This Additional Graduation Work consists of two different fields of research: experimentally determining the linear coefficient of thermal expansion, and by Finite Element Modelling determining salt water diffusion coefficients. The research is centered around the wrapped composi
...
This Additional Graduation Work consists of two different fields of research: experimentally determining the linear coefficient of thermal expansion, and by Finite Element Modelling determining salt water diffusion coefficients. The research is centered around the wrapped composite joint: an innovative technology using a composite wrap connecting steel hollow sections instead of traditional welds.
Linear Coefficient of Thermal Expansion
The Linear Coefficient of Thermal Expansion is determined theoretically and experimentally.
The results show for hand layup composites a respectively theoretical and experimental coefficient of 1.85 *10-5 /°C and 2.15 *10-5/°C in longitudinal and transverse direction and of 4.80 *10-5/°C and 7.82 *10-5/°C in through-thickness direction. For composites produced by a new production method, resulting in a higher fibre volume fraction the respectively theoretical and experimental coefficient is 1.44 *10-5/°C and 1.59 *10-5/°C in longitudinal and transverse direction and 4.01 *10-5/°C and 5.79 *10-5/°C in through-thickness direction.
Especially in through-thickness direction difference between theoretical results are up to 62%. This discrepancy is expected to be caused primarily by defects present in the material, as specimens are produced from mechanically tested material. Defects in the material cause discontinuities, which can lead to a higher apparent CTE as the matrix is less confined by fibres.
Diffusion Modelling
With Finite Element Software Abaqus the rate of diffusion of salt water into composite produced by means of hand layup is determined. Experimental results of submersion tests in salt water under room temperature were used for validation of the models. Material is assumed to have a moisture content of 0 at the start of the experiment. On top of that, a saturation of 0.47% is assumed, based on previous research.
Firstly, the in-plane diffusion coefficient Dx,y is determined. Results of Finite Element analysis are validated against a slice of non-post cured material with least thickness, and thus primary diffusion direction in longitudinal direction. In Abaqus Fickian behaviour is assumed, with a negligible small diffusion in through-thickness direction. This results in Dx,y = 0.08 mm2/day.
Secondly, the in-plane diffusion coefficient is used to determine the through-thickness diffusion coefficient by using experimental weight increase by salt water uptake of End Notched Flexure and Interlaminar Shear coupons as validation for Finite Element models. This results in Dz = 0.03 mm2/day.
Thirdly, the saturation of a bi-material steel composite End Notched Flexure coupon is numerically determined. It is determined that in 4.6 years the coupon would fully saturate.
The research showed big differences in saturation rates of different geometries and conditions (post cured or non-post cured). This can primarily be explained by the assumption material is dry before experimentally submerging. Especially non-post cured material will have a non-zero begin moisture content. By not taking this into account, the moisture content is underestimated.
Additional experiments including drying of material should be executed to with certainty determine the diffusion coefficient and saturation of the material.