SAUDI RESEARCH SCIENCE INSTITUTE 2023

Numerical methods for solving compressible

fluid dynamics problems
Dr. Rasha Al Jahdali, Research scientist
 Numerical discretization
is the Mathematics branch responsible for designing effective ways
1. Importing or
to find numerical solutions to complex Mathematical problems. creating
geometry

Space Discretization
• Efficiency is key.
• …and consistency and stability

4. Solving and
2. Generating
visualizing
mesh
results
Time Discretization
stability and convergence issues

3. Defining
physics of the
problem with
the help of load,
BC and IC
 Importing or creating geometry

Creating geometry for simulation involves designing or generating a virtual

representation of the system or object to be analyzed. This can be done using
approaches like Computer-Aided Design (CAD) Software. The geometry is
created by defining the system's shape, dimensions, and features and may
involve using specialized software or tools. After creating the geometry, it is
important to validate its accuracy, simplify it if necessary, and prepare it for
simulation by checking for errors and ensuring compatibility with the software.
The final step is exporting or integrating the geometry into the simulation
environment, ready for analysis and study.
 Generating mesh

A mesh is one of the fundamental elements of a simulation process in finite element analysis
(FEA). It is a network formed of cells and points (or nodes). It can have almost any shape or
size and is used to solve Partial Differential Equations. Each cell of the mesh represents an
individual solution of the equation, which, when combined with the whole network, results in
a solution for the entire mesh.

Meshing is the method of generating a 2D or 3D grid over a geometry to discretize it and analyze
it with simulation. The grids are defined based on the complexity of geometry.
 Generating mesh

Mesh Discretization
• The first step for numerically solving a set of partial differential equations
(PDEs) is the discretization of the equations and the discretization of the
problem domain.
• Solving the entire problem domain at once is impossible, whereas solving
multiple small pieces of the problem domain is perfectly fine.
• The equations discretization process is related to methods such as the Finite
Difference Method, Finite Volume Method, and Finite Element Method,
whose purpose is to take equations in the continuous form and generate a
Example of a square domain discretization
system of algebraic difference equations. where each small square represents a cell
• The domain discretization process generates a set of discrete cells and, and their vertices nodes

therefore, points or nodes that cover the continuous problem domain.

 Generating mesh

Mesh Elements
• Vertices: These are the individual points in the mesh that define
its geometry. Vertices are typically represented by their 3D
coordinates (x, y, z) in space.
• Edges: Edges connect pairs of vertices and define the geometric
connectivity of the mesh. They represent the lines or curves that
connect adjacent vertices.
• Faces: Faces are formed by a collection of connected edges and
define the surface of the mesh. They can be triangular or Mesh elements: vertex, edge and face.

quadrilateral, depending on the mesh type.

 Mesh Types
Structured Mesh

• Structured Mesh: In a structured mesh, the vertices, edges, and faces follow a regular
pattern or grid, such as a rectangular or hexagonal grid. This type of mesh is well-suited
for regular geometries and can efficiently represent structured domains.

Unstructured Mesh

• Unstructured Mesh: An unstructured mesh does not follow a regular pattern and
allows for arbitrary connectivity between vertices, edges, and faces. It is more flexible
and suitable for representing complex geometries with irregular shapes.

• Hybrid Mesh: A hybrid mesh combines structured and unstructured elements, allowing
for a combination of regular and irregular connectivity patterns. This type of mesh is
often used to optimize the simulation in areas where regularity is present.
 Generating mesh

Common cell shapes

Two-dimensional: There are two types of two-dimensional cell shapes
that are commonly used. These are the triangle and the quadrilateral.
Three-dimensional: The basic 3-dimensional element are
the tetrahedron, quadrilateral pyramid, triangular prism, and hexahedron.
They all have triangular and quadrilateral faces.
Mesh Quality
• Mesh quality refers to the geometric characteristics of the mesh elements
and their impact on the accuracy and efficiency of simulations.
• Good mesh quality ensures that the mesh accurately represents the
underlying geometry and provides reliable results. It involves considerations
such as element shape, size, aspect ratio, smoothness, and boundary
conformity.
• Mesh optimization techniques can be employed to improve mesh quality,
such as refining elements in high-gradient areas, smoothing element shapes,
or repositioning vertices for better alignment with the geometry.
 Defining physics of the problem

When we say that the physics of a problem are fully captured in a simulation, it means that
the simulation accurately represents the fundamental physical principles and behaviors that
govern the system being studied.

The key aspects of capturing the physics in a simulation are:

1.Governing Equations: The simulation is based on mathematical models that represent the
underlying physics of the problem. These models typically involve partial differential equations (PDEs)
or other mathematical formulations that describe how the system evolves over time or in response to
external influences.
2.Material Properties: Accurate representation of the material properties is crucial for capturing the
physics in a simulation. These properties may include parameters like density, elasticity, thermal
conductivity, viscosity, electrical conductivity, and magnetic permeability.
 Defining physics of the problem

3. Boundary and Initial Conditions: Defining appropriate boundary conditions (BCs) and initial
conditions (ICs) is essential. BCs constrain the behavior of the system at its boundaries, while
ICs specify its initial state. These conditions play a vital role in replicating real-world scenarios
and accurately simulating the physics of the problem.

4. Numerical Methods: To solve the governing equations, numerical methods are employed
within the simulation software. These methods discretize the problem domain and
approximate the continuous equations, allowing for efficient computation. Proper selection
and implementation of numerical methods are crucial to ensure accurate and reliable results.
 IC and BC in CFD simulations

Initial Conditions (ICs): Initial conditions are set up to define the fluid flow state at the beginning of
the simulation. They include velocity, pressure, temperature, and species concentrations. The
selection of initial conditions depends on the specific problem being simulated and the available data.
For example, the initial conditions may represent a known, stable flow field in a steady-state
simulation. The initial conditions may be derived from experimental data or a previous simulation time
step in unsteady simulations.
Boundary Conditions (BCs): Boundary conditions are specified at the boundaries of the computational
domain and provide information about the flow behavior at these boundaries. Different types of
boundary conditions are used to model various real-world scenarios. Some common types of
boundary conditions include:

a. Dirichlet Boundary Conditions: These conditions directly specify the values of flow variables at the
boundary. For example, the velocity or pressure may be prescribed at the boundary. Dirichlet
boundary conditions are typically used for inlets and outlets where the flow characteristics are known.
 IC and BC in CFD simulations

b. Neumann Boundary Conditions: Also known as flux boundary conditions, these conditions specify the
rate of change of flow variables normal to the boundary. For example, the gradient of pressure or velocity
may be prescribed at the boundary. Neumann boundary conditions are often used for walls, where the flow
variables are constrained based on the physical properties of the boundary surface.
c. Robin Boundary Conditions: These conditions combine Dirichlet and Neumann conditions and are used
when both the value and the gradient of flow variables are known at the boundary. Robin boundary
conditions are employed in situations where the boundary condition is influenced by both the fluid flow and
the properties of the boundary surface.
d. Periodic Boundary Conditions: These conditions are employed when the computational domain is
periodic, meaning that the flow variables at one boundary are identical to the flow variables at another
corresponding boundary. Periodic boundary conditions are used to simulate flows in infinite domains or to
model periodic phenomena.
e. Symmetry Boundary Conditions: These conditions are used to model symmetry planes in the domain.
They enforce the symmetry property by specifying that the flow variables are symmetrical across the
boundary. Symmetry boundary conditions are employed to reduce computational costs by simulating only a
portion of the actual geometry.
 Solving and visualizing results

Solving and visualizing the results of a simulation are critical steps in the analysis process. After setting up
the simulation and specifying the necessary parameters, the following steps are typically involved:

1.Solving the Simulation: To solve the simulation, the mathematical models and numerical methods
implemented in the simulation software are used. The software iteratively solves the equations that
describe the system's behavior over time or in response to specified conditions. The simulation software
performs calculations and generates numerical results based on the input parameters, boundary conditions,
and initial conditions.
2.Extracting Results: Once the simulation is completed, the software provides output data representing the
system's behavior. This data can include quantities such as displacements, stresses, temperatures,
velocities, fluid flow rates, or any other variables of interest that were defined in the simulation. The output
data is stored in files or databases for further analysis and visualization.
 Solving and visualizing results

3. Visualizing the Results: Visualization techniques are employed to gain insights and interpret the simulation
results. These techniques help in representing complex numerical data more intuitively and understandably.
Visualization tools and software allow for the creation of plots, graphs, contour maps, animations, or three-
dimensional renderings that effectively communicate the behavior and characteristics of the system.