Implementation of Kriging Surrogate Models for Delamination Detection in Composite Structures

Composite structures are prone to impact and fatigue loads in service which in turn causes defects such as de-lamination. This affects the dynamic characteristics of the composite laminates such as its Natural frequency, Mode shapes and its damping characteristics. Hence it can be used as a prominent indicator for structural health monitoring. This paper focusses on the use of Kriging based surrogate model for the prediction of delamination in Carbon fibre reinforced epoxy composite panels with an edge constraint. The change in natural frequency and mode shapes of the damaged and undamaged model has been taken as the primary damage indicator for the present analysis. The size and location of the damage were predicted from the developed surrogate models. The performance of the proposed approach has been studied experimentally and compared it with numerical simulation and has been found satisfactory. The surrogate approach reduces the computational time considerably without compromising the accuracy