In the past, commercial CFD codes have struggled to penetrate the aerospace flow-simulation market that is dominated by a myriad of established "in-house" codes, developed by (and sometimes shared between) the major aerospace companies.

Often the comparison with in-house CFD was unfair as, having borne the significant cost of developing these bespoke tools, aerospace companies were often reluctant to consider discarding them in favor of commercial CFD. More often than not, it was indeed true that general-purpose CFD codes,

developed to tackle a wide range of engineering problems, did not compare favorably with in-house codes on the specific types of problems that the latter were developed specifically to tackle.

STAR-CCM+ was designed to challenge this paradigm. Although a general-purpose code in spirit, STAR-CCM+ was conceived with the explicit requirement that the needs of the aerospace community had to be fully met. Offering accuracy, which is at least equivalent to inhouse aero codes, STAR-CCM+ offers a host of additional benefits and does not require the years of "fine-tuning" which is a feature of many such solutions.

At its heart are two state-of-the-art coupled solvers: an implicit algebraic-multigrid solver and an explicit multistage Runge-Kutta solver. Throughout extensive testing and validation, STAR-CCM+'s coupled solvers have proved exceptionally robust and accurate when dealing with shocks, natural convection and other problems involving a strong coupling between velocity, pressure, and temperature (in both compressible and incompressible flows). CD-adapco's aerospace specialist, Stephen McIlwain, explains" Sometimes when engineers first encounter STAR-CCM+, they are so impressed by its user interface and ability to deal with polyhedral meshes that they completely overlook the fact that STAR-CCM+ is based around not one, but two first-rate coupled solvers".

Unlike many coupled solvers (such as those incorporated within typical in-house CFD codes), STAR-CCM+'s solver is not limited to purely inviscid (Euler) simulations, it is backed up by a comprehensive suite of turbulence and multiphysics models.

STAR-CCM+ also includes a state-of-the-art segregated solver, which is ideal for low-Mach number simulations. Switching between the segregated solver and each of the coupled solvers is a trivial single step process. This means that, for example, the segregated solver can be used to quickly provide a sensible initial condition, from which the coupled solvers can be run to convergence. The major benefit of this is that users can use a single CFD package for all simulation regimes, whether subsonic, transonic or hypersonic. In a presentation to the 2005 STAR European conference, Matt Milne of QinetiQ reported the progress of their evaluation of STAR-CCM+ as a front line CFD code for the simulation of complete civil and military aircraft configurations. He concluded, "The results of this evaluation indicate that when run on high-quality blockstructured meshes, STAR-CCM+ yields high quality results of at least equal accuracy to existing in-house methods and other commercial codes. In particular, excellent agreement has been seen between measured and predicted levels of drag."

In-house CFD codes are typically built upon aging architectures that are not entirely compatible with the latest developments in CFD technology. Since STAR-CCM+ was started from an entirely "blank sheet of paper", it contains no legacy coding, and is able to utilize the very latest developments in solver technology. This manifests itself most clearly in STAR-CCM+'s ability to work with meshes of arbitrary topology and in particular polyhedral meshes.

Previously, one of the major obstacles for aerospace companies in the simulation of complete aircraft configurations was in the construction of high quality computational meshes. Traditionally the choice had to be made between structured hexahedral and unstructured tetrahedral meshes. Hexahedral meshes, while offering unparalleled accuracy, are expensive to construct, relying on a large amount of manual intervention. Unstructured tetrahedral meshes, although much easier to generate, are typically much less accurate and result in significantly higher mesh counts.

Polyhedral meshes offer the best of both worlds, the automatic meshing benefits of tetrahedra combined with near hexahedral accuracy. Extensive testing has shown that a typical polyhedral mesh requires less than one fifth of the cell count of a tetrahedral mesh to deliver an equivalent or improved level of accuracy. Although, aerospace companies will continue to use their own in-house CFD for much of their work for the foreseeable future, STAR-CCM+ offers a solution that is a practical alternative today, and has the credibility to become the basis for a complete replacement in the future.