A numerical and experimental investigation of dynamic fracture in Polyamide 11: the effect of the sample geometry

Abstract

An experimental set-up and a numerical model are proposed to study the rapid crack propagation (RCP) resistance of polyamide 11 (PA11). Pipe and plate samples are studied. The solicitation type, imposed displacements or pressure, of polymer pipes is discussed. The necessity to pre-stress polymer pipe with imposed displacements is highlighted. Indeed, the work of external forces is not negligible for pressurized polymer pipes. A reliable estimate of the dynamic energy release rate GId is in this last case not guaranteed. A new experimental set-up is used to Pre-Stress Pipe Specimen (PS2) in mode I. The crack is initiated artificially with an external impact on a razor blade. A quasi-constant dynamic regime of propagation is then reached on about 20 cm. A finite element procedure is used to estimate GId. Knowing the crack tip location during RCP inertia effects i.e. kinetic energy are quantified. Numerical results reveal a higher dynamic correction factor for a pipe (0.2) than a plate structure (0.9). An important and non negligible part of the stored energy is dissipated by the structure during RCP in pipe structure. Crack tip location as a function of time is measured with the help of a high speed camera during dynamic regime of propagation. The calculated mean crack tip velocity is quasi-constant in PA11 whatever (i) the initial stored energy in the structure, (ii) the sample geometry and (iii) the crack configuration. This velocity is known to be the crack branching velocity (0.6cR). The dynamic energy release rate GId is equal to 1.5±0.1 kJ m-2 for a pipe sample and 9.2±0.7 kJ m-2 for a plate sample at the crack branching velocity. Fracture surface analyses are leaded to explain this significant difference.