding upgrade design
of refilling stations. Even though some experimental data are available for the explosion
characteristics of DME [Mogi and Horiguchi, 2009], we needed to simulate it for the
given 3D structure of the recharging station to get the precise view of the explosion
consequence. We compared the difference of current LPG station with the one using
DME-LPG mixture from the obtained result and investigated how we can reduce the
explosion risk effectively, including installation of additional devices for protection.
The final results are being adopted as part of new safety regulations for the DME-LPG
stations to be operated in Korea.
Also, we are coding a model of explosion by matlab, but it is not perfom properly. So it
is required to revise some code to make accurate results of explosion using our model.

References
BP Statistical Review of World Energy 2009, http://www.bp.com/productlanding.do?cate
goryId=6929&contentId=7044622
Center for Chemical Process Safety (CCPS), 1999, Guidelines for Chemical Process Quantitative
Risk Analysis, 2nd Ed., American Institute of Chemical Engineers.
DNV. Phast User Manual, 2007 (http://www.dnv.com/services/software/products/safeti/
safetihazardanalysis/phast.asp).
Gexcon, 2010, FLACS v9.0 User’s Manual (www.gexcon.com).
Lea, C. J. A, 2002, Review of the State-of-the-Art in Gas Explosion Modelling, Health & Safety
Laboratory.
Mogi, T. and Horiguchi, S. , 2009, Explosion and detonation characteristics of dimethyl ether, J.
Hazardous Materials, 164, 114-119.
Park, K., Mannan, M. S., Jo, Y.-D., Kim, J.-Y., Keren, N. and Wang, Y., 2006, Incident analysis
of Bucheon LPG filling station pool fire and BLEVE, J. Hazardous Materials, 137(1), 62-67.
Jiang, J., Liu, Z. G. and Kim, A. K. , 2001, Comparison of blast prediction models for vapor cloud
explosion, National Research Council Canada Report, NRCC-44715.