Statistical analysis of modelling approaches for CFD simulations of high-pressure natural gas releases

The use of computational fluid dynamics (CFD) in process safety to estimate the risk of a given incidental scenario has become ever more present in common industry practice. The simulation of high-pressure, compressible natural gas jets is often performed by modelling its source with a simpler notional diameter approach, such that the highly computationally expensive nearfield zone need not to be simulated; this is particularly determining when simulating a gas release in complex scenario like liquid natural gas (LNG) regasification plants. In this study, we analysed the structure of compressible and incompressible jets, using Birch 1984 (B84) and Birch 1987 (B87) models. In this work, a study on the positioning of the notional diameter with respect to the real orifice of the released gas is performed, along with a statistical analysis to assess the limits of the simpler model approaches. It was found that no spacing is needed between the virtual and real sources, as the potential core generated by the simpler model is as large as the fully simulated nearfield zone by the compressible model. Additionally, an end-of-transition zone position correlation is reported. The incompressible models can be used instead of the fully compressible model for a wide range of release conditions, with both models providing accurate predictions of axisymmetrical mole fraction, temperature, and velocity profiles between 2.5 and 130 bar of storage pressure at a 1-inch orifice diameter. However, as the diameter increases, B84 is not a viable model for a “full bore” (10-inch diameter size) release at 65 bar. While B84 is reliable, B87 is the superior model for its ability to account for the compressible effects of the expansion. Therefore, B87 should be used when simulating cases where temperature is of particular interest to the user.