pith. sign in

arxiv: 2505.13776 · v1 · submitted 2025-05-19 · 🧮 math.NA · cs.NA· math.OC

Convergence Analysis of an Adaptive Nonconforming FEM for Phase-Field Dependent Topology Optimization in Stokes Flow

Bangti Jin , Jing Li , Yifeng Xu , Shengfeng Zhu This is my paper

Pith reviewed 2026-05-22 13:32 UTC · model grok-4.3

classification 🧮 math.NA cs.NAmath.OC
keywords adaptive finite element methodnonconforming elementstopology optimizationStokes equationsphase-field modelconvergence analysisCrouzeix-Raviart elements
0
0 comments X

The pith

An adaptive nonconforming finite element method converges to optimality conditions for phase-field topology optimization in Stokes flow.

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

The paper introduces an adaptive algorithm that combines phase-field parameterization with nonconforming finite elements to solve topology optimization problems for Stokes flow. Conforming linear elements approximate the phase field, while Crouzeix-Raviart elements handle the velocity and piecewise constants the pressure. The central achievement is a convergence proof showing that minimizers from successive adaptive meshes have a subsequence converging to a solution of the first-order optimality system, with the discrete pressures also converging. This matters for ensuring that numerical designs in fluid optimization are mathematically justified rather than artifacts of discretization.

Core claim

The sequence of minimizers of the discrete problems contains a subsequence that converges to a solution of the first-order optimality system of the continuous problem, and the associated subsequence of discrete pressure fields converges as well. The proof hinges on a new discrete compactness result for nonconforming linear finite elements on adaptively generated meshes.

What carries the argument

The new discrete compactness result for nonconforming linear finite elements over a sequence of adaptively generated meshes, which allows passing to the limit in the variational inequalities and equations defining the optimality system.

If this is right

  • The adaptive algorithm can be applied with confidence that its outputs approximate true optimal configurations.
  • Local mesh refinement is justified without compromising convergence properties.
  • The pressure approximation converges along with the design and velocity fields.
  • Numerical tests confirm faster convergence or better accuracy than uniform mesh refinement for example problems.

Where Pith is reading between the lines

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

  • This compactness technique may apply to adaptive methods for other nonconforming discretizations in optimization or fluid problems.
  • Similar analysis could be developed for different phase-field models or time-dependent flows.
  • Extensions to three-dimensional domains would test the method's scalability in practical engineering settings.

Load-bearing premise

The analysis requires a discrete compactness property to hold for the nonconforming elements on the particular adaptive mesh sequences produced by the refinement strategy.

What would settle it

Finding an adaptive mesh refinement sequence where the discrete minimizers fail to have any convergent subsequence satisfying the continuous optimality system would disprove the convergence claim.

Figures

Figures reproduced from arXiv: 2505.13776 by Bangti Jin, Jing Li, Shengfeng Zhu, Yifeng Xu.

Figure 1
Figure 1. Figure 1: Initial phase-field functions for Examples [PITH_FULL_IMAGE:figures/full_fig_p021_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Numerical results for Example (a) from top to bottom: mesh, optimized design ϕ ∗ k , estimators ηk,1 and ηk,2. The number of vertices on each mesh is 1441, 2954, 6946 and 17354. More precisely, the first adaptive iteration (k = 0) yields a rough shape of the design, and with subsequent adaptive refinements, the algorithm produces improved boundary interfaces of the fluid region (in yellow), due to the high… view at source ↗
Figure 3
Figure 3. Figure 3: The convergence history of the total energy (top) and the volume constraint error (bottom) versus the [PITH_FULL_IMAGE:figures/full_fig_p022_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Numerical results for Example (b) from top to bottom: mesh, optimized designs ϕ ∗ k , and the estimators ηk,1 and ηk,2. The number of vertices on each mesh is 1441, 3070, 7442 and 19364. 22 [PITH_FULL_IMAGE:figures/full_fig_p022_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Numerical results for Example (c) from top to bottom: mesh, optimized designs ϕ ∗ k , and the estimators ηk,1 and ηk,2. The number of vertices on each mesh is 2174, 4973, 11669 and 28468. and uniform refinements, where the refinement steps and iteration numbers of the augmented Lagrangian method for adaptive and uniform refinements are (K, N) = (4, 50) and (K, N) = (3, 50), respectively. For both strategie… view at source ↗
Figure 6
Figure 6. Figure 6: Comparison of optimal designs ϕ ∗ K−1 by adaptive (left within each pair) and uniform (right within each pair) refinements with optimal shapes indicated in yellow for Examples (a)-(c). 23 [PITH_FULL_IMAGE:figures/full_fig_p023_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Numerical results for Example (d) from top to bottom: mesh, optimized designs ϕ ∗ k , and the estimators ηk,1 and ηk,2. The number of vertices on each mesh is 5631, 6929, 9631, 16052 and 24371 [PITH_FULL_IMAGE:figures/full_fig_p025_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Comparison of optimal iso-surfaces with a value of 0 [PITH_FULL_IMAGE:figures/full_fig_p025_8.png] view at source ↗
read the original abstract

In this work, we develop an adaptive nonconforming finite element algorithm for the numerical approximation of phase-field parameterized topology optimization governed by the Stokes system. We employ the conforming linear finite element space to approximate the phase field, and the nonconforming linear finite elements (Crouzeix-Raviart elements) and piecewise constants to approximate the velocity field and the pressure field, respectively. We establish the convergence of the adaptive method, i.e., the sequence of minimizers contains a subsequence that converges to a solution of the first-order optimality system, and the associated subsequence of discrete pressure fields also converges. The analysis relies crucially on a new discrete compactness result of nonconforming linear finite elements over a sequence of adaptively generated meshes. We present numerical results for several examples to illustrate the performance of the algorithm, including a comparison with the uniform refinement strategy.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

1 major / 2 minor

Summary. The manuscript develops an adaptive nonconforming finite-element algorithm for phase-field topology optimization of Stokes flow. Conforming linears approximate the phase field, Crouzeix-Raviart elements approximate the velocity, and piecewise constants approximate the pressure. The central claim is that the sequence of discrete minimizers admits a subsequence converging to a solution of the continuous first-order optimality system, with the associated discrete pressures also converging; the proof rests on a new discrete compactness result for the nonconforming elements on the adaptively generated mesh sequence. Numerical examples compare the adaptive strategy with uniform refinement.

Significance. If the convergence result is established, the work supplies the first rigorous justification for adaptive mesh refinement in this class of phase-field Stokes topology optimization problems, which is practically important for resolving fine-scale features without excessive degrees of freedom. The new discrete compactness statement for Crouzeix-Raviart elements on adaptively refined meshes constitutes an independent technical contribution that could be reused in other nonconforming discretizations of incompressible flow.

major comments (1)
  1. [Analysis section (compactness result)] The discrete compactness result (stated in the analysis section and invoked to pass to the limit in the weak form of the Stokes equations and the variational inequality) is load-bearing for the central convergence claim. The proof must explicitly confirm that the sequence of adaptively generated meshes produced by the marking and bisection algorithm satisfies the mesh-regularity hypotheses (bounded aspect ratios, controlled hanging-node configurations) required for strong L² convergence of the velocity and uniform control of the jumps; these properties are not automatically inherited from standard adaptive strategies and are not verified in the provided sketch.
minor comments (2)
  1. [Section 2] Notation for the discrete spaces and the discrete optimality system should be introduced with explicit reference to the mesh sequence index to avoid ambiguity when passing to the limit.
  2. [Numerical results] The numerical examples would benefit from a table reporting the number of degrees of freedom and the objective value at each adaptive step to quantify the efficiency gain over uniform refinement.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the careful review and for recognizing the significance of the discrete compactness result for Crouzeix-Raviart elements on adaptive meshes. We address the single major comment below and will incorporate the requested verification into the revised manuscript.

read point-by-point responses

  1. Referee: [Analysis section (compactness result)] The discrete compactness result (stated in the analysis section and invoked to pass to the limit in the weak form of the Stokes equations and the variational inequality) is load-bearing for the central convergence claim. The proof must explicitly confirm that the sequence of adaptively generated meshes produced by the marking and bisection algorithm satisfies the mesh-regularity hypotheses (bounded aspect ratios, controlled hanging-node configurations) required for strong L² convergence of the velocity and uniform control of the jumps; these properties are not automatically inherited from standard adaptive strategies and are not verified in the provided sketch.

    Authors: We agree that an explicit verification of the mesh-regularity hypotheses is necessary for the compactness argument to be complete. In the revised version we will insert a new auxiliary result (Lemma 4.3) immediately preceding the discrete compactness statement. The lemma establishes that the sequence of meshes generated by the standard Dörfler marking strategy combined with newest-vertex bisection, starting from a shape-regular initial triangulation, satisfies (i) a uniform bound on the aspect ratios of all elements and (ii) a controlled number of hanging-node configurations per edge. The proof follows the standard arguments of Carstensen et al. (2014) adapted to the nonconforming setting and will be referenced explicitly when the compactness result is invoked to pass to the limit in the Stokes weak form and the variational inequality. This addition removes the gap identified by the referee without altering the overall structure of the convergence proof. revision: yes

Circularity Check

0 steps flagged

No significant circularity; convergence proof relies on newly introduced discrete compactness result

full rationale

The paper's central claim is the convergence of the adaptive nonconforming FEM sequence to a solution of the first-order optimality system for the phase-field Stokes topology optimization problem. This is achieved by establishing a new discrete compactness result for nonconforming linear (Crouzeix-Raviart) elements on adaptively generated meshes, which is presented as an independent contribution within the analysis. No steps reduce by construction to fitted parameters, self-citations, or prior ansatzes; the compactness result supplies the necessary strong convergence and jump control to pass to the limit in the weak form, objective, and variational inequality without circular reduction. The derivation is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard finite-element theory for the Stokes system and phase-field models plus the newly stated discrete compactness result; no free parameters or invented entities appear in the abstract.

axioms (2)
  • domain assumption Existence of a solution to the first-order optimality system for the continuous phase-field topology optimization problem.
    The discrete minimizers are shown to converge to such a solution.
  • standard math The adaptive mesh sequence satisfies the conditions needed for the new discrete compactness property of the nonconforming elements.
    This is explicitly identified as the crucial technical ingredient.

pith-pipeline@v0.9.0 · 5685 in / 1287 out tokens · 47731 ms · 2026-05-22T13:32:25.209128+00:00 · methodology

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

55 extracted references · 55 canonical work pages

  1. [1]

    N. Aage, T. H. Poulsen, A. Gersborg-Hansen, and O. Sigmund. Topology optimization of large scale Stokes flow problems. Struct. Multidiscip. Optim. , 35(2):175– 180, 2008

    work page 2008

  2. [2]

    Ainsworth and J

    M. Ainsworth and J. T. Oden. A Posteriori Error Estimation in Finite Element Analysis . Wiley-Interscience, John Wiley & Sons, New York, 2000

    work page 2000

  3. [3]

    Alexandersen and C

    J. Alexandersen and C. S. Andreasen. A review of topology optimisation for fluid-based problems. Fluids, 5(1):29, 2020

    work page 2020

  4. [4]

    D. H. Alonso, L. F. N. de S´ a, J. S. R. Saenz, and E. C. N. Silva. Topology optimization applied to the design of 2D swirl flow devices. Struct. Multidiscip. Optim. , 58(6):2341–2364, 2018

    work page 2018

  5. [5]

    D. H. Alonso, J. S. R. Saenz, and E. C. N. Silva. Non-newtonian laminar 2D swirl flow design by the topology optimization method. Struct. Multidiscip. Optim. , 62(1):299–321, 2020

    work page 2020

  6. [6]

    Becker and S

    R. Becker and S. Mao. Quasi-optimality of adaptive nonconforming finite element methods for the Stokes equations. SIAM J. Numer. Anal. , 49(3):970–991, 2011

    work page 2011

  7. [7]

    Bonito, C

    A. Bonito, C. Canuto, R. H. Nochetto, and A. Veeser. Adaptive finite element methods. Acta Numer., 33:163– 485, 2024

    work page 2024

  8. [8]

    Borrvall and J

    T. Borrvall and J. Petersson. Topology optimization of fluids in Stokes flow. Int. J. Numer. Meth. Fluids , 41(1):77–107, 2003

    work page 2003

  9. [9]

    S. C. Brenner. Poincar´ e-Friedrichs inequalities for piecewise H1 functions. SIAM J. Numer. Anal. , 41(1):306– 324, 2003. 24 k = 0 k = 1 k = 2 k = 3 k = 4 Figure 7: Numerical results for Example (d) from top to bottom: mesh, optimized designs ϕ∗ k, and the estimators ηk,1 and ηk,2. The number of vertices on each mesh is 5631, 6929, 9631, 16052 and 24371....

    work page 2003

  10. [10]

    S. C. Brenner and L. R. Scott. The Mathematical Theory of Finite Element Methods . Springer, New York, third edition, 2008

    work page 2008

  11. [11]

    Buffa and C

    A. Buffa and C. Ortner. Compact embeddings of broken Sobolev spaces and applications. IMA J. Numer. Anal., 29(4):827–855, 2009

    work page 2009

  12. [12]

    Carstensen, M

    C. Carstensen, M. Feischl, M. Page, and D. Praetorius. Axioms of adaptivity. Comp. Math. Appl. , 67(6):1195– 1253, 2014

    work page 2014

  13. [13]

    Carstensen, D

    C. Carstensen, D. Peterseim, and H. Rabus. Optimal adaptive nonconforming fem for the Stokes problem. Numer. Math., 123(2):291–308, 2013

    work page 2013

  14. [14]

    J. M. Cascon, C. Kreuzer, R. H. Nochetto, and K. G. Siebert. Quasi-optimal convergence rate for an adaptive finite element method. SIAM J. Numer. Anal. , 46(5):2524–2550, 2008

    work page 2008

  15. [15]

    L. Chen. iFEM : an integrated finite element method package in MATLAB. Technical report, University of California at Irvine, 2009. Available at https://github.com/lyc102/ifem

    work page 2009

  16. [16]

    P. G. Ciarlet. The Finite Element Method for Elliptic Problems . SIAM, Philadelphia, PA, 2002

    work page 2002

  17. [17]

    Crouzeix and P.-A

    M. Crouzeix and P.-A. Raviart. Conforming and nonconforming finite element methods for solving the stationary Stokes equations. RAIRO Anal. Num´ er., 7(3):33–76, 1973

    work page 1973

  18. [18]

    E. Dari, R. Dur´ an, and C. Padra. Error estimators for nonconforming finite element approximations of the Stokes problem. Math. Comp., 64(211):1017–1033, 1995

    work page 1995

  19. [19]

    Y. Deng, Y. Wu, and Z. Liu. Topology Optimization Theory for Laminar Flow: Applications in Inverse Design of Microfluidics. Springer, Singapore, 2018

    work page 2018

  20. [20]

    L. C. Evans and R. F. Gariepy. Measure Theory and Fine Properties of Functions . CRC Press, Boca Raton, FL, revised edition, 2015

    work page 2015

  21. [21]

    Evgrafov

    A. Evgrafov. The limits of porous materials in the topology optimization of Stokes flows. Appl. Math. Optim. , 52(3):263–277, 2005

    work page 2005

  22. [22]

    Gantner and D

    G. Gantner and D. Praetorius. Plain convergence of adaptive algorithms without exploiting reliability and efficiency. IMA J. Numer. Anal. , 42(2):1434–1453, 2022

    work page 2022

  23. [23]

    Garcke and C

    H. Garcke and C. Hecht. A phase field approach for shape and topology optimization in Stokes flow. In A. Pratelli and G. Leugering, editors, New Trends in Shape Optimization , pages 103–116. Birkh¨ auser, New York, 2015

    work page 2015

  24. [24]

    Garcke and C

    H. Garcke and C. Hecht. Applying a phase field approach for shape optimization of a stationary Navier-Stokes flow. ESAIM Control Optim. Calc. Var. , 22(2):309–337, 2016

    work page 2016

  25. [25]

    Garcke and C

    H. Garcke and C. Hecht. Shape and topology optimization in Stokes flow with a phase field approach. Appl. Math. Optim., 73(1):23–70, 2016

    work page 2016

  26. [26]

    Garcke, C

    H. Garcke, C. Hecht, M. Hinze, and C. Kahle. Numerical approximation of phase field based shape and topology optimization for fluids. SIAM J. Sci. Comput. , 37(4):A1846–A1871, 2015

    work page 2015

  27. [27]

    Gersborg-Hansen, O

    A. Gersborg-Hansen, O. Sigmund, and R. B. Haber. Topology optimization of channel flow problems. Struct. Multidiscip. Optim., 30(3):181–192, 2005

    work page 2005

  28. [28]

    Girault and P

    V. Girault and P. A. Raviart. Finite Element Methods for Navier-Stokes Equations . Springer-Verlag, Berlin, 1986

    work page 1986

  29. [29]

    J. K. Guest and J. H. Pr´ evost. Topology optimization of creeping fluid flows using a Darcy-Stokes finite element. Int. J. Numer. Meth. Engng. , 66(3):461–484, 2006. 26

    work page 2006

  30. [30]

    Hu and J

    J. Hu and J. Xu. Convergence and optimality of the adaptive nonconforming linear element method for the Stokes problem. J. Sci. Comput. , 55:125–148, 2013

    work page 2013

  31. [31]

    B. Jin, J. Li, Y. Xu, and S. Zhu. An adaptive phase-field method for structural topology optimization. J. Comput. Phys., 206:112932, 27 pp., 2024

    work page 2024

  32. [32]

    B. Jin, J. Li, Y. Xu, and S. Zhu. On the Crouzeix-Raviart finite element approximation of phase-field dependent topology optimization in Stokes flow. Preprint, arXiv:2505.04120, 2025

    work page arXiv 2025

  33. [33]

    B. Jin, F. Wang, and Y. Xu. Adaptive approximations of inclusions in a semilinear elliptic problem related to cardiac electrophysiology. IMA J. Numer. Anal. , pages in press, available as arXiv:2504.04483, 2025

    work page arXiv 2025

  34. [34]

    Jin and Y

    B. Jin and Y. Xu. Adaptive reconstruction for electrical impedance tomography with a piecewise constant conductivity. Inverse Problems, 36(1):014003, 28 pp., 2020

    work page 2020

  35. [35]

    Kossaczky

    I. Kossaczky. A recursive approach to local mesh refinement in two and three dimensions. J. Comp. Appl. Math., 55(3):275–288, 1994

    work page 1994

  36. [36]

    Kreissl, G

    S. Kreissl, G. Pingen, and K. Maute. Topology optimization for unsteady flow. Int. J. Numer. Methods Engrg. , 87(13):1229–1253, 2011

    work page 2011

  37. [37]

    M. G. Larson and F. Bengzon. The Finite Element Method: Theory, Implementation, and Applications . Springer, Heidelberg, 2013

    work page 2013

  38. [38]

    Li and J

    F. Li and J. Yang. A provably efficient monotonic-decreasing algorithm for shape optimization in Stokes flows by phase-field approaches. Comput. Methods Appl. Mech. Engrg. , 398:115195, 24 pp., 2022

    work page 2022

  39. [39]

    J. Li, Y. Xu, and S. Zhu. Adaptive computation of elliptic eigenvalue topology optimization with a phase-field approach. J. Numer. Math. , 2025. https://doi.org/10.1515/jnma-2024-0060

    work page doi:10.1515/jnma-2024-0060 2025

  40. [40]

    Morin, P

    P. Morin, P. H. Nochetto, and K. G. Siebert. Data oscillation and convergence of adaptive fem. SIAM J. Numer. Anal., 38(2):466–488, 2000

    work page 2000

  41. [41]

    Morin, K

    P. Morin, K. G. Siebert, and A. Veeser. A basic convergence result for conforming adaptive finite elements. Math. Models Methods Appl. Sci. , 18(5):707–737, 2008

    work page 2008

  42. [42]

    D. A. Nield and A. Bejan. Convection in Porous Media . Springer, Cham., fifth edition, 2017

    work page 2017

  43. [43]

    R. H. Nochetto, K. G. Siebert, and A. Veeser. Theory of adaptive finite element methods: an introduction. In Multiscale, Nonlinear and Adaptive Approximation , pages 409–542. Springer, Berlin, 2009

    work page 2009

  44. [44]

    L. H. Olesen, F. Okkels, and H. Bruus. A high-level programming-language implementation of topology opti- mization applied to steady-state Navier–Stokes flow. Int. J. Numer. Methods Engrg. , 65(7):975–1001, 2006

    work page 2006

  45. [45]

    Ortner and D

    C. Ortner and D. Praetorius. On the convergence of adaptive nonconforming finite element methods for a class of convex variational problems. SIAM J. Numer. Anal. , 49(1):346–367, 2011

    work page 2011

  46. [46]

    I. P. A. Papadopoulos. Numerical analysis of a discontinuous Galerkin method for the Borrvall-Petersson topology optimization problem. SIAM J. Numer. Anal. , 60(5):2538–2564, 2022

    work page 2022

  47. [47]

    I. P. A. Papadopoulos and E. S¨ uli. Numerical analysis of a topology optimization problem for Stokes flow. J. Comput. Appl. Math. , 412:114295, 21 pp., 2022

    work page 2022

  48. [48]

    L. F. N. S´ a, J. S. Romero, O. Horikawa, and E. C. N. Silva. Topology optimization applied to the development of small scale pump. Struct. Multidiscip. Optim. , 57(5):2045–2059, 2018

    work page 2045

  49. [49]

    Shi and M

    Z. Shi and M. Wang. Finite Element Methods . Science Press, Beijing, 2013

    work page 2013

  50. [50]

    K. G. Siebert. A convergence proof for adaptive finite elements without lower bound. IMA J. Numer. Anal. , 31(3):947–970, 2011. 27

    work page 2011

  51. [51]

    F. Stummel. Basic compactness properties of nonconforming and hybrid finite element spaces. RAIRO Anal. Num´ er., 14(1):81–115, 1980

    work page 1980

  52. [52]

    C.-J. Thore. Topology optimization of Stokes flow with traction boundary conditions using low-order finite elements. Comput. Methods Appl. Mech. Engrg. , 386:114177, 14 pp., 2021

    work page 2021

  53. [53]

    C. Traxler. An algorithm for adaptive mesh refinement in n dimensions. Computing, 59(2):115–137, 1997

    work page 1997

  54. [54]

    Verf¨ urth.A Posteriori Error Estimation Techniques for Finite Element Methods

    R. Verf¨ urth.A Posteriori Error Estimation Techniques for Finite Element Methods . Oxford University Press, Oxford, 2013

    work page 2013

  55. [55]

    Wiker, A

    N. Wiker, A. Klarbring, and T. Borrvall. Topology optimization of regions of Darcy and Stokes flow. Int. J. Numer. Meth. Engng. , 69(7):1374–1404, 2007. 28

    work page 2007