Probabilistic Damage Tolerance Analysis of Debonded Sandwich Composites

A methodology for probabilistic damage tolerance analysis of a debonded composite is developed and demonstrated using a double cantilever beam (DCB) made of sandwich composite materials. An initial debond (crack) length is assumed to exist at the upper face/core interface and the crack remains at or near this interface during the crack propagation phase, and the risk of failure is monitored and managed by using nondestructive inspections and subsequent maintenance actions. An empirical energy-release rate equation of the sandwich DCB based on a modified beam theory was adopted for this study. It is assumed that when the Mode-I stress-intensity factor exceeds the fracture toughness, the stress will cause unstable crack growth and structural failure. Probabilistic distributions are used to treat uncertainties in fracture toughness, crack growth rate, manufacturing and repair quality, and a random process consisting of independent maximum flight loads. To conduct the analyses, an advanced and efficient method, RPI (Recursive Probability Integration), is used which combines random simulations with numerical integration. RPI maintains the generality of random simulations while taking advantage of the efficiency feature of numerical integration. The example study has demonstrated that RPI can be several orders of magnitude more efficient than the traditional Monte Carlo method. In addition, unlike other efficient probabilistic methods, there is no limitation on the number of random variables that can be treated within the risk assessment framework.