Pith. sign in

REVIEW 2 major objections 6 minor 87 references

A trained LSTM reproduces full adhesive force trajectories with ~2% pull-off error

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 13:32 UTC pith:XI2DQBKF

load-bearing objection Solid trajectory-level surrogate study; the FMS resampling at N=120 is the real soft spot and the authors know it. the 2 major comments →

arxiv 2607.19060 v1 pith:XI2DQBKF submitted 2026-07-21 cs.LG cond-mat.softcs.AIphysics.data-anstat.ML

Deep learning-based prediction of time-resolved adhesive forces in viscoelastic Hertzian contacts

Ali Maghami , Merten Stender , Michele Ciavarella , Antonio Papangelo This is my paper
classification cs.LG cond-mat.softcs.AIphysics.data-anstat.ML
keywords viscoelastic adhesionHertzian contactsurrogate modelLSTMsequence-to-sequenceTabor parameteradhesive force trajectorycontact mechanics
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper asks whether a deep sequence model can replace expensive numerical simulation of viscoelastic adhesive contact, where the force at any instant depends on the entire loading history. It answers yes: a two-layer LSTM that receives the displacement history plus the Tabor adhesion-regime parameter predicts the complete time-resolved force trajectory for loading–dwell–unloading protocols. On held-out trajectories spanning Tabor parameter 0.2 to 3.2 and loading/unloading rates over four orders of magnitude, the model reaches a mean-squared error of 5.0e-4, median pull-off-force error of about 2.2%, median hysteresis error of about 1.1%, and median inference time of 0.16 s. The practical stake is that repeated force-trajectory evaluations, once requiring minutes to hours, become essentially instant, enabling real-time control and design optimization for soft gripping.

Core claim

The central claim is that the full force history of an adhesive viscoelastic Hertzian contact—not just scalar detachment metrics—is learnable from the prescribed displacement protocol and the Tabor parameter alone. Because viscoelasticity introduces memory via a Boltzmann convolution, the mapping is path-dependent and requires a stateful sequence model; the paper shows that an LSTM with concatenated Tabor conditioning (256 then 128 units) is sufficient. The model is trained on 12,450 boundary-element trajectories and verified on unseen parameter combinations, on both low- and high-Tabor adhesion regimes, and against an analytical crack-propagation limiting case for rate-dependent effective s

What carries the argument

The fixed-measurement-step (FMS) representation: each variable-length numerical trajectory is spline-resampled to 120 measurement steps (50 loading, 20 dwell, 50 unloading) while preserving a physical-time channel, so heterogeneous simulations become fixed-length sequences usable in batch training. The learned surrogate is a two-layer LSTM (256→128) with the static Tabor parameter concatenated at every step, followed by a time-distributed dense readout; the LSTM's internal cell state carries the fading memory that mirrors the convolution integral. Physics-guided input channels—causally computed velocity and a binary edge indicator marking dwell onset/termination—reduce validation error by ro

Load-bearing premise

The result depends on the fixed-measurement-step grid of 120 points—especially 50 points over unloading—preserving the sharp snap-off events; if that grid aliases fast pull-offs at high Tabor parameter, the reported global errors understate the worst-case trajectory errors in exactly the detachment regime that matters.

What would settle it

Run a dense set of held-out high-Tabor (µ ≈ 3.2), fast-unloading (v̂U ≈ 10³) protocols and compare predicted force trajectories to boundary-element simulations; if the median pull-off-force error in that subset exceeds about 2.2% or the mean force MSE is an order of magnitude above the reported 5e-4, the representation is not resolving snap-off. A complementary check: retrain with N = 240 FMS points and see whether worst-quartile errors shrink materially; if they do not, the error is not resolution-limited.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Force trajectories for unseen loading–dwell–unloading protocols can be predicted in about 0.16 s, roughly three orders of magnitude faster than the boundary-element solver, with a fixed cost independent of physical regime.
  • The surrogate reproduces both short-range (JKR-like) and long-range (DMT-like) adhesion behavior, so it can serve as a rapid detector of pull-off force, pull-off time, and hysteresis across the Tabor range 0.2–3.2.
  • Because the inputs are the displacement history plus Tabor parameter rather than precomputed protocol scalars, the model is positioned to generalize to new loading shapes within the same constitutive family.
  • The trained model can be evaluated at nearby FMS resolutions (100 or 140 steps) without retraining, though accuracy is best at the training resolution.

Where Pith is reading between the lines

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

  • The same FMS-plus-LSTM recipe should transfer to other history-dependent contact quantities—friction hysteresis, multi-asperity adhesion, or contact stiffness—provided a numerical or experimental dataset with variable-length trajectories is available; this is an extension the paper does not itself demonstrate.
  • The residual error concentration at snap-off suggests a cheap, testable improvement: increase FMS resolution in the unloading phase or sample near-instability trajectories more densely; if errors drop sharply, the remaining bottleneck is representation rather than architecture.
  • The 0.16 s inference time opens the door to closed-loop use the paper only gestures at, such as inverting the model to design displacement protocols that achieve a target pull-off force or hysteresis, since gradients through the differentiable surrogate could be used in optimization.
  • A natural stress test is to train on a two-relaxation or power-law viscoelastic model; success would suggest the LSTM learns generic fading-memory structure rather than overfitting the single-relaxation standard linear solid used here.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 6 minor

Summary. The paper presents a deep-learning surrogate for the complete time-resolved adhesive force response of a rigid sphere indenting an adhesive viscoelastic half-space under loading-dwell-unloading protocols. The authors generate 12,450 BEM trajectories spanning four orders of magnitude in loading/unloading rates, dwell times, and Tabor parameters 0.2–3.2, and encode them on a fixed 120-point measurement grid (FMS). They compare 18 sequence-model variants (LSTM, layer-normalized LSTM, CNN+LSTM, TCN, transformer, residual LSTM, each with concatenated, FiLM, or gated Tabor conditioning). The best model, M1-concat (two-layer LSTM with concatenated Tabor conditioning), is reported to achieve held-out MSE 5.0e-4, median pull-off force error ≈2.2%, median hysteresis error ≈1.1%, and median inference time 0.16 s. The model is also checked against the Persson–Brener effective-surface-energy relation in the fully relaxed short-range limit. The central claim is that a fixed-cost neural sequence model can replace repeated BEM evaluations for within-distribution protocols, with accuracy that is controlled and improvable.

Significance. If the FMS representation is shown to preserve the physical content of the BEM trajectories, this is a genuinely useful contribution. It is, to my knowledge, the first trajectory-level surrogate for viscoelastic adhesive Hertzian contacts, going beyond scalar pull-off prediction. The systematic comparison of 18 architectures with three conditioning mechanisms, the quartile-stratified error analysis, the data-efficiency experiment, and the live demonstrator are valuable for future surrogate-model work in contact mechanics. The paper is also appropriately careful to present the surrogate as an amortized tool rather than a replacement for high-fidelity simulation, and it identifies the sharp-transition regime as the main accuracy bottleneck. The main uncertainty concerns whether the fixed measurement-step representation itself introduces bias in exactly the detachment events the surrogate is meant to predict.

major comments (2)
  1. [§2.3, Eq. (6); §3.4, Table 2] The FMS representation is the load-bearing link between the BEM reference and every reported error statistic, yet the paper never measures the error introduced by the FMS resampling itself. All training targets and all test errors (MSE, pull-off force, hysteresis, pull-off time) are computed on the 120-point FMS grid. The native BEM solution is adaptive; a snap-off whose duration is shorter than the local FMS spacing is smoothed by spline interpolation before the target is formed. The manuscript itself identifies exactly this regime: §4 states errors concentrate in 'sharply varying trajectories', §5 lists 'increasing the FMS' as a remedy, and Appendix H shows visible FMS-resolution sensitivity even for sample (8). Since pull-off force and its timing are the quantities most needed for gripping/release, I request an explicit FMS-convergence analysis: take a stratified subset of high-Tabor/
  2. [§2.4, §3.1, Figure 5] All headline numbers are point estimates from one random 80/10/10 split and a single training run of each architecture (fixed seed). Given that hyperparameters were selected on the same validation split, the claim that M1-concat is the best among 18 models and the reported hold-out MSE/median errors need uncertainty quantification. The data-fraction experiment of Figure 5(d) uses five retrainings on nested subsets but keeps validation and test partitions fixed, so it does not address split variability. Please retrain the reference model (and ideally the other five families with concat conditioning) with at least 5 different seeds/splits and report mean ± std for MSE, pull-off error, hysteresis error, and pull-off-time error. Without this, the comparison in Figure 5(b) may be within run-to-run noise, and the 'best architecture' conclusion is under-supported.
minor comments (6)
  1. [Highlights / Abstract] The inference-time claim is inconsistent: the abstract states a median inference time of 0.16 s, while the Highlights state 'about 0.25 s for a full force branch'. Table 1 reports values from 0.12 to 0.26 s. Please unify the phrasing.
  2. [Graphical abstract] Typo: 'less than a seccond' should be 'less than a second'.
  3. [§2.4 / §5] The random split means the test set is drawn from the same continuous parameter distribution and mostly contains interpolated parameter combinations. This is acceptable for an in-range surrogate, but the text should explicitly say that 'unseen parameter combinations' means within the sampled range, and state that extrapolation beyond the Tabor/rate ranges is not claimed. A structured split (e.g., by velocity octave or Tabor band) would strengthen the generalization claims if the authors intend to make them.
  4. [§2.2 / §5] The dwell-time range is stated as [10^-3, 3] in §2.2 but as 'dwell times from 10^-3 to 5' in the conclusions (§5). Please reconcile.
  5. [Table A.3 caption] The caption refers to 'Figure 1(e)' for the peak normalized indentation, but Figure 1 appears to have panels (a)–(d). Check the panel reference.
  6. [§3.2, Figure 5(c)] The Persson–Brener verification is a useful consistency check but not an independent validation: the crack-velocity correspondence and the empirical coefficient α are taken from the authors' own Ref. [19] and its deposited data. The deviations at the highest crack velocities should be quantified numerically in the text, not only described qualitatively, since they are directly connected to the FMS-resolution concern raised above.

Circularity Check

0 steps flagged

No significant circularity: the surrogate training/evaluation chain is self-contained; self-citations are minor and not load-bearing, while FMS aliasing is an accuracy limitation, not a circular reduction.

full rationale

The central derivation is a standard supervised-learning pipeline: inputs are the FMS-resampled displacement history plus the Tabor parameter (Eqs. 6-7), targets are the FMS-resampled BEM force histories, and the model parameters are learned on an 80% training split and evaluated on held-out 10% test splits. No fitted parameter is renamed as a prediction: pull-off force, pull-off time, and hysteresis are extracted a posteriori from the predicted trajectory and compared against BEM references (Table 2, Fig. 7). The FMS representation (Section 2.3) is a resampling choice, not a target-defining fit; the paper itself acknowledges that the fixed N=120 grid may miss sharply varying snap-off events (Section 4: remaining errors concentrate in 'sharply varying trajectories'; Section 5 proposes 'increasing the FMS'; Appendix H shows FMS-resolution dependence). This is an aliasing/accuracy limitation, not circularity. The analytical verification in Figure 5(c) uses an external PB theory curve and self-cited numerical data from Ref. [19], including the empirical coefficient alpha from that paper, but the M1-concat model was not trained on those points and the comparison is an out-of-distribution check of an abstract quantity (Gamma_eff). Self-citations to Refs. [19,29,45] validate the BEM solver and provide benchmark data, but the surrogate's predictions are not forced by those citations; the physical equations in Section 2.1 and Appendix E are stated in the paper. Overall, there is no step where an output is equivalent by construction to an input, and no load-bearing self-citation chain; the score reflects only minor self-citations that are not load-bearing.

Axiom & Free-Parameter Ledger

3 free parameters · 6 axioms · 0 invented entities

The paper contributes a learned surrogate, not new physics. Its central claim rests on the BEM data generator, the choice of constitutive/interface models, the hand-chosen FMS grid, and validation against self-cited analytical data. No new physical entities are postulated.

free parameters (3)
  • FMS resolution (NL,ND,NU)=(50,20,50) = N=120
    Hand-chosen in Section 2.3. Appendix H shows accuracy is resolution-dependent, so this choice directly affects reported error levels.
  • Network and training hyperparameters = LSTM 256->128; lr=1e-4; batch=8; early stopping patience=30
    Selected by validation during tuning; detailed tuning results are not reported (Section 2.6).
  • Empirical stress-calibration coefficient alpha = pi/9
    Taken from Ref. [19] to relate LJ tensile stress to the PB critical stress (Section 3.2). Used only in the analytical verification, not in training.
axioms (6)
  • domain assumption Standard Linear Solid with single relaxation time and k=0.1 captures the relevant viscoelastic adhesion behavior
    Section 2.1 restricts the study to this constitutive model; the paper explicitly disclaims extrapolation to other rheologies.
  • domain assumption Lennard-Jones traction-separation law governs the interface
    Eq. (1) is the standard interface model used throughout the BEM data generation.
  • domain assumption BEM discretisation with overlapping triangles and Newton-Raphson yields accurate ground-truth trajectories
    Section 2.1 and Appendix E; validated only in prior references, no code or data are shipped with this paper.
  • domain assumption Random 80/10/10 split with a fixed seed produces a meaningful held-out generalization test
    Section 2.4. Because parameters are sampled continuously, held-out test cases are mostly interpolations rather than extrapolations.
  • ad hoc to paper FMS spline resampling preserves the physical-time information needed for accurate force prediction
    Section 2.3. The choice of 50/20/50 points is made by the authors and is not derived from the physics.
  • domain assumption Persson-Brener theory and the alpha mapping apply in the relaxed short-range adhesion limit
    Section 3.2. Used for an external analytical check; relies on data and calibration from the authors' own Ref. [19].

pith-pipeline@v1.3.0-alltime-deepseek · 28302 in / 10938 out tokens · 114831 ms · 2026-08-01T13:32:42.789311+00:00 · methodology

0 comments
read the original abstract

Fast prediction of the response of adhesive soft viscoelastic contacts represents a current challenge in soft robotics and for gripping and manipulation tasks. Determining the complete time-resolved force trajectory requires full numerical simulations, whose computational cost is strongly parameter-dependent, making them impractical for real-time application or design-optimization loops. In this work, we overcome this limitation by training a scalar-conditioned, stateful, sequence-to-sequence deep learning model to predict the full force evolution from a prescribed displacement history for both short- and long-range adhesion regimes. The data set spans four orders of magnitude in loading and unloading rates and includes varied dwell times, with the Tabor parameter ranging from $0.2$ to $3.2$. To enable learning across these heterogeneous time scales, we introduce a fixed-measurement-step (FMS) representation that converts variable-length trajectories into fixed-length sequences while preserving their physical-time information. Different architectures were trained, including long short-term memory (LSTM) networks, temporal convolutional neural (TCN) networks, and time-distributed dense layers with three different Tabor-conditioning mechanisms. The models were compared using global waveform and error metrics. We found that the best-performing model has an LSTM architecture with concatenated conditioning, which achieves a held-out mean-squared error of $5.0\times10^{-4}$, a median pull-off-force error of $\approx2.2\%$, and a median hysteresis error of $\approx1.1\%$. For the held-out protocols, the model predicts a complete force trajectory with a median inference time of $0.16$ s. The model is tested across unseen parameter combinations and against analytical limiting cases, providing a rapid surrogate for repeated numerical evaluations with potential use in control-oriented applications.

Figures

Figures reproduced from arXiv: 2607.19060 by Ali Maghami, Antonio Papangelo, Merten Stender, Michele Ciavarella.

Figure 1
Figure 1. Figure 1: Schematic of indentation of a viscoelastic substrate by a rigid spherical indenter: (a) the sphere, substrate [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Representative force-time responses and empirical statistics for the simulated dataset. Panels (a)–(f) [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Fixed-measurement-step (FMS) mapping from the native BEM time series to a fixed-length sequence [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Schematic overview of the neural-network components and representative sequence-to-sequence archi [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Model selection, physical validation, and data-efficiency diagnostics for the M1-concat surrogate. [PITH_FULL_IMAGE:figures/full_fig_p014_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Force–indentation response of the M1-concat surrogate for the twelve representative unseen protocols [PITH_FULL_IMAGE:figures/full_fig_p017_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Distribution and cross-correlation of physically interpretable prediction errors over the held-out test [PITH_FULL_IMAGE:figures/full_fig_p020_7.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

87 extracted references · 28 canonical work pages · 1 internal anchor

  1. [1]

    W. Ji, Z. Qian, B. Xu, G. Chen, D. Zhao, Apple viscoelastic complex model for bruise damage analysis in constant velocity grasping by gripper, Computers and Electronics in Agriculture 162 (2019) 907–920. doi:10.1016/j.compag.2019.05.022

    work page doi:10.1016/j.compag.2019.05.022 2019

  2. [2]

    Z. Liu, F. Yan, Switchable adhesion: on-demand bonding and debonding, Advanced Science 9 (12) (2022) 2200264. doi:10.1002/advs.202200264

    work page doi:10.1002/advs.202200264 2022

  3. [3]

    AutoPeel: Adhesion-aware Safe Peeling Trajectory Optimization for Robotic Wound Care

    X. Liang, Y. Zhang, F. Liu, F. Richter, M. Yip, Autopeel: Adhesion-aware safe peeling trajectory optimization for robotic wound care, arXiv preprint arXiv:2409.14282 (2024). doi:10.48550/arXiv.2409.14282

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2409.14282 2024

  4. [4]

    Ziemer, M

    L. Ziemer, M. Garzke, A. Hand, M. Ben-Larbi, E. Stoll, A. Tighe, Impact of single and combined space environment factors on the performance of elastomer micropatterned dry adhesives, Polymer Testing 149 (2025) 108866

    2025

  5. [5]

    Giordano, R

    G. Giordano, R. B. N. Scharff, M. Carlotti, M. Gagliardi, C. Filippeschi, A. Mondini, A. Pa- pangelo, B. Mazzolai, Mechanochromic suction cups for local stress detection in soft robotics, Advanced Intelligent Systems (2024) 2400254doi:10.1002/aisy.202400254

    work page doi:10.1002/aisy.202400254 2024

  6. [6]

    B. Tao, Z. Gong, H. Ding, Climbing robots for manufacturing, National Science Review 10 (5) (2023) nwad042. doi:10.1093/nsr/nwad042

    work page doi:10.1093/nsr/nwad042 2023

  7. [7]

    K. D. Humfeld, D. Gu, G. A. Butler, K. Nelson, N. Zobeiry, A machine learning frame- work for real-time inverse modeling and multi-objective process optimization of compos- ites for active manufacturing control, Composites Part B: Engineering 223 (2021) 109150. doi:10.1016/j.compositesb.2021.109150

    arXiv 2021

  8. [8]

    H. Guo, Y. Lan, Z. Gao, C. Zhang, L. Zhang, X. Li, J. Lin, A. Elsheikh, W. Chen, Interaction between eye movements and adhesion of extraocular muscles, Acta Biomaterialia 176 (2024) 304–320. doi:10.1016/j.actbio.2024.01.028

    work page doi:10.1016/j.actbio.2024.01.028 2024

  9. [9]

    K. L. Johnson, K. Kendall, A. Roberts, Surface energy and the contact of elastic solids, Proceedings of the royal society of London. A. mathematical and physical sciences 324 (1558) (1971) 301–313. doi:10.1098/rspa.1971.0141

    arXiv 1971

  10. [10]

    B. V. Derjaguin, V. M. Muller, Y. P. Toporov, Effect of contact deformations on the adhesion of particles, Journal of Colloid and interface science 53 (2) (1975) 314–326. doi:10.1016/0021- 9797(75)90018-1

    doi:10.1016/0021- 1975

  11. [11]

    Papangelo, M

    A. Papangelo, M. Tricarico, A. Maghami, Adhesive single and multi-asperity contacts, in: Reference Module in Materials Science and Materials Engineering, Elsevier, 2026. doi:10.1016/B978-0-443-30138-4.00009-1

    work page doi:10.1016/b978-0-443-30138-4.00009-1 2026

  12. [12]

    Tabor, Surface forces and surface interactions, Journal of Colloid and Interface Science 58 (1) (1977) 2–13

    D. Tabor, Surface forces and surface interactions, Journal of Colloid and Interface Science 58 (1) (1977) 2–13. doi:10.1016/0021-9797(77)90366-6. 21

    work page doi:10.1016/0021-9797(77)90366-6 1977

  13. [13]

    Maugis, Adhesion of spheres: the jkr-dmt transition using a dugdale model, Journal of colloid and interface science 150 (1) (1992) 243–269

    D. Maugis, Adhesion of spheres: the jkr-dmt transition using a dugdale model, Journal of colloid and interface science 150 (1) (1992) 243–269. doi:10.1016/0021-9797(92)90285-t

    doi:10.1016/0021-9797(92)90285-t 1992

  14. [14]

    Violano, A

    G. Violano, A. Chateauminois, L. Afferrante, A jkr-like solution for viscoelastic adhesive con- tacts, Frontiers in Mechanical Engineering 7 (2021) 664486. doi:10.3389/fmech.2021.664486

    arXiv 2021

  15. [15]

    Tricarico, M

    M. Tricarico, M. Ciavarella, A. Papangelo, Enhancement of adhesion strength through mi- crovibrations: Modeling and experiments, Journal of the Mechanics and Physics of Solids 196 (2025) 106020. doi:10.1016/j.jmps.2024.106020

    arXiv 2025

  16. [16]

    Tricarico, A

    M. Tricarico, A. Shiferaw, A. Papangelo, Influence of material and geometrical parame- ters on the adhesive performance of vibration-modulated soft contacts, European Journal of Mechanics-A/Solids (2026) 106130doi:10.1016/j.euromechsol.2026.106130

    arXiv 2026

  17. [17]

    Ciavarella, M

    M. Ciavarella, M. Tricarico, A. Papangelo, On the dynamic jkr adhesion problem, Mechanics of Materials 202 (2025) 105252. doi:10.1016/j.mechmat.2025.105252

    arXiv 2025

  18. [18]

    Maghami, M

    A. Maghami, M. Tricarico, M. Ciavarella, A. Papangelo, Viscoelastic amplification of the pull-off stress in the detachment of a rigid flat punch from an adhesive soft viscoelastic layer, Engineering Fracture Mechanics 298 (2024) 109898. doi:10.1016/j.engfracmech.2024.109898

    arXiv 2024

  19. [19]

    Maghami, Q

    A. Maghami, Q. Wang, M. Tricarico, M. Ciavarella, Q. Li, A. Papangelo, Bulk and frac- ture process zone contribution to the rate-dependent adhesion amplification in viscoelastic broad-band materials, Journal of the Mechanics and Physics of Solids 193 (2024) 105844. doi:10.1016/j.jmps.2024.105844

    arXiv 2024

  20. [20]

    Afferrante, G

    L. Afferrante, G. Violano, On the effective surface energy in viscoelastic hertzian contacts, Journal of the Mechanics and Physics of Solids 158 (2022) 104669. doi:10.1016/j.jmps.2021.104669

    arXiv 2022

  21. [21]

    Carbone, C

    G. Carbone, C. Mandriota, N. Menga, Theory of viscoelastic adhesion and friction, Extreme Mechanics Letters 56 (2022) 101877. doi:10.1016/j.eml.2022.101877

    arXiv 2022

  22. [22]

    Mandriota, N

    C. Mandriota, N. Menga, G. Carbone, Adhesive contact mechanics of viscoelas- tic materials, International Journal of Solids and Structures 290 (2024) 112685. doi:10.1016/j.ijsolstr.2024.112685

    arXiv 2024

  23. [23]

    L. Shui, L. Jia, H. Li, J. Guo, Z. Guo, Y. Liu, Z. Liu, X. Chen, Rapid and continuous regulating adhesion strength by mechanical micro-vibration, Nature communications 11 (1) (2020) 1583. doi:10.1038/s41467-020-15447-x

    work page doi:10.1038/s41467-020-15447-x 2020

  24. [24]

    R. A. Schapery, A theory of crack initiation and growth in viscoelastic media: I. theoretical development, International Journal of fracture 11 (1975) 141–159. doi:10.1007/BF00034721

    work page doi:10.1007/bf00034721 1975

  25. [25]

    Greenwood, K

    J. Greenwood, K. Johnson, The mechanics of adhesion of viscoelastic solids, Philosophical Magazine A 43 (3) (1981) 697–711. doi:10.1080/01418618108240402

    work page doi:10.1080/01418618108240402 1981

  26. [26]

    Persson, E

    B. Persson, E. Brener, Crack propagation in viscoelastic solids, Physical Review E 71 (3) (2005) 036123. doi:10.1103/physreve.71.036123

    doi:10.1103/physreve.71.036123 2005

  27. [27]

    Greenwood, The theory of viscoelastic crack propagation and healing, Journal of Physics D: Applied Physics 37 (18) (2004) 2557

    J. Greenwood, The theory of viscoelastic crack propagation and healing, Journal of Physics D: Applied Physics 37 (18) (2004) 2557. doi:10.1088/0022-3727/37/18/011. 22

    doi:10.1088/0022-3727/37/18/011 2004

  28. [28]

    V. L. Popov, A note by kl johnson on the history of the jkr theory, Tribology Letters 69 (4) (2021) 132. doi:10.1007/s11249-021-01511-0

    work page doi:10.1007/s11249-021-01511-0 2021

  29. [29]

    Papangelo, M

    A. Papangelo, M. Ciavarella, Detachment of a rigid flat punch from a viscoelastic material, Tribology letters 71 (2) (2023) 48. doi:10.1007/s11249-023-01720-9

    work page doi:10.1007/s11249-023-01720-9 2023

  30. [30]

    M. S. Ahmad-Abad, A. Maghami, M. Ghalishooyan, A. Shooshtari, A family of min- imum residual displacement methods as nonlinear solution schemes for equilibrium path-following in structural mechanics, Computers & Structures 300 (2024) 107407. doi:10.1016/j.compstruc.2024.107407

    arXiv 2024

  31. [31]

    Papangelo, M

    A. Papangelo, M. Ciavarella, A numerical study on roughness-induced adhesion enhancement in a sphere with an axisymmetric sinusoidal waviness using lennard–jones interaction law, Lubricants 8 (9) (2020) 90. doi:10.3390/lubricants8090090

    work page doi:10.3390/lubricants8090090 2020

  32. [32]

    F. V. Souza, D. H. Allen, Multiscale modeling of impact on heterogeneous viscoelastic solids containingevolvingmicrocracks, InternationalJournalforNumericalMethodsinEngineering 82 (4) (2010) 464–504. doi:10.1002/nme.2773

    work page doi:10.1002/nme.2773 2010

  33. [33]

    K. Guo, Z. Yang, C.-H. Yu, M. J. Buehler, Artificial intelligence and machine learn- ing in design of mechanical materials, Materials Horizons 8 (4) (2021) 1153–1172. doi:10.1039/d0mh01451f

    doi:10.1039/d0mh01451f 2021

  34. [34]

    C. T. Mackay, D. Nowell, Informed machine learning methods for application in engineer- ing: A review, Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science 237 (24) (2023) 5801–5818. doi:10.1177/09544062231164575

    work page doi:10.1177/09544062231164575 2023

  35. [35]

    Y. Lu, X. Hu, Y. Pu, Z. Yu, N. An, S. Tang, X. Guo, A llm-inspired experimental-data- drivenframeworkforviscoelasticsoftstructures, InternationalJournalofMechanicalSciences (2026) 111696doi:10.1016/j.ijmecsci.2026.111696

    arXiv 2026

  36. [36]

    Javadi, A

    S. Javadi, A. Maghami, S. M. Hosseini, A deep learning approach based on a data- driven tool for classification and prediction of thermoelastic wave’s band structures for phononic crystals, Mechanics of Advanced Materials and Structures 29 (27) (2022) 6612–

    2022

  37. [37]

    Kellner, M

    L. Kellner, M. Stender, H. Herrnring, S. Ehlers, N. Hoffmann, K. V. Høyland, et al., Es- tablishing a common database of ice experiments and using machine learning to under- stand and predict ice behavior, Cold regions science and technology 162 (2019) 56–73. doi:10.1016/j.coldregions.2019.02.007

    work page doi:10.1016/j.coldregions.2019.02.007 2019

  38. [38]

    Eshkofti, S

    K. Eshkofti, S. M. Hosseini, The modified physics-informed neural network (pinn) method for the thermoelastic wave propagation analysis based on the moore-gibson- thompson theory in porous materials, Composite Structures 348 (2024) 118485. doi:10.1016/j.compstruct.2024.118485

    arXiv 2024

  39. [39]

    C. Yan, X. Feng, C. Wick, A. Peters, G. Li, Machine learning assisted discovery of new thermoset shape memory polymers based on a small training dataset, Polymer 214 (2021) 123351. doi:10.1016/j.polymer.2020.123351

    arXiv 2021

  40. [40]

    Y.-T. Wang, X. Zhang, X.-S. Liu, Machine learning approaches to rock fracture mechanics problems: Mode-i fracture toughness determination, Engineering Fracture Mechanics 253 (2021) 107890. doi:10.1016/j.engfracmech.2021.107890. 23

    arXiv 2021

  41. [41]

    C. E. Athanasiou, X. Liu, B. Zhang, T. Cai, C. Ramirez, N. P. Padture, J. Lou, B. W. Sheldon, H. Gao, Integrated simulation, machine learning, and experimental approach to characterizing fracture instability in indentation pillar-splitting of materials, Journal of the Mechanics and Physics of Solids 170 (2023) 105092. doi:10.1016/j.jmps.2022.105092

    arXiv 2023

  42. [42]

    R. Yi, D. Georgiou, X. Liu, C. E. Athanasiou, Mechanics-informed, model-free symbolic regression framework for solving fracture problems, Journal of the Mechanics and Physics of Solids (2024) 105916doi:10.1016/j.jmps.2024.105916

    arXiv 2024

  43. [43]

    Perera, V

    R. Perera, V. Agrawal, A generalized machine learning framework for brittle crack problems using transfer learning and graph neural networks, Mechanics of Materials 181 (2023) 104639. doi:10.1016/j.mechmat.2023.104639

    arXiv 2023

  44. [44]

    X. Li, X. Zhang, W. Feng, Q. Wang, Machine learning-based prediction of fracture toughness andpathinthepresenceofmicro-defects, EngineeringFractureMechanics276(2022)108900. doi:10.1016/j.engfracmech.2022.108900

    arXiv 2022

  45. [45]

    Maghami, M

    A. Maghami, M. Stender, A. Papangelo, Pull-off force prediction in viscoelastic adhesive hertzian contact by physics augmented machine learning, International Journal of Solids and Structures 322 (2025) 113584. doi:10.1016/j.ijsolstr.2025.113584

    arXiv 2025

  46. [46]

    Goodbrake, S

    C. Goodbrake, S. Motiwale, M. S. Sacks, A neural network finite element method for contact mechanics, Computer Methods in Applied Mechanics and Engineering 419 (2024) 116671. doi:10.1016/j.cma.2023.116671

    arXiv 2024

  47. [47]

    Motiwale, W

    S. Motiwale, W. Zhang, R. Feldmeier, M. S. Sacks, A neural network finite element approach for high speed cardiac mechanics simulations, Computer Methods in Applied Mechanics and Engineering 427 (2024) 117060. doi:10.1016/j.cma.2024.117060

    arXiv 2024

  48. [48]

    Kalliorinne, R

    K. Kalliorinne, R. Larsson, F. Pérez-Ràfols, M. Liwicki, A. Almqvist, Artificial neural net- work architecture for prediction of contact mechanical response, Frontiers in Mechanical Engineering 6 (2021) 579825. doi:10.3389/fmech.2020.579825

    arXiv 2021

  49. [49]

    Sahin, M

    T. Sahin, M. von Danwitz, A. Popp, Solving forward and inverse problems of contact me- chanics using physics-informed neural networks, Advanced Modeling and Simulation in En- gineering Sciences 11 (1) (2024) 11. doi:10.1186/s40323-024-00265-3

    work page doi:10.1186/s40323-024-00265-3 2024

  50. [50]

    Sahin, D

    T. Sahin, D. Wolff, A. Popp, Physics-informed neural networks for solving contact problems in three dimensions, in: Advances and Challenges in Computational Mechanics, Springer, 2026, pp. 419–431. doi:10.1007/978-3-031-93213-7_33

    doi:10.1007/978-3-031-93213-7_33 2026

  51. [51]

    Stender, M

    M. Stender, M. Tiedemann, D. Spieler, D. Schoepflin, N. Hoffmann, S. Oberst, Deep learning for brake squeal: Brake noise detection, characterization and prediction, Mechanical Systems and Signal Processing 149 (2021) 107181. doi:10.1016/j.ymssp.2020.107181

    arXiv 2021

  52. [52]

    Geier, S

    C. Geier, S. Hamdi, T. Chancelier, P. Dufrénoy, N. Hoffmann, M. Stender, Machine learning- based state maps for complex dynamical systems: applications to friction-excited brake system vibrations, Nonlinear dynamics 111 (24) (2023) 22137–22151. doi:10.1007/s11071- 023-08739-6

    doi:10.1007/s11071- 2023

  53. [53]

    Sattari Baboukani, Z

    B. Sattari Baboukani, Z. Ye, K. G Reyes, P. C. Nalam, Prediction of nanoscale friction for two-dimensional materials using a machine learning approach, Tribology Letters 68 (2020) 1–14. doi:10.1007/s11249-020-01294-w. 24

    work page doi:10.1007/s11249-020-01294-w 2020

  54. [54]

    Y. Kim, C. Yang, Y. Kim, G. X. Gu, S. Ryu, Designing an adhesive pillar shape with deep learning-based optimization, ACS applied materials & interfaces 12 (21) (2020) 24458–24465. doi:10.1021/acsami.0c04123

    work page doi:10.1021/acsami.0c04123 2020

  55. [55]

    D. Son, V. Liimatainen, M. Sitti, Machine learning-based and experimentally validated op- timal adhesive fibril designs, Small 17 (39) (2021) 2102867. doi:10.1002/smll.202102867

    work page doi:10.1002/smll.202102867 2021

  56. [56]

    A. Luo, H. Zhang, K. T. Turner, Machine learning-based optimization of the design of composite pillars for dry adhesives, Extreme Mechanics Letters 54 (2022) 101695. doi:10.1016/j.eml.2022.101695

    arXiv 2022

  57. [57]

    Y. Kim, J. Yeo, K. Park, A. Destrée, Z. Qin, S. Ryu, Designing directional adhesive pillars using deep learning-based optimization, 3d printing, and testing, Mechanics of Materials 185 (2023) 104778. doi:10.1016/j.mechmat.2023.104778

    arXiv 2023

  58. [58]

    C. B. Dayan, D. Son, A. Aghakhani, Y. Wu, S. O. Demir, M. Sitti, Machine learning-based shear optimal adhesive microstructures with experimental validation, Small 20 (2) (2024) 2304437. doi:10.1002/smll.202304437

    work page doi:10.1002/smll.202304437 2024

  59. [59]

    Shojaeifard, M

    M. Shojaeifard, M. Ferraresso, A. Lucantonio, M. Bacca, Machine learning-based optimal design of fibrillar adhesives, Journal of the Royal Society Interface 22 (223) (2025) 20240636. doi:10.1098/rsif.2024.0636

    arXiv 2025

  60. [60]

    D. N. Nguyen, D. Kaminski, P. Subramanian, S. Bateman, T. C. Le, Machine learning for adhesion assessment and prediction, Journal of Adhesion Science and Technology (2026) 1–31doi:10.1080/01694243.2026.2687662

    arXiv 2026

  61. [61]

    Hochreiter, J

    S. Hochreiter, J. Schmidhuber, Long short-term memory, Neural computation 9 (8) (1997) 1735–1780. doi:10.1162/neco.1997.9.8.1735

    doi:10.1162/neco.1997.9.8.1735 1997

  62. [62]

    Chen, Recurrent neural networks (rnns) learn the constitutive law of viscoelasticity, Com- putational Mechanics 67 (3) (2021) 1009–1019

    G. Chen, Recurrent neural networks (rnns) learn the constitutive law of viscoelasticity, Com- putational Mechanics 67 (3) (2021) 1009–1019. doi:10.1007/s00466-021-01981-y

    doi:10.1007/s00466-021-01981-y 2021

  63. [63]

    Koeppe, F

    A. Koeppe, F. Bamer, M. Selzer, B. Nestler, B. Markert, Explainable artificial intelligence for mechanics: physics-explaining neural networks for constitutive models, Frontiers in Materials 8 (2022) 824958. doi:10.3389/fmats.2021.824958

    arXiv 2022

  64. [64]

    George Thuruthel, P

    T. George Thuruthel, P. Gardner, F. Iida, Closing the control loop with time-variant em- bedded soft sensors and recurrent neural networks, Soft Robotics 9 (6) (2022) 1167–1176. doi:10.1089/soro.2021.0012

    arXiv 2022

  65. [65]

    J. Wang, J. Shu, M. M. Alam, Z. Gao, Z. Li, R. K.-Y. Tong, Drift-aware feature learning based on autoencoder preprocessing for soft sensors, Advanced Intelligent Systems 6 (3) (2024) 2300486. doi:10.1002/aisy.202300486

    work page doi:10.1002/aisy.202300486 2024

  66. [66]

    J. Li, J. Luo, F. Zhang, W. Zhou, X. Wei, C. Liao, M. Shou, Modeling of magnetorheo- logical dampers based on a dual-flow neural network with efficient channel attention, Smart Materials and Structures 32 (10) (2023) 105006. doi:10.1088/1361-665x/acf016

    work page doi:10.1088/1361-665x/acf016 2023

  67. [67]

    Karami, H

    M. Karami, H. Lombaert, D. Rivest-Hénault, Real-time simulation of viscoelastic tissue behavior with physics-guided deep learning, Computerized Medical Imaging and Graphics 104 (2023) 102165. doi:10.1016/j.compmedimag.2022.102165. 25

    arXiv 2023

  68. [68]

    Hinrichsen, C

    J. Hinrichsen, C. Ferlay, N. Reiter, S. Budday, Using dropout based active learning and surrogate models in the inverse viscoelastic parameter identification of human brain tissue, Frontiers in Physiology 15 (2024) 1321298. doi:10.3389/fphys.2024.1321298

    arXiv 2024

  69. [69]

    M. L. Williams, Structural analysis of viscoelastic materials, AIAA journal 2 (5) (1964) 785–808. doi:10.2514/3.2447

    work page doi:10.2514/3.2447 1964

  70. [70]

    Mainardi, G

    F. Mainardi, G. Spada, Creep, relaxation and viscosity properties for basic fractional mod- els in rheology, The European Physical Journal Special Topics 193 (1) (2011) 133–160. doi:10.1140/epjst/e2011-01387-1

    doi:10.1140/epjst/e2011-01387-1 2011

  71. [71]

    Ciavarella, J

    M. Ciavarella, J. Joe, A. Papangelo, J. Barber, The role of adhesion in contact mechanics, Journal of the Royal Society Interface 16 (151) (2019) 20180738. doi:10.1098/rsif.2018.0738

    arXiv 2019

  72. [72]

    Johnson, J

    K. Johnson, J. Greenwood, An adhesion map for the contact of elastic spheres, Journal of colloid and interface science 192 (2) (1997) 326–333. doi:10.1006/jcis.1997.4984

    arXiv 1997

  73. [73]

    Greenwood, Adhesion of elastic spheres, Proceedings of the Royal Society of London

    J. Greenwood, Adhesion of elastic spheres, Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences 453 (1961) (1997) 1277–1297. doi:10.1098/rspa.1997.0070

    arXiv 1961

  74. [74]

    Christensen, Theory of viscoelasticity: an introduction, Elsevier, 2012

    R. Christensen, Theory of viscoelasticity: an introduction, Elsevier, 2012

    2012

  75. [75]

    J. Q. Feng, Contact behavior of spherical elastic particles: a computational study of parti- cle adhesion and deformations, Colloids and Surfaces A: Physicochemical and Engineering Aspects 172 (1-3) (2000) 175–198. doi:10.1016/s0927-7757(00)00580-x

    work page doi:10.1016/s0927-7757(00)00580-x 2000

  76. [76]

    K. L. Johnson, Contact Mechanics, Cambridge University Press, Cambridge, 1987. doi:10.1017/cbo9781139171731

    doi:10.1017/cbo9781139171731 1987

  77. [77]

    J. Lin, E. Keogh, S. Lonardi, B. Chiu, A symbolic representation of time series, with implications for streaming algorithms, in: Proceedings of the 8th ACM SIGMOD work- shop on Research issues in data mining and knowledge discovery, 2003, pp. 2–11. doi:10.1145/882082.882086

    arXiv 2003

  78. [78]

    S. Bai, J. Z. Kolter, V. Koltun, An empirical evaluation of generic convolutional and recurrent networks for sequence modeling, arXiv preprint arXiv:1803.01271 (2018). doi:10.48550/arXiv.1803.01271

    Pith/arXiv arXiv doi:10.48550/arxiv.1803.01271 2018

  79. [79]

    D. P. Kingma, J. Ba, Adam: A method for stochastic optimization, arXiv preprint arXiv:1412.6980 (2014). doi:10.48550/arXiv.1412.6980

    Pith/arXiv arXiv doi:10.48550/arxiv.1412.6980 2014

  80. [80]

    bulk and fracture process zone contribution to the rate-dependent adhesion amplification in viscoelastic broad-band materials

    A. Papangelo, Data for “bulk and fracture process zone contribution to the rate-dependent adhesion amplification in viscoelastic broad-band materials” (2024). doi:10.5281/zenodo.13358696. URLhttps://doi.org/10.5281/zenodo.13358696

    work page doi:10.5281/zenodo.13358696 2024

Showing first 80 references.