An analytical method to predict stiffener debonding in postbuckled compression panels

Structural efficiency, production rate and tooling costs for composite structures can all be improved if the current practice of co-curing stiffeners to panels is replaced by secondary bonding of pre-cured stiffeners to pre-cured panels, whilst exploiting postbuckling strength. Component-level Finite Element Analysis (FEA) of stiffener debonding in postbuckled panels is extremely costly due to the small mesh size required to resolve stresses at the local crack front. A new strip-based analytical method for long panels is presented, which calculates the strain energy release rate for debonding using the sensitivity of panel buckling loads to small perturbation of bonded widths, representing damage growth. The predictions of the strip analysis are compared with experimental test and FEA, namely the Cohesive Zone Method (CZM) and Virtual Crack Closure Technique (VCCT). The debonding load predicted by the strip method is within 1% of the CZM for a panel with a co-cured thermoset flange. When a thermoplastic (welded) panel-flange bond is analyzed, the CZM load is 9% lower than the new method, due to the inherent inaccuracy of the CZM for ductile debonding. A welded hat-stiffened panel debonds from beneath the hat, as observed experimentally and by VCCT. In this case, the new method is within 2% of experimental and VCCT debonding loads. The strip method predicts the stability of propagation and is several orders of magnitude faster than CZM and VCCT. It can be used for design and uncertainty quantification at stiffened-panel scale, providing a route to weight saving and structural certification by analysis.