Computational & Technology Resources
an online resource for computational,
engineering & technology publications
Civil-Comp Proceedings
ISSN 1759-3433
CCP: 109
Edited by: Y. Tsompanakis, J. Kruis and B.H.V. Topping
Paper 4

Surrogate Models for Numerical Structural Design with Polymorphic Uncertain Parameters

W. Graf, M. Götz and M. Kaliske

Institute for Structural Analysis, Technische Universität Dresden, Germany

Full Bibliographic Reference for this paper
W. Graf, M. Götz, M. Kaliske, "Surrogate Models for Numerical Structural Design with Polymorphic Uncertain Parameters", in Y. Tsompanakis, J. Kruis, B.H.V. Topping, (Editors), "Proceedings of the Fourth International Conference on Soft Computing Technology in Civil, Structural and Environmental Engineering", Civil-Comp Press, Stirlingshire, UK, Paper 4, 2015. doi:10.4203/ccp.109.4
Keywords: numerical design, surrogate model, metamodel, polymorphic uncertainty.

The design of structures is one of the major tasks for engineers. The objectives of numerical design, computing robust and reliable structures, can be realized by means of analysing different variants, application of optimization tasks, or solving inverse problems. The method of choice depends on the aim of design and the problem.

Numerical structural design should be robust with respect to the polymorphic nature and characteristic of the available information. Uncertainties are inherently present in the resistance of structural materials, environmental and man-imposed loads, boundary conditions, physical and numerical models, and to other types of aleatoric and epistemic uncertainties. Generally, the availability of information in engineering practice is limited. Incomplete, fragmentary, diffuse, and frequently expert specified knowledge leads to imprecision in data.

This paper presents the passive and the active approach for design tasks analysed with optimization or solving the inverse problem. The advantages and disadvantages of each concept are pointed out. The solution of the inverse problem, suitable in the early design stages to detect permissible design spaces is one of the main points of this paper. Especially for early design stages it is necessary to take polymorphic uncertainty into account due to a lack of information. This incorporation yields to increasing numerical effort.

One possibility to solve the design task in appropriate time is to apply surrogate models. This surrogate model generates a metamodel to reproduce the deterministic objective function. In this contribution, the improvement in efficiency is demonstrated by the sophisticated metamodel, Extreme Learning Machine.

The applicability is demonstrated by means of engineering examples.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description