OpenCFD are pleased to announce that they have begun alpha testing of version 1.4 of their OpenFOAM software. Many of the developments in the release are towards further improving the efficiency and robustness of OpenFOAM for large-scale engineering CFD, including low and high speed aerodynamics, multiphase flows and solid stress analysis.

Developments to the solvers (applications) include:

• interface-capturing solver family: incorporate the new multidimensional universal limited explicit solver (MULES) for efficient, unconditionally-bounded solution;

• cavitatingFoam: new solver for fully compressible two-phase flow including cavitation;

• multiphaseInterFoam: an extension of the interFoam solver for an arbitrary number of phases;

• rhopSonicFoam: additional control parameters added;

• solidDisplacementFoam and steadySolidDisplacementFoam: replacement solvers for linear-elastic stress analysis including a solution-acceleration procedure for steady-state problems.

Developments of utilities include:

• new ideasUnvToFoam converter to convert meshes from I-DEAS .unv format to OpenFOAM format;

• ccm26ToFoam: upgraded CCM mesh converter to use CCM version 2.6 library and handle boundary regions better;

• improved zone support to fluentMeshToFoam;

• foamPackSource to pack a source directory, e.g. utility or solver into a .tgz file that removes all superfluous files, e.g. binaries, dependencies, before packing;

• new patchAverage and patchIntegrate utilities to calculate average and integrand of fields over all patches;

• new redistributeMeshPar utility which reads a decomposed mesh and redistributes cells across processors (without reconstructing the mesh on a single processor).

Model development:

• porous media model including power-law inertial and viscous models for either explicit or implicit implementation in any of the pressure-velocity solvers (demonstration example included);

• multiple reference frames (MRF) support for any of the pressure-velocity solvers (demonstration example included);

• flexible and efficient finite volume based mesh motion solvers (replacing the finite element mesh motion solvers);

• implementation of the k-omega SST turbulence model for high-Re external and internal flows;

• generalisation of particle and cloud base classes in the Lagrangian library to make it easier to create more complex, derived entities.

Improvements to boundary conditions include:

• generalisation of total pressure boundary conditions, e.g. for entrainment boundaries, and improvements to corresponding velocity boundary conditions;

• generalisation of wave transmissive boundary conditions for compressible flow;

• mapping plane boundary condition, e.g. for fully-developed inlet conditions, particularly for (but not limited to) LES;

• time varying boundary condition interpolated from data, e.g. from experiment.

A new generalised geometric/algebraic multigrid (GAMG) solver:

• includes runtime selectable geometric and algebraic agglomeration algorithms including MGridGen;

• has an option for agglomeration to be cached;

• operates a flexible V-cycle including pre- and post-smoothing;

• optionally agglomerates coarsest matrix onto master processor and uses direct solver thus significantly reducing the communications overhead on latency-dominated Beowulf clusters;

• can be used to solve both symmetric and asymmetric matrices.

New preconditioned conjugate gradient solver with runtime selectable preconditioner including:

• diagonal incomplete-Cholesky (DIC), equivalent to ICCG;

• GAMG;

• diagonal preconditioning or, optionally, no preconditioning.

New preconditioned bi-conjugate gradient solver with runtime selectable preconditioner including:

• diagonal incomplete-LU (DILU), equivalent to BICCG;

• diagonal preconditioning or, optionally, no preconditioning.

Discretisation schemes:

• improved gradient limiters, including cell and face-based multidimensional forms, particularly useful for improving the accuracy of the general linear upwind scheme;

• new outletStabilised scheme for entrainment or outlet conditions where flow reversal may occur.

Other:

• full implementation of symmetric and spherical tensors to improve consistency, elegance, computational speed and storage, particularly for turbulence models;

• simple parallelised build system with wmake;

• new writeAndEnd control parameter, to program an exit from a time loop;

• implementation of multiple constraints at corners for volField to pointField interpolation;

• an improved version of octree;

• prefetch cache optimisations for all solvers for the Itanium2 processor.