Nonlinear Model Updating of Aerospace Structures via Taylor-Series Reduced-Order Models

Evangelos Papatheou; Jake Hollins; Konstantinos Agathos; Nikolaos D. Tantaroudas

arxiv: 2604.03788 · v3 · submitted 2026-04-04 · 💻 cs.CE · cs.NA· cs.SY· eess.SY· math-ph· math.MP· math.NA

Nonlinear Model Updating of Aerospace Structures via Taylor-Series Reduced-Order Models

Nikolaos D. Tantaroudas , Jake Hollins , Konstantinos Agathos , Evangelos Papatheou This is my paper

Pith reviewed 2026-05-13 17:10 UTC · model grok-4.3

classification 💻 cs.CE cs.NAcs.SYeess.SYmath-phmath.MPmath.NA

keywords nonlinear model updatingreduced-order modelsTaylor-series expansionfinite element modelsprojection basis adaptationaerospace structuresamplitude-dependent dynamics

0 comments

The pith

A Taylor-series reduced-order model combined with basis adaptation lets nonlinear finite element updating match experimental amplitude-dependent frequencies and recover stiffness parameters more accurately than linear methods.

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

The paper develops a nonlinear model updating procedure that starts from the full finite element equations of motion with polynomial stiffness terms and proportional damping. It converts these to a first-order autonomous system, builds second- and third-order Taylor operators for the nonlinear forces, and projects them onto a reduced eigenvector basis. The basis itself is adapted through a generalized Cayley transform on the complex unitary group so that the reduced modes align with measured data. Numerical tests on a wingbox panel show that the resulting low-dimensional model reproduces amplitude-dependent natural frequencies and modal assurance criterion values that remain inaccessible to purely linear updating schemes, while also returning the underlying stiffness parameters with greater accuracy.

Core claim

The central claim is that the nonlinear equations of motion, after Taylor expansion of the internal forces to second and third order and projection onto an adapted biorthogonal eigenvector basis obtained via the generalized Cayley transform, yield a reduced-order model whose parameters can be updated to match experimental measurements at multiple amplitudes, thereby recovering stiffness values that linear updating alone cannot obtain.

What carries the argument

Second- and third-order Taylor operators for the nonlinear internal forces, projected onto a projection basis adapted by the generalized Cayley transform on the complex unitary group.

If this is right

The nonlinear reduced-order model reproduces amplitude-dependent natural frequencies observed in vibration tests.
Modal assurance criterion values between model and experiment improve for nonlinear regimes compared with linear updating.
Stiffness parameters are recovered with measurably higher accuracy than linear schemes alone can achieve.
The framework extends directly to any structure whose nonlinearity can be expressed as a low-order polynomial stiffness term.

Where Pith is reading between the lines

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

The same adaptation step could be applied iteratively as test amplitude increases, allowing progressive refinement of the model.
If higher-order Taylor terms prove necessary, the same projection machinery would still apply with only an increase in the number of retained nonlinear coefficients.
The approach suggests that updating could be performed directly on data from stepped-sine or broadband tests rather than requiring separate linear and nonlinear identification stages.

Load-bearing premise

The assumption that a second- and third-order Taylor-series expansion is sufficient to represent the nonlinear stiffness forces inside the structure.

What would settle it

If the updated nonlinear reduced-order model fails to reproduce the measured amplitude-dependent natural frequencies or returns stiffness parameters no more accurate than a linear scheme on the same wingbox data, the central claim would be refuted.

read the original abstract

Finite element model updating is a mature discipline for linear structures, yet its extension to nonlinear regimes remains an open challenge. This paper presents a methodology that combines nonlinear model order reduction (NMOR) based on Taylor-series expansion of the equations of motion with the projection-basis adaptation scheme recently proposed by Hollins et al. [2026] for linear model updating. The structural equations of motion, augmented with proportional (Rayleigh) damping and polynomial stiffness nonlinearity, are recast as a first-order autonomous system whose Jacobian possesses complex eigenvectors forming a biorthogonal basis. Taylor operators of second and third order are derived for the nonlinear internal forces and projected onto the reduced eigenvector basis, yielding a low-dimensional nonlinear reduced-order model (ROM). The Cayley transform, generalised from the real orthogonal to the complex unitary group, parametrises the adaptation of the projection basis so that the ROM mode shapes optimally correlate with experimental measurements. The resulting nonlinear model-updating framework is applied to a representative wingbox panel model. Numerical studies demonstrate that the proposed approach captures amplitude-dependent natural frequencies and modal assurance criterion(MAC) values that a purely linear updating scheme cannot reproduce, while recovering the underlying stiffness parameters with improved accuracy.

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

This paper sketches a nonlinear extension of model updating for aerospace structures using Taylor-series ROMs plus Cayley basis adaptation, but the abstract alone gives no equations or numbers to check if it actually works.

read the letter

Hi, the main point is that the paper takes the standard linear model updating setup and tries to push it into nonlinear territory for things like wing panels. It uses second- and third-order Taylor expansions to build a reduced-order model of the nonlinear forces, then adapts the projection basis with a generalized Cayley transform so the reduced modes line up better with measurements. The claim is that this recovers amplitude-dependent frequencies and stiffness parameters more accurately than a linear scheme on a wingbox test case. That combination looks like the actual new piece, since the adaptation step itself comes from a co-author's earlier linear paper. The outline is clear enough on paper: start from the equations with polynomial stiffness, turn them into a first-order system, project the Taylor operators, and parametrize the basis change on the complex unitary group. It does a reasonable job naming the open problem and showing a workable path forward without overclaiming the math. The soft spots are straightforward. We have only the abstract, so there are no derivations to inspect for the projected nonlinear operators, no confirmation that second- and third-order terms are enough for the internal forces, and no data, error bars, or MAC values to see whether the reported improvements are real or just from the adaptation fitting. The circularity with the 2026 Hollins work is moderate but real, since any stiffness recovery gains could trace back to that prior procedure rather than the nonlinear wrapper. This is for people who already work on structural dynamics and finite-element updating in aerospace, especially those who need to handle amplitude effects. A reader who wants a concrete method sketch could pull an idea or two from it, but anyone needing reproducible results would have to wait for the full text. It deserves a serious referee because the problem is relevant and the approach sits on established tools; a review could sort out whether the Taylor order and adaptation actually deliver on the claims once the equations and numbers are on the table.

Referee Report

2 major / 1 minor

Summary. The manuscript proposes a nonlinear finite element model updating framework for aerospace structures. It combines Taylor-series expansions (second- and third-order) of the nonlinear internal forces in the equations of motion with a projection-basis adaptation scheme that generalizes the Cayley transform to the complex unitary group. The resulting low-dimensional nonlinear ROM is applied to a wingbox panel model; the abstract claims that numerical studies show the method reproduces amplitude-dependent natural frequencies and MAC values that linear updating cannot, while recovering underlying stiffness parameters with improved accuracy.

Significance. If the numerical demonstrations hold, the work would advance nonlinear model updating by providing an efficient ROM-based route to match experimental modal data under amplitude-dependent conditions. The integration of biorthogonal complex eigenvectors with the generalized Cayley parametrization is a technically interesting extension of prior linear adaptation methods, potentially useful for aerospace applications where nonlinear stiffness effects matter.

major comments (2)

[Abstract] Abstract: the central claim that 'numerical studies demonstrate that the proposed approach captures amplitude-dependent natural frequencies and modal assurance criterion (MAC) values that a purely linear updating scheme cannot reproduce' is load-bearing, yet the abstract supplies no quantitative metrics, error values, frequency-shift tables, MAC comparisons, or exclusion criteria for the wingbox results, preventing assessment of whether the second- and third-order Taylor operators actually suffice or merely overfit.
[Abstract] Abstract: the stiffness-parameter recovery improvement is asserted to result from the nonlinear ROM extension, but the projection-basis adaptation is taken directly from Hollins et al. [2026]; without an explicit isolation of the nonlinear contribution (e.g., a comparison holding the adaptation fixed and varying only the Taylor projection), it is unclear whether the reported accuracy gain is attributable to the new elements or to the prior fitting procedure.

minor comments (1)

[Abstract] The abstract refers to 'polynomial stiffness nonlinearity' and 'Taylor operators of second and third order' without indicating how the truncation order was selected or validated against the target amplitude range for the wingbox.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. We address each major point below and will revise the abstract and numerical studies section to provide greater clarity and quantitative support for our claims.

read point-by-point responses

Referee: [Abstract] Abstract: the central claim that 'numerical studies demonstrate that the proposed approach captures amplitude-dependent natural frequencies and modal assurance criterion (MAC) values that a purely linear updating scheme cannot reproduce' is load-bearing, yet the abstract supplies no quantitative metrics, error values, frequency-shift tables, MAC comparisons, or exclusion criteria for the wingbox results, preventing assessment of whether the second- and third-order Taylor operators actually suffice or merely overfit.

Authors: We agree that the abstract would be strengthened by including specific quantitative metrics. In the revised manuscript we will add concise numerical results from the wingbox studies, such as the observed amplitude-dependent frequency shifts (e.g., percentage deviations from linear predictions) and MAC value improvements (e.g., from 0.XX to 0.YY), together with the associated stiffness-parameter recovery errors. These additions will allow readers to directly evaluate the sufficiency of the second- and third-order Taylor operators. revision: yes
Referee: [Abstract] Abstract: the stiffness-parameter recovery improvement is asserted to result from the nonlinear ROM extension, but the projection-basis adaptation is taken directly from Hollins et al. [2026]; without an explicit isolation of the nonlinear contribution (e.g., a comparison holding the adaptation fixed and varying only the Taylor projection), it is unclear whether the reported accuracy gain is attributable to the new elements or to the prior fitting procedure.

Authors: The projection-basis adaptation follows the generalized Cayley scheme of Hollins et al., yet the core contribution of the present work is the incorporation of Taylor-series nonlinear terms into the reduced-order model. To isolate this effect we will add, in the revised numerical studies, a controlled comparison that holds the adaptation fixed and contrasts results obtained with a purely linear ROM against those obtained with the second- and third-order Taylor ROMs. This will demonstrate that the reported gains in stiffness-parameter accuracy arise specifically from the nonlinear modeling extension. revision: yes

Circularity Check

1 steps flagged

Basis adaptation imported from co-author prior work creates moderate circularity burden on accuracy claims

specific steps

self citation load bearing [Abstract]
"This paper presents a methodology that combines nonlinear model order reduction (NMOR) based on Taylor-series expansion of the equations of motion with the projection-basis adaptation scheme recently proposed by Hollins et al. [2026] for linear model updating. ... The Cayley transform, generalised from the real orthogonal to the complex unitary group, parametrises the adaptation of the projection basis so that the ROM mode shapes optimally correlate with experimental measurements."

The mechanism asserted to enable optimal correlation with measurements and improved parameter recovery is the adaptation scheme imported wholesale from Hollins et al. [2026] (co-author overlap via Jake Hollins). The paper does not re-derive or independently validate the generalised Cayley parametrisation; therefore the numerical studies' demonstration of superiority over linear updating is not self-contained and may be forced by the prior work's fitting procedure.

full rationale

The paper's central claim of improved stiffness recovery and amplitude-dependent frequency/MAC capture rests on combining new Taylor-series NMOR with the projection-basis adaptation from Hollins et al. [2026]. While the nonlinear ROM extension is independent, the load-bearing step of parametrising basis adaptation via generalised Cayley transform to 'optimally correlate' with measurements is cited directly to overlapping-author prior work without re-derivation or external verification here. This creates partial circularity: claimed superiority over linear updating may reduce to the effectiveness of the prior fitting scheme rather than the new nonlinear operators. No other self-definitional or fitted-prediction reductions are visible from the abstract.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard structural dynamics assumptions plus the specific choice of Taylor truncation order and the extension of the Cayley transform; no new physical entities are postulated.

free parameters (1)

stiffness parameters
Underlying stiffness values recovered during the updating process to match experimental measurements on the wingbox model.

axioms (2)

domain assumption Structural equations of motion augmented with proportional Rayleigh damping and polynomial stiffness nonlinearity can be recast as a first-order autonomous system whose Jacobian has complex eigenvectors forming a biorthogonal basis.
Invoked in the abstract as the starting point for deriving the Taylor operators and reduced-order model.
ad hoc to paper Taylor operators of second and third order are sufficient to approximate the nonlinear internal forces for the target amplitude range.
Chosen and applied without further justification or convergence analysis visible in the abstract.

pith-pipeline@v0.9.0 · 5514 in / 1689 out tokens · 53066 ms · 2026-05-13T17:10:54.950632+00:00 · methodology

Nonlinear Model Updating of Aerospace Structures via Taylor-Series Reduced-Order Models

Core claim

What carries the argument

If this is right

Where Pith is reading between the lines

Load-bearing premise

What would settle it

discussion (0)