arxiv: 2603.05008 · v3 · submitted 2026-03-05 · 🧮 math.NA · cs.NA

Recognition: no theorem link

Nitsche methods for constrained problems in mechanics

Tom Gustafsson , Antti Hannukainen , Vili Kohonen , Juha Videman

Authors on Pith no claims yet

Pith reviewed 2026-05-15 15:50 UTC · model grok-4.3

classification 🧮 math.NA cs.NA

keywords Nitsche methodfinite element methodLagrange multiplierequality constraintsinequality constraintssolid mechanicsautomatic differentiationstabilized formulation

0 comments

The pith

Nitsche finite element methods can be rewritten in minimization form to enforce equality and inequality constraints on mechanical quantities.

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

The paper establishes guidelines for deriving Nitsche finite element methods that enforce equality and inequality constraints acting directly on the unknown mechanical field. It begins by formulating the constrained problem as a stabilized finite element discretization of a saddle-point system with a Lagrange multiplier. This system is then recast as a pure minimization problem that preserves the structure needed for nonlinear finite element codes and automatic differentiation. The resulting Nitsche formulation extends the method past its classical use for boundary conditions to general interior or interface constraints in solid mechanics. Numerical examples confirm that the approach retains optimal convergence rates across several test problems.

Core claim

Nitsche methods for constraints are obtained by first discretizing the stabilized saddle-point formulation that incorporates a Lagrange multiplier, then algebraically eliminating the multiplier to arrive at an equivalent minimization problem that can be added directly to existing nonlinear finite element solvers via automatic differentiation.

What carries the argument

Stabilized saddle-point formulation with Lagrange multiplier, algebraically rewritten as a Nitsche-type minimization functional.

If this is right

Constraints on the value of the unknown field can be added to any existing nonlinear finite element code without altering its core minimization structure.
Both equality and inequality constraints are handled by the same algebraic derivation from the stabilized saddle-point system.
Automatic differentiation supplies the consistent linearization needed for Newton-type solvers without manual derivation of Nitsche terms.
The same template applies to a range of solid mechanics problems beyond classical boundary conditions, including interior or interface constraints.

Where Pith is reading between the lines

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

The minimization form may allow Nitsche-type constraints to be inserted into optimization-based solvers or adjoint-based sensitivity analyses with little extra coding.
If the rewriting step preserves convexity properties, the approach could be tested on problems where the underlying energy is known to be non-convex.
The guideline procedure could be applied to derive Nitsche terms for constraints that couple different physical fields, such as displacement and temperature.

Load-bearing premise

The stabilized saddle-point formulation with Lagrange multiplier can be rewritten as a minimization problem that keeps the same stability and convergence behavior for nonlinear mechanical problems and automatic differentiation implementations.

What would settle it

A nonlinear solid mechanics simulation with an active inequality constraint in which the Nitsche method either loses stability or fails to recover the expected convergence rate under mesh refinement.

Figures

Figures reproduced from arXiv: 2603.05008 by Antti Hannukainen, Juha Videman, Tom Gustafsson, Vili Kohonen.

**Figure 2.** Figure 2: Two membrane contact problem convergence rates follow the theoretical linear [PITH_FULL_IMAGE:figures/full_fig_p017_2.png] view at source ↗

**Figure 3.** Figure 3: Condition numbers of the Jacobian in the Newton iterations for the penalty and [PITH_FULL_IMAGE:figures/full_fig_p018_3.png] view at source ↗

**Figure 4.** Figure 4: Numerical solution for the membrane against elastic solid problem using the [PITH_FULL_IMAGE:figures/full_fig_p020_4.png] view at source ↗

**Figure 5.** Figure 5: The convergence rate of the membrane against elastic solid problem follows the [PITH_FULL_IMAGE:figures/full_fig_p021_5.png] view at source ↗

**Figure 6.** Figure 6: The first principal stress for the elastic solid in membrane against elastic solid problem [PITH_FULL_IMAGE:figures/full_fig_p022_6.png] view at source ↗

**Figure 7.** Figure 7: Numerical solution for two plates in contact using the Nitsche method with [PITH_FULL_IMAGE:figures/full_fig_p023_7.png] view at source ↗

**Figure 8.** Figure 8: Two plate contact problem convergence rate follows the theoretical quadratic [PITH_FULL_IMAGE:figures/full_fig_p024_8.png] view at source ↗

**Figure 10.** Figure 10: 24 [PITH_FULL_IMAGE:figures/full_fig_p025_10.png] view at source ↗

**Figure 9.** Figure 9: The corners of the plate are displaced upwards when downwards point load [PITH_FULL_IMAGE:figures/full_fig_p026_9.png] view at source ↗

**Figure 10.** Figure 10: The convergence rate of the Kirchhoff plate with inequality boundary condition [PITH_FULL_IMAGE:figures/full_fig_p027_10.png] view at source ↗

read the original abstract

We present guidelines for deriving new Nitsche Finite Element Methods to enforce equality and inequality constraints that act on the value of the unknown mechanical quantity. We first formulate the problem as a stabilized finite element method for the saddle point formulation where a Lagrange multiplier enforces the underlying constraint. The Nitsche method is then presented in a general minimization form, suitable for adding constraints to nonlinear finite element methods and allowing straightforward computational implementation with automatic differentation. This extends the method beyond classical boundary condition enforcement. To validate these ideas, we present Nitsche formulations for a range of problems in solid mechanics and give numerical evidence of the convergence rates of the Nitsche method.

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.

Desk Editor's Note private letter to a colleague

The paper gives a workable minimization-form Nitsche recipe for general equality and inequality constraints in nonlinear mechanics, with numerical checks that look okay but thin support for stability once the problem goes nonlinear.

read the letter

The main point is that this work shows how to start from a stabilized saddle-point problem with a Lagrange multiplier and rewrite it as an unconstrained minimization whose stationarity produces the Nitsche terms. That form is then easy to drop into nonlinear finite-element codes that use automatic differentiation. They give derivation guidelines and apply the idea to several solid-mechanics constraints beyond the usual boundary conditions. The numerical examples report observed convergence rates on hyperelasticity and contact-type problems, which is useful to see in practice. The implementation angle is the clearest strength: once the minimization is set up, adding the constraint looks straightforward and does not require separate multiplier variables in the final code. The soft spot is the jump to nonlinear problems. The equivalence between saddle-point and minimization is standard for linear cases, but the paper does not supply a general argument that the stabilization parameters chosen from linear analysis remain sufficient when the tangent operator becomes nonlinear or the active set changes with inequalities. The examples are on specific meshes and loads, so they do not rule out parameter retuning or loss of rates in broader use. No full a-priori error analysis appears. This is aimed at people who write or maintain nonlinear finite-element solvers for constrained mechanics and want a flexible way to add Nitsche terms without heavy code changes. It is not a deep theoretical advance, but the recipe is concrete enough that a serious editor should send it to review so the nonlinear stability question can be examined properly.

Referee Report

2 major / 2 minor

Summary. The manuscript presents guidelines for deriving Nitsche finite element methods to enforce equality and inequality constraints acting on mechanical quantities. It begins with a stabilized saddle-point formulation using Lagrange multipliers and rewrites this as a general unconstrained minimization problem. The resulting form is positioned as suitable for direct incorporation into nonlinear finite element methods via automatic differentiation. Numerical examples from solid mechanics (including hyperelasticity and contact) are provided to demonstrate observed convergence rates.

Significance. If the minimization reformulation preserves stability and convergence when applied to nonlinear problems, the work would supply a practical, code-friendly route for adding constraints to existing nonlinear solvers without altering the core Newton loop. The explicit guidelines for deriving new Nitsche terms and the emphasis on automatic differentiation constitute a concrete implementation advantage over classical penalty or Lagrange-multiplier approaches. The numerical results supply initial evidence that optimal rates can be recovered on the tested problems.

major comments (2)

[Abstract and nonlinear formulation] Abstract and formulation section: the claim that the stabilized saddle-point problem can be rewritten as a minimization whose stationarity yields a stable Nitsche method for nonlinear mechanics lacks a general argument. The equivalence is standard for linear problems, but no analysis is supplied showing that the stabilization parameter chosen from the linear case remains sufficient once the tangent operator becomes nonlinear or the active set for inequalities changes.
[Numerical examples] Numerical examples section: convergence rates are reported for specific meshes and loads, yet no information is given on the precise choice or adaptation of the stabilization parameter, mesh refinement strategy, or exclusion of data near singularities. Without these details the observed rates cannot be used to support the general claim of retained optimality for nonlinear problems.

minor comments (2)

The notation used for the stabilization parameter and the Nitsche terms should be made uniform between the linear saddle-point derivation and the nonlinear minimization form to avoid reader confusion.
Figure captions would benefit from explicit statements of the polynomial degree, mesh size sequence, and load magnitude used in each convergence study.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments. We respond point-by-point to the major comments below and indicate the revisions planned for the next version of the manuscript.

read point-by-point responses

Referee: Abstract and formulation section: the claim that the stabilized saddle-point problem can be rewritten as a minimization whose stationarity yields a stable Nitsche method for nonlinear mechanics lacks a general argument. The equivalence is standard for linear problems, but no analysis is supplied showing that the stabilization parameter chosen from the linear case remains sufficient once the tangent operator becomes nonlinear or the active set for inequalities changes.

Authors: We agree that the manuscript supplies only a formal algebraic equivalence between the stabilized saddle-point system and the stationarity condition of the minimization problem; this equivalence holds regardless of linearity because it follows directly from completing the square in the augmented Lagrangian. No general stability or convergence proof is given for the nonlinear regime or for changing active sets. The paper instead positions the minimization form as a practical device for automatic differentiation in existing nonlinear solvers and supports its use through numerical examples. We will revise the abstract and formulation section to state explicitly that the derivation is formal and that retention of optimal rates is observed numerically rather than proved in general. revision: yes
Referee: Numerical examples section: convergence rates are reported for specific meshes and loads, yet no information is given on the precise choice or adaptation of the stabilization parameter, mesh refinement strategy, or exclusion of data near singularities. Without these details the observed rates cannot be used to support the general claim of retained optimality for nonlinear problems.

Authors: We accept the criticism. The revised manuscript will add an explicit subsection describing (i) the concrete formula used to select the stabilization parameter from local mesh size and tangent stiffness estimates, (ii) the uniform h-refinement sequence employed, and (iii) the precise rule for excluding a fixed number of elements adjacent to singularities or active contact zones when computing global error norms. These additions will make the reported rates reproducible and will clarify the scope of the numerical evidence. revision: yes

Circularity Check

0 steps flagged

Derivation proceeds from standard saddle-point to minimization without self-reduction

full rationale

The paper begins with the established stabilized saddle-point formulation using a Lagrange multiplier and rewrites it as an unconstrained minimization problem whose stationarity recovers the Nitsche terms. This equivalence is standard for linear problems and is extended to nonlinear mechanics via the same algebraic rewriting, with stability and convergence supported by numerical examples rather than by construction. No step defines a quantity in terms of the final result, fits a parameter to data that is then relabeled as a prediction, or relies on a self-citation chain whose content is unverified outside the present work. The central claim therefore remains independent of its own outputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The approach rests on the standard assumption that a constrained mechanical problem admits a stable saddle-point formulation that can be discretized and stabilized in the usual finite-element manner; no new free parameters, invented entities, or ad-hoc axioms are introduced in the abstract.

axioms (1)

domain assumption A constrained mechanical problem can be written as a stabilized saddle-point finite-element formulation whose Lagrange multiplier enforces the constraint.
This is the explicit starting point stated in the abstract for deriving the Nitsche method.

pith-pipeline@v0.9.0 · 5409 in / 1178 out tokens · 50831 ms · 2026-05-15T15:50:40.130033+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

36 extracted references · 36 canonical work pages

[1]

J. Nitsche, Über ein Variationsprinzip zur Lösung von Dirichlet- Prob- lemen bei Verwendung von Teilräumen, die keinen R andbedingungen unterworfen sind, in: Abhandlungen aus dem mathematischen Seminar der Universität Hamburg, Vol. 36, Springer, 1971, pp. 9–15

work page 1971
[2]

Courant, Variational methods for the solution of problems of equi- librium and vibrations, Bull

R. Courant, Variational methods for the solution of problems of equi- librium and vibrations, Bull. Amer. Math. Soc. (1943) 1–23

work page 1943
[3]

Babuška, The finite element method with Lagrangian multipliers, Nu- merische Mathematik 20 (3) (1973) 179–192

I. Babuška, The finite element method with Lagrangian multipliers, Nu- merische Mathematik 20 (3) (1973) 179–192

work page 1973
[4]

Gustafsson, R

T. Gustafsson, R. Stenberg, Finite element methods for elastic contact: penalty and Nitsche, Rakenteiden Mekaniikka 58 (2) (2025) 46–58. 28

work page 2025
[5]

Becker, P

R. Becker, P. Hansbo, R. Stenberg, A finite element method for domain decomposition with non-matching grids, ESAIM: Mathematical Mod- elling and Numerical Analysis 37 (2) (2003) 209–225

work page 2003
[6]

Hansbo, M

P. Hansbo, M. G. Larson, Discontinuous Galerkin methods for incom- pressible and nearly incompressible elasticity by Nitsche’s method, Com- puter methods in applied mechanics and engineering 191 (17-18) (2002) 1895–1908

work page 2002
[7]

Burman, S

E. Burman, S. Claus, P. Hansbo, M. G. Larson, A. Massing, CutFEM: discretizing geometry and partial differential equations, International Journal for Numerical Methods in Engineering 104 (7) (2015) 472–501

work page 2015
[8]

Juntunen, R

M. Juntunen, R. Stenberg, Nitsche’s method for general boundary con- ditions, Mathematics of computation 78 (267) (2009) 1353–1374

work page 2009
[9]

Gustafsson, R

T. Gustafsson, R. Stenberg, J. Videman, Nitsche’s method for Kirchhoff plates, SIAM Journal on Scientific Computing 43 (3) (2021) A1651– A1670

work page 2021
[10]

Chouly, P

F. Chouly, P. Hild, A Nitsche-based method for unilateral contact prob- lems: numerical analysis, SIAM Journal on Numerical Analysis 51 (2) (2013) 1295–1307

work page 2013
[11]

Chouly, An adaptation of Nitsche’s method to the Tresca friction problem, Journal of Mathematical Analysis and Applications 411 (1) (2014) 329–339

F. Chouly, An adaptation of Nitsche’s method to the Tresca friction problem, Journal of Mathematical Analysis and Applications 411 (1) (2014) 329–339

work page 2014
[12]

Gustafsson, R

T. Gustafsson, R. Stenberg, J. Videman, Mixed and stabilized finite element methods for the obstacle problem, SIAM Journal on Numerical Analysis 55 (6) (2017) 2718–2744

work page 2017
[13]

Chouly, T

F. Chouly, T. Gustafsson, P. Hild, A Nitsche method for the elastoplas- tic torsion problem, ESAIM: Mathematical Modelling and Numerical Analysis 57 (3) (2023) 1731–1746

work page 2023
[14]

Gustafsson, R

T. Gustafsson, R. Stenberg, J. Videman, Nitsche’s method for the ob- stacle problem of clamped Kirchhoff plates, in: European Conference on Numerical Mathematics and Advanced Applications, Springer, 2017, pp. 407–415. 29

work page 2017
[15]

Fabre, C

M. Fabre, C. Pozzolini, Y. Renard, Nitsche-based models for the unilat- eral contact of plates, ESAIM: Mathematical Modelling and Numerical Analysis 55 (2021) S941–S967. doi:10.1051/m2an/2020063

work page doi:10.1051/m2an/2020063 2021
[16]

Benzaken, J

J. Benzaken, J. A. Evans, R. Tamstorf, Constructing Nitsche’s method for variational problems, Archives of Computational Methods in Engi- neering 31 (4) (2024)

work page 2024
[17]

R.Stenberg, Onsometechniquesforapproximatingboundaryconditions in the finite element method, Journal of Computational and Applied Mathematics 63 (1-3) (1995) 139–148

work page 1995
[18]

H. J. Barbosa, T. J. Hughes, Circumventing the Babuška-Brezzi condi- tion in mixed finite element approximations of elliptic variational in- equalities, Computer methods in applied mechanics and engineering 97 (2) (1992) 193–210

work page 1992
[19]

Burman, P

E. Burman, P. Hansbo, M. G. Larson, R. Stenberg, Galerkin least squaresfiniteelementmethodfortheobstacleproblem, ComputerMeth- ods in Applied Mechanics and Engineering 313 (2017) 362–374

work page 2017
[20]

Chouly, P

F. Chouly, P. Hild, Y. Renard, Finite element approximation of contact and friction in elasticity, Springer, 2023

work page 2023
[21]

Burman, P

E. Burman, P. Hansbo, M. G. Larson, The augmented Lagrangian method as a framework for stabilised methods in computational me- chanics, Archives of Computational Methods in Engineering 30 (2023) 2579–2604

work page 2023
[22]

Gustafsson, Finite element methods for contact problems, Ph.D

T. Gustafsson, Finite element methods for contact problems, Ph.D. the- sis, Aalto University (2018)

work page 2018
[23]

Brezis, Functional analysis, Sobolev spaces and partial differential equations, Vol

H. Brezis, Functional analysis, Sobolev spaces and partial differential equations, Vol. 2, Springer, 2011

work page 2011
[24]

Gustafsson, R

T. Gustafsson, R. Stenberg, J. H. Videman, Nitsche’s method for unilat- eral contact problems, Portugaliae Mathematica 75 (3) (2019) 189–204

work page 2019
[25]

Gustafsson, R

T. Gustafsson, R. Stenberg, J. Videman, Error analysis of Nitsche’s mortar method, Numerische Mathematik 142 (4) (2019) 973–994. 30

work page 2019
[26]

Nitsche Methods for Constrained Problems in Mechanics

V. Kohonen, T. Gustafsson, Numerical experiments for "Nitsche Methods for Constrained Problems in Mechanics" (Apr. 2026). doi:10.5281/zenodo.18221987

work page doi:10.5281/zenodo.18221987 2026
[27]

Chouly, P

F. Chouly, P. Hild, Y. Renard, Symmetric and non-symmetric variants ofnitsche’smethodforcontactproblemsinelasticity: theoryandnumer- ical experiments, Mathematics of Computation 84 (293) (2015) 1089– 1112

work page 2015
[28]

Gustafsson, G

T. Gustafsson, G. D. McBain, scikit-fem: A Python package for finite element assembly, Journal of Open Source Software 5 (52) (2020) 2369

work page 2020
[29]

Bradbury, et al., JAX: Autograd and XLA, Astrophysics Source Code Library (2021) ascl–2111

J. Bradbury, et al., JAX: Autograd and XLA, Astrophysics Source Code Library (2021) ascl–2111

work page 2021
[30]

part 1: Finite element discretization and mixed reformulation, Math

F.BenBelgacem, C.Bernardi, A.Blouza, M.Vohralík, Ontheunilateral contact between membranes. part 1: Finite element discretization and mixed reformulation, Math. Model. Nat. Phenom. 4 (1) (2009) 21–43. doi:10.1051/mmnp/20094102. URLhttps://doi.org/10.1051/mmnp/20094102

work page doi:10.1051/mmnp/20094102 2009
[31]

part 2: a posteriori analysis and numerical experiments, IMA Journal of Numerical Analysis 32 (3) (2012) 1147– 1172

F.BenBelgacem, C.Bernardi, A.Blouza, M.Vohralík, Ontheunilateral contact between membranes. part 2: a posteriori analysis and numerical experiments, IMA Journal of Numerical Analysis 32 (3) (2012) 1147– 1172

work page 2012
[32]

L. A. Caffarelli, The obstacle problem revisited, Journal of Fourier Anal- ysis and Applications 4 (4) (1998) 383–402

work page 1998
[33]

Gustafsson, R

T. Gustafsson, R. Stenberg, J. Videman, On finite element formulations for the obstacle problem–mixed and stabilised methods, Computational Methods in Applied Mathematics 17 (3) (2017) 413–429

work page 2017
[34]

F. K. Bogner, The generation of interelement compatible stiffness and mass matrices by the use of interpolation formulae, in: Proc. Conf. Ma- trix Methods in Struct. Mech., Airforce Inst. Of Tech., Wright Patterson AF Base, Ohio, 1965

work page 1965
[35]

S. A. Nazarov, A. Stylianou, G. Sweers, Hinged and supported plates with corners, Zeitschrift für angewandte Mathematik und Physik 63 (5) (2012) 929–960. 31

work page 2012
[36]

Mlika, Y

R. Mlika, Y. Renard, F. Chouly, An unbiased Nitsche’s formulation of largedeformationfrictionalcontactandself-contact, ComputerMethods in Applied Mechanics and Engineering 325 (2017) 265–288. 32

work page 2017