Nonlinear Modeling of Carbon Fiber Reinforced Composites at Multiple Scales

Simon, Jaan-Willem; Fish, Jacob; Reese, Stefanie; Wagner, Werner

Nonlinear Modeling of Carbon Fiber Reinforced Composites at Multiple Scales

Simon, Jaan-Willem^RWTH*

2018

Verantwortlichkeitsangabevorgelegt von Dr.-Ing. Jaan-Willem Simon

2018

Habilitationsschrift, RWTH Aachen University, 2018

Hauptberichter/Gutachter
Reese, Stefanie (Thesis advisor)^RWTH* ; Wagner, Werner (Thesis advisor) ; Fish, Jacob (Thesis advisor)

Tag der mündlichen Prüfung/Habilitation
2018-02-28

Online
URL: https://publications.rwth-aachen.de/record/755388/files/755388.pdf

Einrichtungen

Lehrstuhl und Institut für Angewandte Mechanik (311510)

Kurzfassung
Composite materials are generally formed by combinations of two or more components to achieve properties that are superior to those of the single constituents. Their use has become very popular in numerous engineering applications in aerospace, automotive, construction, and maritime industry. Hence, the evaluation of the effective mechanics properties of these materials is of increasing industrial and scientific interest.Modeling the mechanical behavior of composite materials can be challenging due to the complexities introduced by their microstructure. In order to set up models which can be applied on the structural level (macro-scale) with an acceptable computational effort and still account for effects resulting from the underlying microstructure (micro-scale), computational multi-scale methods are needed. Such methods range from hierarchical to synergistic and concurrent approaches. While there is no scale separation in (fully) concurrent methods, synergistic ones separate either the length scale — i.e. the volume of the simulated microstructure is smaller than the volume that it represents on the macro level — or the time scale. By contrast, in hierarchical methods both the time and length scales are separated. Here, the entire composite is considered as only one effective material with homogenized properties on the structural level, which results in a significant increase in computational efficiency while sufficient fidelityfor many applications is maintained.Applying multi-scale strategies requires the development of appropriate material models on the different scales. Moreover, these models usually include several material parameters which have to be determined either by real or virtual experiments. In the current work, virtual experiments are performed on representative volume elements (RVEs) in order to numerically obtain the response of the considered microstructure. To this end, the geom-etry of the constituents and their distributions need to be reflected sufficiently accurate, such that the RVE can be considered representative in a statistical sense. In order to illustrate the strategy, several different examples will be shown in the following. All presented results have been obtained by the author and co-workers from his group ”Composite Materials and Structures” at the Institute of Applied Mechanics — headed by Prof. Stefanie Reese — at RWTH Aachen University.

Composite materials are generally formed by combinations of two or more components to achieve properties that are superior to those of the single constituents. Their use has become very popular in numerous engineering applications in aerospace, automotive, construction, and maritime industry. Hence, the evaluation of the effective mechanics properties of these materials is of increasing industrial and scientific interest.Modeling the mechanical behavior of composite materials can be challenging due to the complexities introduced by their microstructure. In order to set up models which can be applied on the structural level (macro-scale) with an acceptable computational effort and still account for effects resulting from the underlying microstructure (micro-scale), computational multi-scale methods are needed. Such methods range from hierarchical to synergistic and concurrent approaches. While there is no scale separation in (fully) concurrent methods, synergistic ones separate either the length scale — i.e. the volume of the simulated microstructure is smaller than the volume that it represents on the macro level — or the time scale. By contrast, in hierarchical methods both the time and length scales are separated. Here, the entire composite is considered as only one effective material with homogenized properties on the structural level, which results in a significant increase in computational efficiency while sufficient fidelityfor many applications is maintained.Applying multi-scale strategies requires the development of appropriate material models on the different scales. Moreover, these models usually include several material parameters which have to be determined either by real or virtual experiments. In the current work, virtual experiments are performed on representative volume elements (RVEs) in order to numerically obtain the response of the considered microstructure. To this end, the geom-etry of the constituents and their distributions need to be reflected sufficiently accurate, such that the RVE can be considered representative in a statistical sense. In order to illustrate the strategy, several different examples will be shown in the following. All presented results have been obtained by the author and co-workers from his group ”Composite Materials and Structures” at the Institute of Applied Mechanics — headed by Prof. Stefanie Reese — at RWTH Aachen University.

Restricted:
PDF