EFFECT OF FIBRE ARCHITECTURE ON THE FALLING WEIGHT

EFFECT OF FIBRE ARCHITECTURE ON THE FALLING WEIGHT IMPACT PROPERTIES OF HEMP/EPOXY COMPOSITES Carlo Santulli Università di Roma – La Sapienza, Italy carlo. santulli@uniroma 1. it Alessandro Caruso ITIS "Alessandro Volta" Sassuolo, Italy

SUMMARY • Aim of this work • Impact in plant fibre composites (IFW) • Impact hysteresis cycles • Mode of failure • Conclusions

AIM OF THIS WORK Comparing falling weight impact (IFW) performance of laminates obtained using non-woven hemp mat or hemp loose fibres (in two configurations: unidirectional and 0/90) by analysing impact hysteresis (force vs. deflection) cycles and relating them to their mode of failure under bidimensional impact loading. In all three configurations the laminates were produced by hand lay-up by introducing the maximum volume of hemp fibres that allowed an effective impregnation by the polymer matrix.

FALLING WEIGHT IMPACT IN PLANT FIBRE COMPOSITES The measurement of some variables from the hysteresis cycle, and partition of impact energy in three terms, namely elastic energy, plastic energy and damping (or rebound) energy, can give indication on the mode of failure of these materials under high strain rate loading.

EXPERIMENTAL COMPOSITE MANUFACTURING Two laminate configurations have been realised using loose hemp fibres, disposing them all with the same orientation (0°) in one case (LU laminate), whilst in the other case layers are laminated with a 0/90° configuration (LC laminate). The third configuration has been obtained using non-woven hemp mat (M laminate). IMPACT TESTING Samples have been impacted from a height of 0. 5 m by a 11. 4 kg mass, so to obtain a 56 Joules impact energy, using a 19. 8 mm diameter impactor. Considering the thickness of the samples (3. 1 mm. ) this energy was deemed more than sufficient to produce full penetration in all the laminates, so that from the area of the hysteresis cycle, the energy absorbed during penetration could be measured.

QUASI-STATIC TESTS (TENSILE AND FLEXURAL) The best performance, but also the largest scattering in properties is obtained from the unidirectional composites (LU). The latter is possibly due to the difficulty in disposing the fibres exactly along the selected direction.

TYPICAL HYSTERESIS CYCLES Hysteresis cycles show a dampened rebound only for LU laminates, whilst in the case of LC and particularly M laminates the rebound energy after penetration is released at quasi-constant rate, with a very limited plastic phase and a hardly detectable load drop.

Measurement of the damping ratio (as from Benevolenski et al. , 2000) in two cases: Dampened rebound phase (LU laminate) (left) and Quasi-linear rebound phase (M laminate) (right) The much larger values and scattering of damping ratio in LC and M laminates indicate that the energy absorbed through rebound may be substantial (although by no means consistent in all the laminates) which would possibly result in a higher impact resistance, provided that no preferential damage propagation through defects is triggered.

VARIABLES MEASURED FROM HYSTERESIS CYCLES a. Maximum load; b. Deflection at maximum load; c. Maximum deflection; d. Load corresponding to a deflection of 3. 1 mm (laminate thickness); e. Linear stiffness; f. Hysteresis energy; g. Damping ratio

MAX LOAD AT THICKNESS DEFLECTION AT MAX. LOAD LINEAR STIFFNESS MAX. DEFLECTION HYSTERESIS ENERGY

RESULTS FROM HYSTERESIS CYCLES ANALYSIS Impact properties appear superior for the non-woven mat reinforced laminates, especially due to the fact that lower deflection is obtained during the impact event, often leading to strong rebound rather than dampened vibration of the plate. Also, in spite of the slightly higher average value of linear stiffness obtained for the LU laminates, M laminates show less scattered values during the linear phase. In contrast, scattering between tests during the plastic and rebound phases is much larger in M and LC laminates than in LU lamnates, since in the latter the mode of impact damage is almost exclusively crack propagation oriented in the fibre direction. Average penetration energy calculated from hysteresis cycles comes out to be around 190 k. J/m 2 for LC laminates, 140 k. J/m 2 for M laminates, and 45 k. J/m 2 for LU laminates. As a comparison from previous hysteresis cycles studies, GMT (glass mat thermoplastics) was in the region of 160 k. J/m 2 (Santulli 2003) and flax/epoxy was around 60 k. J/m 2 (Santulli et al. , 2005).

Unidirectional loose hemp fibre laminate Diametral crack propagation oriented in the fibre direction (some fibre orientation mismatch is noticeable especially at rear) 0°/90° loose hemp fibre laminate Non-negligible amount of spalling (material loss at rear) and therefore ineffectivity of a part of the laminate to withstand impact load Non-woven mat hemp fibre laminates More perfect (quasi-circular) penetration although some slight preferential direction is shown on the non-impacted face.

CONCLUSIONS • This work confirms that hemp fibre reinforced composites have the potential to replace semi-structural fibreglass laminates in terms of impact resistance. • In particular, disposing the reinforcing fibres in mats in place of using them loose, while it reduces the fibre volume introduced in the matrix with some detrimental effect on static properties, makes in contrast the laminates less vibrationprone during the impact event. • This ultimately results in a higher possibility of impactor rebound, even after penetration, and in a higher amount of energy dissipated during impact. • The full study of hysteresis cycles obtained by falling weight impact tests is confirmed to represent a useful tool for the evaluation of falling weight impact resistance in plant fibre composites.