Critical Traverse Velocity Threshold for Through-Cut of 3 mm CFRP by Abrasive Waterjet: a Binary Logistic Model Based on Post-Cut Images

Abstract

Abrasive waterjet (AWJ) trimming of thin carbon fibre-reinforced polymer (CFRP) laminates faces the challenge of defining a reliable process window to ensure full penetration without costly metrology. This study establishes, for the first time, a statistically validated through-cut threshold for 3 mm quasi-isotropic T300/epoxy CFRP under standard industrial AWJ parameters (pressure: 200 MPa, abrasive mass flow rate: 120 g/min, standoff distance: 5 mm). Using only post-cut digital caliper measurements and photographic documentation—a low-cost metrology approach—a binary logistic regression model was calibrated from experimental data at three traverse velocities (300, 500, and 700 mm/min, n = 8 replicates each), yielding empirical through-cut probabilities of 1.00, 0.375, and 0.00, respectively. Firth’s bias-reduced penalised likelihood method was employed to address complete separation in the data, providing stable parameter estimates. The model identifies a critical traverse velocity of V₉₅= 292 mm/min (95 % profile-likelihood:276–308 mm/min) guaranteeing 95 % through-cut probability. This quantifies and validates the empirical shop-floor rule of "≤ 300 mm/min". The estimated transition zone width (~200 mm/min) is markedly narrower than that reported for ductile metals, attributable to the low interlaminar fracture toughness of CFRP, which precipitates abrupt, unstable delamination once jet energy flux falls below a critical level. Cross-validation demonstrated a predictive accuracy of 95.8 %. The study provides a robust, accessible framework for process window calibration in small-to-medium enterprises, bridging the gap between empirical practice and statistically controlled manufacturing for thin CFRP components.