Monitoring bond degradation and predicting failure modes in GFRP-reinforced concrete beams using information entropy analysis of acoustic emission waveforms

Bond degradation at the glass fiber-reinforced polymer (GFRP)-concrete interface poses significant structural risks and therefore requires reliable monitoring techniques. The present study has introduced an information entropy-based approach to evaluate bond deterioration in GFRP-reinforced concrete (RC) beams under flexural loading. The effect of concrete strength (M30/M60 grades), embedment length, surface configurations, and confinement conditions on the bond behavior has been investigated by systematically testing twenty-four specimens. Weighted information entropy has been derived from acoustic emission (AE) parameters including root mean square (RMS), signal strength, and energy. Waveform entropy based on Shannon's and Kullback-Leibler's formulations has captured the uncertainty in AE signal amplitude distributions. Results have shown that energy-based entropy contributes significantly more to the total entropy (55–65 %) in comparison to the RMS and the signal strength entropies (15–25 % each). Energy entropy has been observed to be higher during peak-stress stages, whereas RMS entropy has been observed to be more susceptible to debond initiation. High-strength concrete specimens have exhibited 32.6 % higher bond strength with correspondingly higher entropy at peak stress. Longer embedment lengths of the GFRP rebar have showed 20 % lower bond strength than GFRP bars with shorter embedment lengths with stable entropy up to 30 %-40 % of peak stress. Confinement reinforcement has improved bond strength by 14 % and altered failure modes from brittle splitting to ductile pull-out. The entropy-based methodology has successfully distinguished between pull-out, splitting, and splitting-shear failure modes and have thereby provided reliable indicators for real-time structural health monitoring of GFRP-RC members. • An entropy-based AE approach is proposed for tracking bond degradation in GFRP-RC beams. • Energy and RMS entropy effectively capture damage evolution and failure stages. • Unconfined specimens show higher entropy, indicating brittle bond failures. • Information entropy analysis reveals disorder in AE signals due to interfacial debonding. • The method supports real-time structural health monitoring of GFRP-RC systems.