An automated vibration-based structural damage localization strategy using filter-type feature selection

Alves, Victor; Cury, Alexandre

doi:10.1016/j.ymssp.2023.110145

Cited by 17 publications

(4 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• Kbest: is an approach that selects the k best attributes based on a statistical measure, such as the analysis of variance (ANOVA) used in this study. By defining a value for k, it is possible to choose the k most significant features, which have a more relevant impact on the model's prediction [38,39]; • Random Forest (RF): is a technique that can be applied not only for building classification and regression models but also as a practical feature selection tool [40]. Random Forest performs several independent decision trees, each using different subsets of attributes and random samples from the data set.…”

Section: Feature Selection Techniquesmentioning

confidence: 99%

Feature Selection Techniques for CR Isotope Identification with the AMS-02 Experiment in Space

Borchiellini,

Mano,

Barão

et al. 2024

Particles

View full text Add to dashboard Cite

Isotopic composition measurements of singly charged cosmic rays (CR) provide essential insights into CR transport in the Galaxy. The Alpha Magnetic Spectrometer (AMS-02) can identify singly charged isotopes up to about 10 GeV/n. However, their identification presents challenges due to the small abundance of CR deuterons compared to the proton background. In particular, a high accuracy for the velocity measured by a ring-imaging Cherenkov detector (RICH) is needed to achieve a good isotopic mass separation over a wide range of energies. The velocity measurement with the RICH is particularly challenging for Z=1 isotopes due to the low number of photons produced in the Cherenkov rings. This faint signal is easily disrupted by noisy hits leading to a misreconstruction of the particles’ ring. Hence, an efficient background reduction process is needed to ensure the quality of the reconstructed Cherenkov rings and provide a correct measurement of the particles’ velocity. Machine learning methods, particularly boosted decision trees, are well suited for this task, but their performance relies on the choice of the features needed for their training phase. While physics-driven feature selection methods based on the knowledge of the detector are often used, machine learning algorithms for automated feature selection can provide a helpful alternative that optimises the classification method’s performance. We compare five algorithms for selecting the feature samples for RICH background reduction, achieving the best results with the Random Forest method. We also test its performance against the physics-driven selection method, obtaining better results.

show abstract

Section: Feature Selection Techniquesmentioning

confidence: 99%

Feature Selection Techniques for CR Isotope Identification with the AMS-02 Experiment in Space

Borchiellini,

Mano,

Barão

et al. 2024

Particles

View full text Add to dashboard Cite

show abstract

“…, 2007; Vagnoli et al. , 2017; Alves and Cury, 2023). More specifically, the basic idea of them is that the damage in a structure leads to changes of its physical properties such as stiffness and mass, which alter the structure's dynamic characteristics described by structural modal parameters, mainly including natural frequencies, mode shapes (called displacement modes) and so on (Su et al.…”

Section: Introductionmentioning

confidence: 99%

“…Vibration-based damage localization methods have been developed rapidly in recent years, which are non-destructive and fast and easy to implement (Avci et al, 2021;Catbas et al, 2013;Katunin et al, 2015;Zhao et al, 2007;Vagnoli et al, 2017;Alves and Cury, 2023). More specifically, the basic idea of them is that the damage in a structure leads to changes of its physical properties such as stiffness and mass, which alter the structure's dynamic characteristics described by structural modal parameters, mainly including natural frequencies, mode shapes (called displacement modes) and so on (Su et al, 2016;Chaudhuri, 2008;Roy and Chaudhuri, 2013).…”

Section: Introductionmentioning

confidence: 99%

A Gaussian multi-scale space difference method to multiple slight damage localization based on strain modes

Song,

Wang,

et al. 2023

View full text Add to dashboard Cite

PurposeIn structural health monitoring, localization of multiple slight damage without baseline data is significant and difficult. The purpose of this paper is to discuss these issues.Design/methodology/approachDamage in the structure causes singularities of displacement modes, which in turn reveals damage. Methods based on the displacement modes may fail to accurately locate the slight damage because the slight damage in engineering structure results in a relatively small variation of the displacement modes. In comparison with the displacement modes, the strain modes are more sensitive to the slight damage because the strain is the derivative of the displacement. As a result, the slight variation in displacement data will be magnified by the derivative, leading to a significant variation of the strain modes. A novel method based on strain modes is proposed for the purpose of accurately locating the multiple slight damage.FindingsIn the two bay beam and steel fixed-fixed beams, the numerical simulations and the experimental cases, respectively, illustrate that the proposed method can achieve more accurate localization in comparison with the one based on the displacement modes.Originality/valueThe paper offers a practical approach for more accurate localization of multiple slight damage without baseline data. And the robustness to measurement noise of the proposed method is evaluated for increasing levels of artificially added white Gaussian noise until its limit is reached, defining its range of practical applicability.

show abstract

“…Notwithstanding the many procedures that can reduce ambient effects, further efforts are still needed to improve the damage assessment tools for practical applications (Wang et al 2022). Therefore, several contributions have been proposed by researchers, including different signal processing techniques to analyze and extract relevant information from raw dynamic data (Cardoso et al 2019a, Alves and Cury 2023.…”

Section: Introductionmentioning

confidence: 99%

Novelty detection on a laboratory benchmark slender structure using an unsupervised deep learning algorithm

Finotti,

Silva,

Oliveira

et al. 2023

Lat. Am. j. solids struct.

View full text Add to dashboard Cite

The process that involves the continuous monitoring and analysis of a structure's behavior and performance is known as Structural Health Monitoring (SHM). SHM typically involves the use of sensors and other monitoring devices to collect data, such as displacements, strains, accelerations, among others. These data are analyzed using advanced algorithms and machine learning techniques to identify any signs of abnormal behavior or deterioration. This paper presents a numerical and experimental study of a slender frame subjected to five levels of structural changes under impact loading. The dynamic responses were recorded by four accelerometers and used to build models based on Sparse Auto-Encoders (SAE). Such models can identify each of the five structural states in an unsupervised way. A new strategy to define the hyperparameters of the SAE is presented, which proved to be adequate in the experiments conducted. Finally, the experimental data set is made available to the scientific community to serve as a benchmark for validating SHM methodologies to identify structural changes from dynamic measurements.

show abstract

An automated vibration-based structural damage localization strategy using filter-type feature selection

Cited by 17 publications

References 50 publications

Feature Selection Techniques for CR Isotope Identification with the AMS-02 Experiment in Space

Feature Selection Techniques for CR Isotope Identification with the AMS-02 Experiment in Space

A Gaussian multi-scale space difference method to multiple slight damage localization based on strain modes

Novelty detection on a laboratory benchmark slender structure using an unsupervised deep learning algorithm

Contact Info

Product

Resources

About