A computational framework that supports dynamic fidelity for aeroelastic analysis and optimization is presented. The current research represents some of the recent developments in the Service Oriented Computing EnviRonment (SORCER)[1-6]. From an engineering perspective SORCER is a programming and computing environment that enables one to perform large scale system level engineering analysis and design space exploration that may be done on a geographically distributed heterogeneous computing environment. From a computer science view SORCER is a federated service-to-service (S2S) meta-computing environment that treats service providers as network peers with well-defined semantics of a service-object oriented architecture (SOOA). Here we present the evolution of SORCER from an ExertionOriented Programming paradigm to a Variable Oriented paradigm and its ability to enable dynamic fidelity of engineering responses and their respective sensitivities. In Exertion-Oriented Programming the SORCER framework supports the development of an engineering analysis or design process by enabling the developer to easily combine services on the network to create a process (Exertion) that performs multidisciplinary analysis or design of an engineering system. Hence, the application developer constructs the Exertions (Tasks and Jobs) and their relationships. In Exertion-Oriented Programming the focus is on engineering applications or services and the combination of these services to produce an analysis or result. Variable Oriented (VO) Programming focuses on a specific engineering quantity and views its calculation as a function or a function of functions as opposed to an engineering application or service. This effort describes the development and capability of the VO programming along with a demonstration of the dynamic fidelity by performing aeroelastic analysis with six different fidelities of induced drag. Euler based calculations with a Trefftz-plane, linear panel methods with a Trefftz-plane, standard one-point approximations with Euler and panel methods, and modified one-point approximations with Euler and panel methods. The fidelity can be selected dynamically by the analysis or optimization algorithm based on either accuracy or efficiency requirements.
[1]
Brian Sanders,et al.
Optimal Scheduling of Control Surfaces on a Flexible Wing to Reduce Induced Drag
,
2004
.
[2]
Michael Sobolewski,et al.
Objected-Oriented Metacomputing with Exertions
,
2010
.
[3]
Michael Sobolewski.
Federated Method Invocation with Exertions
,
2007
.
[4]
Max M Munk,et al.
The Minimum Induced Drag of Aerofoils
,
1979
.
[5]
R. Kolonay,et al.
A Comparison of Four Approximation Techniques for An Euler Based Induced Drag Function
,
2008
.
[6]
Michael W. Sobolewski.
Provisioning Object-oriented Service Clouds for Exertion-oriented Programming
,
2011,
CLOSER.
[7]
H. Ashley,et al.
Aerodynamics of Wings and Bodies
,
1965
.
[8]
Michael Sobolewski.
Exertion Oriented Programming
,
2008
.
[9]
Michael Sobolewski.
Metacomputing with Federated Method Invocation
,
2009
.
[10]
Michael W. Sobolewski.
Federated Collaborations with Exertions
,
2008,
2008 IEEE 17th Workshop on Enabling Technologies: Infrastructure for Collaborative Enterprises.