Physics-Informed Single Atom Matching Pursuit: Guided-Waves Wavenumbers and Propagation Distance Estimation for Damage Localization in Structural Health Monitoring Journal: Mechanical Systems and Signal Processing

Sebastian Rodriguez, Borja Ferrandiz, Francisco Chinesta, Nazih Mechbal, Marc Rébillat · 2026

Read Physics-Informed Single Atom Matching Pursuit: Guided-Waves Wavenumbers and Propagation Distance Estimation for Damage Localization in Structural Health Monitoring Journal: Mechanical Systems and Signal Processing by Sebastian Rodriguez, Borja Ferrandiz, Francisco Chinesta, Nazih Mechbal, Marc Rébillat (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-03933.txt.

Introduction

Rodriguez and co-authors present a physics-led method for finding damage in thin structures. Guided waves are sent through a part by a bonded sensor. The waves scatter from edges and from flaws. That scatter can reveal harm, but many wave modes start at once. Mixed modes make the signal hard to read.

Their method, PISACMP, breaks the signal into single wave atoms. Physics rules are built into that split, not learned from data alone. The split yields wave numbers and travel distances. Those distances link the sensor, the flaw, and the receiver. Standard ellipse methods can then place the damage. The team frames the result as a clear, low-cost route to locating harm.

Catalog record

CREATORSebastian Rodriguez, Borja Ferrandiz, Francisco Chinesta, Nazih Mechbal, Marc Rébillat
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE44 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.03933

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Physics-Informed Single Atom Matching Pursuit: Guided-Waves Wavenumbers and Propagation Distance Estimation for Damage Localization in Structural Health Monitoring Journal: Mechanical Systems and Signal Processing Sebastian Rodriguez, Borja Ferrandiz, Francisco Chinesta, Nazih Mechbal, Marc Rébillat Abstract Structural Health Monitoring (SHM) aims at the real-time monitoring of the integrity of engineering structures, with Guided-waves (GWs) providing high sensitivity to damage presence and to ageing effects for thin-walled components. In conventional GW-based SHM, a bonded piezoelectric transducer (PZT) emits a short tone burst that produces an Initial Wave Packet (IWP) propagating through the structure. As this packet interacts with boundaries and potential damages, additional scattered wave packets are produced. A major limitation of such approaches lies in the simultaneous excitation of multiple dispersive GW modes by a single PZT, which significantly complicates signal interpretation and damage monitoring. In this context, this work proposes the Physics-Informed Single Atom Matching Pursuit (PISAMP) method, a signal decomposition method grounded in the physical principles governing wave propagation. In contrast with purely data-driven or numerically intensive techniques, the proposed approach embeds strong physical constraints into a low-dimensional and computationally efficient signal representation. This formulation enables the direct identification of key physically meaningful features, including modal wavenumber functions and propagation distances between actuator, damage and sensors. These extracted features, especially source-damage-sensor distances, allows to subsequently perform damage location using well established Elliptical Localization techniques. The principal novelty of this study lies in integrating wave propagation physics into a compact signal decomposition framework and developing an interpretable damage localization methodology for GW-SHM applications. 1 Introduction In industrial applications, ensuring the structural integrity of critical components is a major engineering challenge. The automatic monitoring of damage is essential to prevent catastrophic failures and reduce maintenance costs. In this context, damage refers to changes in material properties and/or structural geometry that adversely affect the current or future performance of a structure. The systematic process of detecting, localizing, and characterizing such changes is known as Structural Health Monitoring (SHM) [ 26 , 23 , 8 , 5 , 6 ] . Among the various SHM techniques developed for thin metallic and composite structures, approaches based on ultrasonic Guided Waves (GW) generated and sensed by bonded piezoelectric transducers (PZT) have proven particularly effective [ 24 , 25 , 13 , 16 ] . In a typical GW-based SHM configuration, a PZT actuator bonded to the structure emits a tone burst centered around a selected frequency. This excitation generates an initial wave packet (IWP) that propagates through the structure and interacts with boundaries, geometric discontinuities, and potential damages. Each interaction produces additional scattered wave packets, resulting in measured signals that contain overlapping and dispersed contributions from multiple propagation paths. Detecting and isolating damage-related echoes within such complex signals is therefore a central challenge. In this framework, the ability to decompose measured signals into physically interpretable wave packets that can be directly associated with structural features is of paramount importance. Several signal processing strategies have been proposed to address this issue. These approaches can be broadly classified into dictionary-based and non-dictionary-based methods. Dictionary-based techniques include, for instance, the Matching Pursuit Method (MPM) [ 12 ] , which approximates a signal as a linear combination of predefined atoms selected from an over-complete dictionary. In guided waves applications, such dictionaries must be constructed a priori to account for delayed, attenuated, and dispersed versions of both A0 and S0 modes [ 27 , 17 , 1 , 11 , 15 ] . However, accurately precomputing all possible modal contributions for continuous thin structures is impractical, limiting the method’s adaptability and physical consistency. Non-dictionary-based approaches—such as Empirical Mode Decomposition (EMD), Variational Mode Decomposition (VMD), Proper Orthogonal Decomposition (POD), and Hilbert–Huang Transform methods—have also been applied to guided wave analysis [ 4 , 10 ] . Recent work introduced the Rank Reduction AutoEncoder (RRAE), a purely data-driven architecture that automatically extracts features from measured signals to maximize damage detection performance [ 21 ] . While these techniques can effectively separate multi-component signals, their mathematical formulations are often weakly connected to the physics of wave propagation, limiting the physical interpretability of the extracted components. To overcome these limitations, this work introduces a decomposition framework grounded in the mathematical description of guided wave propagation in thin plates. This propagation is modeled in the frequency domain through a wavenumber function that depends on frequency and propagation distance. On one hand, the wavenumber function captures dispersive effects, namely the propagation induced distortion of the Initial Wave Packet (IWP). On the other hand, the propagation distance accounts for the extent to which the signal travels through the solid medium. Therefore, the proposed approach models measured signals as delayed and dispersed impulse responses of the external excitation, thereby directly reflecting the fundamental physics of guided waves propagation. Its primary objective is to estimate not only the modal wavenumber functions of the structure, but also the propagation distances traveled by the different wave packets from the original source to the sensing locations. By embedding physical constraints into a compact and low-dimensional signal representation, the method enables the simultaneous extraction of these physically meaningful parameters while preserving computational efficiency and robustness. The proposed decomposition, referred to as the Physics-Informed Single Atom Matching Pursuit (PISAMP) method, can be viewed as a physics-driven extension of the Single Atom Convolutional Matching Pursuit framework introduced in [ 22 ] . Beyond modal characterization, the method provides estimates of propagation distances between actuator, damage, and sensors. As an illustration of their usefulness for SHM purposes, in this work, these estimated distances are subsequently employed to perform efficient damage location using well-established Elliptical Localization techniques, enabling accurate and physically consistent detection of damages in thin plate structures. The remainder of this paper is organized as follows. Section 2 presents the reference problem addressed in this study, namely the propagation of guided waves in thin plates. Section 3 introduces the proposed physics-informed numerical methodology for estimating both modal wavenumber functions and the propagation distances traveled by wave packets from the source, based solely on simulated signal data. Next Section 4 presents the application of PISAMP for efficient and interpretable damage localization. Section 5 then evaluates the performance of the approach through a series of numerical investigations, including the approximation of both simple and complex signals and its application to damage localization. Finally, Section 6 summarizes the main findings and outlines perspectives for future research. 2 Reference problem: guided-waves propagation in an isotropic homogeneous medium The reference problem considered in the following consists on the processing of guided-waves ultrasonic signals that propagates over thin plates made out of isotropic homogeneous materials as is shown in Fig. 1 . Such a plate merely corresponds to a composite plate commonly used in aeronautic structures. Figure 1 : Wave propagation in a plate. In the low frequency range of such structure, only two wave modes exist, namely the A_{0} and S_{0} modes, and given an input frequency f , the corresponding wavenumbers k_{S_{0}}(f) and k_{A_{0}}(f) can be computed as [ 24 , 25 , 13 , 16 ] : k_{S_{0}}(f)=2\pi f\sqrt{\frac{\rho}{Q}} (1) k_{A_{0}}(f)=\frac{2\pi f}{\sqrt{2}}\sqrt{\left(\frac{\rho}{Q}+\frac{\rho}{G\xi}\right)+\sqrt{\left(\frac{\rho}{Q}-\frac{\rho}{G\xi}\right)^{2}+\frac{1}{\pi f^{2}}\left(\rho+\frac{\rho}{IQ}\right)}} (2) with: G=\frac{E}{2(1+\nu)} , Q=\frac{E}{(1-\nu^{2})} , \xi=\pi^{2}/12 and I=h^{3}/12 . E,\nu,h,\rho denotes the homogenized Young modulus, Poisson modulus, thickness of the plate and density of

Read the complete text