Nonlocal Contact Repulsion Model

• The model is a framework where repulsive responses depend on neighborhood interactions rather than a single local gap, integrating data from multiple contacts or surface regions.
• It unifies approaches from granular mechanics, pressureless viscous flows, and adhesive interfaces by using convolution kernels, global force balances, or overlap-volume energetics.
• The methodology enhances prediction accuracy in applications like granular compaction, cell motility, and nanoswitch stability by capturing many-body interaction effects.

Searching arXiv for the topic and closely related papers to ground the article in current literature. {"^^^^^^^^5query5 contact\" OR ^^^^^^^^5all:\5 repulsion\" OR ^^^^^^^^5all:\5 granular systems\" OR ^^^^^^^^5all:\5 viscous model driven by nonlocal attraction-repulsion\"","max_results":^^^^5all:\5query5^^^^,"sort_by":"relevance"} {"^^^^^^^^5query5 OR id:(&&&^^^^5all:\5^^^^&&&) OR id:(&&&^^^^5 OR all:\5^^^^&&&) OR id:(&&&^^^^5 OR all:\5^^^^&&&) OR id:(&&&^^^^5 OR all:\5^^^^&&&) OR id:(Wang et al., 10 Sep 2025) OR id:(Ferrari et al., 2024) OR id:(Bonetti et al., 2017)","max_results":^^^^5all:\5query5^^^^,"sort_by":"relevance"} The term nonlocal contact repulsion force model denotes, in the cited literature, a family of formulations in which repulsion at contact or near-contact is not determined solely by a pointwise gap, a local overlap, or a local pressure, but instead depends on a broader interaction structure: simultaneous loads on the same particle, convolution kernels over physical space, integral operators on a contact surface, globally balanced surface forces, or the overlap volume of entire bodies. In confined granular mechanics, it appears as a nonlocal correction to Hertz contact that couples all contacts incident on a grain (&&&^^^^5query5^^^^&&&). In pressureless viscous and cell-motility PDEs, it appears as a nonlocal attraction–repulsion term or a saturation-dependent drift reversal (&&&^^^^5 OR all:\5^^^^&&&, Ferrari et al., 2024). In confined surfaces, nanoswitches, and adhesive interfaces, it appears through global force balances, pairwise-summed repulsive interactions, or symmetric nonlocal kernel operators (&&&^^^^5 OR all:\5^^^^&&&, &&&^^^^5 OR all:\5^^^^&&&, Bonetti et al., 2017). In computational contact, it appears as the gradient of an overlap-dependent energy, evaluated by Fiber Monte Carlo (Wang et al., 10 Sep 2025).

^^^^5all:\5^^^^. Conceptual scope and meanings of nonlocality

The cited works do not use a single universal definition of nonlocality. Instead, they assign the label to distinct mechanisms that share one structural feature: the repulsive or contact response at one location depends on data extending beyond that location.

Setting	Nonlocal quantity	Repulsion/contact mechanism
Confined granular systems	PRESERVED_PLACEHOLDER_^^^^5query5^^^^, multi-contact displacements	Contact force depends on all simultaneous contacts on a particle
Pressureless viscous flow	PRESERVED_PLACEHOLDER_^^^^5all:\5^^^^	Singular short-range repulsion plus quadratic attraction
Cell motility	PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^	Drift changes sign when neighborhood is saturated
Confined surfaces in solution	PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^ and global force balance	Repulsion from all nearby surface sites
Adhesive contact	PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^	Contact traction and damage depend on surface neighborhoods
Energy-based computational contact	$$V_c=\int_{\Omega^1(\mathbf{u})\cap\Omega^2(\mathbf{u})}dx$$	Repulsive force is the gradient of overlap-volume energy

In the granular formulation, the classical assumption of independent contacts is explicitly rejected for confined systems: all simultaneous contact forces deform the particle globally, and those mesoscopic deformations modify each contact law (&&&^^^^5query5^^^^&&&). In the pressureless viscous model, the usual local barotropic pressure is replaced by a convolution force $\varrho\nabla(K*\varrho)$, where the singular kernel acts as short-range exclusion and the quadratic term provides confinement (&&&^^^^5 OR all:\5^^^^&&&). In the cell-motility model, repulsion is encoded through a nonlocal saturation coefficient inside the drift rather than through diffusion, so overcrowding reverses the direction of the deterministic transport (Ferrari et al., 2024).

A plausible implication is that “nonlocal contact repulsion” is best understood as a structural category rather than a single constitutive law. The precise mathematical realization depends on whether the problem is posed in discrete mechanics, continuum PDEs, interfacial statistical mechanics, or variational contact computation.

^^^^5 OR all:\5^^^^. Confined granular mechanics and the breakdown of independent contacts

The granular contact formulation of elastic confined spheres is the clearest mechanical prototype of a nonlocal contact repulsion force model. It is developed for smooth elastic spheres, frictionless contact, no adhesion, no gravity, moderate deformations, and confined granular systems, with the explicit statement that independent contacts are only valid for small deformations (&&&^^^^5query5^^^^&&&).

The core constitutive relation is a nonlocal correction of Hertz contact. For two spheres of radii $R_1,R_2$, moduli $E_1,E_2$, and Poisson ratios $\nu_1,\nu_2$, the relative approach PRESERVED_PLACEHOLDER_^^^^5all:\5query5^^^^ and contact force PRESERVED_PLACEHOLDER_^^^^5all:\5all:\5^^^^ satisfy

PRESERVED_PLACEHOLDER_^^^^5all:\5 OR all:\5^^^^

with

PRESERVED_PLACEHOLDER_^^^^5all:\5 OR all:\5^^^^

and

PRESERVED_PLACEHOLDER_^^^^5all:\5 OR all:\5^^^^

Here the shift PRESERVED_PLACEHOLDER_^^^^5all:\55^^^^ is the sum of nonlocal displacements induced by every other contact acting on the same particles. If PRESERVED_PLACEHOLDER_^^^^5all:\5^^^^6, the formulation reduces exactly to Hertz theory: PRESERVED_PLACEHOLDER_^^^^5all:\5^^^^7 The force on a contact edge PRESERVED_PLACEHOLDER_^^^^5all:\5^^^^8 is updated iteratively as

PRESERVED_PLACEHOLDER_^^^^5all:\5^^^^9

This makes the contact law many-body in spirit: the repulsive force at one interface depends on the entire contact network around the particle (&&&^^^^5query5^^^^&&&).

The paper further derives explicit nonlocal surface displacements outside the contact patch. The key mesoscopic ingredient is the normal displacement caused by a force acting elsewhere on the same sphere,

PRESERVED_PLACEHOLDER_^^^^5 OR all:\5query5^^^^

which is superposed with a Hertzian local contact displacement. This construction preserves a Hertz-like local pressure in the contact patch but removes the restriction that contacts are independent (&&&^^^^5query5^^^^&&&).

The corrected formulation adds two higher-order improvements. First, a contact radius correction accounts for nonlocal radial compliance, not only vertical indentation. Second, a curvature correction replaces the one-term Hertz profile by higher-order Taylor terms: PRESERVED_PLACEHOLDER_^^^^5 OR all:\5all:\5^^^^ The paper states that the curvature-corrected formulation converges at the four-term correction and improves agreement with finite elements and experiments up to the onset of contact impingement. The reported increase in applicability is about 5% for die compaction, ^^^^5 OR all:\5^^^^% for hydrostatic compaction, and 6% for die compaction with oblique walls (&&&^^^^5all:\5^^^^&&&).

These granular formulations establish a central principle: nonlocal repulsion at contact need not mean long-range forces in free space. It may instead mean that elastic response at one contact is mediated by deformation fields generated by the entire set of simultaneous contacts on the same grain.

^^^^5 OR all:\5^^^^. Continuum PDE models: nonlocal repulsion as convolution force or drift reversal

In the pressureless viscous compressible Navier–Stokes model, the nonlocal contact repulsion force is introduced by replacing local pressure with an attraction–repulsion convolution term (&&&^^^^5 OR all:\5^^^^&&&). The governing system is

PRESERVED_PLACEHOLDER_^^^^5 OR all:\5 OR all:\5^^^^

with

PRESERVED_PLACEHOLDER_^^^^5 OR all:\5 OR all:\5^^^^

The singular term PRESERVED_PLACEHOLDER_^^^^5 OR all:\5 OR all:\5^^^^ is repulsive at short distance and the quadratic term PRESERVED_PLACEHOLDER_^^^^5 OR all:\55^^^^ is attractive or confining at long range. The paper explicitly interprets this as a nonlocal analogue of contact pressure: the repulsion is geometric and distance-based rather than a pointwise pressure PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^6 (&&&^^^^5 OR all:\5^^^^&&&).

The energy structure is correspondingly nonlocal: PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^7 and

PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^8

For initial data satisfying the stated PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^9 and PRESERVED_PLACEHOLDER_^^^^5 OR all:\5query5^^^^ integrability assumptions, the paper proves the existence of a global weak solution on PRESERVED_PLACEHOLDER_^^^^5 OR all:\5all:\5^^^^, together with an energy inequality, a Bresch–Desjardins-type estimate, and a Mellet–Vasseur-type estimate (&&&^^^^5 OR all:\5^^^^&&&). The main analytical point is that these estimates replace the compactness normally furnished by local pressure.

The cell-motility model encodes nonlocal repulsion differently. In a one-dimensional isolated domain PRESERVED_PLACEHOLDER_^^^^5 OR all:\5 OR all:\5^^^^, the new continuum PDE is

PRESERVED_PLACEHOLDER_^^^^5 OR all:\5 OR all:\5^^^^

with

PRESERVED_PLACEHOLDER_^^^^5 OR all:\5 OR all:\5^^^^

The factor

PRESERVED_PLACEHOLDER_^^^^5 OR all:\55^^^^

is the nonlocal saturation coefficient. If the neighborhood mass is below the threshold PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^6, the drift follows the nonlocal gradient and acts attractively; if the neighborhood mass exceeds PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^7, the coefficient changes sign and the drift reverses, producing repulsion from saturated regions (Ferrari et al., 2024). The paper contrasts this with the APS model, which has nonlocal attraction only, and with the Carrillo-type factor PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^8, which merely reduces mobility in crowded regions rather than reversing drift direction.

Taken together, these two PDE families show that nonlocal contact repulsion can replace local pressure, and it can also be encoded as a sign-changing transport coefficient tied to crowding over a sensing radius. In neither case is repulsion restricted to a local constitutive function of the density at a single point.

^^^^5 OR all:\5^^^^. Confined surfaces, global force balance, and near-contact stabilization

For confined crystal surfaces in solution, the nonlocal contact repulsion model is built into a kinetic Monte Carlo description of contact formation between a periodic solid-on-solid crystal and a flat inert surface (&&&^^^^5 OR all:\5^^^^&&&). A site PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^9 is declared in contact when

PRESERVED_PLACEHOLDER_^^^^5 OR all:\5query5^^^^

and the contact density is

PRESERVED_PLACEHOLDER_^^^^5 OR all:\5all:\5^^^^

Repulsion is taken from electric double-layer theory and is assigned sitewise as

PRESERVED_PLACEHOLDER_^^^^5 OR all:\5 OR all:\5^^^^

The total repulsive free energy is

PRESERVED_PLACEHOLDER_^^^^5 OR all:\5 OR all:\5^^^^

and its force is

PRESERVED_PLACEHOLDER_^^^^5 OR all:\5 OR all:\5^^^^

The confining surface is always placed in mechanical equilibrium: PRESERVED_PLACEHOLDER_^^^^5 OR all:\55^^^^ This is nonlocal in the explicit sense that the confining surface position is determined by the combined contribution of all surface sites, not by a single local height (&&&^^^^5 OR all:\5^^^^&&&).

The same paper couples this repulsion to a finite-ranged attraction constructed to mimic a van der Waals-like PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^6 interaction and finds nontrivial contact morphologies. Stable contacts arise only when thermal fluctuations nucleate them. The stable contact density PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^7 is non-monotonic in pressure and exhibits contact favoring pressure levels (CFPLs). The morphology depends on PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^8: islands for PRESERVED_PLACEHOLDER_^^^^5 OR all:\5^^^^9, bands for $$V_c=\int_{\Omega^1(\mathbf{u})\cap\Omega^2(\mathbf{u})}dx$$^^^^5query5^^^^, and pits for $$V_c=\int_{\Omega^1(\mathbf{u})\cap\Omega^2(\mathbf{u})}dx$$^^^^5all:\5^^^^ (&&&^^^^5 OR all:\5^^^^&&&). A key consequence is that nonlocal repulsion does not simply prevent contact; under the right force balance it organizes when and how contact forms.

At the nanoscale, the nanoswitch model couples electrostatic attraction, elastic restoring force, van der Waals/Casimir attraction, and short-range contact repulsion (&&&^^^^5 OR all:\5^^^^&&&). The equilibrium balance without repulsion is

$$V_c=\int_{\Omega^1(\mathbf{u})\cap\Omega^2(\mathbf{u})}dx$$^^^^5 OR all:\5^^^^

and with repulsion included,

$$V_c=\int_{\Omega^1(\mathbf{u})\cap\Omega^2(\mathbf{u})}dx$$^^^^5 OR all:\5^^^^

The contact repulsion is modeled as a nonretarded exchange repulsion via the $$V_c=\int_{\Omega^1(\mathbf{u})\cap\Omega^2(\mathbf{u})}dx$$^^^^5 OR all:\5^^^^ Lennard–Jones term

$$V_c=\int_{\Omega^1(\mathbf{u})\cap\Omega^2(\mathbf{u})}dx$$5

which, after half-space integration and PFA reduction to a sphere–plate geometry, gives

$$V_c=\int_{\Omega^1(\mathbf{u})\cap\Omega^2(\mathbf{u})}dx$$6

Thus $$V_c=\int_{\Omega^1(\mathbf{u})\cap\Omega^2(\mathbf{u})}dx$$7, much steeper than the electrostatic $$V_c=\int_{\Omega^1(\mathbf{u})\cap\Omega^2(\mathbf{u})}dx$$8 attraction and the short-range van der Waals attraction (&&&^^^^5 OR all:\5^^^^&&&).

The same work gives a stability criterion for reset after switching,

$$V_c=\int_{\Omega^1(\mathbf{u})\cap\Omega^2(\mathbf{u})}dx$$9

with reported thresholds of $\varrho\nabla(K*\varrho)$^^^^5query5^^^^ at $\varrho\nabla(K*\varrho)$^^^^5all:\5^^^^ nm for Au and $\varrho\nabla(K*\varrho)$^^^^5 OR all:\5^^^^ at $\varrho\nabla(K*\varrho)$^^^^5 OR all:\5^^^^ nm for Ni. Its central conclusion is that omitting contact repulsion predicts pull-in and collapse, whereas including it yields a stable finite-separation minimum and cyclic switching (&&&^^^^5 OR all:\5^^^^&&&).

These models share a common structural feature: the force opposing collapse is assembled from an entire near-contact region rather than prescribed as a purely local hard-wall constraint.

5. Integral surface operators and overlap-volume energetics

A different branch of the literature places nonlocal contact repulsion inside a boundary-value problem with internal variables. In the adhesive contact model, the body occupies a bounded Lipschitz domain $\varrho\nabla(K*\varrho)$^^^^5 OR all:\5^^^^ with contact boundary $\varrho\nabla(K*\varrho)$5, and the state variables include displacement $\varrho\nabla(K*\varrho)$6 and adhesive/damage parameter $\varrho\nabla(K*\varrho)$7 (Bonetti et al., 2017). The nonlocal operator is

$\varrho\nabla(K*\varrho)$8

with a symmetric kernel $\varrho\nabla(K*\varrho)$9, $R_1,R_2$^^^^5query5^^^^. The quasistatic momentum balance is coupled to a contact boundary condition in which the traction at $R_1,R_2$^^^^5all:\5^^^^ depends on the state of neighboring points through an integral term, and the adhesive flow rule contains the nonlocal contributions

$R_1,R_2$^^^^5 OR all:\5^^^^

This is a genuine surface nonlocality: contact reaction and damage evolution are both influenced by a neighborhood interaction on $R_1,R_2$^^^^5 OR all:\5^^^^ (Bonetti et al., 2017).

The analytical result is a global-in-time existence theorem. Under the stated assumptions on the domain, tensors, nonlinearities, kernel, and initial data, there exists a global solution $R_1,R_2$^^^^5 OR all:\5^^^^ with

$R_1,R_2$5

together with an energy-dissipation inequality. The proof combines Yosida regularization, Schauder fixed point theory, compactness, and a nonstandard prolongation argument (Bonetti et al., 2017). Here nonlocal repulsion is inseparable from unilateral constraints and damage mechanics.

The overlap-volume formulation moves in an even more geometric direction. Contact is represented by an energy

$R_1,R_2$6

with the linear case $R_1,R_2$7 used in most simulations (Wang et al., 10 Sep 2025). The overlap volume is

$R_1,R_2$8

and the repulsive force is the negative energy gradient,

$R_1,R_2$9

In discrete form,

$E_1,E_2$^^^^5query5^^^^

Because increasing overlap increases the contact energy, the force pushes the bodies apart (Wang et al., 10 Sep 2025).

The computational difficulty is evaluating $E_1,E_2$^^^^5all:\5^^^^ for arbitrary geometries. The paper resolves this using Fiber Monte Carlo (FMC), which samples line segments rather than points so that boundary intersections become differentiable through an implicit parameter $E_1,E_2$^^^^5 OR all:\5^^^^. The authors emphasize that the framework is mesh-independent, eliminates master-slave identification and projection iterations, and can be incorporated into existing numerical solvers such as FEM (Wang et al., 10 Sep 2025). In this setting, nonlocality is geometric and variational: the contact law depends on the entire overlap region, not on a pointwise gap or a local normal projection.

6. Recurring principles, limits of the term, and a useful contrast

Across these works, several recurring principles emerge. First, short-range repulsion is often paired with a second mechanism that prevents the opposite pathology: long-range attraction or confinement in aggregation models, elastic restoring force in nanoswitches, or bond degradation and unilateral constraints in adhesive contact (&&&^^^^5 OR all:\5^^^^&&&, &&&^^^^5 OR all:\5^^^^&&&, Bonetti et al., 2017). Second, many formulations use repulsion not to eliminate contact altogether but to regularize its onset, redistribute it, or stabilize a finite-separation state. The confined-surface model admits nucleation, Ostwald ripening, coalescence, and stable contact morphologies (&&&^^^^5 OR all:\5^^^^&&&); the granular model predicts that even $E_1,E_2$^^^^5 OR all:\5^^^^ can still produce nonzero compressive force in some geometries under confinement (&&&^^^^5query5^^^^&&&); the nanoswitch model restores cyclic operation rather than enforcing infinite separation (&&&^^^^5 OR all:\5^^^^&&&).

A common misconception is that nonlocal contact repulsion must mean the same thing in every field. The cited literature shows otherwise. In one setting it is a superposed elastic deformation field, in another a convolution potential, in another a global force balance over many surface sites, in another a symmetric integral operator on the contact surface, and in another the gradient of an overlap-volume functional (&&&^^^^5query5^^^^&&&, &&&^^^^5 OR all:\5^^^^&&&, &&&^^^^5 OR all:\5^^^^&&&, Bonetti et al., 2017, Wang et al., 10 Sep 2025). The shared feature is dependency on a neighborhood, a contact network, or a global geometric measure.

A second misconception is that any near-contact barrier is automatically a nonlocal contact model. A useful contrast is provided by the conservative Allen–Cahn phase-field lattice Boltzmann model with adaptive near-contact repulsion, which introduces a repulsive thin-film barrier for oppositely oriented nearby interfaces but emphasizes a fully local implementation. It avoids ray tracing and global geometric reconstruction, uses a bounded neighborhood search, and estimates film thickness analytically from the local phase field (&&&^^^^5 OR all:\5^^^^9&&&). This suggests that near-contact repulsion and nonlocal contact repulsion are related but not identical categories.

The literature therefore supports a precise but broad characterization: a nonlocal contact repulsion force model is a formulation in which the repulsive or contact response is mediated by information extending beyond a single local gap variable. Depending on the application, that mediator may be the surrounding contact network, a convolution kernel, a surface integral operator, a global balance of interfacial forces, or an overlap-dependent energy functional.