Designing a Liquid Rocket Engine is one of the hardest tasks in engineering. The heart of the engine is the injector. It must spray fuel and oxygen into the combustion chamber perfectly. If the injector is bad, the engine will vibrate or explode. Real physical testing is very dangerous and costs millions of dollars. Therefore, engineers use CFD simulation to design the injector safely on a computer. We use ANSYS Fluent to model the complex mixing of liquids and gases.

In this report, we perform a Liquid Rocket Engine CFD simulation. We focus on a coaxial injector where fuel and oxidizer mix. The challenge is seeing the tiny droplets. To solve this, we use a special technique called mesh adaptation CFD. This method automatically makes the grid smaller where the liquid breaks up. This Liquid Rocket Engine fluent simulation helps engineers analyze near‑injector liquid–gas interaction and mixing behavior, which are essential prerequisites for efficient combustion. For more on flight physics, please explore our Aerodynamic & aerospace tutorials: https://cfdland.com/product-category/engineering/aerodynamics-aerospace-cfd-simulation/

Figure 1: Example of Liquid Rocket Engine injector

Simulation Process: VOF Multiphase Model and Mesh Adaptation Strategy in Fluent

For this Liquid Rocket Engine CFD project, we created a 2D axisymmetric model. This shape represents a typical coaxial injector. It has a central channel for liquid oxygen and an outer ring for kerosene fuel. We started with a structured grid using square cells. However, a fixed mesh is not good enough for moving liquids because droplets are very small. Therefore, we activated the mesh adaptation CFD feature in ANSYS Fluent. This powerful tool monitors the flow. When it detects a liquid surface, it automatically cuts the cells into smaller pieces. This ensures that the Liquid Rocket Engine ANSYS Fluent solver captures the detailed shape of the liquid–gas interface and liquid ligaments in the near‑injector region.

We configured the physics using the Liquid Rocket Engine fluent simulation settings. We selected the Volume of Fluid (VOF) multiphase model. This model is essential because it tracks the interface between the liquid fuel and the gas in the chamber. Because the spray changes every microsecond, we used a Transient solver. This calculates the flow history step by step. We defined the boundary conditions with specific mass flow rates for the oxygen and kerosene. The CFD Analysis of Liquid Rocket Engine injector relies on the solver to calculate the interaction between the two fluids. The automatic mesh refinement updates the grid at every time step. This dynamic process allows us to simulate the complex breakup of the jet without creating a mesh that is too heavy for the computer.

Figure 2: Computational grid visualization showing Mesh Adaptation where ANSYS Fluent automatically refines cells near the liquid-gas interface.

Post-processing: Analysis of Primary Jet Breakup and Near‑Field Mixing Behavior

This section interprets the engineering data to understand how the injector performs. We analyze the contours to see if the fuel and oxidizer mix well, which is vital for the manufacturer. We first analyze the Volume of Fraction (VOF) Contours in Figure 2 to evaluate the atomization quality. The simulation captures the full history from time T1 to T5. At time T1 and T2, the Liquid Rocket Engine injector pushes out solid streams of fuel. These jets are coherent and have not broken up yet. However, by time T3, the physics change dramatically. We observe the formation of spiral structures and ligaments. This means the liquid sheet is tearing apart due to aerodynamic forces. At time T5, the flow reaches a “quasi-steady” state. The most critical achievement here is the visualization of primary jet breakup and interface disintegration. The detached cyan and green liquid fragments observed at T5 represent highly deformed liquid structures and ligaments rather than fully developed spray droplets. These features indicate the onset of the pre‑atomization stage, which significantly increases the liquid–gas interfacial area and enhances near‑injector mixing, an essential prerequisite for efficient combustion.

Figure 3: Time evolution of Volume Fraction contours (T1–T5), illustrating primary jet breakup, interfacial instability growth, and ligament formation inside the chamber.

Next, we examine the Velocity Magnitude Contour (Figure 4). The colors show the speed of the fluid. The bright red zones at the inlet show the highest speed. The simulation calculates a maximum velocity of 21.6 m/s. This high velocity promotes strong aerodynamic interaction, leading to primary jet breakup and jet spreading. We also observe slower regions (Blue and Cyan) with speeds of 6-10 m/s. These are “vortices.” They swirl the gas around. These vortices are good. They mix the fuel and oxidizer together like a blender. Based on the velocity and volume fraction distribution, a near‑field liquid jet spreading angle of approximately 60–80 degrees is observed. This angle is critical. It determines how the flame spreads in the chamber. A manufacturer uses this Liquid Rocket Engine CFD simulation result to place the igniter in the correct spot. If the angle is too narrow, the flame might damage the bottom of the engine. If it is too wide, it might burn the walls. confirms that the injector generates a stable near‑field jet spreading pattern suitable for further combustion system design. Finally, looking at Figure 2, we can see the result of the mesh adaption CFD. The grid is very dense (lots of small squares) exactly where the liquid droplets are. This proves the technique worked. It allowed us to capture the complex shapes of the droplets without wasting computer memory on the empty gas areas. This makes the design process faster and cheaper.

Figure 4: Velocity magnitude distribution showing the high-speed injection core  and the recirculation vortices that promote mixing.

It should be noted that this study focuses on the near‑injector region and primary breakup mechanisms using the VOF approach, while secondary atomization and discrete droplet dynamics are beyond the scope of the present simulation.