Experimental monitoring and modeling of fatigue damage for 3D-printed polymeric beams under irregular loading

Abstract

Studying the fatigue of 3D-printed (3DP) structures usually requires time-consuming and ad hoc material characterizations. This paper experimentally demonstrates a non-intrusive structural health monitoring (SHM) framework capable of monitoring and modeling fatigue damage of 3DP structures. The experiments emulate realistic working conditions of machinery with polyethylene terephthalate beam parts as a subsystem. The pre-notched specimens with 0-, 45-, and 90-degree raster angles are manufactured and tested under highly irregular fatigue loads. Based on the phase space warping algorithm, the smooth orthogonal decomposition identifies the slow-time higher-order damage feature space, whose leading subspace is used as the estimated scalar-damage-time history. The critical inflection points (CIPs), identified in the higher-order damage feature subspace, delineate the three fatigue crack propagation stages with different rates of damage evolution. The lower-order projection of the CIPs in the damage-time histories can serve as early and late damage indicators, with a maximally 50% ahead of any observable signs of structural degradation from the response. The first CIP indicates the damage as early as at 10% of the total life span irrespective of the raster orientation. Based on a normalized temporal damage model, a hypothesis is posed to ascribe the raster angle-induced damage mechanisms to inter-laminar and cross-laminar fractures, validated by an optical-microscopy-based postmortem analysis. This paper provides an ad hoc alternative for studying fatigue damage with minimal assumptions and prior knowledge of the 3DP structure. Furthermore, the identified damage mechanisms suggest aligning the raster orientation with the applied stress to maximize the machinery’s fatigue life.