Brief technical information of Fastflo

Fastflo is a finite element package for the numerical solution of partial differential equations (PDEs) in two- and three-dimensional regions.

Fastflo is very flexible because the finite element methodology can handle domains with complex shapes, and because it incorporates a high level language, Fasttalk, in which a mathematical notation is used to specify and solve a wide range of PDEs. Users interact with Fastflo via a Graphical User Interface, which offers editing, file management, graphics and hands-on control of the computations.

Fastflo is available for Windows and Unix systems. A Fluids ToolBox, released with Version 3 of Fastflo, provides easy access to advanced Computational Fluid Dynamics (CFD) algorithms. Fastflo has a comprehensive Tutorial Guide and an on-line Reference Manual. 

Fastflo’s main features

• an internal triangular mesh generator for 2D problems, plus ability to import data from external mesh generators
• a range of isoparametric element types: triangles, quadrilaterals, tetrahedra, hexahedra; all these elements are available with linear or
  quadratic interpolation and can be modified by users
• the use of Fasttalk to specify PDEs and the algorithms to solve them; users are not constrained to a fixed menu of techniques and applications
• the ability to solve systems of PDEs with several unknowns, with multiple domains and fixed or moving boundaries
• a selection of sparse matrix solvers, both direct and iterative with pre-conditioning
• a Graphical User Interface to specify problems, direct the solution procedures, and display the results in 2D and 3D
• availability on PCs and UNIX systems
• Tutorial Guide and on-line Reference Manual
• a Fluids ToolBox providing advanced CFD algorithms for laminar and turbulent flow.
• developed since 1991 by mathematicians and engineers from CSIRO, Australia’s largest R&D organisation

Users present their PDE problems to Fastflo via two files, one for the mesh and one for the problem specification. Fastflo uses unstructured meshes, and problems can be solved in complex geometrical shapes. If the PDE is time-dependent, the user may develop an algorithm for timestepping. If the PDE is nonlinear, an appropriate iterative strategy can be implemented. After any necessary timestepping and iterative algorithms have been implemented, the user will have a set of linear PDEs to be solved at each timestep or iteration. In Fastflo, these PDEs can have partial derivatives up to second order and vector or tensor coefficients. A complete set of 38 derivative expressions is available.

Dependence of coefficients on spatial and other variables is managed by a mathematical expression capability with a wide range of operators. A global vector stack is used to store and manipulate the field variables.

Fastflo can handle systems of PDEs. For example, a benchmark problem solved using Fastflo was simulation of 3D turbulent air flow around an automobile. This problem, discussed at the 1996 meeting of the World User Association for Computational Fluid Dynamics, had six variables, namely three velocity components, pressure, turbulent kinetic energy and dissipation. Fastflo’s results compared well with experimental measurements for drag and pressure on the upper surface. Other important features are Fastflo’s moving mesh capability, and capability to specify problems in multiple sub-domains.

Table of operators available in Fastflo:  A complete set of 38 derivative expressions is available.  These are listed, for example, in the technical brochure downloadable as pdf.

Last updated Tuesday September 18, 2007