Speaker
Description
Modern structural engineering increasingly relies on lightweight, high-performance materials. To minimize mass, the use of excessive safety factors is no longer viable, and designs are pushed closer to their operational limits. This necessitates reliable Non-Destructive Testing (NDT) to maintain operational standards. One of the challenges in this domain is the detection of Barely Visible Impact Damage (BVID) in composite structures. BVID is often characterized by a complex morphology that combines subsurface delaminations with surface-breaking cracks.
This work addresses these challenges by developing a comprehensive framework for active Laser Spot Thermography (LST). While LST allows for precise inspections, its defect detection capability depends on the specific excitation parameters. This study proposes a quantitative, numerical simulations-based framework for the selection of the inspection parameters for LST. The framework comprises three different stages.
First, for materials with unknown thermal properties, local thermal diffusivity is experimentally estimated via initial LST measurements. Using known density and specific heat capacity values, the diffusivity is converted into thermal conductivity to define the material properties in the numerical model.
Second, the workflow utilizes a Finite Element Method (FEM) environment where the sample geometry is fully parameterized. It performs 3D thermal simulations under a stationary Gaussian-shaped heat source across a grid of excitation points, mirroring the experimental procedure. This environment determines the effective inspection parameters: laser power, heating time, cooling time, and the spatial scanning step. The selection algorithm balances three constraints: maintaining the surface temperature below a user-defined safety threshold, achieving a required Signal-to-Noise Ratio (SNR), and determining the largest acceptable scanning step. The algorithm iteratively reduces the scanning step size until the specific SNR target is met, thereby minimizing the total inspection duration.
Finally, in the third stage, the determined parameters are applied to the inspection. To ensure consistent defect quantification, the framework employs a unified signal processing pipeline that incorporates Thermal Signal Reconstruction (TSR) and spatial gradient analysis, applied identically to both synthetic and experimental datasets.
The accuracy of the simulation results was validated against experimental data from Carbon Fiber Reinforced Polymers (CFRP) with Flat-Bottom Holes (FBH) and aluminum plates with artificial notches. The comparative analysis demonstrated that the simulation provides conservative SNR estimates (typically underpredicted by approx. 15%) to ensure reliable detection. Subsequently, the methodology was applied to case studies involving a steel sample with artificial cracks and a CFRP sample with BVID damage, where it successfully resolved both the surface cracks and the underlying delamination.
The obtained results indicate that numerical modeling effectively supports NDT planning and is a valuable complement to experimental testing that facilitates precise material characterization and effective detection of defects.
