Pith. sign in

REVIEW 4 major objections 5 minor 54 references

Proper Orthogonal Decomposition-based Model-Order Reduction for Smoothed Particle Hydrodynamics Simulation

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Compressing smoothed particle hydrodynamics with proper orthogonal decomposition beats uniformly thinning the particle count at equal degrees of freedom, and a frozen-coefficient linearization makes it fast.

desk verdict First systematic POD-MOR for SPH with honest limits; the headline 'same DoF' claim is weakened by in-sample training and a loose DoF metric. read the letter →

arxiv 2507.19825 v1 pith:P7Q5AQ6B submitted 2025-07-26 physics.comp-ph

Lidong Fang , Zilong Song , Kirk Fraser , Faisal Habib , Christopher Drummond , Huaxiong Huang This is my paper · ORCID
classification physics.comp-ph MSC 65F2576-1076N0680A1935Q7070-0837M05 PACS 02.70.-c46.15.-x47.11.-j47.85.Dh
keywords smoothedparticlehydrodynamicsmodel-orderreductionproperorthogonaldecompositionLagrangianframeworkfrictionstirspotweldingPODerrorfrozencoefficientsGalerkinprojection
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

This paper tries to establish that proper orthogonal decomposition (POD), the standard tool for model-order reduction of grid-based simulations, also works when the simulation is a smoothed particle hydrodynamics (SPH) run in the Lagrangian frame, where discrete particles carry the solution and move with the flow. Using a two-dimensional friction stir spot welding (FSSW) model that couples the flow and heat equations, the authors claim that reducing the model with POD beats the naive way of reducing it — thinning the particle count — at equal degrees of freedom, with substantially smaller errors. They further claim the reduced model can be accelerated by linearizing the equations and freezing coefficients for several time steps at a time, with only modest added error, and that the reduction carries over to predictive settings with different tool rotation speeds.

What carries the argument

The carrying mechanism is the POD basis — a low-dimensional orthonormal subspace extracted by truncated singular value decomposition of a snapshot matrix of the Lagrangian particle fields (position, velocity, temperature, density) — followed by Galerkin projection of the SPH equations onto that basis. Because the SPH operators are nonlinear and couple different fields, the projection alone does not cut computational cost; the paper's acceleration device is a linearization that freezes every term except the reduced coordinates of one field into a precomputable matrix, so the right-hand side becomes a matrix-vector product, with the frozen entries refreshed every few time steps.

What would settle it

Rerun the Grid-C versus Grid-B comparison at SPH parameters that produce irregular particle motion, for example $\lambda = \zeta = 0$, which the paper shows destroys temporal self-convergence, and test whether the POD-MOR velocity error stays below the uniform-coarsening error or the reduced velocity field diverges; a single such run where the reduced model blows up would falsify the claim that POD-MOR beats uniform particle thinning at equal degrees of freedom.

Watch

Extended reading notes

Core claim

The central claim is that a projection-based reduced model built from POD modes of SPH snapshots reproduces the full SPH solution with much less error than a uniformly coarsened SPH simulation using the same number of degrees of freedom. For the coupled FSSW model, retaining about 40% of the POD modes for position, velocity, temperature, and density keeps total relative error near one percent, and 10% of the modes keeps it near ten percent, while the same degree-of-freedom budget in a coarser particle grid yields errors of several percent upward, with velocity error dominant. The paper further claims the linearization-and-freezing acceleration, which updates the frozen geometric and kernel terms only every $f$ steps, nearly doubles computational efficiency at $f = 2$ for about 25% additional error, with CPU time falling roughly like $f^{-1}$, and that the method stays accurate in predictive settings where the training data come from a different tool rotational speed, provided the training speed is not much slower than the test speed.

Load-bearing premise

The comparison rests on a deliberately regular flow regime produced by hand-tuning two SPH regularization parameters, $\lambda = 2\times10^{-3}$ and $\zeta = 4\times10^{-1}$, chosen in Section 3.3 and Table 3.3; the paper itself warns that POD-MOR is likely to be unstable for the velocity field when the original simulation shows irregular motion, so if realistic friction stir welding flows are irregular, the claimed advantage may not survive.

Editorial extensions

If this is right

  • SPH models in the regular regime can be compressed to roughly half their degrees of freedom at about one percent relative error, a reduction that uniform particle coarsening cannot match at equal cost.
  • The frozen-coefficient linearization makes the reduced model genuinely faster than the full SPH simulation, with mild error growth for freezing windows of a few steps, making parameter studies and inverse problems practical.
  • POD modes trained at one tool rotational speed transfer to nearby speeds, so the offline training cost can be amortized over many online simulations with different parameters.
  • The POD mode visualizations show the reduction works by grouping particles into coherent radial structures, with the leading modes capturing each field's dominant physics (outer position layers, center-weighted velocity, shoulder-generated heat, boundary density drops).

Reading between the lines

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

  • If the advantage survives irregular flows, POD-MOR could replace particle coarsening as the default cheap-surrogate strategy for SPH parameter sweeps, and the freezing trick suggests a general recipe for nonlinear Lagrangian mesh-free operators: precompute the linearized kernels once and refresh them on a slower clock tied to particle redistribution.
  • The consistently dominant velocity error hints that a different projection (for instance least-squares Petrov-Galerkin) or a velocity-enriched basis might push the reduction ratio well past the 40-50% modes reported here — a testable variant the paper does not run.
  • The asymmetry in the predictive tests (training slow, testing fast fails while the reverse works) suggests a practical snapshot-selection rule: bias the training ensemble toward the fastest dynamics in the parameter range of interest.
  • Because the density and temperature modes are almost radial, a problem-adapted per-field basis could beat the uniform mode splits the paper uses and lower the total degree-of-freedom budget further.
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

4 major / 5 minor

Summary. The paper proposes a projection-based model-order reduction approach for smoothed particle hydrodynamics (SPH) simulations, using proper orthogonal decomposition (POD) to construct reduced bases for particle position, velocity, temperature, and density. The method is applied to a two-dimensional friction stir spot welding (FSSW) model that couples flow and heat equations. The authors report that POD-based MOR (POD-MOR) yields significantly smaller errors than uniform particle coarsening at the same degrees of freedom, that linearization and freezing of coefficients accelerates computation with small additional error, and that the method works in predictive settings with different tool rotational speeds.

Significance. If the claims are established, this would be a useful first systematic demonstration of POD-based MOR for SPH in a Lagrangian framework, with potential value for parameter studies and inverse problems. The paper includes careful elements: a relative L2 error indicator, self-convergence tests in time, comparisons at multiple resolutions, and a transparent statement of parameter choices. The visualization of POD modes is also informative. However, the central quantitative claim is currently weakened by the in-sample construction of the POD basis and by an inconsistent definition of 'same degrees of freedom' between POD-MOR and uniform coarsening. The acceleration results based on freezing are promising and are the most robust part of the evidence.

major comments (4)
  1. [Sections 3.5 and 3.6, Tables 3.6 and 3.8] The headline comparison does not establish the 'same degrees of freedom' claim. In POD-MOR, the reduced model is still solved on the full Grid-C particle set (1084 particles); only the number of coefficients in the POD representation is reduced, while neighbor searches, kernel sums, and particle updates still operate over the full particle set. In contrast, the uniform-coarsening run uses 504 or 144 particles, with a different particle layout, different boundary-layer treatment, and a different discrete SPH operator. Moreover, the actual number of POD coefficients used in Table 3.6 is 320 (position) + 320 (velocity) + 504 (density) = 1144, which is not equal to the number of degrees of freedom in the Grid-B simulation (504 particles times the number of state variables). The comparison therefore conflates 'number of POD coefficients' with 'number of SPH particles.' Please restate the comparison in terms of a consistent cost metric, such as actual CPU time or wall-clock time, or define the degrees of freedom precisely and justify why the two reductions are equivalent.
  2. [Sections 3.4-3.6] In the reproducible-setting experiments, the POD basis is built from snapshots of the same full-SPH run that is used as the error reference. Consequently, the reported 'POD error' is essentially the projection error of the training trajectory onto its own leading modes, which is a best-case in-sample error and does not measure the generalization of the reduced model. This is particularly problematic because the central claim in Section 1 and the Abstract is that POD-MOR outperforms uniform coarsening. Please add out-of-sample tests in which the basis is trained on one simulation and evaluated on another (for example, training on Grid-C and testing on Grid-B, or training on a different rotational speed and testing on the reference flow), and report both training and test errors.
  3. [Section 3.3 and Section 4] The numerical evidence is restricted to a deliberately selected parameter regime. The authors choose lambda = 2e-3 and zeta = 4e-1 to obtain regular particle motion and state that 'POD-MOR SPH simulations is likely to be unstable, particularly for the velocity field, when the original SPH simulation shows irregular motion.' This limitation is acknowledged in the conclusion, but the Abstract and Introduction state the advantage of POD-MOR without this qualification. The claim of general effectiveness should be explicitly limited to regular, self-convergent flow regimes, or the method should be tested in at least one irregular regime to document its failure mode.
  4. [Section 3.9] The predictive-setting experiments vary only the tool rotational speed and keep the same Grid-B particle layout, boundary treatment, and discretization in both training and testing. This is a useful first step, but it does not address the central equal-DoF comparison with uniform coarsening, because the comparison with a coarser particle distribution is not performed in the predictive setting. The statement that POD-MOR 'can be applied with general situations with different parameter values' is stronger than what the evidence supports; please either add cross-resolution predictive tests or soften the claim.
minor comments (5)
  1. [Throughout] Typos should be corrected: 'directic projection' (Sections 3.4, 3.5, 3.6) should be 'direct projection'; 'erros' (Introduction) should be 'errors'; 'computaitonal' (Section 3.9) should be 'computational'.
  2. [Table 3.8 caption] The caption says 'Under different prescribed temperature profiles,' but the table reports results for the full coupled model without prescribed temperature; the caption appears to be a copy-paste from Table 3.6 or 3.7 and should be corrected.
  3. [Section 3.3, page 30] The sentence 'POD-MOR SPH simulations is likely to be unstable...' has a subject-verb agreement error; it should read 'POD-MOR SPH simulations are likely to be unstable...'.
  4. [Section 2.3, Equations (2.37)-(2.38)] The reduced momentum and temperature equations are presented in a very dense form with repeated indices and no explicit summation conventions; adding a short explanation of the notation or using tensor notation would help readers verify the Galerkin projection.
  5. [Figure 3.6] The text says 'We observe a good alignment of all velocity, temperature, and density POD error for each initial particle distribution,' but the figure appears to show separate curves; please clarify whether alignment means similar magnitude and whether the curves are plotted on the same scale.

Circularity Check

2 steps flagged · score 6.0 of 10

Partial circularity: the equal-DoF comparison is definitionally loaded, and the reproducible-setting POD errors are in-sample projection errors; predictive tests provide some independent support.

  1. self definitional [Section 3.5, Table 3.6; cf. Section 2.3 (Eq. 2.27)]
    "With the same DoFs, POD achieves less error than uniformly decreasing the particle number, because POD uses a better strategy in picking combination of particles than just uniformly decreasing number of particles in the computational domain Ω."

    The 'same DoFs' comparison equates the number of POD coefficients with the number of particles in Grid-B, while the POD-MOR run is still solved on the full Grid-C particle set. Section 2.3 admits: 'the computation still requires pulling-back to the system of N = |Λ| particles', so the reduced model retains all particle DoFs and only shrinks the coefficient vector. Uniform coarsening, by contrast, actually removes particles and changes the discrete SPH operator. Thus the claimed advantage of POD over uniform coarsening at equal DoF is an artifact of defining DoF asymmetrically, not a measured property of the two reductions.

  2. fitted input called prediction [Section 3.5, Table 3.6 and Section 2.3, Eq. (2.28)]
    "In Table 3.6, we compare the the POD error (between the POD-MOR SPH solution and the original SPH solution of Grid-C) with the coarsening error (between the original SPH solutions of Grid-Band Grid-C)."

    The POD basis is built from snapshots of the same Grid-C simulation that defines the reference solution. Equation (2.28) defines U as the minimizer of the snapshot reconstruction error ||X - X̃||_F, so the reported 'POD error' is the optimal in-sample projection error of the training data onto its own leading modes. Its smallness is therefore guaranteed by construction rather than being evidence of predictive skill. The genuinely predictive test appears only in Section 3.9 with varying rotational speeds, and there no uniform-coarsening comparison is made.

full rationale

The paper's central claim that POD-MOR beats uniform particle coarsening at the same degrees of freedom rests on Table 3.6 and Table 3.8, where the POD basis is trained on the very Grid-C solution used as the reference and where 'DoF' is counted as the number of POD coefficients rather than the actual number of particle degrees of freedom still present in the simulation. Those two features make the reproducible-setting errors close to in-sample projection errors and make the equal-DoF comparison definitionally favorable to POD. However, the paper does not hide the distinction: Section 3.9 explicitly labels the earlier tests as 'reproducible setting' and then provides predictive tests across different tool rotational speeds, which are genuinely out-of-sample and give some independent evidence that POD-MOR and its linearized acceleration can work. The citations to Fraser, Cao, and other co-authored prior work supply the FSW model and SPH parameters, but they are not used to justify the POD claim, so there is no load-bearing self-citation chain. The circularity is partial: the headline superiority over uniform coarsening is supported mainly by the in-sample, asymmetrically defined comparisons, while the predictive demonstrations are narrower and do not re-establish that specific comparison.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

No new physical entities are introduced. Free parameters are the SPH stabilization constants and the mode-count/freezing-step knobs. The main axiomatic burden is that the FSSW physics is captured by the 2D equations and that the chosen regular regime is representative.

free parameters (4)
  • Particle diffusion parameter lambda = 2e-3
    Chosen by hand to minimize maximum radial motion (Table 3.3) and to produce regular flow; the method's stability depends on this choice.
  • XSPH coefficient zeta = 4e-1
    Chosen by hand alongside lambda, below the theoretical upper bound 0.5, to keep particle trajectories regular.
  • POD mode count = 10% to 50% of total DoF depending on experiment
    A user-selected hyperparameter controlling the trade-off between accuracy and compression; not fitted to data, but the central error claims depend on it.
  • Freezing step f = varies, e.g. f=2 or f=5
    Number of time steps between updates of frozen coefficients; selected to trade CPU time against error (Figures 3.11-3.12).
assumptions (4)
  • domain assumption The 2D FSSW model with Arrhenius viscosity and heat sources adequately represents the physical welding process.
    Section 2.1; the entire study is a numerical experiment on this model, so the physical fidelity of the model is assumed.
  • standard math SPH kernel approximation and the standard SPH discrete equations converge to the underlying PDEs.
    Section 2.2; the paper relies on well-established SPH formulations without proving convergence.
  • domain assumption POD basis computed from snapshots of one trajectory spans the solution space for nearby parameters.
    Section 3.9; the predictive setting tests this for a limited range of angular velocities, and the authors find the assumption holds only when the training speed is not slower than the test speed.
  • ad hoc to paper The chosen SPH parameters lambda and zeta yield a representative regular flow regime.
    Section 3.3; the parameters are selected to minimize radial motion, and the authors explicitly note that for irregular motion POD-MOR is likely unstable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Proper Orthogonal Decomposition-based Model-Order Reduction for Smoothed Particle Hydrodynamics Simulation." pith.science (2026). https://pith.science/paper/P7Q5AQ6B

@misc{pith2026250719825,
  author       = {Pith},
  title        = {Pith review of: Proper Orthogonal Decomposition-based Model-Order Reduction for Smoothed Particle Hydrodynamics Simulation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P7Q5AQ6B}},
  note         = {Machine review of arXiv:2507.19825}
}
read the original abstract

In this paper, we present a projection-based model-order reduction (MOR) technique for smoothed particle hydrodynamics (SPH) simulations, which is a mesh-free approach within the Lagrangian framework. Our approach utilizes the proper orthogonal decomposition (POD) technique to generate a subspace basis for the reduction process. The main objective of this study is to conduct an initial exploration of the feasibility of employing POD-based MOR (POD-MOR) in SPH simulations and to quantify the associated POD error. To illustrate the effectiveness of this approach, we consider the friction stir spot welding problem, which involves the coupling of flow equations and heat equation. Our findings reveal that, with the same degrees of freedom, POD-MOR significantly reduces computational error compared to the uniform reduction of particle numbers in SPH simulations. Additionally, the acceleration technique of POD-MOR for SPH simulation via linearization and freezing coefficients has been shown to be effective while keeping the error small. We have also showed the effectiveness of POD-MOR in predictive settings in SPH simulations with different parameter values.

Figures

Figures reproduced from arXiv: 2507.19825 by the authors.

Figure 3
Figure 3. [PITH_FULL_IMAGE:figures/full_fig_p022_3.png] view at source ↗
Figure 3.1
Figure 3.1. Typical initial distribution of smoothed particles. The blue and green particles represent [PITH_FULL_IMAGE:figures/full_fig_p024_3_1.png] view at source ↗
Figure 3
Figure 3. [PITH_FULL_IMAGE:figures/full_fig_p026_3.png] view at source ↗
Figures from the paper (20 more)
Figure 3.2
Figure 3.2. Figure 3.2: The fixed grid for interpolation. 25 [PITH_FULL_IMAGE:figures/full_fig_p026_3_2.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p029_3.png]
Figure 3.3
Figure 3.3. Figure 3.3: The error of velocity, temperature, and density with respect to ∆ [PITH_FULL_IMAGE:figures/full_fig_p030_3_3.png]
Figure 3.4
Figure 3.4. Figure 3.4: The error of velocity, temperature, and density with respect to ∆ [PITH_FULL_IMAGE:figures/full_fig_p031_3_4.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p031_3.png]
Figure 3.5
Figure 3.5. Figure 3.5: The snapshot of smoothed particles at the end of simulation, where particle velocity [PITH_FULL_IMAGE:figures/full_fig_p032_3_5.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p039_3.png]
Figure 3.6
Figure 3.6. Figure 3.6: POD error with coupled model. Initial particle distribution Grid-B and Grid-C, Θ = [PITH_FULL_IMAGE:figures/full_fig_p040_3_6.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p042_3.png]
Figure 3.7
Figure 3.7. Figure 3.7: First six POD modes of position x1. 42 [PITH_FULL_IMAGE:figures/full_fig_p043_3_7.png]
Figure 3.8
Figure 3.8. Figure 3.8: First six POD modes of velocity u1. 43 [PITH_FULL_IMAGE:figures/full_fig_p044_3_8.png]
Figure 3.9
Figure 3.9. Figure 3.9: First six POD modes of temperature. 44 [PITH_FULL_IMAGE:figures/full_fig_p045_3_9.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p046_3.png]
Figure 3.10
Figure 3.10. Figure 3.10: First six POD modes of density. 46 [PITH_FULL_IMAGE:figures/full_fig_p047_3_10.png]
Figure 3.11
Figure 3.11. Figure 3.11: POD error vs. number of freezing steps with coupled model. Initial particle distribution [PITH_FULL_IMAGE:figures/full_fig_p048_3_11.png]
Figure 3.12
Figure 3.12. Figure 3.12: CPU time vs. number of freezing steps with coupled model. Initial particle distribution [PITH_FULL_IMAGE:figures/full_fig_p049_3_12.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p050_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p051_3.png]
Figure 3.13
Figure 3.13. Figure 3.13: POD error vs. ratio of POD modes with coupled model. Initial particle distribution [PITH_FULL_IMAGE:figures/full_fig_p052_3_13.png]
Figure 3.14
Figure 3.14. Figure 3.14: POD error vs. number of freezing steps with coupled model. Initial particle distribution [PITH_FULL_IMAGE:figures/full_fig_p053_3_14.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

54 extracted references · 51 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

  2. [2]

    , author Zahr, M.J

    author Amsallem, D. , author Zahr, M.J. , author Farhat, C. , year 2012 . title Nonlinear model order reduction based on local reduced-order bases . journal International Journal for Numerical Methods in Engineering volume 92 , pages 891--916

    work page 2012

  3. [3]

    , year 2022

    author Babu, V. , year 2022 . title Fundamentals of incompressible fluid flow . publisher Springer

    work page 2022

  4. [4]

    , author Greer, A

    author Battezzati, L. , author Greer, A. , year 1989 . title The viscosity of liquid metals and alloys . journal Acta Metallurgica volume 37 , pages 1791--1802

    work page 1989

  5. [5]

    , author Gugercin, S

    author Benner, P. , author Gugercin, S. , author Willcox, K. , year 2015 . title A survey of projection-based model reduction methods for parametric dynamical systems . journal SIAM review volume 57 , pages 483--531

    2015

  6. [6]

    , author Holmes, P

    author Berkooz, G. , author Holmes, P. , author Lumley, J.L. , year 1993 . title The proper orthogonal decomposition in the analysis of turbulent flows . journal Annual review of fluid mechanics volume 25 , pages 539--575

    work page 1993

  7. [7]

    , year 1985

    author Brookshaw, L. , year 1985 . title A method of calculating radiative heat diffusion in particle simulations . journal Publications of the Astronomical Society of Australia volume 6 , pages 207--210

    work page 1985

  8. [8]

    , author Fraser, K

    author Cao, X. , author Fraser, K. , author Song, Z. , author Drummond, C. , author Huang, H. , year 2022 . title Machine learning and reduced order computation of a friction stir welding model . journal Journal of Computational Physics volume 454 , pages 110863

    work page 2022

Show all 54 references
  1. [9]

    , year 2015

    author Carlberg, K. , year 2015 . title Adaptive h-refinement for reduced-order models . journal International Journal for Numerical Methods in Engineering volume 102 , pages 1192--1210

    2015

  2. [10]

    , author Barone, M

    author Carlberg, K. , author Barone, M. , author Antil, H. , year 2017 . title Galerkin v. least-squares petrov--galerkin projection in nonlinear model reduction . journal Journal of Computational Physics volume 330 , pages 693--734

    2017

  3. [11]

    , author Bou-Mosleh, C

    author Carlberg, K. , author Bou-Mosleh, C. , author Farhat, C. , year 2011 . title Efficient non-linear model reduction via a least-squares petrov--galerkin projection and compressive tensor approximations . journal International Journal for numerical methods in engineering v...

    2011

  4. [12]

    , author Farhat, C

    author Carlberg, K. , author Farhat, C. , author Cortial, J. , author Amsallem, D. , year 2013 . title The gnat method for nonlinear model reduction: effective implementation and application to computational fluid dynamics and turbulent flows . journal Journal of Computational...

    2013

  5. [13]

    , year 2000

    author Chatterjee, A. , year 2000 . title An introduction to the proper orthogonal decomposition . journal Current science , pages 808--817

    2000

  6. [14]

    , author Sorensen, D.C

    author Chaturantabut, S. , author Sorensen, D.C. , year 2010 . title Nonlinear model reduction via discrete empirical interpolation . journal SIAM Journal on Scientific Computing volume 32 , pages 2737--2764

    2010

  7. [15]

    , author Carlberg, K

    author Choi, Y. , author Carlberg, K. , year 2019 . title Space--time least-squares petrov--galerkin projection for nonlinear model reduction . journal SIAM Journal on Scientific Computing volume 41 , pages A26--A58

    2019

  8. [16]

    , author Coombs, D

    author Choi, Y. , author Coombs, D. , author Anderson, R. , year 2020 . title Sns: A solution-based nonlinear subspace method for time-dependent model order reduction . journal SIAM Journal on Scientific Computing volume 42 , pages A1116--A1146

    2020

  9. [17]

    , author Cheung, S.W

    author Copeland, D.M. , author Cheung, S.W. , author Huynh, K. , author Choi, Y. , year 2022 . title Reduced order models for lagrangian hydrodynamics . journal Computer Methods in Applied Mechanics and Engineering volume 388 , pages 114259

    2022

  10. [18]

    , author Mijajlovic, M

    author Djurdjanovic, M. , author Mijajlovic, M. , author Milcic, D. , author Stamenkovic, D. , year 2009 . title Heat generation during friction stir welding process . journal Tribology in industry volume 31 , pages 8

    2009

  11. [19]

    , author Revenga, M

    author Espanol, P. , author Revenga, M. , year 2003 . title Smoothed dissipative particle dynamics . journal Physical Review E volume 67 , pages 026705

    2003

  12. [20]

    , author Chapman, T

    author Farhat, C. , author Chapman, T. , author Avery, P. , year 2015 . title Structure-preserving, stability, and accuracy properties of the energy-conserving sampling and weighting method for the hyper reduction of nonlinear finite element dynamic models . journal Internatio...

    2015

  13. [21]

    , author Kiss, L.I

    author Fraser, K. , author Kiss, L.I. , author St-Georges, L. , author Drolet, D. , year 2018 . title Optimization of friction stir weld joint quality using a meshfree fully-coupled thermo-mechanics approach . journal Metals volume 8 , pages 101

    2018

  14. [22]

    , author St-Georges, L

    author Fraser, K. , author St-Georges, L. , author Kiss, L.I. , year 2016 . title A mesh-free solid-mechanics approach for simulating the friction stir-welding process . journal Joining technologies , pages 27--52

    2016

  15. [23]

    , year 2017

    author Fraser, K.A. , year 2017 . title Robust and efficient meshfree solid thermo-mechanics simulation of friction stir welding . Ph.D. thesis. Universit \'e du Qu \'e bec \`a Chicoutimi

    2017

  16. [24]

    , author Lumley, J.L

    author Holmes, P. , author Lumley, J.L. , author Berkooz, G. , year 1996 . title Turbulence, Coherent Structures, Dynamical Systems and Symmetry . publisher Cambridge University Press

    1996

  17. [25]

    , author Beissel, S.R

    author Johnson, G.R. , author Beissel, S.R. , year 1996 . title Normalized smoothing functions for sph impact computations . journal International Journal for Numerical Methods in Engineering volume 39 , pages 2725--2741

    1996

  18. [26]

    , author Golinval, J.c

    author Kerschen, G. , author Golinval, J.c. , author Vakakis, A.F. , author Bergman, L.A. , year 2005 . title The method of proper orthogonal decomposition for dynamical characterization and order reduction of mechanical systems: an overview . journal Nonlinear dynamics volume...

    2005

  19. [27]

    , author Manzoni, A

    author Lassila, T. , author Manzoni, A. , author Quarteroni, A. , author Rozza, G. , year 2014 . title Model order reduction in fluid dynamics: challenges and perspectives . journal Reduced Order Methods for modeling and computational reduction , pages 235--273

    2014

  20. [28]

    , author Chen, J.S

    author Lee, C.H. , author Chen, J.S. , year 2013 . title Proper orthogonal decomposition-based model order reduction via radial basis functions for molecular dynamics systems . journal International journal for numerical methods in engineering volume 96 , pages 599--627

    2013

  21. [29]

    , author Xu, R

    author Lind, S.J. , author Xu, R. , author Stansby, P.K. , author Rogers, B.D. , year 2012 . title Incompressible smoothed particle hydrodynamics for free-surface flows: A generalised diffusion-based algorithm for stability and validations for impulsive flows and propagating w...

    2012

  22. [30]

    , author Liu, G

    author Liu, M. , author Liu, G. , year 2010 . title Smoothed particle hydrodynamics (sph): an overview and recent developments . journal Archives of computational methods in engineering volume 17 , pages 25--76

    2010

  23. [31]

    , year 2001

    author Lucia, D.J. , year 2001 . title Reduced order modeling for high-speed flows with moving shocks . publisher Air Force Institute of Technology

    2001

  24. [32]

    , author Beran, P.S

    author Lucia, D.J. , author Beran, P.S. , author Silva, W.A. , year 2004 . title Reduced-order modeling: new approaches for computational physics . journal Progress in aerospace sciences volume 40 , pages 51--117

    2004

  25. [33]

    , author Rodriguez, S.N

    author Magargal, L.K. , author Rodriguez, S.N. , author Jaworski, J. , author Iliopoulos, A.P. , author Michopoulos, J.G. , year 2022 . title Lagrangian-to-eulerian mapping for modal decomposition of multiphase flows via smoothed-particle hydrodynamics , in: booktitle AIAA AVI...

    2022

  26. [34]

    , author Balajewicz, M

    author Mojgani, R. , author Balajewicz, M. , year 2017 . title Lagrangian basis method for dimensionality reduction of convection dominated nonlinear flows . journal arXiv preprint arXiv:1701.04343

    2017 arXiv

  27. [35]

    , year 1989

    author Monaghan, J. , year 1989 . title On the problem of penetration in particle methods . journal Journal of Computational physics volume 82 , pages 1--15

    1989

  28. [36]

    , year 1992

    author Monaghan, J.J. , year 1992 . title Smoothed particle hydrodynamics . journal Annual review of astronomy and astrophysics volume 30 , pages 543--574

    1992

  29. [37]

    , year 2002

    author Monaghan, J.J. , year 2002 . title Sph compressible turbulence . journal Monthly Notices of the Royal Astronomical Society volume 335 , pages 843--852

    2002

  30. [38]

    , year 2005

    author Monaghan, J.J. , year 2005 . title Smoothed particle hydrodynamics . journal Reports on progress in physics volume 68 , pages 1703

    2005

  31. [39]

    , author Baratzadeh, F

    author Patil, S. , author Baratzadeh, F. , author Lankarani, H. , year 2017 . title Modeling of the material flow in friction stir spot welding using sph approach . journal Visualization of Mechanical Processes: An International Online Journal volume 7

    2017

  32. [40]

    , year 1901

    author Pearson, K. , year 1901 . title Liii. on lines and planes of closest fit to systems of points in space . journal The London, Edinburgh, and Dublin philosophical magazine and journal of science volume 2 , pages 559--572

    1901

  33. [41]

    , author Willcox, K

    author Peherstorfer, B. , author Willcox, K. , year 2015 . title Dynamic data-driven reduced-order models . journal Computer Methods in Applied Mechanics and Engineering volume 291 , pages 21--41

    2015

  34. [42]

    , author Libersky, L.D

    author Randles, P. , author Libersky, L.D. , year 1996 . title Smoothed particle hydrodynamics: some recent improvements and applications . journal Computer methods in applied mechanics and engineering volume 139 , pages 375--408

    1996

  35. [43]

    , author Vega, J.M

    author Rap \'u n, M.L. , author Vega, J.M. , year 2010 . title Reduced order models based on local pod plus galerkin projection . journal Journal of Computational Physics volume 229 , pages 3046--3063

    2010

  36. [44]

    , author Colonius, T

    author Rowley, C.W. , author Colonius, T. , author Murray, R.M. , year 2004 . title Model reduction for compressible flows using pod and galerkin projection . journal Physica D: Nonlinear Phenomena volume 189 , pages 115--129

    2004

  37. [45]

    , author Reynolds, A.P

    author Seidel, T. , author Reynolds, A.P. , year 2003 . title Two-dimensional friction stir welding process model based on fluid mechanics. journal Science & Technology of Welding & Joining volume 8

    2003

  38. [46]

    , year 1968

    author Shepard, D. , year 1968 . title A two-dimensional interpolation function for irregularly-spaced data , in: booktitle Proceedings of the 1968 23rd ACM national conference , pp. pages 517--524

    1968

  39. [47]

    , year 1987 a

    author Sirovich, L. , year 1987 a. title Turbulence and the dynamics of coherent structures. i. coherent structures . journal Quarterly of applied mathematics volume 45 , pages 561--571

    1987

  40. [48]

    , year 1987 b

    author Sirovich, L. , year 1987 b. title Turbulence and the dynamics of coherent structures. ii. symmetries and transformations . journal Quarterly of Applied mathematics volume 45 , pages 573--582

    1987

  41. [49]

    , year 1987 c

    author Sirovich, L. , year 1987 c. title Turbulence and the dynamics of coherent structures. iii. dynamics and scaling . journal Quarterly of Applied mathematics volume 45 , pages 583--590

    1987

  42. [50]

    , author Fraser, K

    author Stubblefield, G. , author Fraser, K. , author Phillips, B. , author Jordon, J. , author Allison, P. , year 2021 . title A meshfree computational framework for the numerical simulation of the solid-state additive manufacturing process, additive friction stir-deposition (...

    2021

  43. [51]

    , author Brunton, S.L

    author Taira, K. , author Brunton, S.L. , author Dawson, S.T. , author Rowley, C.W. , author Colonius, T. , author McKeon, B.J. , author Schmidt, O.T. , author Gordeyev, S. , author Theofilis, V. , author Ukeiley, L.S. , year 2017 . title Modal analysis of fluid flows: An over...

    2017

  44. [52]

    , author Grant, G

    author Tartakovsky, A. , author Grant, G. , author Sun, X. , author Khaleel, M. , year 2006 . title Modeling of friction stir welding (FSW) process with smooth particle hydrodynamics (SPH) . type Technical Report . SAE Technical Paper

    2006

  45. [53]

    , author Yang, H

    author Yang, H. , author Yang, H. , author Hu, X. , year 2015 . title Simulation on the plunge stage in refill friction stir spot welding of aluminum alloys , in: booktitle 2015 4th International Conference on Mechatronics, Materials, Chemistry and Computer Engineering , organ...

    2015

  46. [54]

    , year 2012

    author Yang, H.G. , year 2012 . title Numerical simulation on friction stir spot welding of alumimum alloy using sph method , in: booktitle 2012 IEEE International Conference on Computer Science and Automation Engineering (CSAE) , organization IEEE . pp. pages 658--660

    2012

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.