Influence of modelling hypotheses on strength assessment of CFRP stepped repairs

Composite stepped repairs can achieve high strength recovery without the addition of bolts or fasteners to the structure. They are therefore a major issue in the field of aerospace composite structure damage repair. However, there is no standardized method to design this type of repairs. Many analytical, semi-analytical and finite element models were proposed throughout the years to predict the strength of stepped repairs, using various hypotheses and simplifications. The aim of this paper is to investigate the influence of modelling hypotheses on stress distribution and strength prediction of composite stepped repairs. Five simplified stepped joint models using macro-element (ME) modelling and finite element method (FE) are compared to a full 3D FE model of a stepped repaired panel. The influence of step length and adhesive fracture toughness was investigated to determine the field of validity of each model. Among FE models, it was shown that modelling the equivalent joint under 2D generalized plain strain gives a very close strength prediction to the 3D stepped repair model while saving computation time. Simplified macro-element models under bar or beam hypotheses are fairly close to the results of FE modelling, but the deviation between those and FE is sensitive to step length and adhesive fracture toughness.