Showing papers on "Surrogate model published in 1998"

Computer Aided Reliability and Robustness Assessment

Abstract: The competitive pressure to shorten product development time has necessitated the automotive industry to rely more on Computer Aided Engineering (CAE) for analyzing and proving product reliability and robustness. The challenge of this approach is the incorporation of product variability, due to manufacturing and customer usage variations in the analysis, requires a massive computation process which may be prohibitive even with today's advanced computers. In this paper, we demonstrate the use of an efficient computational procedure based on optimal Latin Hypercube Sampling (LHS) and a "cheap-to-compute" nonlinear surrogate model using Multivariate Adaptive Regression Splines (MARS) to emulate a computationally intensive complex CAE model. The result of the analysis is the identification of sensitivity of design parameters, in addition to a computationally affordable reliability assessment. Fatigue life durability of automotive shock tower is presented as an example to demonstrate the methodology.

Simulation-driven metamodeling of complex systems using neural networks

Abstract: Simulation of large complex systems for the purpose of evaluating performance and exploring alternatives is a computationally slow process, currently still out of the domain of real-time applications. To overcome this limitation, one approach is to obtain a metamodel of the system, i.e., a surrogate model which is computationally much faster than the simulator and yet is just as accurate. We describe the use of Neural Networks (NN) as metamodeling devices which may be trained to mimic the input-output behavior of a simulation model. In the case of Discrete Event System (DES) models, the process of collecting the simulation data needed to obtain a metamodel can also be significantly enhanced through Concurrent Estimation techniques which enable the extraction of information from a single simulation that would otherwise require multiple repeated simulations. We will present applications of two benchmark problems in the C 3 I domain: A tactical electronic reconnaissance model describing the flight of a reconnaissance aircraft carrying a bearing angle measuring sensor over a radar field in order to detect ground-based radar sites; and an aircraft refueling and maintenance system as a component of a typical Air Tasking Order (ATO). A comparative analysis with alternative metamodeling approaches indicates that a NN captures significant nonlinearities in the behavior of complex systems that may otherwise not be accurately modeled.