(21d) CFD Based Study to Predict Rollover in an Industrial LNG Storage Tank | AIChE

(21d) CFD Based Study to Predict Rollover in an Industrial LNG Storage Tank


Srinivasan, R. - Presenter, Indian Institute of Technology Madras
Karimi, I. - Presenter, National University of Singapore
Natural Gas is transported in the liquefied form under cryogenic conditions, which makes LNG easier to transport and safer to store. LNG is stored in a highly insulated tank, despite which heat ingress from the environment is inevitable. Consequently, the lightest component (methane) evaporates. This results in a vapor stream known as Boil-off-gas (BOG) and increases the density of the remaining LNG. In an LNG terminal, when cargo is unloaded to a storage tank containing LNG of different densities, stratified layers with different liquid densities are formed. These layers further mix due to density equilibration and interface disruption under the effect of static pressure and cross the saturation temperature. This results in a large amount of BOG generation instantly, which leads to high tank pressure and a hazardous situation. This phenomenon of overall mixing and sudden generation of BOG is known as rollover. Over 24 rollover incidents have been reported over the last 60 years (GIIGNL, 2015), including those in La Spezia (Italy, 1971) and Partington (UK, 1993). In order to prevent the significant losses that arise during a rollover, it is important to develop a fundamental understanding of the boil-off-gas generation and the rollover phenomena, which is the focus of this paper.

The rollover phenomenon has been studied both theoretically as well as experimentally over the years. Germeles et al. (1975) proposed a double-diffusion LNG model by approximating methane & other components as a salt-sugar solution (thermohaline correlation) without considering boiling mechanism and turbulence during a rollover event. Heestand et al. (1983) improved this model by considering turbulence convection for heat and mass transfer between the stratified layers and rejected the salt-sugar solution model. Koyama et al. (2007), used a 3D CFD model to study rollover by considering light LNG filling only through the bottom of the tank without including the miscibility and the vapor mass removal from the liquid phase due to evaporation. Li et al. (2015) considered a 2D CFD model and induced stratification artificially to study rollover without considering the heat ingress. Hubert et al. (2019) improved the Li model by considering 3D CFD simulation with heat ingress through tank wall but did not include LNG filling mode of operation. Saleem et al. (2020) considered miscibility factors in the 2D CFD model and formulated stratification by different modes of tank filling. The major shortcoming of the model was to study the rollover under 2D tank without considering the effect of LNG unloading mode over rollover and BOG generation. These studies indicate that a 3D model of the tank that considers the effect of loading and unloading operations can represent the rollover in a more realistic manner. In the existing literature, rollover scenarios have been studied by inducing stratification artificially at the top and bottom of the tank, but the effect of unloading operation over rollover has not been studied. Since this is the most common occurrence in industry and the one of most concern to the LNG terminal operations personnel, we seek to fill these gaps in this work.

In this paper, we develop a CFD model to gain a fundamental understanding of the rollover phenomenon and the associated BOG generation. In this model, a 3D tank is used to store LNG at the receiving terminal. The tank is initialized to hold some initial level of LNG (inventory), while LNG of varying densities are allowed to fill either through the top or bottom of the tank to induce stratification. We consider heat ingress continuously through the walls of the tank, which increases the temperature of the stored LNG under the effect of local pressure (both tank top pressure and static pressure of LNG). This local pressure is used to calculate the bubble and dew point of the LNG mixture. Our model offers the flexibility to consider the LNG in the storage tank to be considered either as miscible or immiscible as necessary. In this paper, we report results from a case study to study rollover during LNG unloading operation. While this is a common operation in industry and the most susceptible to inducing rollover, it has not been previously studied in literature. Stratification is induced by filling a heavy density LNG from the bottom of the tank. Sensitivity studies on various important parameters tank size, inventory, rate of heat ingress, and density difference of LNG are also performed. Results from scaled-down studies (smaller size tank, shorter duration) and from industrial scales will both be presented. The operational benefits to the terminal from the accurate estimates of BOG generation rate and rollover time will also be discussed.


  1. GIIGNL Technical Study Group, 2015, Rollover in LNG Storage Tanks; Technical report.
  2. Germeles, A.E., 1975. A model for LNG tank rollover 11.
  3. Heestand, J., Shipman, C.W., Meader, J.W., 1983. A predictive model for rollover in stratified LNG tanks. AIChE J. 29, 199–207. https://doi.org/10.1002/aic.690290205
  4. Koyama, K., 2007. CFD Simulation on LNG Storage Tank to Improve Safety and Reduce Cost, in: Koyamada, K., Tamura, S., Ono, O. (Eds.), Systems Modeling and Simulation. Springer Japan, Tokyo, pp. 39–43. https://doi.org/10.1007/978-4-431-49022-7_8
  5. Li, Y., Li, Z., Wang, W., 2015. Simulating on rollover phenomenon in LNG storage tanks and determination of the rollover threshold. Journal of Loss Prevention in the Process Industries 37, 132–142. https://doi.org/10.1016/j.jlp.2015.07.007
  6. Hubert, A., Dembele, S., Denissenko, P., Wen, J., 2019. Predicting Liquefied Natural Gas (LNG) rollovers using Computational Fluid Dynamics. Journal of Loss Prevention in the Process Industries 62, 103922. https://doi.org/10.1016/j.jlp.2019.103922
  7. Saleem, A., Farooq, S., Karimi, I.A., Banerjee, R., 2020. CFD Analysis of Stratification and Rollover Phenomena in an Industrial-Scale LNG Storage Tank. Ind. Eng. Chem. Res. 59, 14126–14144. https://doi.org/10.1021/acs.iecr.0c02546