Guidelines for Relative Hazard Ranking of Refrigerants and Siting Considerations for LNG Liquefaction Units

Recent growth in the availability of domestic natural gas has led the industry to push for the development of liquefaction facilities, at both existing import/regasification facilities and new locations.  As a consequence of this evolving need for North American liquefaction capacity, the Federal Regulators (i.e., FERC) have added interpretations and expectations for new accidental release case studies that are specific to liquefaction facilities.  The requirements for LNG facility siting and design are codified in the CFR 193 and NFPA 59A, but many of the exclusion zone analysis requirements are stated broadly and involve some level of expert interpretation. 

Accidental release case studies required for the Resource Report 11 filings must now include LNG and flammable refrigerants.  The hazards posed by refrigerants had not been previously relevant for receiving terminals.  One goal of the required analysis is to calculate the extent of flammable vapor clouds, to demonstrate and confirm that these do not cross the proposed facility’s property boundaries.  Given economic and geographic limitations, there is often a trade-off between liquefaction unit capacity, siting, and even technology.  Flammable refrigerants also introduce the new risk of vapor cloud explosions and associated overpressures.  As a result, refrigerant vapor dispersion analysis is also now required, as is the quantification of the associated overpressure from potential vapor cloud explosions.

However, the analysis for liquefaction units must include both the LNG and the flammable liquid refrigerants.  Two primary liquefaction refrigeration technologies are typically considered: propane pre-cooled mixed refrigerant and mixed refrigerant (PRICO).  We have found that different considerations may drive the flammable vapor exclusion zones due to liquid spills and jetting and flashing of pressurized releases.  The combination of physical properties, mixture composition, inventory, and operating parameters can lead to widely different results.  The tools that are available and approved for performing these analyses impose limitations upon the mixtures and the number of constituents that can be modeled.

This paper will provide case studies and discussion illustrating the dominant parameters for liquefaction unit accidental release modeling.  Guidance for identifying and selecting refrigerant release scenarios will be provided.  Of most interest may be the interaction of the relative hazard distances for LNG and refrigerant vapor dispersion with refrigerant overpressure analysis.