In English

Finite Element Model Evaluation using Factorial Design: Simulation of Chloride diffusion in 3D heterogeneous concrete

HOSSEIN TAHERSHAMSI
Göteborg : Chalmers tekniska högskola, 2017. 34 s.
[Examensarbete på avancerad nivå]

Concrete structures should be capable of withstanding the conditions throughout the serviceability of the structure. Due to the exposure of the external agents arising from the environment e.g. chloride ions, concrete structures will gradually deteriorate. Transport of chloride ions in concrete is a complex phenomenon which is hard to predict and model. However, modelling is highly preferable as a means for assessing and predicting the behaviour of existing concrete structures, such as bridges. The behaviour of such models, nevertheless, need to be evaluated as a prediction tool in order to optimise cost and time of experimental studies. The general aim is to perform a statistical method called "factorial design" on sets of outputs taken from a 3D heterogeneous concrete model. The obtained results are used to observe the main effects and interactions of different factors. Two different cases of chloride transport are investigated: stationary and transient diffusion. The analysis of stationary case shows a proper relation between aggregate content and Interfacial Transition Zone(ITZ) diffusivity. A resolution IV fractional factorial was designed for six factors of the transient model. The data shows that more gravel content increases the time for chloride ions to reach the chloride threshold. Three dominant parameters- gravel content, cement storage capacity and cement diffusivitywere screened to perform a full factorial design. No specific interactions were observed among the three studied parameters. However, the influence of higher order interactions were noticed between the fractional and full factorial design.

Nyckelord: Diffusion, FE simulation, Factorial design, Computational homogenization, Concrete



Publikationen registrerades 2018-01-30.

CPL ID: 254867

Detta är en tjänst från Chalmers bibliotek