Objective LNG membrane tanks represent a future trend in the development of LNG storage tanks. Under seismic actions, the convection and impact of LNG liquid subject these tanks to dynamic pressure that directly applies to the walls of their outer containers—an effect not experienced by conventional LNG full containment tanks. Consequently, it is essential to investigate the distribution patterns of dynamic pressure on LNG membrane tanks.

Methods A numerical simulation method, utilizing the finite element analysis software ANSYS, was employed to model the Hainan and Longkou LNG membrane tanks, with respective capacities of 25×10⁴ m³ and 30×10⁴ m³. Time-history analyses were performed on the 25×10⁴ m³ tank under two seismic conditions: Operating Basis Earthquake (OBE) and Safe Shutdown Earthquake (SSE). The results from the simulation were compared to the calculations based on the classical Housner formula to verify the validity and consistency of the numerical model. Building on this, the dynamic pressure distribution on LNG membrane tanks was analyzed under various conditions, considering the time-history analysis results of the 30×10⁴ m³ tank also subjected to seismic waves during both OBE and SSE conditions. Ultimately, a dynamic pressure distribution formula for LNG membrane tanks was derived through a fitting process.

Results The dynamic pressure of LNG liquid within the tanks was identified as the primary contributor to the maximum acceleration of the tank wall, which occurs when the tanks experience the most unfavorable forces. By analyzing the force conditions of LNG membrane tanks, it can be concluded that: Under the same horizontal angle but at different levels, the proposed dynamic pressure distribution formula for LNG membrane tanks produced results that followed curve trends consistent with numerical simulation results and calculations based on the Housner formula. Specifically, values derived from the proposed formula at each point showed discrepancies of less than 20％when compared to the numerical simulation results. Notably, at lower levels, the results from the proposed formula aligned more closely with the numerical simulations than those obtained from the Housner formula. This closer alignment is attributed to the inclusion of the stiffness effect at the rigid connection between the tank wall and the pile cap. The accuracy of the formula was further validated through comparisons under various horizontal angles at the same level.

Conclusion The proposed dynamic pressure distribution formula for LNG membrane tanks enhances the accuracy of calculating the liquid dynamic pressure experienced by these tanks during seismic actions. This improvement can substantially reduce the conservatism in design while ensuring adequate bearing capacity in the concrete tank wall, thereby illustrating a method to enhance the economic efficiency of the concrete outer container design for LNG membrane tanks. Future studies are anticipated to gather dynamic pressure distribution data that will further refine the adaptability and accuracy of this formula, either through monitoring the dynamic pressure distribution of large-scale LNG membrane tanks in actual applications under seismic conditions or by conducting scale-down model experiments.