ESL373: Computational Laboratory

Experiment 4

Liquid Spray in Cross-Flow

1 Aim
• To simulate and analyze the behavior of a liquid spray in a crossflow within a
rectangular computational domain.

• To analyze and post-process simulation results using a provided finer mesh case-
setup.

2 Introduction
Liquid spray in crossflow is a widely studied phenomenon in various engineering applications,
including gas turbine combustion, spray cooling, and fuel injection systems. The interaction
between the liquid jet and the high-speed crossflow leads to complex aerodynamic effects such
as jet penetration, droplet breakup, and dispersion. Understanding these mechanisms is crucial
for optimizing spray characteristics and improving performance in industrial applications.
The primary challenge in modeling spray in crossflow lies in accurately predicting the
breakup and dispersion of liquid droplets. Experimental studies provide valuable insights, but
numerical simulations using Computational Fluid Dynamics (CFD) offer a more
comprehensive understanding. Among the different modeling approaches, the Lagrangian
approach is widely used due to its ability to track individual particles and provide detailed
breakup dynamics.
To model the spray behavior, a Lagrangian approach is employed using the Discrete Phase
Model (DPM). In this method, the continuous phase (airflow) is treated as a Eulerian field,
while the dispersed phase (liquid droplets) is tracked individually in a Lagrangian framework.
This allows for a detailed analysis of droplet trajectories, size distributions, and velocity pro-
files.

1
3 Turbulence Modeling
The interaction between the liquid jet and the crossflow is highly turbulent, requiring an
appropriate turbulence model to accurately capture the flow physics. In this study, the
standard k-ε turbulence model is used. This model is widely employed due to its balance
between computational cost and accuracy in simulating high-Reynolds-number flows.

3.1 Standard k-ε Model

The k-ε model is based on two transport equations: one for the turbulent kinetic energy (k) and
one for the turbulent dissipation rate (ε). The governing equations are:

𝜕𝑘 𝜈𝑡
+ 𝑈. ∇𝑘 = 𝑃𝑘 − 𝜖 + ∇. ( ∇𝑘) …………………... (1)
𝜕𝑡 𝜎𝑘

𝜕𝜖 𝐶1 𝜖 𝐶2 𝜖 2 𝜈
+ 𝑈. ∇𝜖 = 𝑃𝑘 − + ∇. ( 𝑡 ∇𝜖) ……………... (2)
𝜕𝑡 𝑘 𝑘 𝜎𝜖

where Pk represents the production of turbulence, νt is the turbulent viscosity, and C1, C2, σk,
and σε are empirical constants. This model provides a good approximation for high-speed
turbulent flows and is computationally efficient compared to more complex turbulence
models such as Large Eddy Simulation (LES) or Reynolds Stress Models (RSM).

2
4 Lagrangian Particle Tracking (LPT) Approach
The LPT approach is used to simulate the motion of discrete particles converted from the
Eulerian field. The governing equation for particle evolution is expressed as:

𝑚𝑝 𝑑𝑢𝑝 𝑚𝑝 (𝑢−𝑢𝑝 ) 𝜌𝑝 −𝜌
= + 𝑚𝑝 𝑔 ………………. (3)
𝑑𝑡 𝜏𝑟 𝜌𝑝

where mp, up, and ρp denote the particle mass, velocity, and density, respectively. u and ρ
denote the velocity and density of the carrier phase, respectively. The first term represents the
drag force acting on the particles, while the second term accounts for the gravitational force.
The momentum relaxation time τr of the particle is given by:

4 𝜌𝑝 𝑑𝑝
𝜏𝑟 = ………………………… (4)
3 𝜌𝐶𝐷 |𝑢−𝑢𝑝 |

where CD is the drag coefficient, which is related to the relative Reynolds number Re, dp is the
particle diameter, and µ is the dynamic viscosity of the carrier phase.
Other forces, including Saffman lift, pressure, and buoyancy forces, are typically neglected
in most practical applications.

5 Computational Setup
The simulation is performed in a three-dimensional rectangular computational domain with
dimensions of 30 × 15 × 10 mm. The domain is designed to capture the interaction between
the injected liquid spray and the crossflowing gas stream. A detailed schematic of the
computational setup is provided in the accompanying figures.

5.1 Meshing Strategy

Meshing plays a crucial role in the accuracy and computational efficiency of the simulation. A
structured hexahedral mesh is preferred for better accuracy and stability. However, in complex
regions, an unstructured tetrahedral mesh may be used to capture finer details of spray breakup
and dispersion. The mesh refinement is applied near the injection region and in the wake of
the spray to ensure a higher resolution of droplet interactions. The grid independence study is
conducted to balance accuracy and computational cost.
3
6 Procedure
6.1 Geometry Creation in ANSYS Workbench
To create the computational domain, follow these steps:
1. Open ANSYS Workbench and create a new project. Open ANSYS Workbench and drag
the Geometry module into the project schematic. Double-click on Geometry and select
SpaceClaim to launch the geometry editor.
2. Drag and drop the Geometry module into the workspace.
3. Double-click on Geometry to open SpaceClaim, the geometry modeling tool.
4. In SpaceClaim, select the Rectangle tool from the sketching options.
5. Draw a rectangle with dimensions 30 mm (length) × 15 mm (width), make sure to keep in
mind the origin (either at the center or at the corner).

Figure 1: Rectangular Domain

6. Use the Pull tool to extrude the rectangle to a height of 10 mm, creating a 3D rectangular
domain.

Figure 2: 3D Domain
4
 7. Ensure the domain is properly aligned and positioned within the workspace.
8. Save the geometry and close SpaceClaim.
Next, we will proceed with meshing the domain in the following section.

6.2 Meshing
This section provides a step-by-step guide on setting up the mesh in ANSYS Meshing.

6.2.1 Step 1: Opening Meshing

After finishing the design, return to ANSYS Workbench and click on the Meshing option. The
geometry will be automatically imported into ANSYS Meshing.
6.2.2 Step 2: Naming the Faces
Once the geometry is imported, name the faces as follows:

• Left face: Input Air

• Rightmost face: Output Air

• Top face: Top Wall

• Bottom face: Bottom Wall

• Front face: Front Wall

• Back face: Back Wall

Figure 3: Opening the Meshing Module in ANSYS Workbench

5
6.2.3 Step 3: Changing Physics Preference to CFD
Go to Physics Preference and change the setting from Mechanical to CFD.

Figure 4: Changing Physics Preference to CFD

6.2.4 Step 4: Setting Up the Mesh

Under the Mesh section:

• Right click and Select Methods.

• Choose Automatic Meshing.

6.2.5 Step 5: Generating the Mesh

Set the element size:

• Enter 7e-4 in the Element Size field.

• Click Generate Mesh.

6
 Figure 5: Selecting Automatic Meshing
7 Simulation Setup in ANSYS Fluent
This section details the step-by-step procedure to set up the simulation for the liquid spray
in a crossflow using ANSYS Fluent. The primary objective is to analyze the interaction of a
high-speed airflow with an injected liquid jet.

7.1 Setting up the Solver

1. Open ANSYS Fluent and choose a pressure-based solver.

2. Set the solver to Transient to capture the time-dependent nature of the spray breakup
and crossflow interaction.

Figure 6: Pressure based solver.

7
7.2 Defining the Models
1. Under the Viscous Model section:

• Select k-ε turbulence model.

• Choose the Standard option.

• Enable Enhanced Wall Treatment to better resolve near-wall effects.

7.3 Boundary Conditions

1. Inlet (Airflow): Velocity Inlet with a speed of 50 m/s.

2. Outlet: Pressure Outlet set to atmospheric pressure.

3. Walls (Top, Bottom, Front, and Back): No-slip condition applied.

Figure 7: Selection of turbulence model in ANSYS Fluent.

8
7.4 Solution Methods
1. Select Coupled scheme for pressure-velocity coupling.

2. Under Spatial Discretization:

• Pressure: PRESTO!

• Momentum: First Order Upwind

• Turbulence: First Order Upwind

7.5 Solution Controls (Under-Relaxation Factors)

1. Set under-relaxation factors as:

• Pressure: 0.3

• Momentum: 0.7

• Turbulent Kinetic Energy: 0.8

• Turbulent Dissipation Rate: 0.8

7.6 Solution Initialization and Time-Stepping

1. Perform Hybrid Initialization.

2. Set the time step size to 5 × 10−4 s for capturing transient effects.

3. Number of time steps: 2000 (adjust as per the simulation needs).

4. Let the simulation run until you achieve convergence

5. After convergence is achieved stop the simulation and follow the Discrete phase steps.

This completes the setup of the simulation in ANSYS Fluent, ensuring all necessary models
and parameters are correctly defined before proceeding to the injection setup.

7.7 Defining the Injection Setup

1. Enable Discrete Phase Model (DPM) and select Interaction with Continuous Phase.

2. Under Tracking Parameters:

• Maximum Number of Steps: 500

• Step Length Factor: 5

9
 3. Enable Breakup Model under Physical Models.

4. Set DPM Iterations Interval to 10.

7.8 Defining the Injection Properties

1. Create a new injection and name it Spray.

2. Select Injection Type: Plane Orifice Atomizer.

3. Enter Number of Streams: 100.

4. Set Material: Water-Liquid.

7.9 Defining Injection Position and Properties

1. Adjust X, Y, Z positions so that the injection cone is centered in the domain.

2. Use Show Mesh to visualize and refine placement.

3. Set position vectors:

• Y-Position: 1

• X, Z-Positions: 0

4. Enter Mass Flow Rate: 1 g/s.

5. Define Start and Stop Time:

• Start Time: Current time step (from solution console)

• Stop Time: Total time steps × Time Step Size

10
 Figure 8: Injection setup visualization

7.10 Injector Specifications

• Injector Inner Diameter: 0.5 mm

• Injector Orifice Length: 10 mm

• Constant A: 4.9

• Azimuthal Start Angle: 0°

• Azimuthal End Angle: 360°

7.11 Physical Models

1. Momentum Exchange: Dynamic drag law.

2. Breakup Model: Select Kelvin-Helmholtz Rayleigh-Taylor (KH-RT) model.

This completes the setup of the injection process in ANSYS Fluent. The next steps involve
defining post-processing techniques and data analysis for spray characteristics.

11
 Figure 9: Injection setup visualization

8 Post-Processing in ANSYS CFD-Post

This section details the steps for post-processing the simulation results in ANSYS CFD-Post to
analyze spray characteristics, including particle tracking and visualization. There should be data
available in your system with finer mesh resolution and more parcels for better spray modeling.
Those datasets may be used for your post-processing exercise.
8.1 Exporting Solution Data

1. In ANSYS Fluent, navigate to the Export section.

2. Select Solution Data and Particle History Data to save relevant results for post-processing.

3. Save the exported data files in an accessible location.

8.2 Loading Results in CFD-Post

1. Open ANSYS CFD-Post.

2. Click on Load Results and select the previously saved Fluent solution data.

3. Import the particle tracking data:

(a) Navigate to File → Import.

(b) Select Particle Track File.

(c) Load the saved Particle History Data.

12
8.3 Creating a Plane for Visualization
1. In CFD-Post, click on Insert → Location → Plane.

2. Set the plane type to XY Plane to visualize the spray distribution effectively.

8.4 Generating Particle Tracks

1. Click on the Particle Track option from the top menu.

2. Assign a name to the particle track.

3. Under the Geometry settings:

(a) Set Max Tracks to 5000.

4. Configure the color settings:

(a) Under Mode, select Variable.

(b) Choose the appropriate Variable for visualization (e.g., particle diameter, velocity,
Weber number, breakup time).

(c) Set Range to Local.

(d) Select Linear for the color scale.

(e) Use the Default color map.

(f) Enable Smooth Contour for a refined visualization.

13
 Figure 10: Particle track

Figure 11: Geometry settings

14
 Figure 12: Color setting

8.5 Configuring Symbol Representation

1. Navigate to the Symbol section.

2. Enable Show Symbol by checking the corresponding box.

3. Under Symbol Type, select Ball for clear particle representation.

4. In the Size settings:

(a) Choose Particle Diameter as the size parameter.

(b) Select Scale Type as Relative or Absolute.

(c) Adjust the scale within the suitable range (e.g., 0-1).

5. Ensure Show Track is enabled.

15
8.6 Applying and Analyzing Results
1. Click Apply to generate the particle track visualization.

2. Review the generated output tracks to analyze particle behavior.

This completes the particle tracking analysis in CFD-Post. The next section will cover
contour plot generation for further insights.
8.7 Generating Contour Plots
1. Click on the Contour Plot option from the top menu.

2. Assign a name to the contour plot.

3. Configure the settings:

• Under Domain, select All Domains.

• Under Location, select Plane.

– If a plane is not created, refer to the previous section to create one.

• Under Variable, choose the desired parameter for visualization (e.g., pressure,
velocity, turbulence kinetic energy, etc.).

4. Click Apply to generate the contour plot.

Figure 13: Color setting

16
8.8 Report Writing
For report submission, all students may submit typed reports, incorporating the generated results
within the same document. The final report should be uploaded on Moodle.
• For each of the results generated during the post-processing exercise of the
simulation experiment, provide your own interpretation.
• Additionally, the report must include example(s) of spray in cross-flow applications
in the energy sector. There should be a short discussion on the relevance of spray in
cross-flow for the application you are emphasizing.
• All references used in the report should be cited appropriately in a standard
referencing format. Blind use of AI tools, such as LLMs, Perplexity, ChatGPT is
strongly discouraged. If found penalty in marks will be there. Wikipedia pages cannot
be included as references.
• Ensure that all data, images, and research findings are properly attributed to their
sources.
• What modifications can be made to make this specific spray in cross-flow experiment
to be more meaningful for the ESL373 course?

Teaching Assistants:
Mr. Baki Mayank ([email protected])
Mr. Gaurav Bajpai ([email protected])

Course Instructor and Faculty-in-Charge of Experiment 4:

Prof. Kaushik Saha ([email protected])

17