A finite element approach to predicting vibrothermographic heat detections of fully embedded delaminations within composite plates

Abstract

Vibrothermography is a Nondestructive Evaluation (NDE) technique which is particularly well suited for locating discontinuities, such as cracks and sub-surface delaminations within composite structures. The Vibrothermographic method relies on high-amplitude vibrational excitations within a structure to cause frictional rubbing along flaw surfaces. Frictional heat energy dissipated between flaw surfaces diffuses through the material of the structure to a surface where heat signatures are monitored using infra-red thermographic camera technologies. Vibrothermographic testing methods, developed in the late 1970's with the advent of improved IR camera technologies, have been somewhat slow to develop, due in part to a lack of understanding of the physics surrounding damaged areas of structures. The focus of this research is on the use of Finite Element (FE) Analysis, using ANSYS® software, to simulate the Vibrothermographic method. One goal of the research is to produce algorithms predicting the likely-hood of thermal detection of numerous delamination flaws within a fiberglass/epoxy composite plate. The FE modeling attempts accurately characterize harmonic responses surrounding the "damaged" regions (flaw areas) of the plate. Harmonic response analyses are performed on the FE plate model, and nodal harmonic displacement data is subsequently utilized within transient analysis sub-modeling procedures. Each of the nine delamination flaws are analyzed within these sub-modeling procedures and the rates at which frictional heat energy dissipation, due rubbing, occur at the flaw faces are derived. Each delamination sub-model is subjected to a high-resolution frequency sweep analysis, where frequency dependent heat generation data is collected over a frequency bandwidth of 800-30000Hz. The latter portion of the paper discusses the use of a novel finite element approach to estimating the likely hood of thermal detections for the flaws. The approach uses generic finite element model to predict thermal detection times for an expansive set of flaw and testing parameters. A response surface method of curve-fitting is applied to the collected detection time data, and the resulting equation is subsequently used to predict the heat-signature detectability for each flaw. The results of this method are ultimately compared to empirical vibrothermographic frequency-sweep data, and a preliminary assessment of the viability of the approach is made.