The properties of fluids and gases (Fluid Dynamics) have been studied for many years; understanding and predicting their behavior is important to many fields, from medicine to nuclear power to automobile design to jets. The field of Computational Fluid Dynamics (CFD) uses computing to evaluate the models described above on the object under study.
The governing laws of Fluid Dynamics are partial differential equations; typically these equations are solved by breaking up the physical domain into very small volumes, called cells. The collection of cells that fill in the domain is called the mesh or grid. A software package, called the solver code, reads the grid and setup information, solves the equations and produces results files containing the predicted flow properties (such as flow speed, temperature, species concentrations, etc.) for each cell. Since flow is dynamic, unsteady calculation requires running the solver repeatedly for many small time steps.
Thus the three steps to obtaining a CFD solution are:
- Pre-processing (building the mesh, setting operating parameters)
- Solving (running the solver code) and then
- Post-processing results files to obtain useful information. Pre- and post-processing are usually interactive, while running the solver code is a batch process, often requiring large compute servers working for many hours and producing huge results files.
Post-processing is of vital importance, where results files are made useful, through it we
- Gain understanding of critical flow features
- Report engineering metrics such as forces or maximum temperature, and
- Validate the mesh, setup parameters, quality of the solution.
This is the “pay-off”. It is here that Intelligent Light and its FieldView software and services ensure that users and enterprises get the most from their investments in hardware, software and manpower for CFD.
“The purpose of computing is insight not numbers.”
C. Hastings, Approximations for Digital Computers, 1955