Abstract
The Caelus library is a general purpose numerical framework for the solution of the continuum mechanics problems. The framework is a derivative of the OpenFOAM® library with an emphasis on multi-platform support, improved accuracy, and robustness. As such, Caelus is an open source library developed and distributed under the GPL license agreement. Several papers and talks have been presented the past few years over the progression of Caelus. This talk will provide a summary of the key improvements along with a status of latest release, v7.04 Caelus has been released bi-annually since the inaugural release, v4.10, in October 2014. The versions are number after the year and month of release – where the leading 1 is dropped for simplicity – following a convention similar to Ubuntu. v4.10 was solely aimed at multi-plaform support. The wmake build system from OpenFOAM was replaced in favour of the widely used Python-based build system Scons (www.scons.org). This included a reorganization of the directory and file layout and clean up of unused, untested, or knowingly incorrect library components. Since the initial release, there has been a strong emphasis on low level improvements to, among many other things, discretization numerical schemes and the linear solver architecture. An example of these improvements is emphasized by the steady-state, RANS solution on the DrivAer, an publically available external automotive aerodynamics model, was computed using a SIMPLEC-based solver from Caelus on hybrid prism/tetrahedral (right) and tetrahedral only (left) meshes. The tetrahedral only mesh is identical to the hybrid mesh but with the boundary layer prisms split into component tetrahedra. The differences in surface pressure contours are nearly imperceptible while the computed drag for both solutions was different from the experimental value by less than 3%. It was not possible to obtain a solution on such tetrahedral meshes in early versions of Caelus nor in current versions of OpenFOAM. Two papers presented at the recent ASME Fluids Engineering Division Summer Meeting in Hawaii, USA outline the extensive verification and validation of the incompressible flow.