3D CFD

A boundary condition defines how the computational domain interacts with its surroundings. It specifies, for example, how fluid enters, leaves or behaves at solid walls. Typical boundary conditions include inlet, outlet, wall, symmetry plane or periodic boundary condition. Incorrect boundary conditions can lead to physically wrong results even with a high-quality mesh.

Boundary layer resolution describes how finely the mesh represents the flow close to walls. Within the boundary layer, velocity, turbulence quantities and temperature often change strongly over a very small distance. Good boundary layer resolution is crucial for wall friction, pressure loss, heat transfer, flow separation and aerodynamic forces. Depending on the modelling approach, the boundary layer is either resolved near the wall or approximated using wall functions.

Cell size describes the spatial resolution of the computational mesh in a CFD simulation. Small cells can capture local gradients, boundary layers, vortices, narrow gaps or strong temperature changes more accurately. At the same time, smaller cells increase the cell count, runtime and memory demand. A good cell size distribution is therefore always a compromise between accuracy and computational efficiency.

A CFD simulation calculates the behaviour of fluids such as air, water, oil, coolant or gases using numerical methods. Typical results include velocity, pressure, temperature, pressure loss, mass flow rate, forces and heat transfer. The simulation is based on physical conservation equations that are solved on a computational mesh. In industrial development, CFD is used to evaluate and optimize components and systems before they are built or tested.

The computational domain is the spatial region that is calculated in a CFD simulation. It can be the inside of a pipe, a vehicle engine bay, a cooling air duct or the surroundings of a vehicle. The choice of domain influences boundary conditions, runtime and result quality. A domain that is too small can artificially affect the flow, while an unnecessarily large domain increases computational effort.

Computational fluid dynamics is the field that investigates fluid flows using mathematical models and computer-based numerical methods. It combines fluid mechanics, thermodynamics, mathematics, numerical methods and software engineering. The goal is to make flow fields and thermal processes in technical systems predictable by simulation. In practice, the term is often used synonymously with CFD.

The computational mesh divides the domain of a CFD simulation into many small cells or elements. Within these cells, quantities such as pressure, velocity, temperature or turbulence variables are calculated numerically. The mesh resolution determines how well local gradients, boundary layers, vortices or narrow gaps are captured. A mesh that is too coarse may miss important flow effects, while an overly fine mesh greatly increases runtime and memory demand.

Conservation equations state that physical quantities such as mass, momentum and energy must be balanced within a system. In CFD, these equations are applied to many small cells of the computational mesh. This creates a large system of equations that is solved numerically. The quality of a CFD simulation strongly depends on whether these conservation equations are modelled physically and solved in a numerically stable way.

Conservation of energy describes the balance of internal energy, kinetic energy, heat and work in a system. In CFD, it is required when temperature changes, heat transfer, compressibility, viscous heating or chemical reactions are relevant. For purely isothermal flows, the energy equation may sometimes be omitted. In cooling, combustion, heat exchangers and high-speed flows, however, it is essential.

Conservation of mass means that mass is neither created nor destroyed in a closed system. For fluid flows, this means that mass flow rates at inlets, outlets and within the domain must be balanced consistently. In CFD, conservation of mass is especially important for pressure loss analyses, flow distribution, leakage paths and cooling circuits. Errors in the mass balance are a strong warning sign of numerical or modelling problems.

Conservation of momentum describes how the motion of a fluid changes due to forces. These include pressure forces, viscous forces, inertial forces and, if relevant, external forces such as gravity or rotation. In CFD, momentum conservation is the basis for velocity fields, pressure distributions, shear stresses and flow forces. It explains, for example, why pressure losses occur or why a flow separates from a wall.

Conjugate heat transfer describes the coupled calculation of fluid flow and heat conduction in solid components. It makes it possible to simulate how a fluid heats or cools a component and how heat spreads through the material. CHT is important when wall temperatures cannot simply be prescribed, but result from the interaction between the fluid and the solid. Typical examples include cooling channels, radiators, battery cold plates, engine components and turbomachinery.

The continuity equation is the mathematical form of conservation of mass. It ensures that no non-physical mass is created or lost in the simulation. For incompressible flows, it simplifies to the requirement that the velocity field must be divergence-free. In CFD solvers, the continuity equation is closely linked to pressure-velocity coupling.

Continuum mechanics describes matter as a continuously distributed medium rather than modelling individual molecules. This assumption makes it possible to describe flows using fields such as pressure, temperature, density and velocity. For most technical CFD applications, this approach is highly suitable. Only at very small scales, in highly rarefied gases or in molecular effects does the continuum assumption reach its limits.

Discretization influences accuracy, stability, computational time and possible numerical errors in a simulation.

The energy equation describes how energy is transported, stored and converted in a flow. It is used when temperatures, heat transfer, compression, expansion or thermal sources must be considered. In cooling and thermal management simulations, it is indispensable. It couples the flow field with the temperature field and enables the assessment of thermal loads.

The finite difference method replaces derivatives in differential equations with differences between neighbouring grid points. It is mathematically intuitive and historically an important basis of numerical flow simulation. For simple and structured grids, it can be very efficient. For complex industrial geometries, however, the finite volume method is often preferred because it handles unstructured meshes more flexibly.

The finite element method divides a domain into small elements and approximates the unknown quantities using shape functions. It is especially well known from structural mechanics and strength analysis. Flow and heat transfer problems can also be solved using the finite element method. In industrial CFD, however, it is often less dominant than the finite volume method, depending on the software and application.

The finite volume method is a numerical method in which the computational domain is divided into small control volumes. The conservation equations are balanced over each of these volumes. This makes the method particularly suitable for flow problems where conservation of mass, momentum and energy must be maintained accurately. Many industrial CFD codes use the finite volume method as their central discretization approach.

A flow simulation describes how a fluid moves through a geometry or around a component. It can reveal acceleration, dead-water regions, recirculation, flow separation and pressure losses. Depending on the engineering question, the simulation can be steady-state, transient, isothermal or coupled with heat transfer. Flow simulations are used in automotive engineering, motorsport, energy technology, mechanical engineering and plant engineering.

Fluid dynamics is the branch of fluid mechanics that deals with fluids in motion. It focuses on flow velocities, accelerations, pressure fields, vortices, turbulence and energy transfer. In CFD, fluid dynamics provides the central physical background for calculating technical flows. The term is mainly used when the motion of the fluid is the primary focus.

Fluid mechanics is the science of the behaviour of liquids and gases. It describes how pressure, velocity, density, viscosity and forces are related in fluids at rest or in motion. It forms the physical basis for CFD simulation, aerodynamics, hydraulics, cooling and many mechanical engineering applications. Without an understanding of fluid mechanics, CFD results are difficult to assess correctly.

A grid convergence study systematically investigates how the simulation result changes as the mesh is refined. It is used to estimate the discretization error and assess the numerical quality of a simulation. Unlike a simple plausibility check, it specifically evaluates the influence of the grid on the solution. For critical development decisions, a grid convergence study is an important proof of simulation robustness.

A heat transfer simulation investigates how heat is transported by conduction, convection or thermal radiation. In technical systems, these mechanisms often occur simultaneously and strongly influence the temperature distribution. Heat transfer simulation is especially important for cooling concepts, heat exchangers, thermally loaded components and fluid-flow components. It helps evaluate cooling performance, component temperatures and thermal safety margins.

An initial condition describes the starting state of a simulation. It defines initial values such as pressure, velocity, temperature, turbulence quantities or concentrations in the computational domain. In steady-state simulations, it mainly affects stability and convergence speed. In transient simulations, it can also significantly influence the calculated time history and transition behaviour.

Mesh generation is the process of creating the surface and volume mesh for a CFD simulation. The CAD geometry is prepared so that a numerically computable mesh can be generated from it. The goal is a mesh that resolves the relevant flow regions sufficiently well while remaining computationally efficient. Good mesh generation is often more important for result quality than later adjustments of individual solver settings.

A mesh independence study checks whether a CFD result still depends strongly on the chosen mesh resolution. Simulations with different mesh refinements are compared. If key results such as pressure loss, force, mass flow rate or heat transfer change only slightly with further refinement, the result is considered largely mesh-independent. Such studies improve the reliability and traceability of CFD results.

Mesh quality describes how suitable a computational mesh is for a stable and accurate numerical calculation. Important criteria include cell distortion, aspect ratio, orthogonality, smoothness of cell size transitions and boundary layer resolution. Poor mesh quality can lead to convergence problems, non-physical results or increased numerical diffusion. Therefore, checking mesh quality is a fixed part of a reliable CFD workflow.

The momentum equation describes the motion of a fluid under the influence of forces. It accounts for pressure gradients, viscous effects, acceleration and external forces. In its viscous form, it is closely related to the Navier-Stokes equations. In industrial CFD simulations, it is one of the central equations for calculating pressure fields, velocities, forces and losses.

The Navier-Stokes equations describe conservation of momentum in viscous fluids. Together with the continuity equation and, where required, the energy equation, they form the main mathematical basis of CFD. They balance pressure forces, inertial forces, viscous forces and external forces. Since these equations usually cannot be solved analytically for real technical flows, CFD solves them numerically.

A prism layer is an ordered cell layer directly adjacent to solid walls. It is used to resolve the near-wall flow region more accurately, especially boundary layers, wall friction and heat transfer. Prism layers are important in industrial CFD because many relevant losses and thermal effects occur directly at surfaces. Their height, number and growth rate must match the turbulence model and the intended y+ range.

A surface mesh represents the geometry of a component or computational domain using small surface cells. It captures edges, curvatures, openings, wall surfaces and CAD details for the CFD model. A clean surface mesh is the basis for a stable and high-quality volume mesh. Errors in the surface mesh, such as gaps, intersections or highly distorted faces, can compromise the entire simulation.

A thermal simulation calculates temperatures, heat flows and heat distribution in components or complete systems. It is used to identify hot spots, thermal overload, warm-up behaviour or cooling demand. When coupled with CFD, it can also assess the influence of air or liquid flow on the temperature field. Typical applications include battery cooling, power electronics, engine cooling, heat exchangers and vehicle thermal management.

A volume mesh fills the computational domain with three-dimensional cells. The flow equations are solved numerically within these cells, and the result quantities are computed there. Cell shape and cell distribution must match the geometry and expected flow behaviour. Targeted mesh refinement is especially important in boundary layers, narrow gaps, flow deflections or regions with strong gradients.

The wall distance y+ is a dimensionless quantity describing the location of the first cell centre relative to the wall. It relates the physical wall distance to local flow velocity, density and viscosity. In CFD, y+ indicates whether the near-wall flow is resolved directly or should be modelled using a wall function. An unsuitable y+ range can significantly distort wall friction, heat transfer and flow separation behaviour.

A wall function is a model for the near-wall region of a turbulent flow. It replaces the very fine direct resolution of the viscous sublayer with an empirical or semi-empirical wall treatment. This can significantly reduce the cell count and make industrial CFD simulations more efficient. Wall functions are only reliable when the mesh, y+ range, turbulence model and flow situation are consistent.