January 2013 Summaries

Aerospace manufacturers are increasingly relying on advanced engineering simulations to enhance product development across various systems such as spacecraft, aircraft, and jet engines, where 100% reliability is essential. These simulations, though complex and time-consuming, enable engineers to iterate designs early in the process without the need for physical prototypes. However, the limitations of high-performance computing (HPC) infrastructure can hinder the ability to fully optimize designs, as engineers face capacity constraints and virtual queues. An aerospace OEM addressed these challenges by partnering with Rescale, which provided access to a comprehensive suite of simulation tools and on-demand HPC resources at a reduced cost. This partnership allowed the customer to conduct multiple-parameter sweeps and designs of experiments more efficiently, significantly reducing computational time and uncovering critical insights that would have been missed with local computations. Rescale's platform dynamically provisions hundreds of processors, ensuring secure and isolated data processing, resulting in over 95% reduction in runtime and significant savings in engineer-hours.

For automotive manufacturers and suppliers, product development is a costly and time-consuming endeavor, often hindered by demanding customer expectations, strict regulations, and competitive pressures. Advanced engineering simulations play a crucial role in the early stages of vehicle development by allowing engineers to iterate designs without physical prototypes, although these simulations typically require significant computational resources and time. A Tier 1 automotive supplier addressed these challenges by using Rescale to optimize the profile of a piston skirt in an internal combustion engine, aiming to minimize power loss due to contact friction under various load conditions. The supplier conducted over 150 simulations per load condition using a fully parametric 3D model, accounting for thermal and mechanical deformation, with proprietary code executed across a dynamically provisioned 32-processor cluster. This approach significantly reduced simulation runtime by over 90% and saved approximately 120 hours of engineering manpower compared to local computing, while maintaining high-security standards for data handling.