This study investigates the propagation behavior of Lamb waves in a steel beam at varying inter-sensor distances, employing CWT, FFT, and STFT to evaluate their effectiveness for time–frequency and energy-based signal characterization. Results show that the Lamb-wave time-of-flight exhibits minimal variation despite changes in sensor spacing, yielding a nearly constant group velocity of approximately 4540 m/s, which confirms the stable propagation of the S₀ mode. Multiscale wavelet analysis reveals that nearly 90% of the total energy is concentrated in Scale‑1, while the combined contribution of Scales 2–5 remains below 10%, indicating a dominant, low-dispersion single-mode response.

Quantitative CWT-based indicators also show consistent trends with increasing distance. Wavelet energy increases by about 20%, whereas entropy decreases from 0.4908 to 0.4651, reflecting stronger energy localization in Scale‑1 and improved separation of the S₀ mode from background noise and secondary modes. The RMS value increases from 0.2199 to 0.239, suggesting reduced attenuation and lower dispersion along the propagation path. The reduction in kurtosis further indicates diminished impulsive peaks, increased waveform smoothness, and an enhanced signal-to-noise ratio. Overall, the findings demonstrate that CWT provides superior capability over FFT and STFT in analyzing energy evolution, attenuation characteristics, and time–frequency dynamics of guided waves. These advantages establish CWT as a robust quantitative tool for structural health monitoring and guided-wave analysis.