Project Title:

A Parallel Computing Environment for Probabilistic Response Analysis of High-Temperature

04.01-0018
911218
A Parallel Computing Environment for Probabilistic Response Analysis of High-Temperature
Composites
Applied Research Associates
6404 Falls of Neuse Road, #200
Raleigh
NC
27615
Robert H.
Sues
919-876-0018
LeRC
NAS3-26576
042

04.01-0018
911218

Abstract:

A Parallel Computing Environment for Probabilistic Response Analysis of High-Temperature
Composites
A parallel processing environment consisting of software strategies and optimal hardware
configurations for probabilistic simulation of the response of high-temperature composite
structures will be developed. Probabilistic composite mechanics (PCM) problems have
many inherent levels of both coarse- and fine-grained parallelism. However, the software
strategies needed to achieve large-scale parallelism do not exist. Moreover, current
parallel processor configurations may not be efficient for all cases. Developing
an efficient parallel processing environment for PCM problems will make these computationally
intensive methods practical for tailoring high-temperature structural composites.
The ability to tailor these composites and meet reliability-based design criteria
will contribute to making application of high-temperature composites in aerospace
propulsion structures possible. Phase I will identify the multiple levels of parallelism
in PCM problems and investigate innovative software strategies that can exploit this
parallelism while minimizing parallel processing overhead. Two sample problems will
then be executed on two different parallel architectures. The results will be used
to formulate recommendations for developing optimal parallel processing environments
(software and hardware) for PCM problems.
This hardware-software package will be used to reduce the need for costly testing
of numerous possible composite design configurations to many different load environments.
Commercial application would be in aerospace, automotive, offshore oil, nuclear power,
and construction industries.
parallel processing, probabilistic mechanics, high-temperature composites, structural
reliability