Physics-Informed Machine Learning for Structural Damage Diagnosis in Aluminium Plates - EWSHM 2024 11th European Workshop on Structural Health Monitoring

Aug 5, 2026

Upload by Oliver

Damage diagnosis under Structural Health Monitoring (SHM) is essential for identifying, localizing, and estimating defect extent in structures. Lamb waves are effective for thin-walled components, yet many approaches depend on extracting damage indices from measured signals, requiring heavy post-processing and reducing diagnostic information. This work introduces a physics-informed machine-learning method using a Physics-Informed Neural Network (PINN) to predict aluminium-plate density from displacement measurements, embedding governing PDEs into the loss for explainable, index-free damage detection.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Abstract

Introduction

Structural Health Monitoring and guided Lamb waves

Limitations of damage-index extraction and black-box models

Physics-Informed Neural Networks and PDE-based loss functions

Methodology

Theoretical Background

e-Journal of Nondestructive Testing-ISSN [1435-4934-www.ndt.net](1435-4934-www.ndt.net) EWSHM 2024 11th European Workshop on Structural Health Monitoring Physics-Informed Machine Learning for Structural Damage Diagnosis in Aluminium Plates Lucio PINELLO 1, Luca LOMAZZI1, Jacopo GHELLERO 1, Marco GIGLIO 1, Francesco CADINI 1 1 Department of Mechanical Engineering, Politecnico di Milano, Milan, 20156, Italy, e mail: [lucio.pinello@polimi.it](lucio.pinello@polimi.it) luca.lomazzi@polimi.it jacopo.ghellero@polimi.it [marco.giglio@polimi.it](marco.giglio@polimi.it) [francesco.cadini@polimi.it](francesco.cadini@polimi.it) Abstract. Damage diagnosis plays a crucial role in Structural Health Monitoring (SHM) by facilitating the identification, localization, and estimation of the extent of defects in structures. Lamb waves, known for their sensitivity to defects, are widely employed in SHM methods for thin-walled structures. Most of those traditional methods require extracting damage indices from Lamb wave signals. This operation involves substantial post-processing and implies that part of the diagnostic information is lost. To solve those limitations and improve the damage diagnosis accuracy, machine learning methods have recently been proposed in the literature. However, the reluctance of the industrial sector to adopt conventional black-box models due to their lack of explainability poses a challenge. This study proposes a physics-informed machine-learning approach to address the limitations of standard black-box methods. Particularly, a Physics-Informed Neural Network (PINN) is implemented to predict the density of an aluminium plate based on measurements of plate displacements caused by Lamb wave excitation. This is made possible by the implementation of a specific loss function, which leverages physical knowledge in the form of the partial differential equation governing Lamb waves. Predicting the plate density based on measured displacements eliminates the need for artificial damage indices, utilizing the density variation itself to detect and localize damage. Additionally, the outputs of the PINN, rooted in physics equations, offer enhanced explainability compared to standard black-box models. The versatility of this framework extends to predicting material properties distributions for components, and efforts will be directed towards adapting the method for composite materials, where the approach may pose additional challenges. Keywords: corrosion monitoring, SHM, transfer learning, neural network This work is licensed under CC BY 4.0 Media and Publishing Partner [https://doi.org/10.58286/29742](https://doi.org/10.58286/29742) Introduction Structural Health Monitoring (SHM) is a monitoring strategy that relies on a sensor network permanently installed on the structure or component of interest to allow its continuous monitoring, significantly reducing the time between two consecutive inspections. Among the several non-destructive techniques (NDT) used in the SHM framework, ultrasonic-guided waves, and in particular the Lamb waves, proved to be effective for thinwalled structures [1][2][3][4] by exploiting a piezoelectric (PZT) sensor network for both wave excitation and reception. Focusing on Lamb waves, they have been extensively used for imaging and tomographic methods [5] [6] [7] . However, these methods require the extraction of features from the signals to obtain damage indexes (DIs) [1][8][9] . Similarly, machine learning (ML) algorithms are being applied due to the necessity of having real-time, or almost real-time, performance for SHM purposes. However, conventional ML approaches often are supervised methods and do not solve the need for a pre-processing phase to extract damage features from signals [10][11][12] . Therefore, there is a need for alternatives belonging to the unsupervised framework scheme. It is in this framework that Physics-Informed Neural Networks (PINNs) are gaini

FAQs

Why are Lamb waves widely used in structural health monitoring for thin-walled structures?

Lamb waves are ultrasonic-guided waves sensitive to defects in slender structures like plates. Their guided nature makes them effective for SHM in thin-walled geometries.

What limitation of traditional SHM approaches motivates this study?

Many methods rely on extracting damage indices from Lamb wave signals, which requires substantial post-processing and can lead to loss of diagnostic information. Conventional machine learning also faces adoption barriers due to limited explainability.

How does the proposed PINN approach perform damage detection without damage indices?

The Physics-Informed Neural Network predicts aluminium plate density directly from measured displacement fields. By using a loss function informed by the governing PDE of Lamb waves, the density variation is used to detect and localize damage while improving explainability.

Aug 5, 2026

Upload by Oliver