Introduction to LNG Storage and the Role of Computational Fluid Dynamics

Liquefied natural gas (LNG) is natural gas that has been cooled to approximately -162°C at atmospheric pressure, reducing its volume by a factor of about 600. This phase change makes long-distance transportation and large-scale storage economically viable. Storage tanks for LNG are critical infrastructure in the supply chain, typically designed as double-walled cryogenic tanks with perlite insulation or vacuum insulation to minimize heat ingress. Even small amounts of heat leakage can cause evaporation, known as boil-off gas (BOG), which must be managed to maintain pressure and avoid losses.

Understanding the complex fluid dynamics and heat transfer within these tanks is essential for ensuring safety, structural integrity, and operational efficiency. Computational Fluid Dynamics (CFD) offers a cost-effective and detailed method to simulate the behavior of LNG under various operating conditions. ANSYS Fluent is one of the most widely used CFD solvers in the energy sector, capable of modeling multiphase flows, phase change, turbulence, and conjugate heat transfer with high fidelity. This article provides an in-depth look at how engineers and researchers use ANSYS Fluent to simulate LNG flow in storage tanks, the key phenomena captured, and the practical benefits of such simulations.

Fundamentals of LNG Behavior in Storage Tanks

LNG in a storage tank is not a homogeneous, static fluid. Several complex phenomena occur simultaneously:

  • Thermal stratification – Heat entering through the tank walls and bottom creates vertical temperature gradients. Warmer, less dense LNG rises to the top, leading to stratified layers. If not properly managed, stratification can result in sudden rollover events where layers mix violently, releasing large amounts of BOG and overpressuring the tank.
  • Sloshing – During filling, emptying, or due to external disturbances such as seismic events or ship motion in the case of floating LNG (FLNG) storage, the liquid surface can experience sloshing. This dynamic motion affects structural loads, insulation performance, and BOG generation.
  • Boil-off gas generation – Continuous heat influx causes LNG to evaporate at the liquid-vapor interface. BOG must be vented, reliquefied, or used as fuel. The rate of BOG depends on insulation efficiency, tank geometry, fill level, and ambient conditions.
  • Natural convection currents – Density differences due to temperature variations drive slow convective flows within the liquid. These flows transport heat and can either promote or suppress stratification.

CFD simulations must capture all these coupled physics to provide reliable predictions. ANSYS Fluent’s ability to handle multiphase (liquid-vapor) systems, phase change models (Evaporation-Condensation or Lee model), and various turbulence models makes it well-suited for this task.

CFD Modeling Approach in ANSYS Fluent

A systematic approach is required to set up a credible LNG storage tank simulation. The following subsections detail the key steps.

Geometry and Meshing Considerations

The first step is to create a geometric representation of the tank. For typical large-scale storage tanks (cylindrical with a domed roof and flat or conical bottom), the geometry can be created in ANSYS DesignModeler or imported from CAD. Symmetry can often be exploited: a 2D axisymmetric model is sufficient for many thermal stratification studies, while full 3D models are needed for sloshing or asymmetric filling scenarios.

Meshing is critical. A high-quality mesh with boundary layer refinement near walls and the liquid-vapor interface is essential to capture sharp gradients. For multiphase simulations with the Volume of Fluid (VOF) method, cell sizes near the interface must be fine enough to resolve interface curvature. ANSYS Fluent’s meshing tools (such as the Fluent Meshing workflow) allow for polyhedral, hexcore, or cut-cell meshes that balance accuracy and computational cost. Typical cell counts range from a few hundred thousand for 2D models to tens of millions for large 3D cases.

Material Properties and Multiphase Modeling

LNG is a mixture primarily composed of methane (85–95%) with smaller fractions of ethane, propane, butane, and nitrogen. Material properties such as density, viscosity, specific heat, thermal conductivity, and latent heat of vaporization vary with temperature and composition. For accurate simulations, these properties must be defined as functions of temperature (e.g., using polynomial fits or property tables). ANSYS Fluent supports user-defined functions (UDFs) for custom property models.

Regarding multiphase modeling, two primary approaches exist:

  • Volume of Fluid (VOF) – Best for tracking the sharp interface between liquid LNG and vapor BOG. VOF is suitable for sloshing, filling, and boiling where interface dynamics are important.
  • Eulerian-Eulerian (multifluid) model – Better suited when vapor is dispersed as bubbles (e.g., during rapid boiling) or when interphase mass transfer is significant. This model solves separate conservation equations for each phase.

For many LNG storage simulations, the VOF model in conjunction with a phase change model is the preferred choice because it explicitly tracks the free surface and can handle large density ratios.

Boundary Conditions and Heat Transfer

Boundary conditions must represent the actual tank environment. Typical settings include:

  • Walls – Heat transfer through insulation layers. Instead of modeling the entire insulation thickness, a user-defined wall boundary with a specified heat transfer coefficient (based on insulation material and thickness) is common. Alternatively, conjugate heat transfer can be used if the insulation geometry is included.
  • Bottom and roof – The bottom is often in contact with a concrete foundation with ground temperature. The roof is exposed to ambient air or has additional insulation. Convective and radiative boundary conditions can be applied.
  • Inlet/outlet – For filling or emptying simulations, mass flow rate or velocity inlets are defined. For pressure management, pressure outlets may be used.
  • Initial conditions – The tank is typically initialized with a defined fill level, temperature profile (e.g., a saturated liquid at storage temperature), and vapor volume fraction above the interface.

Heat transfer across the liquid-vapor interface is dominated by phase change, which is modeled using the Lee model (evaporation-condensation). This model uses a user-defined coefficient to represent the rate of mass transfer as a function of the deviation from saturation temperature.

Turbulence and Phase Change Models

Flow inside LNG tanks is usually in the low Reynolds number regime for natural convection, but sloshing and filling can introduce turbulence. The k-omega SST model is a robust choice for blended laminar/turbulent flows, as it can transition appropriately near walls. For inherently unsteady simulations like sloshing, Large Eddy Simulation (LES) may be used for high accuracy but at increased computational cost.

Phase change modeling in ANSYS Fluent for LNG typically uses the Lee model, which computes the mass transfer rate based on the heat flux at the interface and the latent heat. The model requires specifying a relaxation factor (time scale) that should be calibrated against experimental or empirical data. Too high a value can cause numerical instability; too low can underpredict BOG rates.

Solution Setup and Convergence

Simulations are inherently transient because LNG tanks operate over long periods. Time steps are chosen to resolve the fastest relevant physics—typically on the order of 0.1 to 1 second for sloshing, and 1 to 10 seconds for stratification studies. The PISO algorithm is often used for pressure-velocity coupling in VOF simulations. Convergence is monitored via residuals, plus key variables such as total BOG generation rate or temperature at a monitoring point.

Parallel computing is almost always necessary; ANSYS Fluent scales well on multiple CPU cores and can also leverage GPU acceleration for some solvers.

Key Phenomena Captured by CFD Simulations

Thermal Stratification and Rollover Prevention

One of the most important outcomes of LNG storage CFD is predicting the development of thermal stratification over time. By simulating days or weeks of real time, engineers can determine whether the stored LNG will develop stable layers or become prone to rollover. CFD helps optimize insulation design—for example, by identifying areas of high heat flux that increase stratification. Operators can also simulate corrective actions like mixing jets or recirculation pumps to disrupt stratification before dangerous levels are reached.

Case studies have shown that CFD can accurately predict the time to rollover within 5–10% of experimental observations, making it a reliable tool for both design and operational planning.

Sloshing Dynamics and Structural Integrity

For LNG storage on floating platforms (FLNG or LNG carriers), sloshing is a major design concern. CFD with VOF can simulate wave impact on tank walls and dome, providing pressure loads that inform structural analysis (often using ANSYS Mechanical). Sloshing simulations must account for the coupled effects of liquid motion, potential gas pocket compression, and roof structural response. By running a series of simulations at different fill levels and excitation frequencies, engineers can identify resonant conditions and design baffles or membrane reinforcements to mitigate risks.

Boil-off Gas (BOG) Management

BOG generation is a key economic and safety parameter. CFD can predict BOG rates as a function of ambient temperature, tank design, fill level, and insulation condition. This allows engineers to size reliquefaction systems or vent valves correctly. Moreover, simulations can test the impact of operating strategies such as varying the fill rate or using heat exchangers inside the tank to reduce BOG. In many LNG terminals, CFD-derived BOG predictions are used to optimize the send-out schedule and minimize energy waste.

Applications and Benefits in Industry

The use of ANSYS Fluent for LNG storage simulation has become standard practice in several areas:

  • Design optimization – New tank designs are evaluated virtually before construction. Parametric studies on insulation thickness, tank aspect ratio, and internal components (like spray nozzles) help reduce costs and improve safety.
  • Safety analysis – Regulators require demonstration that tank pressure will remain within safe limits under all credible scenarios (e.g., simultaneous filling, ambient temperature spikes). CFD provides the necessary evidence, often replacing costly physical testing.
  • Operational decision support – Operators use validated CFD models to determine maximum fill rates, predicted BOG rates, and times when mixing should be performed to prevent stratification.
  • Floating LNG (FLNG) design – Sloshing and motion-induced phenomena are critical for FLNG vessels. CFD coupled with ship motion analysis helps ensure cargo tanks withstand wave-induced loads throughout their service life.

Challenges and Best Practices in LNG CFD Simulation

Despite its capabilities, simulating LNG in ANSYS Fluent presents challenges:

  • Multicomponent LNG – Real LNG is a mixture, but most simulations assume pure methane. To account for composition effects, the species transport model must be added, increasing complexity.
  • Phase change model calibration – The Lee model relaxation factor is empirical. Best practice is to validate against known BOG data from similar tanks before using the model predictively.
  • Computational cost – Long transient simulations (days or weeks of physical time) require efficient time-stepping and often reduced-order modeling or hybrid approaches. Using dynamic mesh adaptation to refine only near the interface can save significant CPU time.
  • Radiation heat transfer – At cryogenic temperatures, radiation from the warmer inner tank wall to the cold LNG can be non-negligible. The Discrete Ordinates (DO) model or even a simple view-factor method should be included for accuracy.

A recommended workflow is to start with a simplified 2D axisymmetric model to validate physical models and convergence settings, then scale up to 3D for specific phenomena. Always perform grid independence and time-step sensitivity studies to ensure results are trustworthy.

Conclusion

Computational Fluid Dynamics in ANSYS Fluent provides a powerful and practical method for simulating the complex flow and thermal behavior of LNG inside storage tanks. From predicting thermal stratification and rollover to quantifying BOG generation and sloshing loads, CFD offers insights that directly improve safety, efficiency, and design innovation. While challenges such as multicomponent modeling and computational expense remain, ongoing advances in solver algorithms and hardware continue to make these simulations more accessible and accurate. For engineers and researchers in the energy sector, mastering LNG CFD in ANSYS Fluent is an essential capability that delivers tangible value across the entire lifecycle of LNG storage facilities.