Parametric Optimization of Frequency-Selective Thermal Excitation for Depth-Dependent Defect Detection in GFRP Composites

Despite significant advances in infrared thermography, most existing studies are confined to single defect types or excitation modes. This work pioneers a coupled spectrum-thermal response analysis of four excitation modalities: continuous heating, Absolute-Value Sinusoidal, pulsed (square/triangular/sawtooth wave), and step heating, revealing that their "optimality" depends on defect characteristics (type/depth) and detection objectives. Focusing on inclusion defects in GFRP composites, a COMSOL-based 1D model demonstrates excellent agreement with experimental data (r≈0.999, MAE≈0.072°C, RMSE≈0.068°C).Key findings indicate that continuous heating suits deep defects while pulsed excitation favors shallow ones, with square waves (1s period, 50% duty cycle) proving optimal. Fourier series decomposition elucidates the underlying frequency-selective matching mechanism, establishing a theoretical framework for optimizing thermal excitation parameters in composite nondestructive testing.