Chapter Four · failure evidence
What Multiscale Modeling got wrong, from 33 dissertations
The records describe multiscale, continuum, and discrete modeling strategies evaluated across various materials and mechanical systems. Many approaches failed or were rejected due to continuum assumptions missing microstructural features, excessive computational cost of fine-scale simulations, and numerical or kinematic formulation issues. These records come from PhD theses at 16 institutions, 2021 to 2026. Each links to its thesis. They were extracted by language models reading the full text, so treat each as a lead to read, not a verdict.
Continuum and homogeneous models fail to capture discrete microstructural and atomic mechanisms
Homogeneous continuum formulations and macro-only calibrations fail when discrete physics such as dislocation spacing, grain boundary pileup, force chains, atomic layering, and crystallographic slip dominate behavior. Consequently, these models miscalculate yield stress and elastic modulus while failing to reproduce grain-scale crack growth and atomic-scale dynamics.
Tried and failed
calibrating multiscale models using only macroscopic data applied to crystal plasticity constitutive modeling. Outcome: did not generalise. Reason: Macroscopic polycrystal curves lacked orientation-specific constraints needed to accurately resolve single-crystal responses.
Modeling of overloads in cyclic loading · Cranfield
Tried and failed
continuum elasticity modeling applied to curved surface colloidal crystal widths. Outcome: did not generalise. Reason: overestimates elastic modulus by orders of magnitude because discrete dislocation spacing governs width instead of continuum stress
Geometric Templating and Frustration of Two-Dimensional Colloidal Crystals · Harvard
Tried and failed
Linear elastic fracture mechanics and empirical power laws applied to microstructurally sensitive short fatigue crack propagation. Outcome: did not generalise. Reason: Homogeneous continuum assumptions fail to capture grain-scale oscillatory growth rates and crystallographic slip-dependent scatter
Modelling of microstructure sensitive short crack growth in Ni single crystals · Imperial
Tried and failed
size-independent yield strength in microstructure modeling applied to fatigue life prediction with variable grains. Outcome: no signal. Reason: Standard continuum finite element models lack intrinsic length scales and miss grain boundary dislocation pileup effects.
Multiscale Modeling of Fatigue and Fracture in Polycrystalline Metals, 3D Printed Metals, and Bio-inspired Materials · Virginia Tech
Tried and failed
continuum scaling laws applied to granular interaction with high-aspect-ratio geometries. Outcome: did not generalise. Reason: discrete grain-size nonlocal effects dominate continuum behavior at small feature dimensions
Lost to a baseline
Homogeneous continuum simulations failed to capture the smooth post-yield stress-strain transition and over-predicted yield stress compared to heterogeneous RVE simulations.
Considered and rejected
Considered and rejected: Rejected continuum / multifluid mixture (Eulerian-Eulerian) models for crystal-rich magmas because they fail to capture jamming, non-affine strain, discrete force chains, and grain-scale trajectories.
On the kinematic and dynamic evolution of a magma mush: from liquidus to solidus · ResearchWorks
Considered and rejected
Considered and rejected: Rejected local continuum modeling/direct numerical simulation for nanofluid suspension due to inability to resolve atomic layering and microscopic structural changes.
Non-local methods in transport phenomena · UT Austin
Considered and rejected
Considered and rejected: Replacing the discrete CNT structure with a totally continuum isotropic hollow or solid cylinder, rejected because it fails to capture chirality, curvature, and dynamic behaviors.
Study on the Mechanical Properties of Carbon Nanotube Coated‒Fiber Multi-Scale (CCFM) Hybrid Composites · IRIS - POLITO - prod
Considered and rejected
Considered and rejected: Rejected the conventional dielectric continuum model for remote phonon scattering in atomically thin 2D heterostructures because it treats materials as a continuum, ignores atomic discreteness, produces fictitious scattering for sub-layer confined phonon wavelengths, and neglects polar-optical phonon (POP) coupling.
Ab initio electronic transport : from two-dimensional semiconductors to nanoscale metals · UT Austin
Fully resolved fine-scale simulations and concurrent multiscale couplings incur prohibitive computational costs
Direct numerical simulation, fine 3D continuum meshes, and fully resolved pore-scale models become computationally intractable for large domains or high cell counts. Concurrent multiscale approaches that resolve microdomains at every macroscale boundary point or mesh thin boundary layers across extreme length scale differences impose unfeasible computational burdens.
Considered and rejected
Considered and rejected: Rejected discrete cell-based vertex models in favour of continuum finite element modelling to enable computational scalability to organisms with high cell numbers.
Biofidelic simulations of embryonic joint growth and morphogenesis · Imperial
Considered and rejected
Considered and rejected: Rejected high-fidelity 3D solid continuum finite element models for online filtering because explicit solvers with fine meshes and tiny time steps exacerbate computational burden and convergence issues
Input-State Estimation of Inelastic Structural Systems: Theoretical Framework and Experimental Validation · Georgia Tech
Considered and rejected
Considered and rejected: Rejected fully resolved pore-scale Navier-Stokes CFD across full chips due to excessive computational intractability, adopting micro-continuum DBS modeling instead.
Brine Drying and Salt Precipitation Dynamics in Porous Media and Its Impact on Hydro-Thermal-Mechanical Properties · Research Repository UCD
Considered and rejected
Considered and rejected: Finite Volume Method (FVM), Finite Element Method (FEM), and Molecular Dynamics (MD) rejected for microfluidic systems due to intensive meshing/poor cell quality in curved channels, continuum assumption failure, FSI computational demand, and prohibitive molecular computational costs
Considered and rejected
Considered and rejected: Rejected full 3D continuum finite element models for larger textile components due to excessive DOF and computational cost; adopted shell-element multi-scale embedding instead.
Modelling and simulation concepts for fibre reinforced polymers · DSpace-CRIS at TU Wien
Tried and failed
full concurrent multiscale direct numerical simulation applied to multiscale fluid-structure interface dynamics. Outcome: infeasible cost. Reason: simulating fine-scale microdomains at every macroscale boundary point each timestep is computationally prohibitive
Modeling Droplet Impingement Dynamics on Micropillar-Arrayed Viscoelastic Substrates Through Microgeometry-Free and Multiscale Methods · Virginia Tech
Tried and failed
direct resolving of thin boundary layers in FEM applied to multiscale whole-device continuum simulations. Outcome: infeasible cost. Reason: mesh aspect ratio constraints across extreme length scale differences made computation intractable
Considered and rejected
Considered and rejected: Rejected standard concurrent multiscale models (CMMs) due to prohibitive computational cost for engineering-scale applications.
A plasma-wall interaction model for the erosion of materials under ion bombardment · Georgia Tech
Spatial discretization and cross-scale interface coupling cause numerical instabilities and loss of resolution
Coupling discrete to continuum domains across fine meshes creates unhomogenized property mismatches at interfaces. In addition, local multiscale damage formulations and coarse virtual element methods suffer from pathological mesh sensitivity, premature strain localization, and inaccurate secondary stress or velocity fields.
Tried and failed
discrete to continuum multiscale coupling with fine mesh applied to dynamic wave transmission in solids. Outcome: worse than baseline. Reason: unhomogenized property mismatch across the interface at fine scales
Modeling friction and wear using an adaptive discrete-continuum coupling · EPFL
Tried and failed
local continuum damage formulation in multiscale modeling applied to composite plate failure simulation. Outcome: unstable. Reason: pathological mesh sensitivity and premature strain localization during multiscale damage evolution
Multiscale modelling of the material degradation and failure in complex composite structures · Imperial
Tried and failed
multiscale virtual element method applied to porous media coupled mechanics. Reason: coarse discretization fails to accurately resolve localized secondary derived fields like stresses and velocities
Multiscale aeroelastic modelling in porous composite structures · University of Nottingham Repository
Simplified dimensional and linear formulations fail to represent complex geometries and nonlinear dynamic effects
One-dimensional phase-field continuum models and simplified single-scale synthetic textures neglect higher-dimensional nucleation phenomena and scalable rough topographies. Standard linear strain formulations also fail to capture size-dependent geometric softening and asymmetric resonant bandwidth under large dynamic loads.
Tried and failed
one-dimensional phase-field continuum modeling applied to intercalation-based resistance switching dynamics. Outcome: did not generalise. Reason: neglects higher-dimensional nucleation phenomena and interfacial surface effects
Considered and rejected
Considered and rejected: Single-scale prismatic pillared textures and synthetic nested-circle fractal models were rejected because they fail to represent scalable multiscale rough topographies produced by electrodeposition/etching.
Analysis of Interfacial Processes on Non-Wetting Surfaces · Virginia Tech
Considered and rejected
Considered and rejected: Rejected standard linear strain formulation in continuum flexoelectricity because it fails to capture size-dependent geometric softening and asymmetric resonant bandwidth under large dynamic loads
Atomistic and continuum scale models for flexoelectric nanostructures and composites · Leibniz Universität Hannover Repository
Left open by the authors
Problems the authors named and did not get to.
Left open
Develop multiscale mathematical models to predict mechanical properties of crystal-inspired architected lattices from unit-cell features to mesostructures. Blocker: Lack of specific mathematical formulation, targets, or validation datasets defined in the thesis summary
Novel crystal-inspired architected materials design guided by metallurgical hardening mechanisms · Imperial
Left open
Develop condensed multiscale models for the vibroacoustic analysis of multi-layered porous composite structures. Blocker: Lacks specific mathematical formulation details and target benchmarks for model condensation.
Multiscale aeroelastic modelling in porous composite structures · University of Nottingham Repository
Left open
Develop a continuum model approach capturing both fractional quantum Hall and Wigner crystal physics in shallow moiré potentials. Blocker: None
Pushing the frontier of quantum many-body simulation using classical computers · Cornell
Left open
Develop constitutive continuum models that incorporate biological microstructure to distinguish cell mechanics from extracellular matrix behavior. Blocker: No specific constitutive formulation, microstructure geometry, or validation target is defined.
Left open
Calibrate systematic a priori selection rules for continuum domain lengths and solid element mesh sizes in mixed-dimension beam-column models across structural configurations. Blocker: Lack of predefined calibration benchmarks, target error thresholds, or formal parameterization protocols for general structural configurations.
Left open
Extend the FEniCS continuum multiphysics model to incorporate volumetric growth mechanics for simulating evolving surface morphology in tumor growth. Blocker: None
A continuum multiphysics approach for the reconfiguration and healing of contractile microtissues · Cornell
Left open
Perform theoretical multiscale modeling of dynamic catalyst surface heterogeneity, carbon solubility, and sulfur promoter effects during carbon nanotube growth. Blocker: Lack of concrete methodological formulation or specific benchmark targets.
Progress in Computational Methods and Mechanistic Insights on the Growth of Carbon Nanotubes · MIT
Left open
Develop a multiscale framework coupling nanoscale molecular dynamics parameters of damaged hair surfaces to continuum mechanics models of macroscale hair detanglement. Blocker: Lack of specific multiscale coupling algorithms, continuum formulation details, and sensory perception validation data
Left open
Couple the discrete discontinuous basis projection method with QM-MM multiscale frameworks for simulating complex interfaces. Blocker: Lack of implementation details and interface specifications for the multiscale QM-MM integration
Hybrid Functionals in Real-Space Density Functional Theory · Georgia Tech
Left open
Prove theoretical velocity and flux convergence properties for channel-aware coordinate dilation in multiscale PDE problems. Blocker: Lacks specific mathematical formulation, targets, or proof strategies to guide theoretical development
Numerical multiscale methods: from homogenization to milestoning · UT Austin
Checking a claim in this area?
We can run the same search on any method or claim. If nothing turns up, we will say so, and that proves nothing on its own.