Computational Fluid Dynamics
Due to the advancements in almost every field CFD and its application is a rapidly developing discipline due to the continuous development in the capabilities of commercial software and the growth of computer power. The system has been a success and therefore CFD is already widely used in industry and its application is set to spread.
We have learned that the physics which governs fluids is relatively simple. the laws of motion and thermodynamics. However, for designing a system of this scope the solutions are very complex and this makes analytical methods largely unusable for industrial applications. An approach to solve such complex dilemmas is to replace the problem with a number of smaller less complex problems. With the advent of computers this approach became practical and in the late 1960s Computational Fluid Dynamics was born.
What does the CFD do To explain the idea behind CFD lets cite an example of an airplane. As the plane moves along the air must move out of its way. This means that the airplane has to make its way shearing through the wind. The way in which the air flows depends on the plane’s shape. The flow can be smooth but more likely it will contain vortices, shockwaves and other disturbances. We will have to consider the air as the fluid present. To model the behavior of the fluid the volume is split into many smaller sub-volumes, called a mesh. A mesh can be simply the same sub-volume repeated throughout the space or, more usually, it can be molded around the object that is being modeled and so can be complicated.
There are problems associated with the meshing mechanism as it easy to say the word splitting but in reality it is extremely difficult. Therefore the skill is to produce a mesh with exactly the right sub-volumes. If the sub volumes are too big then the solution will have errors. if the volumes are too small then the calculation will take too long to be useful in a design process.
The past tells that in 1990 CFD was an activity for owners of CRAY super-computers, but by 2000 the airflow is analyzed by starting from an initial flow, which can be either a guess at the solution or a specific initial condition. This has lead to a change such that using this initial flow the conservation equations are used to predict the flow a short time later. Every time a new prediction is then made from the newly calculated flow. This is a simpler way to solve the evolution of the airflow.
In the rapid changing world most applications of CFD involve steady flows because they need less computer control. This explanation of using the conservation equations as a flow solver is a generalization. there are other more complex solvers. However, these methods fabricate similar outputs and, more importantly the draw back is, CFD systems look the same to a user regardless of the solution technique.
After having a flow solution, the user is presented with the flow at every point in the mesh. The last phase of the CFD progression is to extract from these data the information that the user actually wants.
However the dynamic solutions suffer from one big draw back that is accuracy. For many applications, such as building design, the level of accuracy required is very low and CFD results are more than sufficient. CFD usually gets the qualitative picture correct, which is useful in helping people understand what