Chapter 16
Rolls-Royce Corporation
High-Performance Computing
Impact on Industrial Design Cycle
Todd Simons
High-performance computing (HPC) oers several key capabilities for
industries involved in advanced manufacturing. Advanced design and
simulation soware provides engineers with the ability to analyze complex
geometries and design components that meet challenging design require-
ments that could not otherwise be accomplished. HPC enables higher-
delity analysis to improve and accelerate the design process providing
improved products. Advanced analysis also supports the development of
advanced products by reducing or eliminating physical tests through the
use of virtual testing. is capability requires ongoing soware develop-
ment because of the ever-changing technologies in hardware and soware
used for large-scale parallel computing. ese activities require support
from national policies on HPC to advance science and technology and
ultimately benet society.
16.1 Introduction 233
16.2 Impacting the Design Cycle 234
16.3 Physical Testing and Risk Reduction 237
16.4 Soware Development for Industrial Problems 239
16.5 National HPC Policy and Industry 240
16.6 Conclusion 241
234 Industrial Applications of High-Performance Computing
e design process for components in high-performance applications is
a multidisciplinary eort requiring expertise in multiple areas, including
structural analysis, thermal analysis, aerodynamics, materials and process
modeling, dynamics, ling (part life analyses), manufacturing, and cost
engineering. e design process must balance design objectives against
requirements from these dierent disciplines, and these constraints are
oen at odds with each other. In aerospace, one of the primary design
objectives is minimizing fuel consumption. Fuel costs are one of the larg-
est operating expenses for airlines. Moreover, more fuel-ecient engines
reduce greenhouse gas emissions from ight. One way to tackle fuel
consumption is reducing weight. Minimizing material to reduce weight
makes it more dicult to meet the requirements for satisfying strength
and ling. In a typical design cycle, a team of engineers takes a component
through an iterative process where they study the trade-os for dierent
compromises that will ultimately meet all of the design requirements to
produce an acceptable part. e numerical simulations used in detailed
analysis can be time consuming and be a pacing item during the design
process. Improvements in designs happen incrementally as the design
team explores and analyzes the design space available. ese design eorts
cannot continue indenitely because of project deadlines and cost con-
straints. Decreasing the time required for each design iteration allows the
design team to more thoroughly explore the design space and improve the
quality of the design. ese time constraints make HPC and computing
capacity a critical resource for design.
Gas turbine engine design begins with lower-delity models to explore
the design space eciently. Initial performance models start with one-
dimensional (1D) models to represent the major components in the gas
turbine engine. ese models use empirical relationships, historical
information, and simple models to approximate the behavior of potential
designs. ese models are used for evaluating engine subsystems such
as compressors, combustors, and turbines, to size these components for
materials used in fabrication, ow rates, pressures, air velocities, size, and
number of airfoils. As the design process matures, higher-delity models
are introduced. Two-dimensional (2D) models are introduced to evaluate
aerodynamics on a plane at the mid-section of the gas turbine engine.
ese lower-order models are used to eciently move the design in the
right direction. Simplied 3D models are used in the early design process

Get Industrial Applications of High-Performance Computing now with the O’Reilly learning platform.

O’Reilly members experience live online training, plus books, videos, and digital content from nearly 200 publishers.