Liquefied Natural Gas Vapor Hazard Mitigation with Expansion Foam

May 3, 2016 - However, an LNG spill may cause a catastrophic incident as demonstrated by the Cleveland East Ohio Gas explosion, which killed 130 peopl...
0 downloads 0 Views 5MB Size
Download PDF
Article pubs.acs.org/IECR

Liquefied Natural Gas Vapor Hazard Mitigation with Expansion Foam Using a Research-Scale Foam Generator Bin Zhang, Brian Harding, Yi Liu, and M. Sam Mannan* Mary Kay O’Connor Process Safety Center, Artie McFerrin Department of Chemical Engineering, Texas A&M University, College Station, Texas 77843-3122, United States ABSTRACT: The flammable vapor cloud is the primary hazard caused by a liquefied natural gas (LNG) spill on land. If it is not properly mitigated, an ignition of the vapor cloud will result in a fire or explosion hazard. High expansion foam is recommended by NFPA 11 and NFPA 471 for LNG spill hazards mitigation. This work studied the physical interaction of the LNG and expansion foam system using a foam generator and a foam test apparatus that are built in-house. The performance of the foam generator was characterized in terms of the foam expansion ratio and generation rate. The temperature profile in the foam zone quantitatively confirmed the warming effect of foam on LNG vapor. The foam breaking rate was determined at the initial stage and steady state. The boil-off effect was studied quantitatively with more details using this novel foam generator. The vapor channel formation and vapor concentration were also investigated for the mitigation effect.

1. INTRODUCTION Natural gas has become more popular as a fuel and a feedstock for the chemical industry. This is because it has an abundance of reserves and emits much less carbon dioxide, nitrogen oxides, sulfur dioxide, and particulates compared with other fossil fuels.1 In the USA, natural gas was previously imported, but in the future, it will be exported due to the production of shale gas. Many export terminals have been approved or proposed.2 The liquefaction of natural gas provides benefits for both storage and transportation, because liquefied natural gas (LNG) has a higher energy density and a decreased volume. However, an LNG spill may cause a catastrophic incident as demonstrated by the Cleveland East Ohio Gas explosion, which killed 130 people.3 One of the main hazards is the LNG vapor hazard after a spill on land, which is the vapor cloud on the ground level generated from the LNG spillage pool. The spilled LNG continues to vaporize due to a low boiling temperature of 112 K. 4 Because the volume expands approximately 600 times during LNG vaporization, a huge vapor cloud can be generated by a small spill. LNG vapor exhibits dense gas behavior at its boiling temperature, and the vapor cloud will remain at ground level until it is heated enough to increase its buoyancy. Because the vapor cloud is flammable between a volumetric concentration of approximately 5% and 15%, a fire or explosion may occur if the vapor hazard is not properly mitigated. High expansion foam is recommend by NFPA 11 and NFPA 471 to mitigate LNG spill hazards.5,6 Previous work has proved that high expansion foam is effective for mitigating LNG vapor hazard.7,8 These works focused on determining the vapor concentration reduction due to high expansion foam application. However, the mitigation mechanism has barely been studied. It was commonly believed that the mitigation © 2016 American Chemical Society

effect only relied on the warming effect of foam, which increases LNG vapor temperature.9 Zhang et al.’s work concluded that foam could also reduce the vaporization rate of LNG due to the blanketing effect to control the hazard.4 The lack of understanding of the physical interaction between the foam and LNG system demands additional experimental work. In this work, a foam generator was designed and built on the basis of the schematic from NFPA 11.5 Improvements were made to better meet the needs of lab-scale tests. A foam test apparatus was designed and constructed to study the temperature profile in the foam zone, visualization of foam and vapor interaction using a camera, vaporization rate, and vapor concentration. To study the vapor hazard of LNG vapor clouds, liquid nitrogen (LN2) was used as a safe analog for LNG. This substitution is acceptable because liquid nitrogen has properties similar to LNG and follows a precedent set by previous work.4,10 Tests using liquid nitrogen (LN2) were conducted by applying high expansion foam, and the findings of the physical interaction between the foam and LNG system were presented and discussed.

2. EXPERIMENT AND METHODOLOGY 2.1. Experimental Setup. The experimental setup consists of a foam generator and a foam test apparatus as shown in Figure 1. The foam generator was designed to study foam application for LNG hazard mitigation, as well as chemical decontamination foam in another work. Using the foam generator, the foam solution from the tank was pumped to Received: Revised: Accepted: Published: 6018

November 28, 2015 April 2, 2016 May 3, 2016 May 3, 2016 DOI: 10.1021/acs.iecr.5b04535 Ind. Eng. Chem. Res. 2016, 55, 6018−6024

Article

Industrial & Engineering Chemistry Research

Figure 1. Experimental setup in the lab. Figure 3. Schematic diagram of foam test apparatus.

the spray nozzle through the stainless steel hose. The main body of the foam generator is shown in Figure 2. Prior to order to provide accurate measurements at room temperature. On top of the heat chamber, there was a double-containment system made of aluminum to contain liquid nitrogen. The inner container (0.84 m × 0.84 m × 0.1 m) was used to contain liquid nitrogen, and the outer container had a trench-shaped rim to support the foam fence. Salt water was poured into the trench to seal the small gap between the foam fence and the outer container, forcing nitrogen vapors to rise through the foam layer. Salt was added to the water to decrease the melting point due to the freezing-point depression effect of electrolytes. The sponge between the two containers helped prevent salt water from freezing after liquid nitrogen was released into the inner container. The foam fence was a rectangular box made of four pieces of polycarbonate sheets. It was put on top of the outer container to contain expansion foam during the tests. There was a 2.5 cm diameter hole on one side of the foam fence, which allowed the cryogenic hose to be inserted in order to release liquid nitrogen in some tests. Ten pairs of thermocouples (T type TJC300 series, Omega Engineering, USA) were installed inside the foam fence to measure the vapor and foam temperature. An oxygen sensor (MF010-1-LC3, Honeywell, USA) was used on top of the foam fence to monitor the oxygen level, which could indicate the nitrogen concentration indirectly. There was a scale tape attached on each side of the foam fence to indicate the foam height. The foam height in the foam fence was recorded using a video camera during the tests. 2.2. Summary of Tests. There were 17 liquid nitrogen tests with high expansion foam application. The foam concentrate used in this work was C2 High Ex Foam from Tyco fire protection products. The foam solution was prepared by mixing foam concentrate with water according to the recommended proportion ratio, i.e., 2%. The expansion foam ratio, defined as the volumetric ratio between foam and foam solution, ranged from 330 to 680 in this work. Since the foam generator was installed horizontally, the foam was shot horizontally and the deflector and gravity directed the foam into the foam fence. The foam usually hit the middle of one side of the fence and dropped to the bottom. Then, the foam started to spread from the center to the extremities, meaning

Figure 2. Main body of the foam generator.

reaching the spray nozzle, the foam solution went through a pressure gauge, which is used to indicate the flow pressure. The spray nozzle was installed inside of a transparent air cylinder and was positioned in order to spray foam solution onto a conical screen at one end of the air cylinder. At the other end of the air cylinder, the air was dragged and pushed in the direction of the screen by a fan. An iris damper was used to control air flow rate. The foam was generated when air passed through the screen and was entrained in the foam solution to form bubbles. The finished foam was discharged at the end of the air cylinder with a deflector to control the angle of release. More information about design details and improvements compared with the design in NPFA 11 was illustrated in another work.11 The foam test apparatus consists of three main parts: the balance, the liquid nitrogen container, and the foam fence (shown in Figure 1). The detailed information is shown in Figure 3. The balance (Scale WPT/4 300 C7, RadWag, Poland) was located under the fence and the liquid nitrogen container in order to measure the mass of the liquid nitrogen and expansion foam. The mass information can be used to calculate the vaporization rate of liquid nitrogen and the foam expansion ratio. A chamber was built above the balance using insulation material, which contained a heating sheet to control the heat input for vaporization in case it was needed. Besides the chamber, the polystyrene plate between the balance and heat chamber also helped prevent the balance from being heated in 6019

DOI: 10.1021/acs.iecr.5b04535 Ind. Eng. Chem. Res. 2016, 55, 6018−6024

Article

Industrial & Engineering Chemistry Research

the expansion ratio reduced with increasing pump pressure, because high pump pressure resulted in high flow rate of the foam solution. The effect of damper opening was also shown. The damper was set as 12.5%, 25%, 50% and 100% open. If all the air was entrained in the foam, a 100% damper opening should have produced the highest expansion ratio foam at a given pressure, because maximum air flow rate was provided with a 100% damper opening. However, the results indicated that a 50% damper opening generated foam had the highest expansion ratio in most cases. The reason is that the maximum air flow rate provided by the fan was too large for the foam solution to entrain all the air in the foam at the screen. Some of the air escaped into the atmosphere without generating foam. The range of expansion ratio using the current setup was from 300 to 850, which can be extended in either direction by switching the spray nozzle. The volumetric foam generation rate is mainly determined by the total volume of air entrained in the foam. Similarly, a 50% damper opening resulted in the highest foam generation rate in most cases as shown in Figure 5. Due to the increase of foam

the corner was usually covered last. These liquid nitrogen tests were conducted using two methods. Method A released liquid nitrogen into the inner container without the foam fence. Once the inner container was filled after approximately 30 min, the foam fence was put on followed by foam application. Method B had the foam fence in position before liquid nitrogen release. The cryogenic hose was inserted through a hole on one side of the foam fence to release the liquid nitrogen. Once the inner container bottom was fully covered by liquid nitrogen (approximately 4 kg of liquid nitrogen within about a 5 min release), the high expansion foam was applied. Liquid nitrogen continued to be released during and after foam applications until we exhausted the liquid nitrogen in the storage tank. The release process usually took approximately 30 to 60 min. The vaporization rate was higher, and more vapor interacted with the foam using Method B, because the inner container was not fully cooled and provided larger heat flux for vaporization at the time of first foam application. The flash of liquid nitrogen during the release was also a contributing factor for more vapor in Method B.

3. RESULTS AND DISCUSSIONS 3.1. Performance of Foam Generator. The quality of foam can be characterized by two parameters, expansion ratio and foam generation rate. When the foam was produced to fill the foam fence, the mass of the foam was measured using the balance at the bottom. Since approximately 98% of the foam solution was water, water density was used to calculate the foam solution volume. Given the volume of the foam fence, the expansion ratio can be calculated. Because liquid nitrogen continued vaporizing throughout the trials, it was very hard to measure the foam mass accurately for tests where liquid nitrogen was present in the container. Therefore, the same foam test, i.e., using the same foam solution and same foam generator conditions, was conducted without liquid nitrogen to determine the foam mass. The time to apply foam was recorded, and therefore, the foam generation rate was obtained by dividing the foam fence volume with foam application time. The effects of pump pressure and damper opening on expansion ratio are shown in Figure 4. The data points on the left of the dash line were obtained using a small spray nozzle (BETE WL 1), but the equivalent pressure for a large spray nozzle (BETE WL 1 1/2) was used in this figure in order to plot the data together. The expansion ratio is basically the volumetric ratio of entrained air and foam solution. Generally,

Figure 5. Foam generation rate at various pump pressures and damper openings.

solution on the screen, increasing the pump pressure tended to increase the foam generation rate at low pump pressures; however, more foam solution on the screen did not help entrain more air to increase the foam generation rate at high pump pressures. The foam generation rate using the current setup ranged from 1.26 to 2.17 m3/min. The information provided in this section aims to show the capabilities of the foam generator built in-house to produce expansion foam. Additional work has been published to systematically discuss the design details of the foam generator, foam properties, and factors that impact foam properties, such as spray nozzle size, expansion ratio, and foam solution age.11 3.2. Warming Effect. The warming effect of foam on vapor temperature was studied through the temperature profile measured in the foam fence as shown in Figure 6. The measurements from several elevations were selected and shown, which were 0.18 m (2d and 2u), 0.38 m (3d and 3u), 0.58 m (4d and 4u), 1.60 m (9d and 9u), and 1.83 m (10d and 10u) above the bottom of the liquid nitrogen container. At each elevation, a pair of thermocouples was used and the installation method is shown in Figure 7. This installation used gravity to measure foam and vapor temperatures, since foam drains downward and vapor rises upward. The upward thermocouple can be easily assessed from the top, which aims to measure the foam temperature. The downward thermocouple can be easily

Figure 4. Foam expansion ratio at various pump pressures and damper openings. 6020

DOI: 10.1021/acs.iecr.5b04535 Ind. Eng. Chem. Res. 2016, 55, 6018−6024

Article

Industrial & Engineering Chemistry Research

to continuous vaporization. The heat transfer between foam and the remaining vapor increased the temperature, which was later decreased due to a continuous supply of vapor from the pool. The downward thermocouples indicated a lower temperature, since rising vapor had a larger effect on downward thermocouples and the foam had a larger effect on upward thermocouples. Previous work used thermocouples to measure the temperature profile in the foam zone;8,10 however, the foam zone is a multiphase medium, and one thermocouple at one location could not measure the temperature of both foam and vapor. This work proposes a novel method to measure the temperature of both foam and vapor at the same elevation. The measured temperature difference between two thermocouples provides a quantitative evidence of heat transfer between foam and vapor for the first time. 3.3. Vapor Channel and Foam Breaking Rate. The formation of vapor channels in the foam is an important physical interaction in the system. After the foam application, the liquid nitrogen pool was smothered for a very short period of time. The vapor accumulated due to continuous vaporization and found a way out later. Figure 8a shows the opening at the top of the foam zone. Figure 8b,c shows the vapor channels formed in the middle and at the corner of the foam fence, respectively. The vapor was always trying to escape from the weakest location, typically the corner, since it was usually hard to reach and filled with foam last. Even though foam was reapplied in many tests to cover the existing channels, vapor channels would often form at the same location for newly applied foam. In general, the locations of vapor channels were quite random. The diameter of the vapor channels varied between tests, which could be as large as approximately 0.08 m as shown in Figure 8b (the distance between two adjacent blue lines was 0.15 m). High expansion foam breaks naturally or due to external effects, such as wind, radiation, and the interaction with liquid nitrogen and vapor. The foam stability was believed to be very important for the mitigation effect, since stable foam tends to have a smaller boil-off effect on the vaporization rate and a better warming effect on vapor temperature.4 The foam breaking process was recorded using a video camera in this work. The foam height information was obtained by analyzing the video with MATLAB code, which was developed using a similar method for jet fire flame boundary determination.12 In

Figure 6. Temperature profile in a test conducted using Method A. From the bottom to the top, the curves are 2d, 2u, 3d, 3u, 4d, 4u, 9d, 9u, 10d, and 10u.

Figure 7. Installation of thermocouples.

assessed from the bottom, which aims to measure the vapor temperature evolving from the pool. Figure 6 shows the temperature profile for the first foam application of a test conducted using Method A, in which the foam expansion ratio was 590. After the foam fence was put on top of the liquid nitrogen container, the temperature in the foam fence decreased slowly, since the vaporization rate was relatively small after the container was cooled by the liquid nitrogen. The foam application temporarily increased the vaporization rate due to the initial boil-off effect,4 which caused the temperature to decrease at all thermocouples. After the thermocouples were fully covered by high expansion foam, the old vapor was displaced by foam and fresh LN2 vapor rose from the pool due

Figure 8. Formation of vapor channel. 6021

DOI: 10.1021/acs.iecr.5b04535 Ind. Eng. Chem. Res. 2016, 55, 6018−6024

Article

Industrial & Engineering Chemistry Research

foam breaking rate can guide the reapplication of foam during emergency response to ensure effective mitigation. 3.4. Boil-off Effect. The boil-off effect has been discussed in some of the previous work,4,10 which is the effect of foam on increasing the vaporization rate. Takeno et al.’s work had to uncouple the boil-off and warming effects given the design of the test apparatus.10 Zhang et al.’s work allowed them to study the boil-off effect through small scale tests in a wind tunnel, where convection and radiation can be manually applied.4 This work continues to investigate the boil-off effect with a larger scale setup together with the warming effect. The mass curve and corresponding vaporization rate during a test are shown in Figure 11. The blue curve is a typical mass curve of tests using

this work, only the upper boundary of the foam zone was determined as the foam height. The foam heights for two selected tests are shown in Figures 9 and 10. These two tests

Figure 9. Foam breaking rate in a test conducted using Method A.

Figure 11. Mass curve and vaporization rate (1, release liquid nitrogen; 2, cover the lid; 3, remove the lid; 4, put foam fence on; 5, apply foam; 6, vaporization after foam application). Figure 10. Foam breaking rate in a test conducted using Method B.

Method A, which was measured at 1 Hz. The orange curve is the corresponding vaporization rate calculated from every 60 s. Liquid nitrogen was released until approximately 40 kg was accumulated in the container. A polystyrene lid was used to cover the liquid nitrogen container so that natural convection could be eliminated from the pool. The lid was removed later to compare the vaporization rate of liquid nitrogen. The purpose was to study if natural convection played a role to vaporize liquid nitrogen. With the results from a few tests, there was no significant effect of natural convection on vaporization rate. After the foam fence was put on, liquid nitrogen vaporized slowly prior to foam application. This vaporization rate was labeled as “a”. After high expansion foam was applied, the peak vaporization rate was labeled as “b”. When the vaporization rate started to stabilize, the vaporization was labeled “c”. Similar information was gathered for three foam applications in 17 tests. The difference between “a” and “b” is the initial boil-off effect “d”. The difference between “a” and “c” is the boil-off effect at the steady state “e”. The boil-off effect is shown in Figure 12. The initial boil-off effect decreased with an increasing number of foam applications. Each foam application could reduce approximately 0.88 g m−2s−1 of the initial boil-off effect. The duration of each initial boil-off decreased as well, which was 92, 63, and 52 s on average. The boil-off effect at steady state was much smaller than the initial boil-off effect, which can only be observed during the first foam application. The study conducted in the wind tunnel could only observe the initial boil-off effect during the first foam application, and there was no steady state boil-off effect.4 Regarding the initial boil-off effect, this novel foam generator could direct foam application into the test apparatus, which was not possible in the previous work.4 The directed foam application allowed better observation of the initial boil-off effect; therefore, the

were conducted using different methods as described previously. The foam expansion ratio was 590 for the Method A test and 680 for the Method B test, which are fairly similar expansion ratios. They both showed similar trends in terms of foam breaking rate, with three breaking rate regimes. The breaking rate was large directly after the first foam application, which was followed by a transition regime before the foam breaking rate reached a steady state. At the steady state, the foam breaking rate was smaller. The foam was reapplied when the foam height was below 1.5 m; however, the reapplication did not interfere with the steady state to cause a large foam breaking rate. The initial high foam breaking rate was probably caused by the vigorous interaction of foam with liquid nitrogen and nitrogen vapor at the beginning of the first foam application. Since the test conducted using Method B had a higher vaporization rate, the foam breaking rate was a little larger in the initial regime; however, the difference of vaporization rate did not affect the foam breaking rate at steady state as shown in Table 1. The initial foam breaking rate was about 8 times the rate at the steady state. The foam breaking rate reported in the work will be very useful to determine the boil-off effect, since the boil-off effect was caused by the water drainage of the foam breaking process. Also, the Table 1. Foam Breaking Rate test

test method

initial rate (mm/min)

1 2

A B

28 31.8

steady state 1 steady state 2 steady state 3 (mm/min) (mm/min) (mm/min) 4.2

2.9 4

3.7 3.7 6022

DOI: 10.1021/acs.iecr.5b04535 Ind. Eng. Chem. Res. 2016, 55, 6018−6024

Article

Industrial & Engineering Chemistry Research

Figure 12. Boil-off effect of foam application (error bars are standard deviation).

initial boil-off effect was observed for the second and third foam application. Regarding the steady state boil-off effect, it was not observed in the previous work.4 Because the foam fence was only 0.3 m high, there was natural convection without foam application. Compared with natural convection, the steady state boil-off effect was too small to be observed. In this work, the height of the foam fence was 1.83 m. Therefore, there was almost no natural convection effect on the liquid pool, and the steady state boil-off effect was observed. With the use of this novel foam generator, the boil-off effect was studied in much greater detail. The results indicated that the boil-off effect was small. It attenuated with the number of foam applications in terms of magnitude and time duration. 3.5. Oxygen Measurement. Oxygen concentration on top of the foam fence is a key parameter to study, since it indirectly indicates the concentration of nitrogen in the air during the test. The oxygen concentration in air is 20.9%. When nitrogen evolved from the liquid nitrogen container and rose to the top of the foam fence, nitrogen replaced oxygen and reduced the oxygen concentration. One oxygen sensor was fixed to the top of the foam fence as shown in Figure 3, and it has a measurement range of between 0.1% and 25% oxygen volumetric concentration. The sensor location was close to the center of foam fence, but was randomly picked, since there is no way to predict the vapor channel location. The installation of the oxygen sensor is shown in Figure 13. An additional

Figure 14. Oxygen measurement in a test conducted using Method A.

Figure 15. Oxygen measurement in a test conducted using Method B.

the first foam application, the oxygen concentration quickly dropped to approximately 13%, because the initial boil-off effect increased the vaporization rate of liquid nitrogen. When the first foam application was finished, the oxygen concentration returned to the normal level in the air, because the foam blanket blocked the vapor pathway. After approximately 1 min, nitrogen vapor pushed through the foam and found an exit at the top. The location of the vapor exit was random, but usually at locations where foam could not be easily reached, such as the four corners of the foam fence and the area below the oxygen sensor. When the vapor exit was not below the oxygen sensor, the change of oxygen concentration could not be measured, as shown in Figure 14, after the first foam application. The oxygen concentration returned to the normal level in the air after the initial boil-off effect. For the second and third foam application, there was a smaller initial drop of oxygen concentration due to the initial boil-off effect, which was consistent with results in Section 3.4; i.e., the initial boil-off effect attenuated with successive foam applications. The oxygen concentration returned to the normal level in the air until a vapor exit was formed below the oxygen sensor; e.g., the oxygen level dropped to approximately 16% at around 9000 s in Figure 14. Figure 15 shows an oxygen measurement in a test conducted using Method B. Because there was a continuous release of liquid nitrogen, more vapor was evolving from the bottom. The oxygen concentration rapidly decreased to 3% after the release of liquid nitrogen. Similarly, the oxygen concentration increased to approximately 20% with a full foam application to block the vapor pathway. The oxygen concentration began to drop

Figure 13. Installation of oxygen sensor.

protection screen was used to shield the oxygen sensor from contacting the foam, because the water of the foam may damage the sensor. However, the holes on the screen still allowed good vapor circulation around the sensor. The oxygen concentrations from two selected tests are shown in Figures 14 and 15. The test in Figure 14 was conducted using Method A. The oxygen concentration decreased slowly after the foam fence was put on top of the container for liquid nitrogen, since the vaporization rate of liquid nitrogen was slow after the container was cooled. During 6023

DOI: 10.1021/acs.iecr.5b04535 Ind. Eng. Chem. Res. 2016, 55, 6018−6024

Article

Industrial & Engineering Chemistry Research

LNG on the Atmospheric Diffusion of Vaporized Gas. J. Loss Prev. Process Ind. 1996, 9, 125. (11) Harding, B.; Zhang, B.; Liu, Y.; Chen, H.; Mannan, M. S. Improved Research-Scale Foam Generator Design and Performance Characterization. J. Loss Prev. Process Ind. 2016, 39, 173. (12) Laboureur, D. M.; Gopalaswami, N.; Zhang, B.; Liu, Y.; Mannan, M. S. Experimental Study on Propane Jet Fire Hazards: Assessment of the Main Geometrical Features of Horizontal Jet Flames. J. Loss Prev. Process Ind. 2016, 41, 355.

immediately after the vapor channel was formed, since much more vapor was rising from the bottom than in the test conducted using Method A. The oxygen concentration was as low as 6% before the second foam application. After the second foam application, the release of liquid nitrogen stopped; therefore, the oxygen concentration began to increase and approached approximately 20%.

4. CONCLUSION This work used a novel foam generator designed and built inhouse. The performance of the foam generator was characterized in terms of foam expansion ratio and generation rate. With the current setup, the range of the foam expansion ratio was from 300 to 850, and the foam generation rate ranged from 1.26 to 2.17 m3/min. The temperature profile above the liquid pool indicated a temperature increase immediately after foam application. A temperature difference between the foam and vapor was measured by a pair of specially designed thermocouples. The formation of the vapor channel was discussed in terms of vapor channel location and size. The foam breaking rate was obtained through video analysis for tests conducted using two different methods. The boil-off effect was studied in greater detail, which confirmed that the boil-off effect was small and did not last long. The oxygen concentration on top of the foam fence was studied, which depended on the scenario of liquid nitrogen release and vapor channel location.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: (979) 862-3985. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This research was sponsored by the Mary Kay O’Connor Process Safety Center at Texas A&M University. REFERENCES

(1) Energy Information Administration. Natural Gas 1998 Issues and Trends; Energy Information Administration: Washington, DC, 1999. (2) Federal Energy Regulatory Commission. LNG; http://www.ferc. gov/industries/gas/indus-act/lng.asp (accessed Jan 1, 2015). (3) Cleveland East Ohio Gas explosion; http://en.wikipedia.org/wiki/ Cleveland_East_Ohio_Gas_explosion (accessed Jan 1, 2015). (4) Zhang, B.; Liu, Y.; Olewski, T.; Vechot, L.; Mannan, M. S. Blanketing Effect of Expansion Foam on Liquefied Natural Gas (LNG) Spillage Pool. J. Hazard. Mater. 2014, 280, 380. (5) National Fire Protection Association. NFPA 11: Standard for Low-, Medium, and High-Expansion Foam; National Fire Protection Association: Quincy, MA, 2010. (6) National Fire Protection Association. NFPA 471: Recommended Practice for Responding to Hazardous Materials Incidents; National Fire Protection Association: Quincy, MA, 2002. (7) University Engineers Inc. An Experimental Study on the Mitigation of Flammable Vapor Dispersion and Fire Hazards Immediately Following LNG Spills on Land; American Gas Association: Arlington, VA, 1974. (8) Yun, G.; Ng, D.; Mannan, M. S. Key Observations of Liquefied Natural Gas Vapor Dispersion Field Test with Expansion Foam Application. Ind. Eng. Chem. Res. 2011, 50, 1504. (9) Hiltz, R. Foam Blanketing: The Use of Foam to Mitigate the Vapor Hazard of Spilled Volatile Chemicals. In Prevention and control of accidental releases of hazardous gases; John Wiley & Sons: New York, 1993; pp 216−231. (10) Takeno, K.; Ichinose, T.; Tokuda, K.; Ohba, R.; Yoshida, K.; Ogura, K. Effects of High Expansion Foam Dispersed onto Leaked 6024

DOI: 10.1021/acs.iecr.5b04535 Ind. Eng. Chem. Res. 2016, 55, 6018−6024