Skip to Content
SensorsSensors
  • Article
  • Open Access

18 November 2026

36 Pages

Physics-Driven Separation of Overlapping Lamb Wave Packets of the Same Mode

,
,
,
,
,
,
and
1
School of Intelligent Manufacturing and Equipment, Jiangmen Polytechnic, Jiangmen 529000, China
2
School of Electronic Information and Electrical Engineering, Yangtze University, Jingzhou 434023, China
3
National Demonstration Center for Experimental Electrotechnics and Electronics Education, Yangtze University, Jingzhou 434023, China
4
School of Mechanical Engineering, Yangtze University, Jingzhou 434023, China
This article belongs to the Section Physical Sensors

Abstract

Separating overlapping Lamb wave packets of the same mode is challenging because dispersion broadens each arrival, so defect-reflected waves often overlap with the primary before it has fully decayed. This paper develops an instantaneous-frequency (IF)-based framework for this problem. The analytic IF of two superposed asymmetric chirplets is derived, revealing three characteristic contamination mechanisms that distort the observed frequency signature under strong overlap. These mechanisms motivate a sequential estimation strategy that first estimates and subtracts the primary packet from its uncontaminated leading edge, after which it recovers the secondary packet from the residual. A set of physical priors is introduced to protect the estimator against the identified contamination mechanisms rather than to reduce the dimensionality of the parameter space. The method is physics-driven in the sense that the chirplet parameters carry direct physical meaning and the estimation is constrained by priors derived from dispersion, causality, and energy ordering; no pre-built dictionary or learned features are used. The sequential structure is conceptually extensible to more packets, although the theoretical and experimental validations focus primarily on two packets with a three-packet simulation providing preliminary evidence for extension. Numerical experiments suggest a conservative working boundary at normalized packet separation ΔT≈2.5, above which the arrival-time error is typically a few microseconds. In a representative notch-scattering experiment on a 2 mm aluminum plate, ten independent measurements showed that the arrival time remains highly stable (τ2=0.092±0.001 ms), whereas the chirp-rate estimate exhibits large variance under residual overlap. A defect-free control experiment confirmed that the extraction of the secondary packet is a physical response rather than an artifact of over-fitting. For the well-separated edge-reflection cases (130–150 mm excess path), the derived group velocities are within 5% of the adopted reference value. Arrival-time estimates are the most reliable output of the method; chirp-rate and envelope-width estimates are less stable and should not be interpreted as equally accurate physical measurements.

Article Metrics

Citations

Article Access Statistics

Multiple requests from the same IP address are counted as one view.