Computational Fluid Dynamics (CFD) (2024)

We offer you custom computational fluid dynamics - from standard cases, to special events, because the virtual reality saves considerable costs.

Two common CFD fields of application are the air stream around entire buildings. In the road space or space conditions, turbulence and increased wind speeds can occur by poor urban planning and architectural structure, which affects the comfort of the pedestrians.

The second major application field of simulations is the indoor air stream, for what we calculate air velocities and temperatures depending on the position and type of air inlet and outlet openings. Often thermal loads or thermal radiation matters in rooms, that we need to take into account as well.

Results of our simulations are cost-saving concepts, that are not only examined by calculation to meet the requirements, but also ensure the desired comfort criteria for the rooms. A positive side effect of that backup function is usually the reduction of operating costs.

Results of our computational fluid dynamic simulations:

• analysis and evaluation of user comfort by using a 2D or 3D flow simulation
• evaluation of indoor air streams to prevent draughts
• analyses of glass roofs on top of atriums and malls and the implementation of smoke and heat exhaust ventilation systems
• planning and optimization of procedural air streams
• planning and fluid mechanical optimization of HVAC systems and components
• calculation and analysis of the spatial distribution of inside air, radiant and operative temperatures

The experiments in virtual reality offer significant advantages: They are – if competent consultancy is provided - cost-effective and completely safe. In 1:1 scale, flow processes can be tested on the computer and thus produces a resilient picture of posterior conditions under varying parameters.

What is CFD?

Foundation for computational fluid dynamics is the conservation laws of physics of mass, momentum and energy, the "Navier-Stokes-Equations”. In addition to this, the use of empirical data describes the turbulence or heat propagation on surfaces.
The “Navier-Stokes-Equations” described the behavior of frictional fluids, such as water or air, already in the first half of the 19th Century, which embodies a system of nonlinear partial differential equations of second order, that are solved by the CFD programs iteratively. In order to solve the equation, you need thousands of volumes or computational boxes, the so-called box models. The modeling can be provided as two- or three-dimensional, stationary, or non-stationary.

Direct numerical simulation is the direct and simultaneous solution of the “Navier-Stokes-Equations” at each point of the air stream, in all directions, and at all times. This procedure reaches quickly several billions of cells to calculate, which exceed the computing power of today's computers, and is only limited to a few exceptions.

Today the most common procedures use the RANS turbulence modeling (Reynolds Averaged Navier Stokes), which describes the kinetic energy and the isotropic dissipation rate with two equations and provides sufficiently accurate results. Not only the airstream is simulated numerically, but also thermal and other physical interactions.

The four steps of a simulation

For the solution of partial differential equations usually serves the “Finite Volume Method” (FVM): The investigated section is converted into a specific number of cells or control volumes. In each cell equations for mass, momentum and energy are applied.

Step 1: The geometrical preparation

The examined spatial air flow situation is represented in a “3D-CAD” program that has an interface or an exchange format to the software for what the CFD equations are calculated. All adjacent surfaces by the CAD model receive an alphanumeric identifier, in order to substantiate them with boundary conditions.

Step 2: The geometrical interconnection – To choose control volumes

The observed internal volume is filled with tetrahedra, hexahedra or other base bodies. The creation of these control volumes requires the compliance to certain rules. The ratio of length and width should not be greater than a 1:5 scale, as it may cause unstable conditions in the simulation. In addition to this, all boundary conditions are defined in this step. This includes the surface texture of the perfused volume, the inlet and outlet apertures, or the temperature of the examined fluid (usually air).

Step 3: The numerical solution

Prerequisite for the successful calculation, is the selection of one or more suitable models for the stream. The type and form of flow (stationary/non-stationary/laminar/turbulent), type of conductor (pipe, channels, etc.) and type of fluid (friction/viscous) have to be distinguished. The numerical approximation procedure for partial differential equations reaches its solution iteratively. Once undercutting a defined approximation error, it terminates the process. In order to check the solution, physical model parameters are approached as a general rule.

Step 4: Evaluation of the results

For the output of the simulation results, mass flows or function graphs are common, which reflect the resulting parameters (such as flow rate or pressure) on an axis of ordinates, while position or time is shown on the abscissa. Particularly vivid are certainly colorful two- or three-dimensional graphics that reveal through a color spectrum, pressure or flow rate of the fluid in the examination area. Of course we are able to animate the results and compile a short movie, to show the flow processes in three dimensions and most vividly.

Pros and Cons of CFD

Numerical simulations cannot and should not replace real experiments. Both serve the most exact and economic prediction of flow processes, which is expected in planned facilities or buildings or to optimize the existing ones. In some cases, simulations complement real experiments, but by now they replace many other experiments.

Advantages of virtual experiments with CFD:

• specific physical boundary conditions or effects can be considered in isolation
• simulations provide at any point (of the pattern) measured data, experiments in contrast only a few selected points
• many flow parameters are gathered, which are not accessible in experiments
• in the beginning of the planning process, a variety of prototypes are simulated in order to quickly gather information for further planning
• simulations are able to contribute to a greater understanding of the problem than experiments
• the costs are usually much lower compared to experiments

Disadvantages of virtual experiments with CFD:

• errors may occur due to simple flow models or simplified boundary conditions
• possible uncertainties caused by too little computing values ​​per cell and hence therefore resulting interpolation errors
• computation time may extend for large models
• the costs may be much higher due to wrong consulting compared to experiments

When considering and analysing the disadvantages it becomes clear that a great deal depends on the user of the simulation program. Usable and valid simulation results require a lot of experience and expertise. The consultancy on whether CFD or real experiments are preferable, should be done neutral and solely influenced by the responsibility for the result.