Generalized Bending: Extended Theories

• Generalized bending is an umbrella framework that expands classical curvature theories by incorporating anisotropy, non-holonomic kinematics, and enriched energy functionals.
• It refines models for bilayers, shells, beams, and rods to capture large deformations, size effects, and symmetry-induced couplings with validated numerical methods.
• Applications span engineered metamaterials, biological membranes, and advanced structural elements, offering enhanced predictive power and practical design insights.

Generalized bending denotes a family of extensions of classical bending descriptions in which curvature response is enlarged beyond small-angle, isotropic, or purely geometric settings. In the cited literature, the term refers to discrete bilayer actuation on arbitrary initial shapes, anisotropic and fiber-enhanced shell theories, beam and rod models with higher-order or non-holonomic kinematics, generalized Willmore- and Helfrich-type energies, homogenized bending-torsion limits from $3D$ elasticity, and symmetry-induced couplings such as axial-bending in architected lattices (Alben, 2011, Duong et al., 2020, Savitha et al., 2022, Crespo et al., 17 Jul 2025, Friedrich, 2019, Zhong et al., 2024). This suggests that generalized bending is best understood as an umbrella notion: bending remains the central curvature-driven response, but the admissible kinematics, constitutive variables, topology, or ambient geometry are broadened.

1. Taxonomy of generalization

Across the literature, generalization proceeds along several distinct axes. Some works generalize the geometry of the reference configuration, for example from rectangles to arbitrary planar bilayer shapes or from initially flat shells to initially curved ones (Alben, 2011, Savitha et al., 2022). Others generalize the kinematics, introducing in-plane fiber curvature, micro-distortion fields, or higher-order tensors beyond displacement alone (Duong et al., 2020, Crespo et al., 17 Jul 2025). A third class generalizes the energy functional, replacing isotropic Helfrich or Willmore integrands by anisotropic, nonlocal, or lower-order-perturbed functionals (Yang, 2017, Friedrich, 2019). A fourth class generalizes the ambient or support structure, as in branched bending on piecewise totally geodesic complexes or bending angles computed in curved spacetime backgrounds (Monroe, 23 Apr 2026, Carvalho et al., 2021).

Setting	Generalization	Representative formulation
Bilayers and shells	Arbitrary shape, anisotropy, in-plane fiber bending	Discrete bilayer energy; $W_{\text{new}}$; $\bar{\boldsymbol{b}}$
Rods and beams	Large angles, variable properties, non-holonomic fields	$\frac{d\theta}{ds}$-equation; $u,P,N$ framework
Surface energies	Anisotropic or generalized curvature functionals	$U(\Sigma)$; generalized Willmore equation
Lattices and other geometries	Symmetry-breaking couplings, branched support	Axial-bending coupling; branched bending complexes

The resulting terminology is therefore domain-dependent. In continuum mechanics, generalized bending usually means a constitutive or kinematic enrichment of plate, shell, rod, or beam theory. In geometric analysis, it usually means a broader curvature functional or a broader class of admissible surfaces. In other domains, it can denote generalized deflection mechanisms, such as branched bending deformations or finite-distance bending angles.

2. Bilayers, plates, and shells

A direct mechanical realization appears in actuated bilayers with arbitrary initial planar shape. A triangular mesh is used to model the central surface of the substrate layer, with stretching energy

$$E_s = \frac{1}{2} C_s \sum_{i, j} \left( \left| \mathbf{r}_i - \mathbf{r}_j \right| - d_{eq} \right)^2,$$

bending energy

$$E_b = C_b \sum_{\alpha, \beta} \left( 1 - \mathbf{n}_\alpha \cdot \mathbf{n}_\beta \right),$$

and total bilayer energy

$E_T = E_s + E_{s,a} + 2 E_b.$

The model assumes equal thickness $h$ and elasticity for substrate and actuated layers, incorporates an actuation strain $W_{\text{new}}$0, and is minimized by the LM-BFGS quasi-Newton method with analytical gradients (Alben, 2011). For rectangular bilayers, the predicted equilibrium curvature is $W_{\text{new}}$1, which matches simulation in the bulk. For simple polyiamonds, the reported behaviors include bending along the shortest or longest extent, inward bending at $W_{\text{new}}$2-degree corners, curvature intensification near boundaries, partitioned bending, and nearly conical zones. Principal-curvature maps and minimum-curvature direction fields are used to identify these regimes.

In generalized Kirchhoff-Love shell theory with embedded fibers, the principal extension is the addition of in-plane bending to the classical stretching and out-of-plane bending modes. The central new kinematic object is the in-plane curvature tensor

$W_{\text{new}}$3

where $W_{\text{new}}$4 is the in-surface director orthogonal to the fiber direction. Relative out-of-plane and in-plane curvature changes are then

$W_{\text{new}}$5

The internal power identifies a symmetric effective stress tensor and symmetric in-plane and out-of-plane moment tensors; the weak form is given explicitly for rotation-free computational formulations (Duong et al., 2020). The computational extension employs isogeometric NURBS shape functions to satisfy the required $W_{\text{new}}$6-continuity, includes a stabilization scheme for possible material instability due to fiber compression, and shows that in-plane bending stiffness is essential for mesh-independent finite shear-band widths in bias-extension simulations (Duong et al., 2021).

A further constitutive generalization is the nonlinear anisotropic shell bending model formulated directly in surface form,

$W_{\text{new}}$7

with

$W_{\text{new}}$8

This model is based on the principal curvatures and their directions in the initial configuration, admits different bending moduli along those directions, and can handle large deformations and initially curved surfaces (Savitha et al., 2022). Five elementary nonlinear bending test cases are reported—rigid body rotation, counter-bending, inflation, pure bending, and torsion—and only the proposed model passes all of them. By contrast, the Koiter model is stated to be accurate only for initially planar shells at small deformations, while Canham and Helfrich models are reported as not initially stress-free for curved shells.

3. Rods, tubes, and beam theories

For arbitrary rods under self-weight and external loading, a continuous general bending equation is derived from a segment-wise force balance: $W_{\text{new}}$9 Its integrated solution is expressed through

$\bar{\boldsymbol{b}}$0

This framework allows arbitrary spatial dependence of mass, cross-section, and strength, does not rely on small-angle approximations, and was applied to iron rulers, rods with equally spaced PVC “leaves,” and palm fronds (Abdullah, 2016). The data report excellent match for homogeneous rods at fixed angles $\bar{\boldsymbol{b}}$1, $\bar{\boldsymbol{b}}$2, and $\bar{\boldsymbol{b}}$3, and good agreement for the palm-frond analogues.

For thin cylindrical tubes, generalized bending appears as a non-affine isometric mode. When the tube length satisfies $\bar{\boldsymbol{b}}$4 with pinch persistence length $\bar{\boldsymbol{b}}$5, small normal indentation can be accommodated without in-plane stretching. The bending energy of the non-affine response is

$\bar{\boldsymbol{b}}$6

and the local bending stiffness peaks at the center and decreases by a universal factor of $\bar{\boldsymbol{b}}$7 at either end (Efrati et al., 2016). This bell-shaped stiffness profile is proposed as a direct way to determine the bending modulus of very thin tubes shorter than their pinch persistence length.

Higher-order and generalized-continuum beam theories extend classical Euler-Bernoulli and Timoshenko models by adding defect-sensitive fields. In the generalized Euler-Bernoulli beam, the energy density is

$\bar{\boldsymbol{b}}$8

and the disclination density is linked to the third derivative,

$\bar{\boldsymbol{b}}$9

The static equilibrium equation is

$\frac{d\theta}{ds}$0

This model exhibits curvature linked to a third-order derivative of the displacement, but no torsion. The generalized Timoshenko beam introduces three scalar fields and supports both curvature and torsion, with

$\frac{d\theta}{ds}$1

Its internal energy includes $\frac{d\theta}{ds}$2, $\frac{d\theta}{ds}$3, $\frac{d\theta}{ds}$4, and $\frac{d\theta}{ds}$5 terms (Crespo et al., 24 Apr 2025).

A broader hierarchical formulation introduces the macroscopic displacement $\frac{d\theta}{ds}$6, the micro-distortion tensor $\frac{d\theta}{ds}$7, and the third-order tensor $\frac{d\theta}{ds}$8. Three regimes are distinguished: holonomic, semi-holonomic, and non-holonomic. The holonomic case reduces to a higher-order Euler-Bernoulli beam; the semi-holonomic case generalises the Timoshenko beam; and the non-holonomic case naturally incorporates both dislocations and disclinations. The holonomic and semi-holonomic models are described as singular limits of the non-holonomic model through large penalty coefficients (Crespo et al., 17 Jul 2025).

Finally, homogenization from $\frac{d\theta}{ds}$9 nonlinear elasticity yields a generalized bending-torsion rod theory without any periodicity assumption. In the bending regime,

$u,P,N$0

and the $u,P,N$1-limit has the form

$u,P,N$2

The limiting quadratic form is defined abstractly by a relaxation problem and captures anisotropic, inhomogeneous, and non-periodic generalized bending-torsion response (Marohnic et al., 2014).

4. Pure bending measures and generalized curvature energies

A central conceptual issue is whether bending can be measured independently of stretching. For soft plates, this is addressed by introducing bending-neutral deformations, finite incremental changes of shape that bear no further bending. Their geometric compatibility condition is

$u,P,N$3

A tensorial pure measure of bending is then proposed: $u,P,N$4 It is invariant under bending-neutral deformations, symmetric, positive-definite, and reduces in the rod limit to Antman’s pure bending measure (Virga, 2023). The same source states explicitly that mean curvature $u,P,N$5 and Gaussian curvature $u,P,N$6 are not invariant under bending-neutral deformations. This directly challenges the common identification of bending with the present-surface curvatures alone in extensible plates.

For lipid vesicles, an anisotropic, scale-invariant curvature energy extends the Canham model: $u,P,N$7 This recovers Canham when $u,P,N$8. The paper proves the genus-dependent lower bound

$u,P,N$9

shows that there is a unique minimizer, up to scaling, over embedded ring tori for the anisotropic energy, and states that elevated anisotropy favors a transition from spherical-, to ellipsoidal-, and then to biconcave-shaped surfaces (Yang, 2017). By contrast, the presence of spontaneous curvature in the Helfrich energy is reported to obstruct the existence of a minimizer over embedded ring tori.

Generalized Willmore theory broadens bending functionals still further. A critical point solves

$U(\Sigma)$0

and a representative functional is

$U(\Sigma)$1

The existence theory is carried out in the class of haunted, branched, immersed bubble trees; area-constrained minimizers exist by direct minimization, and area-volume constrained membranes exist under

$U(\Sigma)$2

Using regularity arguments linked to work of Mondino and Rivière, critical points are smooth away from finitely many points (Friedrich, 2019).

5. Size effects, symmetry breaking, and computation

In cylindrical bending of generalized continua, the relaxed micromorphic model is distinguished by its thin-specimen limit. Classical linear elasticity gives

$U(\Sigma)$3

so bending stiffness vanishes as $U(\Sigma)$4. By contrast, Cosserat, gradient elasticity, and classical micromorphic models are reported to exhibit unphysical unbounded bending stiffness as $U(\Sigma)$5. The relaxed micromorphic model instead satisfies

$U(\Sigma)$6

so the stiffness remains finite for arbitrarily thin plates (Rizzi et al., 2020). The paper identifies this boundedness as a key advantage for physical interpretation and parameter identification.

In $U(\Sigma)$7 lattices, generalized bending appears as a symmetry-induced mechanical coupling. By extending symmetry breaking to mirror and inversion symmetries, a broader range of couplings is identified beyond axial-twist, most notably axial-bending (AB) coupling. The generalized micropolar setting uses displacement and rotation fields

$U(\Sigma)$8

with strain and curvature tensors

$U(\Sigma)$9

The paper states that chirality is not a prerequisite for AB or even AT coupling in $$E_s = \frac{1}{2} C_s \sum_{i, j} \left( \left| \mathbf{r}_i - \mathbf{r}_j \right| - d_{eq} \right)^2,$$0 lattices, and that the strength of these couplings is only weakly correlated with the degree of chirality (Zhong et al., 2024). This is an explicit correction to a common design intuition in metamaterials.

For numerical treatment of generalized plate bending formulations, a unified abstract operator framework is written as

$$E_s = \frac{1}{2} C_s \sum_{i, j} \left( \left| \mathbf{r}_i - \mathbf{r}_j \right| - d_{eq} \right)^2,$$1

It covers primal, mixed, and ultraweak formulations, both continuous and discrete, and allows flexible continuity restrictions across elements. Applied to the Kirchhoff-Love plate model, it yields three primal hybrid and two mixed hybrid methods, and incorporates Morley, Zienkiewicz triangular, and Hellan-Herrmann-Johnson elements as conforming hybrid schemes (Heuer, 2024). The numerical results reported in the data include optimal lowest-order convergence, $$E_s = \frac{1}{2} C_s \sum_{i, j} \left( \left| \mathbf{r}_i - \mathbf{r}_j \right| - d_{eq} \right)^2,$$2 energy convergence and $$E_s = \frac{1}{2} C_s \sum_{i, j} \left( \left| \mathbf{r}_i - \mathbf{r}_j \right| - d_{eq} \right)^2,$$3 $$E_s = \frac{1}{2} C_s \sum_{i, j} \left( \left| \mathbf{r}_i - \mathbf{r}_j \right| - d_{eq} \right)^2,$$4-convergence for the Morley-type method, and $$E_s = \frac{1}{2} C_s \sum_{i, j} \left( \left| \mathbf{r}_i - \mathbf{r}_j \right| - d_{eq} \right)^2,$$5 or $$E_s = \frac{1}{2} C_s \sum_{i, j} \left( \left| \mathbf{r}_i - \mathbf{r}_j \right| - d_{eq} \right)^2,$$6 convergence for the HHJ-type moment variable depending on enrichment.

6. Cross-disciplinary uses of the term

The term also appears outside structural mechanics. In semiflexible polymers, bending stiffness enters as an energetic penalty

$$E_s = \frac{1}{2} C_s \sum_{i, j} \left( \left| \mathbf{r}_i - \mathbf{r}_j \right| - d_{eq} \right)^2,$$7

and a $$E_s = \frac{1}{2} C_s \sum_{i, j} \left( \left| \mathbf{r}_i - \mathbf{r}_j \right| - d_{eq} \right)^2,$$8 coarse-grained homopolymer undergoes stiffness-driven ground-state changes: compact globules for $$E_s = \frac{1}{2} C_s \sum_{i, j} \left( \left| \mathbf{r}_i - \mathbf{r}_j \right| - d_{eq} \right)^2,$$9, rod-like bundles for $$E_b = C_b \sum_{\alpha, \beta} \left( 1 - \mathbf{n}_\alpha \cdot \mathbf{n}_\beta \right),$$0, toroidal structures for $$E_b = C_b \sum_{\alpha, \beta} \left( 1 - \mathbf{n}_\alpha \cdot \mathbf{n}_\beta \right),$$1, and large toroids to open loops for $$E_b = C_b \sum_{\alpha, \beta} \left( 1 - \mathbf{n}_\alpha \cdot \mathbf{n}_\beta \right),$$2 (Aierken et al., 2023). The study uses extended two-dimensional replica-exchange Monte Carlo in temperature and $$E_b = C_b \sum_{\alpha, \beta} \left( 1 - \mathbf{n}_\alpha \cdot \mathbf{n}_\beta \right),$$3.

In DNA minicircles, bending is coupled to twisting and bubble formation through a mesoscopic Hamiltonian,

$$E_b = C_b \sum_{\alpha, \beta} \left( 1 - \mathbf{n}_\alpha \cdot \mathbf{n}_\beta \right),$$4

For the smallest minicircle of $$E_b = C_b \sum_{\alpha, \beta} \left( 1 - \mathbf{n}_\alpha \cdot \mathbf{n}_\beta \right),$$5-bps, whose bending angle is $$E_b = C_b \sum_{\alpha, \beta} \left( 1 - \mathbf{n}_\alpha \cdot \mathbf{n}_\beta \right),$$6, base-pair disruptions occur with larger probability and a sizeable untwisting is obtained with the helical repeat showing a step-like increase at $$E_b = C_b \sum_{\alpha, \beta} \left( 1 - \mathbf{n}_\alpha \cdot \mathbf{n}_\beta \right),$$7 (Zoli, 2014). The reported interpretation is that fluctuational openings release the stress due to bending of the molecule backbone.

In gravitational lensing by Casimir wormholes, the generalized finite-distance bending angle is

$$E_b = C_b \sum_{\alpha, \beta} \left( 1 - \mathbf{n}_\alpha \cdot \mathbf{n}_\beta \right),$$8

with GUP-corrected variants studied for the KMM, DGS, and type II models (Carvalho et al., 2021). In the Benjamin-Ono equation, soliton bending refers to curved spatiotemporal trajectories obtained from a three-wave tau function,

$$E_b = C_b \sum_{\alpha, \beta} \left( 1 - \mathbf{n}_\alpha \cdot \mathbf{n}_\beta \right),$$9

with the parameter $E_T = E_s + E_{s,a} + 2 E_b.$0 controlling the direction and degree of bending of a W-shaped dark soliton (Singh et al., 2020).

A geometric-topological extension is branched bending in finite-volume hyperbolic manifolds. Here the deformation is supported on a piecewise totally geodesic complex of $E_T = E_s + E_{s,a} + 2 E_b.$1-dimensional faces meeting along $E_T = E_s + E_{s,a} + 2 E_b.$2-dimensional branching loci. In the hyperbolic setting,

$E_T = E_s + E_{s,a} + 2 E_b.$3

while in the projective setting,

$E_T = E_s + E_{s,a} + 2 E_b.$4

For the Borromean rings complement, the paper reports

$E_T = E_s + E_{s,a} + 2 E_b.$5

and

$E_T = E_s + E_{s,a} + 2 E_b.$6

recovering a special case of a theorem due to Menasco and Reid and extending classical bending as introduced by Johnson and Millson (Monroe, 23 Apr 2026).

Taken together, these developments show that generalized bending is not a single formalism but a recurrent strategy: one starts from a classical bending theory and enlarges its admissible geometry, constitutive space, or support class so that anisotropy, microstructure, topology, nonlocality, or finite-distance effects can be represented explicitly. In mechanics, this enlargement is often motivated by the need to separate stretching from bending or to admit size effects and defects; in geometric analysis, by minimization and regularity questions for broader curvature energies; and in other fields, by the need to describe bending-like deflection phenomena beyond the classical setting.