Underpinning the simulation based design optimisation approach in the Engineering Design systems is the High Performance Computing (HPC). Here, HPC refers to the integrated environment comprising typically of a large number of rack-mounted commodity based processors (Intel/AMD) with very high speed interconnect, served by a high performance parallel filing system and other peripheral components such as pre and post-processing nodes and management nodes. HPC is used primarily for highly parallel Computational fluid dynamics (CFD) and Finite element analysis (FEA) applications.
The requirements for high fidelity modelling of complex geometries with multi-physics and multi-disciplinary approaches demand extreme computational power. Within a design cycle, these modelling and analysis have to be accomplished within an acceptable and challenging time scale and accuracy; and the platforms on which these simulations are performed have to be robust, stable and reliable in their availability and usability. HPC is the only computational resource that can meet such high demands.
In Rolls-Royce, from single components and sub-systems design/analysis/optimisation through to whole engine modelling all rely heavily on HPC. HPC capability is therefore key to supporting the engineering design community in pushing the boundaries of design envelopes further, enabling new opportunities to be explored, resulting in more efficient, more environmentally friendly and more competitive products. The modelling and simulation technologies have been developed and refined to such an extent as to allow some engine rig tests to be taken out thus reducing cost and risk at product development stage. HPC, together with powerful simulation and analysis tools are also able to provide reliable and responsive support to fleet and in-service modification. Without doubt, HPC is now central to engineering activities across all business sectors globally in Rolls-Royce, from Aerospace to Nuclear, Energy and Marine.
There has been multi-million pound investment in HPC over the last few years, resulting in thousands of processors and peta-bytes of near and long term storage being available to the engineering communities. To meet the ever growing hunger for this computational power from the engineering community, and to support some important research and development work, we continue to explore opportunities such as using externally hosted on-demand facilities as well. We work closely with some of the leading supercomputing centres in the US and Europe to exploit state-of-the-art HPC technology developed in these institutions so as to optimise both our design methodologies and process, leading to more competitive engine products.