arxiv: 2604.05413 · v1 · submitted 2026-04-07 · 📡 eess.SP

Recognition: 2 theorem links

· Lean Theorem

Operator-Theoretic Energy Functionals for Impulse-Excited Nonstationary Signal Analysis

Tahir Cetin Akinci

Authors on Pith no claims yet

Pith reviewed 2026-05-10 19:54 UTC · model grok-4.3

classification 📡 eess.SP

keywords impulse response analysisenergy concentration indextime-frequency representationsdefect detectionnonstationary signalsoperator theorystructural monitoringclassification performance

0 comments

The pith

An operator-theoretic Energy Concentration Index applied to impulse responses detects defects by quantifying localized time-frequency energy shifts.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces a framework that treats impulse-excited signals as finite-energy elements in the L2 Hilbert space and analyzes their time-frequency content through bounded linear operators associated with continuous frames. A nonlinear Energy Concentration Index is defined to measure energy localization in chosen regions of the time-frequency plane, with its continuity properties ensuring that parameter changes yield detectable variations. This index supports construction of a statistical separability measure, from which the compact IMRED detector is derived for classification tasks. Experiments on impulse-excited ceramics show the method achieves an AUC of 0.908, outperforming standard Fourier and wavelet approaches in separating defective from intact samples.

Core claim

Measured responses are modeled as finite energy impulse responses perturbed by stochastic disturbances in the Hilbert space L2(R). Time frequency representations are formulated as bounded linear analysis operators associated with continuous frames. The nonlinear Energy Concentration Index ECI quantifies localized transform domain energy, and its boundedness and continuity link small physical variations to measurable changes in energy distribution. This enables the Impulse Based Multi Resolution Energy Detector IMRED, which demonstrates strong discriminative capability with an AUC of 0.908 in ceramic defect detection experiments.

What carries the argument

The Energy Concentration Index, a nonlinear functional on time-frequency representations that computes localized energy over selected regions, which enables the link from structural perturbations to statistical classification via its continuity and boundedness properties.

If this is right

Variations in damping and resonant frequency produce systematic changes in time frequency coefficients and localized energy concentration.
The IMRED statistic achieves clearer class separation than global Fourier band energy measures and non optimized wavelet band aggregation.
The framework establishes a direct relationship between impulse response modeling, localized energy geometry, and statistical decision mechanisms.
The boundedness of the functional ensures small physical variations produce measurable changes enabling statistical separability.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Extending the operator view could allow for adaptive selection of time-frequency regions based on expected defect types.
The continuity property suggests the method might track gradual degradation in real-time structural health monitoring.
Similar energy functionals could apply to other nonstationary signals in fields like acoustics or biomedical engineering.

Load-bearing premise

The boundedness and continuity of the Energy Concentration Index ensure that small physical variations in system parameters produce measurable changes in localized energy distribution.

What would settle it

Repeating the ceramic impulse excitation experiments with controlled defects and finding that the IMRED statistic yields an AUC no higher than 0.75 or fails to outperform Fourier measures would falsify the discriminative capability claim.

Figures

Figures reproduced from arXiv: 2604.05413 by Tahir Cetin Akinci.

**Figure 1.** Figure 1: illustrates the characteristic damped oscillatory behavior of an impulse–excited system, providing a clear visual counterpart to the analytical model introduced in Section II. The gradual decay and sustained oscillations shown in the figure reflect the influence of the parameters A, α, and ω on the nominal transient response. This baseline depiction establishes the reference behavior against which subseque… view at source ↗

**Figure 2.** Figure 2: Comparison of healthy and defective impulse responses illustrating [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Figure 3: Time–frequency energy maps of the healthy (a) and defective (b) impulse responses, illustrating how damping and frequency perturbations reshape the [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

**Figure 4.** Figure 4: (a) Time–frequency region Ω used in the definition of the energy concentration index, and (b) ROC curve illustrating the classification performance of the IMRED functional. with first and second-order statistics µh = E[Zh], µd = E[Zd], σ 2 h = Var(Zh), σ2 d = Var(Zd). To quantify class discrimination induced by localized energy geometry, we define the separability functional J(Ω) = (µd − µh) 2 σ 2 d + σ 2 … view at source ↗

read the original abstract

This study presents an operator theoretic framework for defect detection in impulse excited nonstationary systems. Measured responses are modeled as finite energy impulse responses perturbed by stochastic disturbances and represented in the Hilbert space L2(R). Time frequency representations are formulated as bounded linear analysis operators associated with continuous frames, enabling a consistent description of how structural perturbations redistribute transient signal energy. Within this formulation, a nonlinear Energy Concentration Index ECI is introduced to quantify localized transform domain energy over selected regions of the time frequency plane. The boundedness and continuity of the functional ensure that small physical variations in system parameters produce measurable changes in localized energy distribution. This property enables the construction of a statistical separability functional that links multi resolution energy geometry to classification performance. Based on these results, a compact Impulse Based Multi Resolution Energy Detector IMRED is derived. The analysis shows that variations in damping and resonant frequency produce systematic changes in time frequency coefficients and localized energy concentration. Experimental validation using impulse excited ceramic measurements demonstrates that the proposed descriptor captures defect induced structural differences with strong discriminative capability. The resulting IMRED statistic achieves an AUC of 0.908 and provides clearer class separation than global Fourier band energy measures and non optimized wavelet band aggregation. These results establish a direct relationship between impulse response modeling, localized energy geometry, and statistical decision mechanisms, providing a mathematically grounded basis for energy driven defect detection in structural monitoring applications.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper gives a Hilbert-space operator framing for energy concentration in impulse responses and derives an IMRED detector that hits AUC 0.908 on ceramic data, but the step from boundedness to that performance number stays unquantified.

read the letter

The core contribution is an operator-theoretic treatment of time-frequency energy for nonstationary impulse responses. Responses are modeled in L2(R) as finite-energy signals plus disturbances, then analyzed with bounded linear operators coming from continuous frames. From that setup the authors define a nonlinear Energy Concentration Index (ECI) over chosen time-frequency regions and build the Impulse Based Multi Resolution Energy Detector (IMRED) around it. The boundedness and continuity of ECI are used to argue that small shifts in damping or resonant frequency will produce detectable changes in localized energy, which in turn supports classification. On the ceramic impulse measurements the IMRED statistic separates classes with AUC 0.908 and does better than global Fourier band energy or non-optimized wavelet aggregation. That experimental result is the clearest payoff and shows the descriptor can capture defect-induced differences in practice. The framing itself is internally consistent and extends existing frame methods without obvious contradictions. The soft spot is the missing quantitative link. The abstract asserts that continuity of the functional guarantees measurable changes, yet no modulus of continuity, Lipschitz constant, or explicit comparison against the variance of the stochastic disturbances appears. Without that, it is difficult to attribute the AUC gain to the operator properties rather than to how the time-frequency regions were selected. Reproducibility is also limited because full derivation steps, region-selection protocol, and controls are not visible. This work is aimed at engineers doing structural health monitoring or non-destructive testing with impulse-excited signals. A reader who already works with time-frequency features or operator methods for energy localization will see a concrete application and a new detector. It is not a foundational theoretical paper, but the combination of the operator model with a real-data classification result is specific enough to merit referee attention. I would send it for peer review so the derivations and experimental details can be checked directly.

Referee Report

3 major / 1 minor

Summary. The manuscript proposes an operator-theoretic framework for analyzing impulse-excited nonstationary signals to detect defects. Responses are modeled as finite-energy impulse responses in L2(R) perturbed by stochastic disturbances; time-frequency representations are cast as bounded linear analysis operators from continuous frames. A nonlinear Energy Concentration Index (ECI) is defined to quantify localized energy over selected time-frequency regions, with its boundedness and continuity asserted to translate small changes in damping or resonant frequency into measurable energy shifts. From this, the Impulse Based Multi Resolution Energy Detector (IMRED) statistic is derived. Experimental results on impulse-excited ceramic measurements report that IMRED achieves an AUC of 0.908 and clearer class separation than global Fourier band energy or non-optimized wavelet aggregation.

Significance. If the operator construction, continuity claims, and experimental attribution hold, the work would supply a mathematically grounded link between frame-theoretic energy functionals and statistical defect classification in structural monitoring. The explicit use of continuous-frame operators and the derivation of a compact statistic from them are positive features that could support reproducible extensions, but the absence of any modulus of continuity, sensitivity bounds, or controls currently limits the result's impact.

major comments (3)

[Abstract] Abstract: the claim that 'the boundedness and continuity of the functional ensure that small physical variations in system parameters produce measurable changes in localized energy distribution' is load-bearing for the AUC result, yet no Lipschitz estimate, modulus of continuity, or explicit comparison of ECI variation against the variance of the stochastic disturbances is supplied.
[Abstract] Abstract: the link from the operator model to the reported IMRED classification performance (AUC 0.908) appears to involve selection of time-frequency regions, but the independence of this selection from the fitted performance metric is not demonstrated, leaving open the possibility that separability arises from post-hoc tuning rather than the asserted functional properties.
[Abstract] Abstract: the experimental validation states that IMRED 'provides clearer class separation than global Fourier band energy measures and non optimized wavelet band aggregation,' but no quantitative error analysis, cross-validation procedure, or control for dataset-specific artifacts is described, preventing attribution of the AUC gain to the operator-theoretic construction.

minor comments (1)

[Abstract] The abstract refers to 'non optimized wavelet band aggregation' without specifying the wavelet family, decomposition level, or aggregation rule used as baseline.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments. We address each major comment below and will incorporate revisions to strengthen the presentation of the operator-theoretic claims and experimental details.

read point-by-point responses

Referee: [Abstract] Abstract: the claim that 'the boundedness and continuity of the functional ensure that small physical variations in system parameters produce measurable changes in localized energy distribution' is load-bearing for the AUC result, yet no Lipschitz estimate, modulus of continuity, or explicit comparison of ECI variation against the variance of the stochastic disturbances is supplied.

Authors: We agree that an explicit quantitative bound would make the continuity claim more precise and directly support the reported AUC. The manuscript establishes continuity of the ECI via the boundedness of the analysis operator and the frame operator properties, but does not supply a modulus of continuity or a direct comparison against disturbance variance. In the revised manuscript we will add a proposition deriving a Lipschitz constant for the ECI under bounded perturbations of damping and resonant frequency, together with a numerical comparison using the observed variance of the stochastic term in the ceramic data. revision: yes
Referee: [Abstract] Abstract: the link from the operator model to the reported IMRED classification performance (AUC 0.908) appears to involve selection of time-frequency regions, but the independence of this selection from the fitted performance metric is not demonstrated, leaving open the possibility that separability arises from post-hoc tuning rather than the asserted functional properties.

Authors: The time-frequency regions are chosen according to the known resonant modes of the ceramic specimens, as determined from the physical model and preliminary spectral analysis of the impulse responses. To demonstrate that this choice is not post-hoc, the revision will include an explicit description of the physical selection rule and an ablation table comparing AUC values obtained with the physically motivated regions against those obtained when the regions are instead optimized on the training data. revision: yes
Referee: [Abstract] Abstract: the experimental validation states that IMRED 'provides clearer class separation than global Fourier band energy measures and non optimized wavelet band aggregation,' but no quantitative error analysis, cross-validation procedure, or control for dataset-specific artifacts is described, preventing attribution of the AUC gain to the operator-theoretic construction.

Authors: We acknowledge that the experimental section would benefit from greater detail on validation and controls. The revision will expand the experimental validation to specify the cross-validation procedure used to obtain the AUC, report variability measures across repetitions, and describe the normalization and artifact-control steps applied to the impulse-excited measurements. These additions will allow clearer attribution of the observed separation to the proposed energy-concentration functional. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation is self-contained with independent experimental validation

full rationale

The paper constructs an operator-theoretic model using continuous frames on L2(R), defines the nonlinear ECI functional on localized time-frequency regions, asserts its boundedness/continuity from the analysis-operator properties, derives the IMRED statistic from that geometry, and reports an independent experimental AUC of 0.908 on ceramic impulse-response data. No equation or claim reduces a reported performance metric to a fitted parameter by construction, no self-citation supplies a load-bearing uniqueness theorem, and the experimental separability is presented as external confirmation rather than a tautological output of the definitions. The theoretical steps remain independent of the classification numbers.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 2 invented entities

Abstract-only review; specific free parameters, axioms, or invented entities cannot be audited in detail. The framework invokes standard Hilbert-space L2(R) and continuous frames as background.

invented entities (2)

Energy Concentration Index (ECI) no independent evidence
purpose: Quantify localized energy in time-frequency regions
Nonlinear functional introduced to measure energy redistribution due to defects
Impulse Based Multi Resolution Energy Detector (IMRED) no independent evidence
purpose: Classify defects from energy geometry
Compact detector derived from the separability functional

pith-pipeline@v0.9.0 · 5541 in / 1273 out tokens · 60570 ms · 2026-05-10T19:54:40.440340+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation washburn_uniqueness_aczel unclear
Definition 1: The Energy Concentration Index (ECI) of a signal x(t)∈L2(R) over Ω is defined as ECIΩ(x)=∫_Ω |Wx(a,b)|² db da/a².
IndisputableMonolith/Foundation/AbsoluteFloorClosure absolute_floor_iff_bare_distinguishability unclear
The boundedness and continuity of the functional ensure that small physical variations in system parameters produce measurable changes in localized energy distribution.

Reference graph

Works this paper leans on

38 extracted references · 29 canonical work pages

[1]

Instantaneous frequency identification for nonstationary signals of time-varying structures using enhanced synchroextracting wavelet transform and dynamic optimiza- tion,

Y . Jiang, X.-Y . Wang, X.-L. Zhang,et al., “Instantaneous frequency identification for nonstationary signals of time-varying structures using enhanced synchroextracting wavelet transform and dynamic optimiza- tion,”J. Low Freq. Noise Vib. Act. Control, vol. 43, 2024, doi: 10.1177/14613484241233392

work page doi:10.1177/14613484241233392 2024
[2]

Detection and classification of nonstationary transient signals using sparse approximations and Bayesian networks,

N. Wachowski and M. R. Azimi–Sadjadi, “Detection and classification of nonstationary transient signals using sparse approximations and Bayesian networks,”IEEE/ACM Trans. Audio, Speech, Lang. Process., vol. 22, no. 12, pp. 1750–1764, Dec. 2014, doi: 10.1109/TASLP.2014.2348913

work page doi:10.1109/taslp.2014.2348913 2014
[4]

Modal parameter identification using improved random decrement and stochastic subspace methods under non-stationary excitation,

J. Wu, J. Hu, M. Ma, C. Zhang, Z. Ma, C. Zhou, and G. Sun, “Modal parameter identification using improved random decrement and stochastic subspace methods under non-stationary excitation,”Mech. Syst. Signal Process., vol. 205, p. 110987, 2024, doi: 10.1016/j.ymssp.2023.110987

work page doi:10.1016/j.ymssp.2023.110987 2024
[5]

Hybrid time–frequency method for vibration fatigue damage analysis under non-stationary non- Gaussian random excitation,

W. Lei, Y . Jiang, and J. Zhang, “Hybrid time–frequency method for vibration fatigue damage analysis under non-stationary non- Gaussian random excitation,”Mech. Syst. Signal Process., 2025, doi: 10.1016/j.ymssp.2025.112630

work page doi:10.1016/j.ymssp.2025.112630 2025
[7]

Synchrosqueezed wavelet transforms: An empirical mode decomposition-like tool,

I. Daubechies, J. Lu, and H.-T. Wu, “Synchrosqueezed wavelet transforms: An empirical mode decomposition-like tool,”Appl. Comput. Harmon. Anal., vol. 30, no. 2, pp. 243–261, 2011, doi: 10.1016/j.acha.2010.08.002

work page doi:10.1016/j.acha.2010.08.002 2011
[8]

Adaptive variational mode decomposition for multicomponent non-stationary signal analysis,

S. Yu, X. Zhang, and Y . Chen, “Adaptive variational mode decomposition for multicomponent non-stationary signal analysis,”Signal Process., vol. 208, p. 108007, 2023, doi: 10.1016/j.sigpro.2023.108007

work page doi:10.1016/j.sigpro.2023.108007 2023
[9]

2024 , issn =

T. Nguyen and M. Khoa, “Stochastic time–frequency modeling of non-stationary structural responses under random excitation,” Mech. Syst. Signal Process., vol. 205, p. 111045, 2024, doi: 10.1016/j.ymssp.2023.111045

work page doi:10.1016/j.ymssp.2023.111045 2024
[10]

Three- dimensional defect reconstruction in carbon fiber–reinforced composites with temporally non-uniform pulsed thermography data,

L. Gahleitner, G. Mayr, P. Burgholzer, and U. Cakmak, “Three- dimensional defect reconstruction in carbon fiber–reinforced composites with temporally non-uniform pulsed thermography data,”NDT&E Int., vol. 154, p. 103363, 2025

2025
[11]

Ultrasonic characterization of degradation in lithium-ion batteries using guided waves,

Y . Zhang, M. Feng, and K. Wang, “Ultrasonic characterization of degradation in lithium-ion batteries using guided waves,”Ultrasonics, vol. 134, p. 107036, 2024, doi: 10.1016/j.ultras.2023.107036

work page doi:10.1016/j.ultras.2023.107036 2024
[12]

On the Consistent Classification and Treatment of Uncertainties in Structural Health Monitoring Applications,

A. Kamariotis, K. Vlachas, V . Ntertimanis, I. Koune, A. Cicirello, and E. Chatzi, “On the Consistent Classification and Treatment of Uncertainties in Structural Health Monitoring Applications,”ASCE– ASME J. Risk Uncertainty Eng. Syst., Part B, vol. 11, no. 1, p. 011108, 2025, doi: 10.1115/1.4067140

work page doi:10.1115/1.4067140 2025
[13]

Vertex-frequency graph signal processing: A comprehensive review,

L. Stankovi ´c, D. Mandic, M. Dakovi ´c, B. Scalzo, M. Brajovi ´c, and E. Sejdi ´c, “Vertex-frequency graph signal processing: A comprehensive review,”Digit. Signal Process., vol. 102, p. 102756, 2020

2020
[14]

Spectral analysis for current and temperature measurements in power cables,

S. Tas ¸kın, S. S ¸eker, M. Karahan, and T. C. Akinci, “Spectral analysis for current and temperature measurements in power cables,”Electric Power Components and Systems, vol. 37, no. 4, pp. 415–426, 2009, doi: 10.1080/15325000902740852

work page doi:10.1080/15325000902740852 2009
[15]

Frequency-domain characterization of aging effects in XLPE power cables using dielectric response analysis,

H. Zhou, Y . Liu, and J. Li, “Frequency-domain characterization of aging effects in XLPE power cables using dielectric response analysis,” IEEE Trans. Power Del., vol. 35, no. 6, pp. 2634–2643, 2020, doi: 10.1109/TPWRD.2020.2981234

work page doi:10.1109/tpwrd.2020.2981234 2020
[16]

A. V . Oppenheim and A. S. Willsky,Signals and Systems, 2nd ed., Prentice Hall, 1997

1997
[17]

Transient vibration feature extraction using time-varying filtering empirical mode decomposi- tion,

Y . Huang, Z. Zhang, and J. Yang, “Transient vibration feature extraction using time-varying filtering empirical mode decomposi- tion,”Mech. Syst. Signal Process., vol. 168, p. 108653, 2022, doi: 10.1016/j.ymssp.2021.108653. 11

work page doi:10.1016/j.ymssp.2021.108653 2022
[18]

Incipient fault detection in power distribution systems: A time–frequency embedded deep-learning-based approach,

Q. Li, Y . Zhang, and H. Liu, “Incipient fault detection in power distribution systems: A time–frequency embedded deep-learning-based approach,”IEEE Trans. Instrum. Meas., vol. 72, Art. no. 2507914, 2023, doi: 10.1109/TIM.2023.3271234

work page doi:10.1109/tim.2023.3271234 2023
[19]

Arc fault detection based on time– frequency characteristics and deep learning in low-voltage electrical systems,

J. Zhang, Y . Zhou, and H. Wang, “Arc fault detection based on time– frequency characteristics and deep learning in low-voltage electrical systems,”IEEE Trans. Plasma Sci., vol. 51, no. 2, pp. 345–356, 2023, doi: 10.1109/TPS.2023.3234567

work page doi:10.1109/tps.2023.3234567 2023
[20]

Time-frequency analysis of arc welding current,

T. C. Akinci, “Time-frequency analysis of arc welding current,”Mechan- ics, vol. 85, no. 5, pp. 66–71, 2010

2010
[21]

The defect detection in ceramic materials based on time- frequency analysis by using the method of impulse noise,

T. C. Akinci, “The defect detection in ceramic materials based on time- frequency analysis by using the method of impulse noise,”Archives of Acoustics, vol. 36, no. 1, pp. 77–85, 2011, doi: 10.2478/v10168-011- 0007-y

work page doi:10.2478/v10168-011- 2011
[22]

Iterative synchrosqueezing- based general linear chirplet transform for time-frequency feature extraction,

Y . Liu, H. Xiang, Z. Jiang, and J. Xiang, “Iterative synchrosqueezing- based general linear chirplet transform for time-frequency feature extraction,”IEEE Trans. Instrum. Meas., vol. 72, Art. no. 3506711, 2023, doi: 10.1109/TIM.2022.3232090

work page doi:10.1109/tim.2022.3232090 2023
[23]

Fault diagnosis method of planetary gearboxes based on multi-scale wavelet packet energy entropy and extreme learning machine,

R. Meng, J. Zhang, M. Chen, and L. Chen, “Fault diagnosis method of planetary gearboxes based on multi-scale wavelet packet energy entropy and extreme learning machine,”Entropy, vol. 27, no. 8, p. 782, 2025, doi: 10.3390/e27080782

work page doi:10.3390/e27080782 2025
[24]

Improving the readability of time-frequency and time-scale representations by the reassignment method,

F. Auger and P. Flandrin, “Improving the readability of time-frequency and time-scale representations by the reassignment method,”IEEE Trans. Signal Process., vol. 43, no. 5, pp. 1068–1089, May 1995, doi: 10.1109/78.382394

work page doi:10.1109/78.382394 1995
[25]

A signal-dependent time–frequency representation: optimal kernel design,

R. G. Baraniuk and D. L. Jones, “A signal-dependent time–frequency representation: optimal kernel design,”IEEE Trans. Signal Process., vol. 41, no. 4, pp. 1589–1602, Apr. 1993, doi: 10.1109/78.212733

work page doi:10.1109/78.212733 1993
[26]

LeCun, Y

Y . LeCun, Y . Bengio, and G. Hinton, “Deep learning,”Nature, vol. 521, pp. 436–444, May 2015, doi: 10.1038/nature14539

work page doi:10.1038/nature14539 2015
[27]

Support vector machine-based classifica- tion scheme for myoelectric control applied to upper limb,

M. A. Oskoei and H. Hu, “Support vector machine-based classifica- tion scheme for myoelectric control applied to upper limb,”IEEE Trans. Biomed. Eng., vol. 55, no. 8, pp. 1956–1965, Aug. 2008, doi: 10.1109/TBME.2008.919734

work page doi:10.1109/tbme.2008.919734 1956
[28]

Analysis of gait dynamics of ALS disease and classification using artificial neural networks,

O. Akgun, A. Akan, H. Demir, and T. C. Akinci, “Analysis of gait dynamics of ALS disease and classification using artificial neural networks,”Tehni ˇcki vjesnik, vol. 25, no. 5, pp. 1495–1502, 2018, doi: 10.17559/TV-20160914144554

work page doi:10.17559/tv-20160914144554 2018
[29]

Power quality disturbances detection and classification based on deep convolution auto-encoder networks,

P. Khetarpal, N. Nagpal, M. S. Al-Numay, P. Siano, Y . Arya, and N. Kassarwani, “Power quality disturbances detection and classification based on deep convolution auto-encoder networks,”IEEE Access, vol. 11, pp. 46026–46038, 2023, doi: 10.1109/ACCESS.2023.3274732

work page doi:10.1109/access.2023.3274732 2023
[30]

Deep learning and its applications to machine health monitoring,

R. Zhao, R. Yan, Z. Chen, K. Mao, P. Wang, and R. X. Gao, “Deep learning and its applications to machine health monitoring,” Mech. Syst. Signal Process., vol. 115, pp. 213–237, Jan. 2019, doi: 10.1016/j.ymssp.2018.05.050

work page doi:10.1016/j.ymssp.2018.05.050 2019
[31]

Convolutional neural network for multi-channel vibrational analysis applied to bearing fault detection,

L. E. S. Salgado Solano, O. A. C ´ardenas, L. M. F. Nava, and F. G. Casta ˜neda, “Convolutional neural network for multi-channel vibrational analysis applied to bearing fault detection,” inProc. 22nd Int. Conf. Electrical Engineering, Computing Science and Au- tomatic Control (CCE), Mexico City, Mexico, 2025, pp. 1–6, doi: 10.1109/CCE67728.2025.11271968

work page doi:10.1109/cce67728.2025.11271968 2025
[32]

Detection of power quality disturbances using wavelet-based convolutional transformer network,

D. H. Chiam, K. H. Lim, and K. H. Law, “Detection of power quality disturbances using wavelet-based convolutional transformer network,” inProc. 2022 Int. Conf. Green Energy, Computing and Sustainable Technology (GECOST), Miri Sarawak, Malaysia, 2022, pp. 150–154, doi: 10.1109/GECOST55694.2022.10010655

work page doi:10.1109/gecost55694.2022.10010655 2022
[33]

Power quality disturbances recognition via lightweight deep learning framework integrated with time-frequency analysis embedded,

J. Song, Y . Wang, Y . Gao, H. Ye, D. Fang, X. Wang, Z. Hu, and S. Lu, “Power quality disturbances recognition via lightweight deep learning framework integrated with time-frequency analysis embedded,”IEEE Trans. Instrum. Meas., vol. 74, pp. 1–14, 2025, Art. no. 3560914, doi: 10.1109/TIM.2025.3615304

work page doi:10.1109/tim.2025.3615304 2025
[34]

D. J. Inman,Engineering Vibration, 4th ed., Pearson, 2014

2014
[35]

Response of dynamical systems to nonstationary inputs,

L. Galleani, “Response of dynamical systems to nonstationary inputs,” IEEE Trans. Signal Process., vol. 60, no. 11, pp. 5775–5786, Nov. 2012, doi: 10.1109/TSP.2012.2214032

work page doi:10.1109/tsp.2012.2214032 2012
[36]

S. S. Rao,Mechanical Vibrations, 5th ed., Prentice Hall, 2011

2011
[37]

doi:10.2172/249299

S. W. Doebling, C. R. Farrar, M. B. Prime, and D. W. Shevitz, “Damage identification and health monitoring of structural and mechanical systems: A literature review,” Los Alamos National Laboratory, Rep. LA-13070- MS, 1996, doi: 10.2172/249299

work page doi:10.2172/249299 1996
[38]

B. P. Lathi,Linear Systems and Signals, 2nd ed., Oxford Univ. Press, 2005

2005
[39]

Mallat,A Wavelet Tour of Signal Processing, 3rd ed., Academic Press, 2009

S. Mallat,A Wavelet Tour of Signal Processing, 3rd ed., Academic Press, 2009

2009
[40]

Papoulis and S

A. Papoulis and S. U. Pillai,Probability, Random Variables, and Stochastic Processes, 4th ed., McGraw-Hill, 2002

2002