Pith. sign in

REVIEW 3 major objections 6 minor 62 references

Specimen design for material parameter identification using topology optimization

T0 review · 3 major / 6 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read This paper shows that specimen geometry can be designed by maximizing the expected information gain of a mechanical test, and that such optimized specimens yield more accurate material parameter identification than conventional or arbitrary

desk verdict A genuinely new coupling of Bayesian D-optimal design with topology optimization for hyperelastic specimen design; the displacement-based results hold up, but the EIG approximation is unvalidated and the image-based route is dropped for the anisotropic case. read the letter →

arxiv 2607.16865 v1 pith:6BGTZTSB submitted 2026-07-18 cond-mat.mtrl-sci

Adeline Wihardja , Kaushik Bhattacharya This is my paper · ORCID
classification cond-mat.mtrl-sci MSC 74P1062K05
keywords BayesianoptimalexperimentaldesigntopologyoptimizationmaterialparameteridentificationhyperelasticityexpectedinformationgainD-optimalityadjointsensitivityanisotropy
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

In material testing, constitutive parameters are inferred from experiments, but conventional specimen shapes (dogbones, notches) sample few deformation states, and identifying anisotropic models requires many tests. This paper proposes designing the specimen geometry itself as a topology optimization problem, using the Bayesian expected information gain as the objective: the design that maximizes the predicted information from a uniaxial test is found by gradient-based optimization over a material density field. The method is demonstrated on isotropic and fiber-reinforced hyperelastic models, where the optimized specimens produce heterogeneous deformation fields that excite multiple fiber orientations, leading to higher expected information gain and lower parameter inference errors than arbitrary designs. If it works in practice, a single well-designed specimen could replace several conventional experiments.

What carries the argument

The expected information gain (EIG), approximated by the Laplace/Gauss-Newton D-optimal criterion 1/2 log det(I + C0 H(ρ)), with H the parameter-to-observable sensitivity (Hessian of the data misfit) evaluated at the prior mean. This criterion is optimized over a density field ρ via SIMP interpolation and adjoint-based gradient computation, coupling the design to the mechanical equilibrium equations.

What would settle it

Run the same OED for a hyperelastic model with a bimodal posterior (or with a prior mean far from the true parameters), computing the true EIG by nested Monte Carlo sampling for both the topology-optimized design and several conventional designs; if the optimized design's true EIG is not higher than the alternatives, the local D-optimal approximation is inadequate.

Watch

Extended reading notes

Core claim

The paper's central claim is that the locally D-optimal Bayesian design criterion—maximizing 1/2 log det(I + C0 H(ρ)), where H is the Gauss-Newton Hessian of the data misfit at the prior mean—can be used as the objective of a density-based topology optimization to generate specimen geometries that are maximally informative for constitutive parameter identification. Optimizing this criterion over a SIMP density field, with sensitivities computed by adjoint equations, produces designs with multiple load-bearing struts oriented in different directions for anisotropic materials, which generate heterogeneous deformation states in a single uniaxial test. The paper reports that these optimized desi

Load-bearing premise

The framework's objective replaces the true expected information gain with a local Gaussian/Laplace approximation evaluated at the prior mean; if the posterior is multimodal or the true parameters lie far from the prior mean, the optimized design may not be the most informative.

Editorial extensions

If this is right

  • Optimized specimens yield higher expected information gain and lower inference errors than conventional or arbitrary designs (Tables 2–3).
  • A single uniaxial test on an optimized specimen can recover anisotropic fiber angles and moduli with errors reduced relative to initial guesses or arbitrary designs.
  • Optimizing under the D-optimal criterion also improves the posterior mean-square error (A-optimal criterion) for the tested cases.
  • Designs are similar across ~15% prior variations, suggesting one-shot design is robust for these models.
  • The displacement-based OED route is more stable than image-based, pointing to a practical workflow for experimental implementation.

Reading between the lines

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

  • Since the D-optimal approximation is exact only for linear maps, its accuracy for strongly nonlinear hyperelastic responses with possible multimodal posteriors remains untested; a nested Monte Carlo validation of the true EIG for the optimized designs would be a useful check.
  • Because the Hessian is evaluated at the prior mean, the method implicitly trusts the prior; in a truly unknown regime, sequential design (updating the prior with each experiment) would be needed to correct a poor initial guess.
  • The observed design robustness across priors may break down for parameters with weakly coupled sensitivities; a parameter-wise information budget could reveal which parameters actually drive the optimized geometry.
  • The computational cost of several hours per design suggests that surrogate models or reduced-order methods would be needed to extend the approach to 3D specimens or history-dependent materials.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes a framework for designing specimen geometries that maximize information gain for identifying hyperelastic material parameters, coupling Bayesian optimal experimental design with density-based topology optimization. The design objective is the local Laplace/Gauss-Newton D-optimal criterion of Eq. (18), with the Hessian of the negative log-likelihood evaluated at the prior mean (Eqs. 15–16). Adjoint equations (A1–A2) are derived for the design gradient, and the method is demonstrated on isotropic and fiber-reinforced hyperelastic models using synthetic data. The authors report that optimized specimens yield higher expected information gain and more accurate parameter recovery than a set of arbitrary designs (Tables 2–3).

Significance. If the claims are supported, the framework is a worthwhile contribution to material parameter identification: it replaces intuition-based specimen selection with a systematic, PDE-constrained optimization over a high-dimensional design space, and the adjoint-based sensitivity derivation is a concrete technical contribution. The synthetic recovery tests, using reference parameters different from the prior, are an appropriate first validation step. However, the central claim rests on a Laplace approximation to the expected information gain that is not independently verified, and the statistical evidence is based on single noise realizations. These gaps need to be addressed before the stated conclusions can be fully accepted.

major comments (3)
  1. [§2.2, Eqs. (8), (15)–(16); §3.3, Eq. (18)] The design objective replaces the true EIG with a local Laplace/Gauss-Newton approximation in which the Hessian H is evaluated at the prior mean P0 (Eqs. 15–16). Table 2 then compares optimized and arbitrary designs using this same OEIG, so the statement that the optimized specimen has 'better expected information gain' is, at this stage, a statement about the proxy, not about the true information gain. For a nonlinear, possibly multimodal posterior, the proxy may not track the true EIG. I request an independent check for at least the fiber-reinforced case: estimate the true EIG (e.g., by nested Monte Carlo or by MCMC posterior sampling using a small number of map evaluations) for the optimized design and for one or two arbitrary designs, and report the comparison. This is load-bearing for the central claim.
  2. [§4.2, Table 3] The synthetic recovery experiments report percentage errors for a single noise realization at each setting. Because the data are noisy, a single draw may not be representative, and it is unclear whether the improvements over arbitrary designs (e.g., C1 error 4.4% vs. 4.4% for A2 in the last block) are statistically meaningful. Please repeat the experiment with multiple independent noise realizations (e.g., 10–20 seeds) and report means and standard deviations, or a paired comparison, for the optimized and the non-optimized designs. The claim that optimized designs 'lead to more accurate parameter inference' requires this type of evidence.
  3. [§4.2, image-based formulation] The text states that the image-based formulation 'often led to non-convergence' for anisotropic materials and that the anisotropic results are therefore confined to the displacement-based formulation. The image-based recovery results in Table 3 use displacement-optimized designs, not image-optimized ones. Consequently, the paper does not demonstrate the image-based branch of the framework for the main anisotropic target. Please either include a stabilized image-based design example or clearly delimit the conclusions to the displacement-based formulation. This is important because the introduction discusses image-based methods as a motivation, and the conclusion claims that 'direct image-based inference is valuable.'
minor comments (6)
  1. [Eq. (31)] The fiber direction vectors are written as a=(cos θ1, sin θ2, 0) and b=(cos θ2, sin θ2, 0). The second component of a should presumably be sin θ1; please correct.
  2. [Table 1 caption] The heading 'V arious priors' contains a typo; it should read 'Various priors.'
  3. [Algorithm 1] The stopping condition '∥OEIG,k+1 − OEIG,k1∥' appears to have a typo: the subscript 'k1' should likely be 'k'; also use consistent norm notation.
  4. [§4.2, text near Table 2] The sentence 'These are also shown in Figure 2' seems to refer to the anisotropic results of Figures 3–4 or Table 2, not Figure 2 (isotropic). Please update the cross-reference.
  5. [§3.2, Eq. (14)] The displacement-based Φd omits the image weighting ρ^{2p} present in Eq. (13); this may be intentional (displacements are defined on the body) but should be stated explicitly to avoid confusion.
  6. [§2.2] In Eq. (8), the notation H(P_MAP) is used, but in the design phase (Eqs. 15–16) H is evaluated at P0. The distinction between the two should be clarified in the text, since the approximation error is one of the main concerns.

Circularity Check

1 steps flagged · score 6.0 of 10

Headline EIG comparison is circular by construction; parameter-recovery claim retains independent support.

  1. fitted input called prediction [Section 3.3 Eq. (19); Section 4.2, Table 2 paragraph]
    "ρbest = arg min ρ −OEIG(ρ, u, uP) such that u satisfies (11) for ρ. ... We compare the performance of our optimized design with a number of arbitrary designs in Table 2. In each case, the optimized design has the highest expected information gain."

    The design ρbest is defined as the maximizer of OEIG in Eq. (19). Table 2 then evaluates the optimized design by comparing its OEIG value with those of arbitrary designs and reports that it is highest. This is true by construction, not as an empirical finding: the evaluation metric is exactly the objective being optimized. The reported OEIG increase from 9.5 to 11.3 is similarly a by-product of optimization convergence. The independent synthetic parameter-recovery results in Table 3 use a different metric (error of the MAP estimate against P_true) and are not circular, but they do not validate the specific claim that the optimized design has 'better expected information gain'.

full rationale

The framework is not generally circular: Eq. (18) defines an approximate D-optimal objective, Eq. (19) optimizes it, and the synthetic-data validation in Table 3 uses reference parameters P_true different from the prior P0 and measures MAP recovery error, so the inference-accuracy claim is independent of the optimized objective. However, the paper's headline comparison in Table 2—that the optimized design has the highest OEIG—is a restatement of the optimization problem: since ρbest is chosen to maximize OEIG, comparing OEIG(ρbest) with OEIG(arbitrary designs) is tautological. The same applies to the OEIG increase from initial to final designs in Section 4.2. The unvalidated Laplace/Gauss-Newton approximation in Eqs. (8) and (15)–(16), and the admitted non-convergence of the image-based formulation in Section 4.2, are substantive limitations but are not circularity. Self-citation [58] supports the image-based inference route, but displacement-based recovery provides an independent route in Table 3, so the self-citation is not load-bearing for the central claim. Overall, the 'more accurate parameter inference' claim has independent content, while the 'better expected information gain' claim reduces to the optimization objective by construction.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central claim rests on a small number of hand-chosen priors, noise covariances, algorithmic parameters, and the local Laplace approximation of the EIG. There are no invented physical entities. The model-form and noise assumptions are standard for this literature, but they are not validated against real experiments.

free parameters (6)
  • SIMP penalty exponent schedule = p: 1 → 3, +0.1 every 20 MMA iterations
    Algorithm parameter chosen by hand (standard topology optimization practice) to drive designs to solid/void; affects the optimized topology but not the physical model.
  • Prior means and covariance for material parameters = P0 = {C1=4, C4=8, C6=8, κ=13 MPa, θ1=π/4, θ2=−π/6}; σ=20% of mean
    User-specified Bayesian prior; the whole design objective (18) is evaluated at this prior, so results depend on it. Chosen to represent plausible soft-tissue values, not fitted to data.
  • Displacement noise covariance Γ_u = (1e-5)^2 I (constant) or (1e-6 t)^2 I (increasing)
    Hand-assumed Gaussian uncorrelated noise level for DIC displacement data; directly scales H and therefore the design. Real DIC noise is correlated, as the paper itself notes (Section 3.2).
  • Force noise covariance Γ_f = (0.025t)^2 (isotropic) / (0.01t)^2 (anisotropic)
    Hand-assumed load-cell noise that grows with time; scales the force term in H.
  • Image noise covariance Γ_i = 0.01^2 I
    Hand-assumed Gaussian noise in grayscale intensity for image-based formulation; used for isotropic design and for image-based recovery.
  • Volume fraction constraint V_f = 0.6 (edge constraint), 0.5 (grip constraint)
    Upper bound on material volume in each design problem, chosen by hand; affects feasible topologies and thus the optimized specimen.
assumptions (5)
  • domain assumption The chosen hyperelastic stored-energy forms (28) and (31) correctly describe the material over the deformation range tested.
    Sections 3.1 and 4. The entire parameter identification and design are conditional on this model form; in synthetic experiments it holds by construction, but for real materials it is an untested assumption.
  • domain assumption Additive Gaussian noise with known covariance and a Gaussian prior; no model error.
    Section 2.1 (Eq. 2) and Eqs. (30)-(32). The posterior covariance and EIG formulas require exact Gaussian noise/prior; DIC noise is acknowledged to be correlated, so this is an idealization.
  • domain assumption Local Bayesian D-optimal (Laplace/Gauss-Newton) approximation of EIG for nonlinear forward maps.
    Section 2.2: 'for G nonlinear, (8) is obtained by linearizing about the current estimate and using the Gauss-Newton approximation for H.' The design (19) is optimal only within this approximation; no check against full EIG is reported.
  • standard math Forward solutions u and u_P are continuous/differentiable with respect to the density field ρ, so adjoint sensitivities are valid.
    Section 3.4: 'assume that the solutions of the forward problems F1 and F2 are continuous with respect to ρ.' Needed for Eq. (25).
  • ad hoc to paper SIMP interpolation and filtering lead to meaningful solid-void designs that can be manufactured and tested.
    Section 3.5; SIMP is a relaxation with hand-chosen exponent schedule and filter radius; the final topology is an approximation of the original binary design problem.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Specimen design for material parameter identification using topology optimization." pith.science (2026). https://pith.science/paper/6BGTZTSB

@misc{pith2026260716865,
  author       = {Pith},
  title        = {Pith review of: Specimen design for material parameter identification using topology optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6BGTZTSB}},
  note         = {Machine review of arXiv:2607.16865}
}
read the original abstract

Constitutive relations close the equations of continuum mechanics, and serve as a surrogate for a material in the design and engineering process. They are often specified in a parameterized form with parameters identified by experiment. In this paper, we propose a framework for identifying experimental configurations that are maximally informative for constitutive model discovery. The framework strongly couples modeling and experimentation: the model leverages high-dimensional data from full-field measurements, while the current uncertainty in the model guides the design of future experiments. We formulate this goal by integrating Bayesian optimal experimental design with topology optimization. The Bayesian design criterion quantifies expected information gain, which drives the topology optimization of the specimen geometry.

Figures

Figures reproduced from arXiv: 2607.16865 by the authors.

Figure 1
Figure 1. (a) Reference domain of 0.036 × 0.012 m2 , with finite element mesh consisting of 96×32 elements. An example density assignment to initiate the optimization can be seen in the scale bar. (b) Representative of speckle pattern to perform image-based optimal designs (15 × 15 pixels within each mesh), with 4 × 4 pixels per speckle. domain to be fixed at ρ = 1 (solid), as shown in [PITH_FULL_IMAGE:figures/full_fig_p013_1.png] view at source ↗
Figure 2
Figure 2. Optimal design for isotropic material. (a) Initial design (density assignment) [PITH_FULL_IMAGE:figures/full_fig_p014_2.png] view at source ↗
Figure 3
Figure 3. Optimal design for fiber-reinforced hyperelastic model in the displacement formu [PITH_FULL_IMAGE:figures/full_fig_p016_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Optimal design for fiber-reinforced hyperelastic model in the displacement formu [PITH_FULL_IMAGE:figures/full_fig_p016_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

62 extracted references · 2 linked inside Pith

  1. [1]

    VIC2D.https://www.correlatedsolutions.com

  2. [2]

    A. Akerson. Optimal structures for failure resistance under impact.Journal of the Mechanics and Physics of Solids, 172:105172, 2023

    2023

  3. [3]

    Alexanderian

    A. Alexanderian. Optimal experimental design for infinite-dimensional bayesian inverse problems governed by pdes: A review.Inverse Problems, 37(4):043001, 2021

    2021

  4. [4]

    Alexanderian, P

    A. Alexanderian, P. J. Gloor, and O. Ghattas. On bayesian a-and d-optimal experi- mental designs in infinite dimensions. 2016

    2016

  5. [5]

    Alexanderian, N

    A. Alexanderian, N. Petra, G. Stadler, and O. Ghattas. A-optimal design of experi- ments for infinite-dimensional bayesian linear inverse problems with regularized\ell 0- sparsification.SIAM Journal on Scientific Computing, 36(5):A2122–A2148, 2014

    2014

  6. [6]

    Alexanderian, N

    A. Alexanderian, N. Petra, G. Stadler, and O. Ghattas. A fast and scalable method for a-optimal design of experiments for infinite-dimensional bayesian nonlinear inverse problems.SIAM Journal on Scientific Computing, 38(1):A243–A272, 2016

    2016

  7. [7]

    Andrade-Campos, J

    A. Andrade-Campos, J. Aquino, J. M. Martins, and B. Coelho. On the design of innovative heterogeneous sheet metal tests using a shape optimization approach.Metals, 9(3):371, 2019

    2019

  8. [8]

    A. N. Annaidh, K. Bruy` ere, M. Destrade, M. D. Gilchrist, and M. Ott´ enio. Character- ization of the anisotropic mechanical properties of excised human skin.Journal of the mechanical behavior of biomedical materials, 5(1):139–148, 2012

    2012

Show all 62 references
  1. [9]

    Arndt, W

    D. Arndt, W. Bangerth, M. Feder, M. Fehling, R. Gassm¨ oller, T. Heister, L. Heltai, M. Kronbichler, M. Maier, P. Munch, J.-P. Pelteret, S. Sticko, B. Turcksin, and D. Wells. Thedeal.IIlibrary, version 9.4.Journal of Numerical Mathematics, 30(3):231–246, 2022

    2022

  2. [10]

    Astruc, F

    L. Astruc, F. Turquier, M. Brieu, and T. Hoc. Microscopic characterization of the anisotropic mechanical behavior of human linea alba and anterior rectus sheath: con- tributions of collagen and elastin.Materials & Design, 259:114853, 2025

    2025

  3. [11]

    Atkinson, A

    A. Atkinson, A. Donev, and R. Tobias.Optimum experimental designs, with SAS, volume 34. OUP Oxford, 2007

    2007

  4. [12]

    Avril, M

    S. Avril, M. Bonnet, A.-S. Bretelle, M. Gr´ ediac, F. Hild, P. Ienny, F. Latourte, D. Lemosse, S. Pagano, E. Pagnacco, et al. Overview of identification methods of mechanical parameters based on full-field measurements.Experimental Mechanics, 48(4):381–402, 2008

    2008

  5. [13]

    H. Axer, D. G. v. Keyserlingk, and A. Prescher. Collagen fibers in linea alba and rectus sheaths: I. general scheme and morphological aspects.Journal of Surgical Research, 96(1):127–134, 2001. 21

    2001

  6. [14]

    Barroqueiro, A

    B. Barroqueiro, A. Andrade-Campos, J. Dias-de Oliveira, and R. Valente. Design of mechanical heterogeneous specimens using topology optimization.International Journal of Mechanical Sciences, 181:105764, 2020

    2020

  7. [15]

    M. P. Bendsøe. Optimal shape design as a material distribution problem.Structural optimization, 1(4):193–202, 1989

    1989

  8. [16]

    M. P. Bendsøe and O. Sigmund. Material interpolation schemes in topology optimiza- tion.Archive of applied mechanics, 69(9):635–654, 1999

    1999

  9. [17]

    M. P. Bendsoe and O. Sigmund.Topology optimization: theory, methods, and applica- tions. Springer Science & Business Media, 2013

    2013

  10. [18]

    Bhattacharya, L

    K. Bhattacharya, L. Cao, and A. Stuart. Optimal experimental design for reliable learning of history-dependent constitutive laws.arXiv preprint arXiv:2603.12365, 2026

    2026 arXiv

  11. [19]

    Bonnet and A

    M. Bonnet and A. Constantinescu. Inverse problems in elasticity.Inverse problems, 21(2):R1–R50, 2005

    2005

  12. [20]

    Bornert, F

    M. Bornert, F. Br´ emand, P. Doumalin, J.-C. Dupr´ e, M. Fazzini, M. Gr´ ediac, F. Hild, S. Mistou, J. Molimard, J.-J. Orteu, et al. Assessment of digital image correlation measurement errors: methodology and results.Experimental mechanics, 49(3):353–370, 2009

    2009

  13. [21]

    B. Bourdin. Filters in topology optimization.International journal for numerical meth- ods in engineering, 50(9):2143–2158, 2001

    2001

  14. [22]

    Bui-Thanh, O

    T. Bui-Thanh, O. Ghattas, J. Martin, and G. Stadler. A computational framework for infinite-dimensional bayesian inverse problems part i: The linearized case, with applica- tion to global seismic inversion.SIAM Journal on Scientific Computing, 35(6):A2494– A2523, 2013

    2013

  15. [23]

    Chaloner and I

    K. Chaloner and I. Verdinelli. Bayesian experimental design: A review.Statistical science, pages 273–304, 1995

    1995

  16. [24]

    B. Chen, B. Starman, M. Haliloviˇ c, L. A. Berglund, and S. Coppieters. Finite element model updating for material model calibration: A review and guide to practice.Archives of Computational Methods in Engineering, 32(4):2035–2112, 2025

  17. [25]

    T. Chu, J. B. Estrada, and S. H. Bryngelson. Bayesian optimal design accelerates discovery of soft material properties from bubble dynamics.Computational Mechanics, 76(2):431–447, 2025

    2025

  18. [26]

    S. L. Cotter, M. Dashti, J. C. Robinson, and A. M. Stuart. Bayesian inverse problems for functions and applications to fluid mechanics.Inverse problems, 25(11):115008, 2009

    2009

  19. [27]

    Dashti and A

    M. Dashti and A. M. Stuart. The bayesian approach to inverse problems.arXiv preprint arXiv:1302.6989, 2013. 22

    2013 arXiv

  20. [28]

    Etling and R

    T. Etling and R. Herzog. Optimum experimental design by shape optimization of specimens in linear elasticity.SIAM Journal on Applied Mathematics, 78(3):1553–1576, 2018

    2018

  21. [29]

    M. E. Gurtin, E. Fried, and L. Anand.The mechanics and thermodynamics of continua. Cambridge university press, 2010

    2010

  22. [30]

    Haber, L

    E. Haber, L. Horesh, and L. Tenorio. Numerical methods for experimental design of large-scale linear ill-posed inverse problems.Inverse Problems, 24(5):055012, 2008

    2008

  23. [31]

    W. K. Hastings. Monte carlo sampling methods using markov chains and their appli- cations. 1970

    1970

  24. [32]

    Hild and S

    F. Hild and S. Roux. Digital image correlation: from displacement measurement to identification of elastic properties: A review.Strain, 42:69–80, 2006

    2006

  25. [33]

    A. G. Holzapfel. Nonlinear solid mechanics ii. 2000

    2000

  26. [34]

    G. A. Holzapfel and R. W. Ogden. Constitutive modelling of arteries.Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 466(2118):1551– 1597, 2010

    2010

  27. [35]

    Huan and Y

    X. Huan and Y. M. Marzouk. Simulation-based optimal bayesian experimental design for nonlinear systems.Journal of Computational Physics, 232(1):288–317, 2013

    2013

  28. [36]

    Joshi, P

    A. Joshi, P. Thakolkaran, Y. Zheng, M. Escande, M. Flaschel, L. De Lorenzis, and S. Kumar. Bayesian-euclid: Discovering hyperelastic material laws with uncertainties. Computer Methods in Applied Mechanics and Engineering, 398:115225, 2022

    2022

  29. [37]

    M. P. Juniper and M. Yoko. Generating a physics-based quantitatively-accurate model of an electrically-heated rijke tube with bayesian inference.Journal of Sound and Vi- bration, 535:117096, 2022

    2022

  30. [38]

    Kawabata, M

    S. Kawabata, M. Matsuda, K. Tei, and H. Kawai. Experimental survey of the strain energy density function of isoprene rubber vulcanizate.Macromolecules, 14(1):154–162, 1981

    1981

  31. [39]

    Kullback and R

    S. Kullback and R. A. Leibler. On information and sufficiency.The annals of mathe- matical statistics, 22(1):79–86, 1951

    1951

  32. [40]

    Leclerc, J.-N

    H. Leclerc, J.-N. P´ eri´ e, S. Roux, and F. Hild. Integrated digital image correlation for the identification of mechanical properties. InInternational conference on com- puter vision/computer graphics collaboration techniques and applications, pages 161–

  33. [41]

    D. V. Lindley. On a measure of the information provided by an experiment.The Annals of Mathematical Statistics, 27(4):986–1005, 1956. 23

    1956

  34. [42]

    Q. Long, M. Scavino, R. Tempone, and S. Wang. Fast estimation of expected informa- tion gains for bayesian experimental designs based on laplace approximations.Computer Methods in Applied Mechanics and Engineering, 259:24–39, 2013

    2013

  35. [43]

    N ´ ı Annaidh, K

    A. N ´ ı Annaidh, K. Bruy` ere, M. Destrade, M. D. Gilchrist, C. Maurini, M. Ott´ enio, and G. Saccomandi. Automated estimation of collagen fibre dispersion in the dermis and its contribution to the anisotropic behaviour of skin.Annals of biomedical engineering, 40(8):1666–1678, 2012

    2012

  36. [44]

    Pierron and M

    F. Pierron and M. Gr´ ediac.The virtual fields method: extracting constitutive mechani- cal parameters from full-field deformation measurements. Springer Science & Business Media, 2012

    2012

  37. [45]

    Pierron and M

    F. Pierron and M. Gr´ ediac. Towards material testing 2.0. a review of test design for iden- tification of constitutive parameters from full-field measurements.Strain, 57(1):e12370, 2021

    2021

  38. [46]

    Pukelsheim.Optimal design of experiments

    F. Pukelsheim.Optimal design of experiments. SIAM, 2006

    2006

  39. [47]

    D. E. Ricciardi, D. T. Seidl, B. T. Lester, A. R. Jones, and E. Jones. Bayesian op- timal experimental design for constitutive model calibration.International Journal of Mechanical Sciences, 265:108881, 2024

    2024

  40. [48]

    R. S. Rivlin and D. Saunders. Large elastic deformations of isotropic materials vii. experiments on the deformation of rubber.Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences, 243(865):251–288, 1951

    1951

  41. [49]

    Sigmund and J

    O. Sigmund and J. Søndergaard Jensen. Systematic design of phononic band–gap materials and structures by topology optimization.Philosophical Transactions of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences, 361(1806):1001–1019, 2003

    2003

  42. [50]

    A. F. Smith and G. O. Roberts. Bayesian computation via the gibbs sampler and related markov chain monte carlo methods.Journal of the Royal Statistical Society: Series B (Methodological), 55(1):3–23, 1993

    1993

  43. [51]

    Souto, A

    N. Souto, A. Andrade-Campos, and S. Thuillier. A numerical methodology to design heterogeneous mechanical tests.International Journal of Mechanical Sciences, 107:264– 276, 2016

    2016

  44. [52]

    A. M. Stuart. Inverse problems: a bayesian perspective.Acta numerica, 19:451–559, 2010

    2010

  45. [53]

    M. A. Sutton, J. J. Orteu, and H. Schreier.Image correlation for shape, motion and deformation measurements: Basic concepts, theory and applications. Springer Science & Business Media, 2009. 24

    2009

  46. [54]

    Svanberg

    K. Svanberg. The method of moving asymptotes—a new method for structural opti- mization.International journal for numerical methods in engineering, 24(2):359–373, 1987

    1987

  47. [55]

    L. R. Treloar. Stress-strain data for vulcanised rubber under various types of deforma- tion.Transactions of the Faraday Society, 40:59–70, 1944

    1944

  48. [56]

    Ucinski.Optimal measurement methods for distributed parameter system identifica- tion

    D. Ucinski.Optimal measurement methods for distributed parameter system identifica- tion. CRC press, 2004

    2004

  49. [57]

    Vigliotti, G

    A. Vigliotti, G. Cs´ anyi, and V. Deshpande. Bayesian inference of the spatial distribu- tions of material properties.Journal of the Mechanics and Physics of Solids, 118:74–97, 2018

    2018

  50. [58]

    Wihardja and K

    A. Wihardja and K. Bhattacharya. Constitutive relations from images.Journal of Applied Mechanics, 92(8):081009, 2025

    2025

  51. [59]

    A. D. Woodbury and T. J. Ulrych. A full-bayesian approach to the groundwater inverse problem for steady state flow.Water resources research, 36(8):2081–2093, 2000

  52. [60]

    K. Wu, P. Chen, and O. Ghattas. A fast and scalable computational framework for large- scale high-dimensional bayesian optimal experimental design.SIAM/ASA Journal on Uncertainty Quantification, 11(1):235–261, 2023

    2023

  53. [61]

    J. Yang, L. Hazlett, A. Landauer, and C. Franck. Augmented Lagrangian Digital Volume Correlation (ALDVC).Experimental Mechanics, 60:1205–1223, 2020

    2020

  54. [62]

    Yoko and M

    M. Yoko and M. P. Juniper. Optimal experiment design with adjoint-accelerated bayesian inference.Data-Centric Engineering, 5:e17, 2024. Appendix A Derivation for sensitivity and adjoint PDEs We seek to find the gradient ofOEIG with respect toρ. The details of this gradient cal...

    2024

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.