Investigation of the Samples Size Effects on Hybrid Surrogate Model Component Surrogates for Crashworthiness Design Chongqing University

Format:

Conference/Event

Author/Creator:

Chen, Chen, author.

Contributor:

Dong, Wenxiang

Yu, Huili

Zhan, Zhenfei

Zhao, Hui

Conference Name:

WCX World Congress Experience (2018-04-10 : Detroit, Michigan, United States)

Language:

English

Physical Description:

1 online resource

Place of Publication:

Warrendale, PA SAE International 2018

Summary:

AbstractSurrogate model based design optimization has been widely adopted in automotive industry. Hybrid surrogate model with multiple component surrogates is considered to be a better choice when simulating highly non-linear responses in vehicle crashworthiness analysis. Currently, the number of component surrogates has to be decided before-hand when constructing of a hybrid surrogate model. This paper conducts a comparative study on the performances of three popular hybrid modeling methods including heuristic computation strategy, and two kinds of optimal weighted surrogates. The effects of samples size on the number of individual surrogates that should be included into the final hybrid surrogate models for crashworthiness responses are investigated. Different hybrid modeling techniques and multiple validation criteria are evaluated. Some observations and conclusions on the selection of component surrogates in hybrid surrogate modeling are given in the end

Notes:

Vendor supplied data

Publisher Number:

2018-01-1028

Access Restriction:

Restricted for use by site license