Performance Analysis of Impact-Echo Detection for Tunnel Lining Voids Based on Combined Numerical and Experimental Results

ABSTRACT With the increasing number of operating tunnels, voids behind linings have become a frequent defect that can trigger leakage, cracking, and local instability. Impact-echo (IE) testing offers a rapid and nondestructive option for detecting such defects, but its parameter matching and quantitative accuracy remain challenging in multilayer composite lining systems. This study combines finite-element simulations with laboratory model tests to quantitatively evaluate how monitoring point location, void depth and size, impact loading characteristics, and multi-void interaction affect the dominant spectral peak and depth inversion accuracy. Numerical and experimental results agree well, supporting the applicability of IE for lining-void detection and revealing the parameter-matching characteristics of IE responses. Void depth dominates the peak-frequency variation: depth inversion errors are under 10 % at central monitoring points but increase near void boundaries because of multipath interference. When the excitation wavelength is approximately twice the void depth, the inversion error can be limited to within 5 %. Larger voids improve energy coupling and stabilize the dominant peak; when the side length exceeds 50 cm, the inversion error drops less than 5 %. In multi-void scenarios, shallow adjacent voids introduce stronger lateral reflections, leading to local frequency elevation and reduced inversion stability. Based on these findings, practical recommendations are provided for point placement, impact-frequency selection, and multi-point spectral fusion to improve reliability in complex tunnel inspections.