In-plane elastic and interlaminar properties of composite laminates are commonly obtained through separate experiments. In this paper, a simultaneous identification method for both properties using a single experiment is proposed. The mechanical properties of laminates were treated as uncertainties and Bayesian inference was employed with measured strain-load curves in compression tests of laminates with embedded delamination. The strain-load curves were separated into two stages: the pre-delamination stage and the post-delamination stage. Sensitivity analysis was carried out to determine the critical properties at different stages, in order to alleviate the ill-posed problem in inference. Results showed that the in-plane Young's modulus and shear modulus in elastic properties are dominant in the pre-delamination stage, and the interlaminar strength and type I fracture toughness in interlaminar properties are dominant in the post-delamination stage. Five times of property identification were carried out; the maximum coefficient of variation of identified properties was less than 1.11%, and the maximum error between the mean values of the identified properties and the ones from standard experiments was less than 5.44%. The proposed method can reduce time and cost in obtaining multiple mechanical properties of laminates.
Keywords: delamination; dynamic Bayesian network; laminate; property identification; sensitivity analysis.