Pith. sign in

REVIEW 4 major objections 5 minor 53 references

Formula-Guided Machine Learning for Ground Vibration Propagation and Attenuation Modeling

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A hybrid iterative loop that alternates least-squares fitting with a symbolic-regression model discovers a compact, frequency-dependent Bornitz attenuation law from field vibration data, with a mean absolute error of…

desk verdict A solid, carefully executed case study that delivers a site-specific attenuation formula, but the headline accuracy claims rest on in-sample fit and the 'discovered law' needs out-of-sample validation before it earns that name. read the letter →

arxiv 2505.13870 v3 pith:JUE6XLKX submitted 2025-05-20 physics.app-ph

Pei-Yao Chen , Chen Wang , Fang Yan , Chao-Yang Zhang , Xiang-Yu Tan , Guo-Ping Lin , Jian-Sheng Fan This is my paper · ORCID
classification physics.app-ph
keywords groundvibrationattenuationBornitzformuladiscoverysymbolicregressionmachinelearningfrequency-dependentdampingfieldtestingvibration-sensitivefacilities
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper aims to show that a frequency-dependent ground-vibration attenuation law can be discovered automatically from field measurements, rather than assumed from theory or delivered as a black box. It locks the overall shape of attenuation to the classical Bornitz power-exponential decay law, then runs an iterative loop in which a symbolic-regression model proposes functional forms for the two damping coefficients while ordinary least squares pins down their numerical values. Applied to a 336-meter measurement line at a synchrotron facility in Beijing under 1–100 Hz sinusoidal excitation, the loop converges in four rounds to a closed-form formula with mean absolute error $3.13\times10^{-7}\,\mathrm{s}^2$ and relative mean absolute error $90.47\%$. If the claim holds, site-specific interpretable vibration-prediction formulas can be produced quickly for vibration-sensitive scientific infrastructure.

What carries the argument

The engine is the frequency-parameterized Bornitz equation $A(r,f)=A_0(r_0/r)^{n(f)}\exp(-\alpha(f)(r-r_0))$, which keeps the classical multiplicative separation of geometric spreading and material absorption while making both coefficients functions of frequency. The method alternates two cheap steps: fix $\alpha(f)$, solve for the optimal $n(f)$ at each frequency by unweighted least squares on the linearized log-amplitude relation, then use an intelligent formula generation model—a generative symbolic-regression network that emits candidate formulas token by token—to fit a readable expression to those per-frequency values; then fix $n(f)$ and repeat for $\alpha(f)$. Four iterations starting from $n(f)=0$ reduce the fitting error and terminate in explicit formulas for both coefficients, so the final object is an inspectable equation rather than a trained network.

What would settle it

Apply the published $n(f)$ and $\alpha(f)$ to a second, geologically different site, or to a held-out set of frequencies and distances from the same site, and check whether relative errors grow systematically with distance or amplitude; alternatively, fit a model in which $n$ also depends on $r$ and see whether it substantially improves the original data—either outcome would show the frequency-only Bornitz separation is not a transferable law.

Watch

Extended reading notes

Core claim

The central claim is that the hybrid iterative fitting procedure yields, from field data alone, an explicit frequency-dependent Bornitz law of the form $A(r,f)=A_0(r_0/r)^{n(f)}\exp(-\alpha(f)(r-r_0))$, with $n(f)$ a piecewise expression and $\alpha(f)$ a compact expression in frequency, giving a mean absolute error of $3.13\times10^{-7}\,\mathrm{s}^2$ and a relative mean absolute error of $90.47\%$ on the measured data. The paper further claims that this law outperforms both a calibrated fixed-parameter Bornitz formula and a prior frequency-dependent Bornitz variant in MAE and RMAE, that its reported minimum geometric damping coefficient of $0.47$ is consistent with the theoretical lower bound of $0.5$, and that its structure agrees with finite-element simulations showing $n$ is essentially independent of distance and grows with frequency. A log-normal residual model supplies distance-dependent $\sigma$ bands that contain $98.7\%$ of the data within $3\sigma$, giving the deterministic formula a probabilistic error envelope.

Load-bearing premise

The procedure assumes the attenuation law has the exact multiplicative form $A = A_0 (r_0/r)^{n} e^{-\alpha(r-r_0)}$ and that the two decay coefficients $n$ and $\alpha$ depend only on frequency, not on distance, vibration amplitude, soil layering, or source character; if attenuation is more entangled than that, the discovered formula is only an interpolation over the measured distances and frequencies at one site.

Editorial extensions

If this is right

  • Site-specific attenuation formulas can be derived automatically from a single field campaign, replacing manual calibration of the damping coefficients.
  • Because the final law is a closed-form equation, vibration impact assessments can trace predictions to physically meaningful geometric and material damping coefficients.
  • The discovered formula beats calibrated fixed-parameter Bornitz and earlier frequency-dependent formulas in both MAE and RMAE on this site, while black-box models that fit slightly better produce physically impossible negative amplitudes.
  • A distance-dependent probabilistic error band can be attached to the deterministic formula, giving engineers a quantified confidence interval for vibration limits.
  • The iterative scheme is not tied to one site; the same loop can be rerun wherever a vibration-sensitive facility is planned or monitored.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the frequency-only separation is an approximation rather than an exact law, the derived formula is best read as a compact interpolation over the measured 1–100 Hz and 2–336 m grid; extrapolating to other distances, frequencies, amplitudes, or soil columns would likely require refitting or additional dependence on soil properties.
  • The discovered frequency dependence of the coefficients invites a physical mapping against independent soil measurements: $\alpha(f)$ could be compared with viscoelastic damping models, and $n(f)$ with wave-front geometry, to test whether the fitted coefficients carry genuine material meaning.
  • The same alternating least-squares-plus-symbolic-regression loop could be applied to other physically anchored formulas, for example with coefficients depending on source depth or excitation amplitude, giving a general recipe for transparent predictive laws in engineering geophysics.
  • Because the symbolic-regression step is stochastic, different runs may converge to different but statistically equivalent expressions; reporting the spread of candidate formulas would clarify how much of the discovered form is forced by the data and how much is a modeling choice.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes a hybrid method that combines the classical Bornitz attenuation formula with an intelligent formula generation model (a sequence-to-sequence approach) to derive frequency-dependent geometric and material damping coefficients, n(f) and alpha(f), from field vibration measurements at the High Energy Photon Source (HEPS) site. The final expressions, Eq. (11)–(13), are reported with MAE 3.13e-7 s2 and RMAE 90.47% (Section 3.2). The authors argue that the structure is physically justified by energy-conservation arguments and by an ABAQUS/Explicit finite-element simulation that reproduces the increasing trend of n(f) with frequency. They then construct a distance-dependent error band, Eq. (18), using a log-normal assumption, and compare the formula's fitting accuracy with calibrated fixed-parameter Bornitz and Yang formulas, XGBoost, and a DNN (Table 2). The paper claims that the proposed method balances interpretability and accuracy and provides a transferable framework for vibration assessment at similar facilities.

Significance. If the central claim were established, the paper would offer a useful contribution: an interpretable, site-specific formula for ground-vibration attenuation that is frequency-dependent, based on a rich field dataset (64 sensors, 1–100 Hz, distances to 336 m). The strengths are the careful field experiment, the sensor consistency checks, the explicit iterative fitting procedure, the comparison with several baselines, and a probabilistic error model. However, the significance is diminished by a central methodological gap: all reported accuracies are in-sample, and the error model is calibrated and evaluated on the same residuals. The paper is honest about some of these limitations (e.g., the FE values 'do not carry direct physical significance'), but the abstract and conclusions state a stronger claim of 'high-precision' and 'predictive accuracy' than the evidence supports. The contribution is therefore better framed as a demonstration of a fitting methodology with an in-sample description of the HEPS dataset, not as a validated predictive law.

major comments (4)
  1. [Section 3.2] The reported MAE (3.13e-7 s2) and RMAE (90.47%) are computed on the same data used to fit n(f) and alpha(f); no distance range, frequency band, or second site is held out. The claim that Eq. (11)–(13) is a 'high-precision' attenuation law is therefore supported only by training-set error. Please provide a held-out evaluation, e.g., leave-one-sensor-out, a random 20% test split, or an external site, and report the test-set MAE/RMAE. Without such validation, the formula is a local interpolation over the measured r–f grid rather than a demonstrated predictive law.
  2. [Section 4.1] The error-band model sigma(r) in Eq. (18) is fitted to the residuals of Eq. (11)–(13), and then the same residuals are used to report that 71.5% of points fall within the 1-sigma range. This is a self-consistency check, not a predictive calibration. Under an assumed normal (or log-normal) distribution, fitting sigma to the data will by design produce coverage close to the nominal level when evaluated on those same data. Please validate the error band on independent data (e.g., a held-out set) or at least state explicitly that the reported coverage is an in-sample property and re-frame the claim accordingly. A chi-square or probability-integral-transform test on held-out data would be a more meaningful check.
  3. [Section 3.3] The finite-element simulation is used to support the assumption that n(f) depends on frequency and not on distance, but the text itself states that the simulated n values 'do not carry direct physical significance' because of the difficulty of modeling real, heterogeneous soil. The simulation only reproduces an increasing trend at four frequencies (1, 5, 10, 20 Hz) in a homogeneous elastic half-space, so it does not validate that the Bornitz structure with n and alpha depending only on f holds in the actual HEPS soil. The paper should temper the claim that the FE results 'support our assumption' and should discuss what additional data (e.g., a second site, varying source amplitudes, or a layered soil model) would be needed to test transferability of the discovered formula.
  4. [Section 4.2] The comparison with XGBoost and DNN is not a fair predictive comparison as presented. The ML models are trained on 75% of the data, but the evaluation set is not specified; if Table 2 reports errors on the training data for the ML models, that would understate their generalization error, while the proposed formula is evaluated in-sample on all data. Please specify the exact evaluation protocol (training/test split for ML and the identical split for the Bornitz/Yang calibrations) and report test-set metrics for all methods. Without this, the claimed 'significant advantages' of the proposed formula over black-box models are not established.
minor comments (5)
  1. [Throughout] There are several notation inconsistencies: Eq. (1) uses A1 and r1, while the text and later equations use Ar and r; the subscripts in Eq. (9) are garbled (e.g., 'f f' and 'Sfx'); and the text around Eq. (2)–(4) refers to 'A0 and r0' and 'A1 and r1' in ways that are not consistently defined. Please unify the notation.
  2. [Section 3.1] Fig. 3(b) is described as an 'on-site consistency test' but the caption and text refer to 'Fig. 2(b)' in one place; please correct the cross-reference.
  3. [Section 4.1] The Shapiro–Wilk test is described, but the p-values are not reported; only W statistics are shown in Fig. 12. Reporting p-values (or at least stating the sample size per r) would allow the reader to judge whether the normality assumption is actually supported at each distance.
  4. [Section 4.1] The sentence 'The distribution of these data points does not exhibit any clear pattern ... This finding demonstrates the robust generalization capability of the formula within the tested range' is an overinterpretation: a random-looking pattern of large errors does not demonstrate generalization, which requires a held-out evaluation.
  5. [Section 2.3] The description of the iterative algorithm would benefit from a formal statement of the optimization objective and a convergence criterion. As written, it is unclear when the iterations stop and whether the final result depends on the starting point n(f)=0.

Circularity Check

2 steps flagged · score 6.0 of 10

Error-band coverage is a self-consistency check, and the headline MAE/RMAE are training-set residuals rather than independent predictions.

  1. fitted input called prediction [Section 3.2, Eq. (11)–(13), and Table 2 in Section 4.2]
    "The fitting results of each iteration are substituted into Equation (5) to calculate the mean absolute error (MAE) and relative mean absolute error (RMAE) of the formula... where the MAE is 3.13×10-7 s2 and the RMAE is 90.47%."

    The coefficients n(f) and alpha(f) in Eqs. (12)–(13) are obtained by iterative least-squares-style fitting on the full HEPS dataset (64 sensors, 1–100 Hz, reference r0=2.1 m). The MAE and RMAE quoted after Eq. (13) are computed on that same dataset, so they measure the fitting residual, not a held-out prediction. Table 2 compares formula-based methods in-sample as well; because the proposed formula has frequency-dependent coefficients and the Bornitz/Yang comparators do not, its lower MAE is the expected effect of extra flexibility rather than independent evidence of predictive superiority. Calling the fitted values 'predicted' (Fig. 16) does not turn the training loss into an out-of-sample test.

  2. self definitional [Section 4.1, Eq. (18), after Fig. 14]
    "Based on the fitted results, the 1-2-3-σ range for different r values is determined, as illustrated in Fig. 14. The results show that 71.5% of the data points fall within the 1-σ range, 94.2% of the data points fall within the 2-σ range, and 98.7% of the data points fall within the 3-σ range, demonstrating the effectiveness of the fitting model."

    The residual distribution is built from the same data that the coverage claim then evaluates: 'for each r, we assume μ=0 and use data from different f values to obtain an unbiased estimate of σ', and Eq. (18) is fitted to those σ(r) estimates. The reported percentages count how many of those same residuals fall inside quantiles of the fitted normal error model. With σ estimated from the residuals, roughly 68%, 95%, and 99.7% coverage is expected for a normal model by construction, so the quoted percentages are a self-consistency check, not an independent validation of the error model.

full rationale

The paper's derivation is mostly transparent: Eq. (11)–(13) are produced by an alternating least-squares / symbolic-regression fit to the measured HEPS amplitudes, and the paper often calls the error the 'fitting error'. That transparency does not remove the circularity in how the results are presented. The headline accuracy (MAE 3.13e-7 s2, RMAE 90.47%) is computed on the very dataset used to determine n(f) and alpha(f), and Table 2 compares all formula-based methods in-sample, so the numerical advantage over fixed-coefficient Bornitz and Yang formulas is partly a flexibility effect rather than a demonstrated predictive gain. The error-band analysis is more explicitly circular: sigma(r) is estimated from the residuals and then used to report 1-, 2-, and 3-sigma coverage of those same residuals, making the coverage a self-consistency check. The finite-element simulation in Section 3.3 is independent of the fitted constants, but it is only a trend-level homogeneous-elastic check, and the paper itself says the simulated coefficient values 'do not carry direct physical significance'. The self-citation to Chen et al. [31] for the formula-generation model is a normal method citation and is not load-bearing for the fitted result. Overall, the fitted formula is not circular qua formula; the circularity lies in presenting training residuals and self-fitted error coverage as validation. The central predictive claim therefore remains unestablished until tested on held-out distances, frequencies, or a second site.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central formula rests on several fitted parameters and domain assumptions. Most numbers in the final equations are fit to the HEPS dataset. No new physical entities are introduced.

free parameters (5)
  • Piecewise coefficients of n(f) = 5.66, 0.38, 0.00027, 1.64e-8, exponent 0.035, threshold 23 Hz
    Fitted in iterative rounds by linear regression and symbolic regression to minimize MAE on HEPS data (Table 1, Eq. 12).
  • Coefficients of alpha(f) = 25.9, 0.035
    Fitted in the same iterative process (Table 1, Eq. 13).
  • Breakpoint between piecewise regimes of n(f) = 23 Hz
    The threshold separating the two n(f) expressions is chosen during symbolic regression, not derived from theory.
  • Coefficients of sigma(r) error band = 1, 0.0025, 0.62
    Fitted to the per-distance standard deviations of residuals (Eq. 18).
  • Coherence exclusion threshold = 0.9
    Data points with coherence below this value are removed before fitting, a hand-set threshold (Section 3.1).
assumptions (5)
  • domain assumption Bornitz formula A = A0 (r0/r)^n exp(-alpha(r-r0)) is the correct structural model for ground vibration attenuation.
    The paper uses this as the base model and only makes n and alpha frequency-dependent (Section 2.1).
  • domain assumption Geometric and material damping coefficients n and alpha depend only on frequency f, not on distance r.
    This is required for the per-frequency linear regressions and the functional forms in Eq. (5); the FE check only tests the distance-independence in a homogeneous medium.
  • domain assumption The intelligent formula generation model from Chen et al. [31] can find valid, parsimonious expressions for n(f) and alpha(f).
    The whole pipeline relies on this external model, which is cited as in press and not shipped or fully described.
  • domain assumption The log-normal error model with zero mean in log space is a valid stochastic model for the residuals at each distance.
    Selected by Shapiro-Wilk comparisons on the same data (Section 4.1); the model is then used to define the error band.
  • ad hoc to paper The finite-element model with elastic parameters (density 1500 kg/m3, E=5000 kPa, Poisson ratio 0.4) is representative enough to validate the trend of n(f).
    These material parameters are chosen arbitrarily for a homogeneous half-space and are not calibrated to the HEPS site, yet the FE results are used to support the physical form of the formula.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Formula-Guided Machine Learning for Ground Vibration Propagation and Attenuation Modeling." pith.science (2026). https://pith.science/paper/JUE6XLKX

@misc{pith2026250513870,
  author       = {Pith},
  title        = {Pith review of: Formula-Guided Machine Learning for Ground Vibration Propagation and Attenuation Modeling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JUE6XLKX}},
  note         = {Machine review of arXiv:2505.13870}
}
read the original abstract

Understanding the propagation and attenuation patterns of ground vibrations is critical for evaluating the impact of environmental disturbances on large-scale scientific facilities. However, complex site conditions often result in intricate vibration behaviors, limiting the accuracy of traditional predictive methods. This study proposes a hybrid iterative fitting method that integrates machine learning with the Bornitz formula through an intelligent formula generation model. The method enables the automatic derivation of high-precision, interpretable ground vibration attenuation formulas from experimental data. A case study was conducted at the High Energy Photon Source in Beijing, where field tests were performed to collect vibration data. Using the proposed approach, an attenuation formula describing ground vibration propagation was derived. The physical validity of the model was further verified via finite element simulations. A probabilistic analysis was then employed to estimate computational errors. Comparative evaluations with black-box machine learning models and empirical formulas from previous studies demonstrate that the proposed method offers significant advantages in both interpretability and accuracy. These findings provide a valuable framework for vibration impact assessment and mitigation in other large-scale scientific infrastructure projects.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

53 extracted references · 42 canonical work pages

  1. [1]

    With the advancement of technology, the infrastructure supporting high -precision technologies demands stringent control over vibrations

    Introduction The propagation and attenuation of vibrations have important applications in fields such as transportation [1,2], pipelines [3,4] and environmental engineering [5–7]. With the advancement of technology, the infrastructure supporting high -precision technologies demands stringent control over vibrations. Ground vibrations, a common form of vib...

    work page

  2. [2]

    Combining traditional machine learning methods with intelligent formula generation model to discover propagation and attenuation laws based on the Bornitz model, leading to the development of a ground vibration propagation formula that is frequency-dependent

    work page

  3. [3]

    Using the proposed method, an attenuation formula with high accuracy and strong interpretability was derived from the experimental data

    Field experiments were conducted at HEPS using a vibrator to generate sinusoidal excitation forces in the frequency range of 1 –100 Hz. Using the proposed method, an attenuation formula with high accuracy and strong interpretability was derived from the experimental data

    work page

  4. [4]

    Providing an explanation of the physical rationality of the formula based on theoretical derivation and numerical simulation

    work page

  5. [5]

    The organization of the remaining sections of the paper is as follows

    Employing probabilistic analysis methods to estimate the error range of the fitting formula and 5 comparing the proposed method with black -box machine learning models and previous empirical formulas, demonstrating its significant advantages in terms of accuracy and interpretability. The organization of the remaining sections of the paper is as follows. S...

    work page

  6. [6]

    Formula-Guided Machine Learning Method 2.1. Bornitz Model The Bornitz formula [22] is a classical model used to describe ground vibration propagation, which approximates the nonlinear behavior of soil by introducing geometric and material damping coefficients. However, the original form of the Bornitz formula is overly simplistic and does not accommodate ...

    work page

  7. [7]

    Field test on HEPS The High Energy Photon Source (HEPS), as shown in Fig

    Experiment and results 3.1. Field test on HEPS The High Energy Photon Source (HEPS), as shown in Fig. 2, is the first fourth -generation synchrotron light source facility being built in China [32]. Located in the northern core area of Huairou Science City in Beijing, this advanced facility serves as a high -performance, high-energy synchrotron radiation l...

    work page

  8. [8]

    Analysis and comparison 4.1. Error analysis Owing to the varying initial amplitudes of the vibration sources at different frequencies and the order - of-magnitude differences arising from the attenuation of vibrations with distance, the MAE, when measured in absolute terms, fails to adequately reflect the fitting error of the formula. Additionally, the co...

    work page

Show all 53 references
  1. [9]

    The method is validated through surface vibration measurements conducted at the HEPS in Beijing

    Conclusions In this study, we propose an intelligent method for fitting analytical formulas that describe ground vibration propagation and attenuation. The method is validated through surface vibration measurements conducted at the HEPS in Beijing. The main contributions of th...

  2. [10]

    This approach enables the efficient and rapid derivation of propagation and attenuation formulas based on the Bornitz model

    A formula discovery method integrating traditional machine learning techniques with intelligent formula generation models was proposed. This approach enables the efficient and rapid derivation of propagation and attenuation formulas based on the Bornitz model

  3. [11]

    The results lay a solid foundation fo r evaluating vibration impacts on precision instruments

    High-fidelity vibration experiments are presented at the HEPS site, providing reliable data for understanding ground vibration propagation. The results lay a solid foundation fo r evaluating vibration impacts on precision instruments

  4. [12]

    The validity of the formula's structure was justified through theoretical analysis and numerical modeling

    Using the proposed method, a propagation attenuation formula suitable for ground vibrations at the HEPS site was discovered. The validity of the formula's structure was justified through theoretical analysis and numerical modeling

  5. [13]

    A probabilistic model was employed to analyze the errors of the discovered formulas, providing a bounded error range model. 27

  6. [14]

    A comparative analysis was performed against previously proposed formulas and black -box models, dem onstrating that the proposed method achieves both high accuracy and interpretability, making it suitable for ground vibration prediction and assessment. This research offers a ...

  7. [15]

    of 12375150

    Acknowledgements The authors gratefully acknowle dge the financial support provided by the General Program of National Natural Science Foundation of China, under fund approval No. of 12375150. 28

  8. [16]

    Jin, D.J

    Q. Jin, D.J. Thompson, D.E.J. Lurcock, E. Ntotsios, The shadow effect on the ground surface due to vibration transmission from a railway tunnel, Transportation Geotechnics 23 ( 2020) 100335. https://doi.org/10.1016/j.trgeo.2020.100335

    2020

  9. [17]

    Zhang, S

    X. Zhang, S. Zhou, C. He, H. Di, J. Si, Experimental investigation on train -induced vibration of the ground railway embankment and under -crossing subway tunnels, Transportation Geotechnics 26 ( 2021) 100422. https://doi.org/10.1016/j.trgeo.2020.100422

    2021

  10. [18]

    Zhang, M

    H. Zhang, M. Qin, K. Liao, K. Wang, G. He, Pipe-soil vibration characteristics of natural gas pipelines during the pigging process, Journal of Natural Gas Science and Engineering 95 (2021) 1041 48. https://doi.org/10.1016/j.jngse.2021.104148

    2021

  11. [19]

    X. Cui, Y . Gao, J. Muggleton, Y . Liu, Superimposed imaging of acoustic wave reflections for the detection of underground nonmetallic pipelines, Mechanical Systems and Signal Processing 209 (2024) 111127. https://doi.org/10.1016/j.ymssp.2024.111127

    2024

  12. [20]

    J. Wang, X. Xiao, L. Peng, J. Wang, Y . He, X. Sheng, Modelling metro-induced environmental vibration by combining physical-numerical and deep learning methods, Mechanical Systems and Signal Processing 220 (2024) 111687. https://doi.org/10.1016/j.ymssp.2024.111687

    2024

  13. [21]

    Y . He, X. Sheng, J. Han, Y . Zhang, X. Xiao, Modeling ground vibration excited by a high-speed train running along a viaduct railway using a new structure-soil interaction method, Science of The Total Environment 956 (2024) 177136. https://doi.org/10.1016/j.scitotenv.2024.177136

    2024

  14. [22]

    Z. Li, M. Ma, K. Liu, B. Jiang, Performance of rubber -concrete composite periodic barriers applied in attenuating ground vibrations induced by metro trains, Engineer ing Structures 285 (2023) 116027. https://doi.org/10.1016/j.engstruct.2023.116027

    2023

  15. [23]

    Xia, Y .M

    H. Xia, Y .M. Cao, G. De Roeck, Theoretical modeling and characteristic analysis of moving -train induced ground vibrations, Journal of Sound and Vibration 329 (2010) 819 –832. https://doi.org/10.1016/j.jsv.2009.10.007

    2010 doi

  16. [24]

    Herbut, A study of the reduction of ground vibrations by an active generator, Soil Dynamics and Earthquake Engineering 88 (2016) 328–344

    A. Herbut, A study of the reduction of ground vibrations by an active generator, Soil Dynamics and Earthquake Engineering 88 (2016) 328–344. https://doi.org/10.1016/j.soildyn.2016.06.015

    2016 doi

  17. [25]

    Q. He, Y . Yang, J. Yang, S. Zhu, W. Zhai, Experimental investigation of dynamic features and propagation law of ground vibrations induced by a suspended monorail train, Soil Dynamics and Earthquake Engineering 183 (2024) 108789. https://doi.org/10.1016/j.soildyn.2024.108789

    2024

  18. [26]

    Yang, L.T

    Y .B. Yang, L.T. Xie, Y .Z. Liu, J. Li, Analytical solution of ground vibration of a half-space induced by sub- and super-critical harmonic moving loads: Theory and application, Soil Dynamics and Earthquake Engineering 187 (2024) 109019. https://doi.org/10.1016/j.soildyn.2024.109019

    2024

  19. [27]

    M. Wu, B. Wang, Z. Ba, K. Dai, J. Liang, Propagation attenuation of elastic waves in multi -row infinitely periodic pile barriers: A closed -form analytical solution, Engineering Structures 315 (2024) 118480. https://doi.org/10.1016/j.engstruct.2024.118480

    2024

  20. [28]

    S. Wang, T. Ma, W. Zhang, Statistical characterization and attenuation of vibration effect for continuous asphalt pavement survey using ground -penetrating radar, Automation in Construction 173 (2025) 10605 7. https://doi.org/10.1016/j.autcon.2025.106057

    2025

  21. [29]

    Zarei, J

    Z. Zarei, J. Sadeghi, A. Sarkar, Evaluation of heavy-vehicle-induced vibrations running on asphalt pavements, Construction and Building Materials 358 (2022) 129399. https://doi.org/10.1016/j.conbuildmat.2022.129399

    2022

  22. [30]

    C. Wang, L. Song, Z. Yuan, J. Fan, State -of-the-art AI-based computational analysis in civil engineering, Journal of Industrial Information Integration 33 (2023) 100470. https://doi.org/10.1016/j.jii.2023.100470. 29

    2023

  23. [31]

    C. Yang, C. Wang, B. Wu, F. Zhao, J. Fan, L. Zhou, Settlement estimation during foundation excavation using pattern analysis and explainable AI modeling, Automation in Construction 166 (2024) 105651. https://doi.org/10.1016/j.autcon.2024.105651

    2024

  24. [32]

    C. Yang, C. Wang, F. Zhao, B. Wu, J. Fan, Y . Zhang, Smart virtual sensing for deep excavations using real-time ensemble graph neural networks, Automation in Construction 172 (2025) 106040. https://doi.org/10.1016/j.autcon.2025.106040

    2025

  25. [33]

    J. Cao, T. Wang, L. Shang, J. Wang, C.-M. V ong, C. Yin, X. Huang, A novel distance estimation algorithm for periodic surface vibrations based on frequency band energy percentage feature, Mechanical Systems and Signal Processing 113 (2018) 222–236. https://doi.org/10.1016/j.ym...

    2018 doi

  26. [34]

    Monjezi, M

    M. Monjezi, M. Ghafurikalajahi, A. Bahrami, Prediction of blast -induced ground vibration using artificial neural networks, Tunnelling and Underground Space Technology 26 (2011) 46 –50. https://doi.org/10.1016/j.tust.2010.05.002

    2011 doi

  27. [35]

    W. Liu, R. Liang, H. Zhang, Z. Wu, B. Jiang, Deep learning based identification and uncertainty analysis of metro train induced ground -borne vibration, Mechanical Systems and Signal Processing 189 (2023) 110062. https://doi.org/10.1016/j.ymssp.2022.110062

    2023

  28. [36]

    Verma, T.N

    A.K. Verma, T.N . Singh, Intelligent systems for ground vibration measurement: a comparative study, Engineering with Computers 27 (2011) 225–233. https://doi.org/10.1007/s00366-010-0193-7

    2011 doi

  29. [37]

    Bornitz, Über die Ausbreitung der von Großkolbenmaschinen erzeugten Bodenschwingungen in die Tiefe, Springer-Verlag, 2013

    G. Bornitz, Über die Ausbreitung der von Großkolbenmaschinen erzeugten Bodenschwingungen in die Tiefe, Springer-Verlag, 2013

    2013

  30. [38]

    Kim, J.-S

    D.-S. Kim, J.-S. Lee, Propagation and attenuation characteristics of various ground vibrations, Soil Dynamics and Earthquake Engineering 19 (2000) 115–126. https://doi.org/10.1016/S0267-7261(00)00002-6

    2000 doi

  31. [39]

    X. Ren, J. Wu, Y . Tang, J. Yang, Propagation and attenuation characteristics of the vibration in soft soil foundations induced by high -speed trains, Soil Dynamics and Earthquake Engineering 117 (2019) 374 –383. https://doi.org/10.1016/j.soildyn.2018.11.004

    2019 doi

  32. [40]

    D. Niu, Y . Deng, H. Mu, J. Chang, Y . Xuan, G. Cao, Attenuation and propagation characteristics of railway load-induced vibration in a loess area, Transportation Geotechnics 37 (2022) 100858. https://doi.org/10.1016/j.trgeo.2022.100858

    2022

  33. [41]

    Petersen, M.L

    B.K. Petersen, M.L. Larma, T.N. Mundhenk, C.P. Santiago, S.K. Kim, J.T. Kim, Deep symbolic regression: Recovering mathematical expressions from data via risk -seeking policy gradients, in: 2020. https://openreview.net/forum?id=m5Qsh0kBQG (accessed February 26, 2024)

    2020

  34. [42]

    Vastl, J

    M. Vastl, J. Kulhánek, J. Kubalík, E. Derner, R. Babuška, SymFormer: End-to-end symbolic regression using transformer-based architecture, (2022). https://doi.org/10.48550/arXiv.2205.15764

    2022 doi arXiv

  35. [43]

    Tenachi, R

    W. Tenachi, R. Ibata, F. Diakogiannis, Deep Sy mbolic Regression for Physics Guided by Units Constraints: Toward the Automated Discovery of Physical Laws, The Astrophysical Journal 959 (2023) 99. https://doi.org/10.3847/1538-4357/ad014c

    2023 doi

  36. [44]

    Hochreiter, J

    S. Hochreiter, J. Schmidhuber, Long Short -term Memory, Neura l Computation 9 (1997) 1735 –80. https://doi.org/10.1162/neco.1997.9.8.1735

    1997 doi

  37. [45]

    Vaswani, N

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A.N. Gomez, Ł. Kaiser, I. Polosukhin, Attention is all you need, in: Proceedings of the 31st International Conference on Neural Information Processing Systems, Curran Associates Inc., Red Hook, NY , USA, 2017: pp. 6000–6010

    2017

  38. [46]

    P.-Y . Chen, C. Wang, J.-S. Fan, Multivariate engineering formulas discovery with knowledge -based neural network, Computer-Aided Civil and Infrastructure Engineering n/a (n.d.). https://doi.org/10.1111/mice.13448

    doi

  39. [47]

    Y . Jiao, G. Xu, X.-H. Cui, Z. Duan, Y .-Y . Guo, P. He, D.-H. Ji, J.-Y . Li, X.-Y . Li, C. Meng, Y .-M. Peng, S.-K. Tian, J.-Q. Wang, N. Wang, Y .-Y . Wei, H.-S. Xu, F. Yan, C.-H. Yu, Y .-L. Zhao, Q. Qin, The HEPS project, J Synchrotron Rad 25 (2018) 1611–1618. https://doi.or...

    2018 doi

  40. [48]

    https://www.smartsolo.com.cn/xl-IGU.html (accessed May 1, 2025)

    IGU SmartSolo, (n.d.). https://www.smartsolo.com.cn/xl-IGU.html (accessed May 1, 2025)

    2025

  41. [49]

    Amirikas, A

    R. Amirikas, A. Bertolini, W. Bialowons, H. Ehrlichmann, D. Elektronen-Synchrotron, GROUND MOTION & COMPARISON OF V ARIOUS SITES, (2005)

    2005

  42. [50]

    https://www.donglingtech.com/ (accessed May 1, 2025)

    Suzhou Dongling vibration test instrument Co., Ltd, (n.d.). https://www.donglingtech.com/ (accessed May 1, 2025)

    2025

  43. [51]

    Yang, Attenuation of Ground Vibrat ion Induced by Dynamical Machinery, in: 1981

    S. Yang, Attenuation of Ground Vibrat ion Induced by Dynamical Machinery, in: 1981. https://www.semanticscholar.org/paper/Attenuation-of-Ground-Vibration-Induced-by- Yang/0d43b1b3b7315076a6a07dd92f0a54446ee0d613 (accessed March 13, 2025)

    1981

  44. [52]

    Yang, Evaluation of Man -Made Ground Vibration s, in: 1995

    X.-J. Yang, Evaluation of Man -Made Ground Vibration s, in: 1995. https://www.semanticscholar.org/paper/Evaluation-of-Man-Made-Ground-Vibrations- Yang/2d24d0a917c0550e1a252751e22b2c1b484ecf81 (accessed March 13, 2025)

    1995

  45. [53]

    https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.minimize.html (accessed March 13, 2025)

    minimize — SciPy v1.15.2 Manual, (n.d.). https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.minimize.html (accessed March 13, 2025)

    2025

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.