On Efficient FE Simulation of Pulse Infrared Thermography for Inspection of CFRPs

Abstract

This paper reports an investigation on efficient finite element (FE) simulation of pulse infrared thermography of CFRPs. The very short duration of pulse heating as well as the layered structure and the anisotropic thermal properties of CFRPs impose specific constraints to the FE analysis of this transient thermal problem. Hence, for FE based optimization of pulse thermographic experiment, appropriate selection of FE parameters is critical so that the computational intensity is reduced without compromising the defects' quantification. Various modelling aspects including non-uniform heat flux induced by the flash lamp, element formulation and appropriate mesh sizing of out-of-plane and in-plane directions and the defect modelling approach are discussed. The effect of the modelling approach on the defect quantitative parameters and post-processing of calculated thermal images is investigated. Alternative modelling approaches are proposed and compared in terms of their reliability for defect quantification and their computational cost. IntroductionOptical infrared thermography (IRT) has been attracting significant industrial interest to be developed as a nondestructive testing (NDT) technique with trusted reliability. An optical source radiates heat onto the surface of the test piece and an infrared camera provides full-field contactless measurement of the surface temperature evolution. This temperature evolution is interpreted for the detection and quantification of damage [1]. Lock-in and pulse thermography are the two common approaches of optical IRT. In lock-in thermography (LT) the excitation power is harmonically modulated at a particular frequency while for pulse thermography (PT) a very short heat pulse is emitted to the sample's surface with a flash lamp. PT enables excitation of a broad range of frequencies through a single excitation cycle and provides more information about the subsurface defects compared to LT, though with reduced SNR.Simulation of IRT is of great importance as it provides support for model-based development, implementation and optimization of the experiment, and finite element FE analysis has proved its robustness in serving this purpose [2][3][4][5]. Optical IRT is based on thermal wave propagation through the depth of the test piece and its interaction with subsurface defects. Hence, the FE results should ideally be comparable with the promises of the IRT experiment and have to (i) provide fast and reliable simulation of the experiment to inspect a relatively large area, (ii) adequately define the through depth thermal behavior and (iii) provide thermal images similar to those measured by the infrared camera.CFRPs have wide structural application and are prone to various defect types e.g. fiber breakage, matrix cracking and delamination. The strongly anisotropic thermal conductivity and multi-layer structure of CFRPs significantly affect the thermal wave propagation which should be appropriately taken into account in the FE model. This paper reports an investigation on FE an...

Performance of frequency and/or phase modulated excitation waveforms for optical infrared thermography of CFRPs through thermal wave radar: A simulation study

“…The top ply is modelled by a graded mesh of two element layers (such that the thickness of the top layer is half of the bottom layer) and the lower plies are modelled with a single element layer. The graded mesh of the top ply is needed for adequate calculation of the surface temperature during fast thermal events [37] (e.g. when the heating amplitude experiences sudden changes during a modulated waveform).…”

“…Temperature-dependent thermal properties are assumed for the unidirectional CFRP [35] with room temperature ( ) diffusivity 25℃ = 5.39 × and along and perpendicular to the fibers, 10 -6 m 2 /s = = 6.16 × 10 -7 m 2 /s respectively. The non-uniform heating induced by the optical source is modelled by a Gaussian distribution of the heat flux throughout the sample's surface in -plane [37], as follows:…”

“…Moreover, thermal interactions of the sample's external faces with the ambient, through radiation and convection, are taken into account. For more details concerning the FE model, the reader is referred to [37].…”

Novel discrete frequency-phase modulated excitation waveform for enhanced depth resolvability of thermal wave radar

“…The radiative heat flux emitted to the sample's surface is calculated by assuming a Gaussian distribution of the heating intensity and a uniform emissivity coefficient of 0.9. For more details concerning the FE model the reader is referred to [25]. The assumed material properties of glass fiber and epoxy resin and resultant diffusivities calculated for a GFRP with 50% volume fraction of glass are given in Table 1.…”