Infoscience

Thesis

Experimental and Numerical Studies of Mode I and Mode II Delamination of Polymer Composites with Embedded Optical Sensors

Over the past decades the use of composite materials has enormously increased, especially in the aeronautical, automotive, and energy production industries. These materials allow to build lighter and larger structures which are more efficient. However, by introducing composite materials into load bearing structures, new modes of failure have to be understood to further improve design and guarantee the safety during the whole life time of such parts. Since composite materials are often produced in layered structures, they are prone to delamination where a crack propagates between the layers. The mechanical tests which are used today to measure the fracture properties cannot account for all processes associated with delamination. Namely, intact fibres that link the fracture surfaces, so called bridging fibres, can strongly influence the outcome of such tests, however, they cannot be quantitatively measured. In this work, a semi-experimental method was developed and used to study delamination tests and identify the contribution of bridging to toughness. For this, the strain distribution around the crack tip was measured with embedded optical sensors, so called fibre Bragg gratings (FBG). A new methodology based on multiplexed FBGs was developed and allowed to acquire a quasi-continuous strain distribution at relatively high rates. The results were then used in an inverse identification method to determine parameters which characterise delamination and bridging. Mode I delamination was studied with the double cantilever beam test in monotonic and fatigue loading. Using the above mentioned method the closing tractions due to bridging fibres were identified and their contribution to the resistance against crack propagation was determined. Compared to the monotonic loading, the contribution of bridging in fatigue was found to be about 30% higher. With a cohesive zone finite element model which was accounting for the bridging tractions, the onset and propagation of the delamination were correctly predicted. The bridging was found to contribute by 50% to the total energy release rate (ERR). In a similar way mode II delamination was studied with a four point end notched flexure test and the ERR as well as friction coefficients were identified using the measured strain distribution. While bridging was found negligible, the ERR in mode II was three times higher than the initiation value of mode I delamination. Finally, in a mixed mode bending test the modes I and II were combined so that each one was contributing by 50% to the delamination. The initiation value was found to be about 15% higher than the one of mode I, although with a large scatter, while the propagation value was only marginally higher. The fibre bridging, which formed during the delamination, was characterised from the changes of strain measured with the embedded optical sensors. The proposed method of strain measurements with embedded optical sensors and inverse identification offers an interesting alternative to existing methods used to characterise delamination. It was successfully used to identify fibre bridging without assumptions on the length of the bridging zone and opens a new way to study micromechanics of fracture in laminated materials.

Related material