arXiv subjects

Manolis Veveakis

Publications and source records attributed to Manolis Veveakis.

At least 19 recordsLinked to original sources

arXiv2026

Is Seismic Forecasting Possible with Physics-based AI?

Slow slip events within subduction zones offer a unique window into earthquake prediction. The subducting plate drives dehydration reactions in the fault, causing cyclical slip and observable surface displacements. Earthquake footprints can then be identified in these displacement series through coupling the multi-physics governing the subduction process with regional seismic activity. However, data noise and traditional filtering methods obscure the underlying mechanisms. Here, we alleviate this constraint with our physics-based attractor. By accounting for the physics of subduction paired with AI-assisted manifold detection, we are able to predict an earthquake in New Zealand's Hikurangi trench one week early. Additionally, predictability limits extend to 5-6 weeks with decadal repeatability, pointing to the fundamental determinism of the suggested mechanism through which physics-based seismic forecasting is possible.

physics.geo-ph↗

arXiv2026

Learning Interpretable PDE Representations for Generative Reconstructions with Structured Sparsity

Scientific measurements are often bottlenecked by suboptimal conditions, whether that be noise, incomplete spatial coverage, or limited resolution, rendering accurate field reconstruction a difficult task. We introduce LatentPDE, a latent diffusion framework designed to simultaneously resolve sparse-observation reconstruction and super-resolution. While existing physics-guided diffusion models typically rely on soft loss penalties or uninterpretable representations, our approach enforces physical compliance by constructing an inherently interpretable latent space. Specifically, we parameterize the latent variables directly as the coefficients and source terms of an assumed governing PDE. In doing so, LatentPDE is able to reliably reconstruct dynamics across highly disparate and structured data gaps. Empirical results on diverse configurations demonstrate that our model achieves high-fidelity recovery at any desired resolution while also tracking the underlying predictive uncertainty.

cs.LG↗

arXiv2026

Brittle-to-ductile fracturing transition: A chemo-mechanical phase-field framework

In chemically reactive environments, the mechanical integrity of geomaterials is fundamentally compromised by solid matrix dissolution. In this study, we propose a fully coupled chemo-mechanical phase-field framework to capture the dynamic interplay between mineral dissolution and fracture propagation. A key feature of the proposed model is the dynamic coupling of local mass removal to the fracture length scale, while also incorporating the damage-accelerated reaction-diffusion processes. Our results capture the development of an enlarged fracture process zone driven by chemical mass removal. This chemically induced widening blunts the sharp crack tip, alleviating the near-tip stress concentrations and causing a pronounced degradation in material stiffness before failure. Furthermore, we reveal a distinct ductilization effect, characterized by a more gradual accumulation of damage and a delayed onset of macroscopic failure. We show that the transition between brittle and ductile failure modes is dictated by the competing timescales of chemical degradation and mechanical deformation. Highly acidic environments enhance matrix dissolution and promote ductile fracture, whereas rapid mechanical loading limits chemical interaction and preserves brittle failure mode.

cond-mat.mtrl-sci↗

arXiv2025

Predicting cement microstructure and mechanical properties in hydrating cement paste with a Phase-Field model

Predicting the evolving microstructure of hydrating cement is essential for understanding and modeling its mechanical property development. Physics-based continuum approaches offer a rigorous framework for capturing the thermodynamics of dissolution and precipitation processes at the microstructural scale. In this work, we present an adapted Phase-Field (PF) model for cement hydration that resolves key physical inconsistencies in existing PF formulations by introducing a revised free-energy potential and distinct equilibrium constants for clinker dissolution and hydrate precipitation. The resulting PF framework reproduces microstructural evolution, yielding realistic porosity levels and continuous phase boundaries in close agreement with experimental observations. The predicted hydrated microstructures are subsequently used in a computational homogenization scheme to evaluate the elastic response of the material. The PF-derived mechanical properties show good agreement with experimental trends, supporting the ability of the proposed framework to consistently link hydration chemistry, microstructure formation, and the resulting mechanical response.

cond-mat.mtrl-sci↗

arXiv2025

Probabilistic Spatial Interpolation of Sparse Data using Diffusion Models

The large underlying assumption of climate models today relies on the basis of a "confident" initial condition, a reasonably plausible snapshot of the Earth for which all future predictions depend on. However, given the inherently chaotic nature of our system, this assumption is complicated by sensitive dependence, where small uncertainties in initial conditions can lead to exponentially diverging outcomes over time. This challenge is particularly salient at global spatial scales and over centennial timescales, where data gaps are not just common but expected. The source of uncertainty is two-fold: (1) sparse, noisy observations from satellites and ground stations, and (2) internal variability stemming from the simplifying approximations within the models themselves. In practice, data assimilation methods are used to reconcile this missing information by conditioning model states on partial observations. Our work builds on this idea but operates at the extreme end of sparsity. We propose a conditional data imputation framework that reconstructs full temperature fields from as little as 1% observational coverage. The method leverages a diffusion model guided by a prekriged mask, effectively inferring the full-state fields from minimal data points. We validate our framework over the Southern Great Plains, focusing on afternoon (12:00-6:00 PM) temperature fields during the summer months of 2018-2020. Across varying observational densities--from swath data to isolated in-situ sensors--our model achieves strong reconstruction accuracy, highlighting its potential to fill in critical data gaps in both historical reanalysis and real-time forecasting pipelines.

stat.AP↗

arXiv2025

Learning Latent Hardening (LLH): Enhancing Deep Learning with Domain Knowledge for Material Inverse Problems

Advancements in deep learning and machine learning have improved the ability to model complex, nonlinear relationships, such as those encountered in complex material inverse problems. However, the effectiveness of these methods often depends on large datasets, which are not always available. In this study, the incorporation of domain-specific knowledge of the mechanical behavior of material microstructures is investigated to evaluate the impact on the predictive performance of the models in data-scarce scenarios. To overcome data limitations, a two-step framework, Learning Latent Hardening (LLH), is proposed. In the first step of LLH, a Deep Neural Network is employed to reconstruct full stress-strain curves from randomly selected portions of the stress-strain curves to capture the latent mechanical response of a material based on key microstructural features. In the second step of LLH, the results of the reconstructed stress-strain curves are leveraged to predict key microstructural features of porous materials. The performance of six deep learning and/or machine learning models trained with and without domain knowledge are compared: Convolutional Neural Networks, Deep Neural Networks, Extreme Gradient Boosting, K-Nearest Neighbors, Long Short-Term Memory, and Random Forest. The results from the models with domain-specific information consistently achieved higher $R^2$ values compared to models without prior knowledge. Models without domain knowledge missed critical patterns linking stress-strain behavior to microstructural changes, whereas domain-informed models better identified essential stress-strain features predictive of microstructure. These findings highlight the importance of integrating domain-specific knowledge with deep learning to achieve accurate outcomes in materials science.

cs.LG↗

arXiv2024

Importance of precipitation on the slowdown of creep behaviour induced by pressure-solution

Pressure-solution is a chemo-mechanical process, involving dissolution at grain/asperity contacts and precipitation away from them. It induces a compaction in time of rocks and sediments. The present study investigates numerically the impact of precipitation on the slowdown of creep behavior induced by pressure-solution. A recently published framework, called the Phase-Field Discrete Element Model, is carefully calibrated against existing indentation experiments and validated for other rate-limiting scenarios. It is shown that when precipitation is relatively slow, the slowdown of pressure-solution is due to a chemical mechanism (accumulation of solute concentration within the pore space), whereas, at fast precipitation, the slowdown is due to a mechanical mechanism (stress reduction at the contact).

cond-mat.mtrl-sci↗

arXiv2024

Chaotic Slow Slip Events in New Zealand from two coupled slip patches: a proof of concept

Recent studies showed that seemingly random Slow Slip Events (SSEs) can display chaotic patterns within the largest source of seismic hazards in New Zealand, the Hikurangi subduction zone. Some irregular SSE occurrences are therefore not arbitrary but behave with short-term predictability. However, the forecasting challenge persists as observations remain too short and noisy to constrain purely data-driven solutions, calling for a physics-based modelling approach. Here we propose a physical model of two coupled oscillators, each capturing the behaviour of a single slow-slip patch, for the deep Kaimanawa and the shallow East Coast SSEs respectively. The simplified model successfully reproduces the type of chaotic behaviour observed at the Global Navigational Satellite System station in Gisborne, yielding SSEs of appropriately varying amplitude and duration. Those results reveal that the multi-physics response of the shear zone strongly controls the underlying system, even before accounting for any geometrical complexity or distribution of material properties.

physics.geo-ph↗

arXiv2024

Stress state evolution of a cemented granular material subjected to bond dissolution by Discrete Element Modeling

Rock weathering is a common phenomenon in most engineering applications, such as underground storage or geothermal energy. This work offers a discrete element modelization of the problem considering cohesive granular material and debonding effect. Oedometer conditions are applied during the weathering and the evolution of the coefficient of lateral earth pressure, a proxy of the state of stress, is tracked. Especially, the influence of the degree of cementation, the confining pressure, the initial value of k0 and the history of load are investigated. It has been emphasized that the granular media aims to reach an attractor configuration. And the grain reorganization occurring is divided into two main phenomena: the collapse of the unstable chain forces (stable only thanks to the cementation) and the softening of the grains.

cond-mat.mtrl-sci↗

arXiv2024

Chemical homogenization for non-mixing reactive interfaces in porous media

Porous media, while ubiquitous across many engineering disciplines, is inherently difficult to characterize due to their innate stochasticity and heterogeneity. The key for predicting porous material behavior comes down to the structuring of its microstructure, where the linkages of microstructural properties to mesoscale effects remain as one of the key questions in unlocking understanding of this class of materials. One proposed method of linking scales comes down to using Minkowski functionals -- geometric morphometers that describe the spatial and topological features of a convex space -- to draw connections from microstructural form to mesoscale features. In this work, chemical equilibrium and kinetics on a microstructure surface were explored, with Minkowski functionals used as the basis for relating microstructural geometry to chemical performance. Using surface CRNs to model chemical behavior -- a novel asynchronous cellular automaton -- linkages were found between the Minkowski functionals and equilibrium equilibrium constant, as well as properties related to the dynamics of the system's reaction quotient.

cond-mat.soft↗

arXiv2024

Path Signatures and Graph Neural Networks for Slow Earthquake Analysis: Better Together?

The path signature, having enjoyed recent success in the machine learning community, is a theoretically-driven method for engineering features from irregular paths. On the other hand, graph neural networks (GNN), neural architectures for processing data on graphs, excel on tasks with irregular domains, such as sensor networks. In this paper, we introduce a novel approach, Path Signature Graph Convolutional Neural Networks (PS-GCNN), integrating path signatures into graph convolutional neural networks (GCNN), and leveraging the strengths of both path signatures, for feature extraction, and GCNNs, for handling spatial interactions. We apply our method to analyze slow earthquake sequences, also called slow slip events (SSE), utilizing data from GPS timeseries, with a case study on a GPS sensor network on the east coast of New Zealand's north island. We also establish benchmarks for our method on simulated stochastic differential equations, which model similar reaction-diffusion phenomenon. Our methodology shows promise for future advancement in earthquake prediction and sensor network analysis.

cs.LG↗

arXiv2023

InSAR-Informed In-Situ Monitoring for Deep-Seated Landslides

This work focuses on assessing the fidelity of Interferometric Synthetic Aperture Radar (InSAR) as it relates to subsurface ground motion monitoring, as well as understanding uncertainty in modeling active landslide scarp displacement for the case study of the in situ monitored El Forn deep seated landslide in Canillo, Andorra. We used the available Sentinel 1 data on the Alaska Satellite Facility (ASF) Vertex platform to create deformation velocity maps and time series of the El Forn landslide scarp. We compared the performances of InSAR data from the recently launched European Ground Motion Service (EGMS) platform and the ASF Vertex Platform in a time series comparison of displacement in the direction of landslide motion with in situ borehole based measurements from 2019 to 2021, suggesting that ground motion detected through InSAR can be used in tandem with field monitoring to provide optimal information with minimum in situ deployment. While identification of active landslide scarps may be possible via the use of EGMS platform, the intents and purposes of this work are in assessment of InSAR as a monitoring tool. Based on that, geospatial interpolation with statistical analysis was conducted to better understand the necessary number of in situ observations needed to lower error on a remote sensing recreation of ground motion over the entirety of a landslide scarp, suggesting between 20 to 25 total observations provides the optimal normalized root mean squared error for an ordinarily kriged model of the El Forn landslide scarp.

physics.geo-ph↗

arXiv2023

Frictional weakening of a granular sheared layer due to viscous rolling revealed by Discrete Element Modeling

Considering a 3D sheared granular layer modeled with discrete elements, it is well known the rolling resistance significantly influences the mechanical behavior. Even if the rolling resistance role has been deeply investigated as it is commonly used to represent the the roughness of the grains and the interparticle locking, the role of rolling viscous damping coefficient has been largely overlooked so far. This parameter is rarely used or only to dissipate the energy and to converge numerically. This paper revisits the physical role of those coefficients with a parametric study of the rolling friction and the rolling damping for a sheared layer at different shear speeds and different confinement pressures. It has been observed that the damping coefficient induces a frictional weakening. Hence, competition between the rolling resistance and the rolling damping occurs. Angular resistance aims to avoid grains rolling, decreasing the difference between the angular velocities of grains. Whereas, angular damping acts in the opposite, avoiding a change in the difference between the angular velocities of grains. In consequence, grains keep rolling and the sample strength decreases. This effect must be considered to not overestimate the frictional response of a granular layer.

physics.geo-ph↗

arXiv2023

A Phase-Field Discrete Element Method to study chemo-mechanical coupling in granular materials

This paper presents an extension of the discrete element method using a phase-field formulation to incorporate grain shape and its evolution. The introduction of a phase variable enables an effective representation of grain geometry and facilitates the application of physical laws, such as chemo-mechanical couplings, for modeling shape changes. These physical laws are solved numerically using the finite element method coupled in a staggered scheme to the discrete element model. The efficacy of the proposed Phase-Field Discrete Element Model (PFDEM) is demonstrated through its ability to accurately capture the real grain shape in a material subjected to dissolution only and compute the stress evolution. It is then applied to model the phenomenon of pressure solution involving dissolution and precipitation in granular materials at the microscale and enables to reproduce the creep response observed experimentally. This framework contributes to the enhanced understanding and simulation of complex behaviors in granular materials and sedimentary rocks for many geological processes like diagenesis or earthquake nucleation.

cond-mat.mtrl-sci↗

arXiv2021

Morphometric description of strength and degradation in porous media

The influence of the microstructural geometry on the behavior of porous media is widely recognized, particularly in geomaterials, but also in biomaterials and engineered materials. Recent advances in imaging techniques, such as X-ray microcomputed tomography, and in modeling make it possible to capture the exact morphometry of the microstructure with high precision. However, most existing continuum theories only partially account for the morphometry. We propose here a unifying approach to link the strength of porous materials with the necessary and sufficient microstructural information, using Minkowski functionals, as per Hadwiger's theorem. A morphometric strength law is inferred from synthetic microstructures with a wide range of porosities and heterogeneities, through qualitative 2D phase-field simulations. Namely, the damage is modeled at the microstructural level by tracking the solid-pore interfaces under mechanical loading. The strength is found to be best described by an exponential function of the morphometers, thus generalizing early works on metals and ceramics. We then show that the predictiveness of this relationship extends to real porous media, including rocks and bones.

cond-mat.soft↗

arXiv2020

Strain localization regularization and patterns formation in rate-dependent plastic materials with multiphysics coupling

Strain localization is an instability phenomenon occurring in deformable solid materials which undergo dissipative deformation mechanisms. Such instability is characterized by the localization of the displacement or velocity fields in a zone of finite thickness and is generally associated with the failure of materials. In several fields of material engineering and natural sciences, estimating the thickness of localized deformation is required to make accurate predictions of the evolution of the physical properties within localized strain regions and of the material strength. In this context, scientists and engineers often rely on numerical modeling techniques to study strain localization in solid materials. However, classical continuum theory for elasto-plastic materials fails at estimating strain localization thicknesses due to the lack of an internal length in the model constitutive laws. In this study, we investigate at which conditions multiphysics coupling enables to regularize the problem of strain localization using rate-dependent plasticity. We show that coupling the constitutive laws for deformation to a single generic diffusion-reaction equation representing a dissipative state variable can be sufficient to regularize the ill-posed problem under some conditions on the softening parameters in the plastic potential. We demonstrate in these cases how rate-dependent plasticity and multiphysics coupling can lead to material instabilities depicting one or several internal length scales controlled by the physical parameters resulting in the formation of regular or erratic patterns. As we consider a general form of the equations, the results presented in this study can be applied to a large panel of examples in the material engineering and geosciences communities.

physics.geo-ph↗

arXiv2020

Adaptive stabilized finite elements: Continuation analysis of compaction banding in geomaterials

Under compressive creep, visco-plastic solids experiencing internal mass transfer processes have been recently proposed to accommodate singular cnoidal wave solutions, as material instabilities at the stationary wave limit. These instabilities appear when the loading rate is significantly faster than the capability of the material to diffuse internal perturbations and lead to localized failure features (e.g., cracks and compaction bands). This type of solution, generally found in fluids, has strong nonlinearities and periodic patterns. Due to the singular nature of the solutions, the applicability of the theory is currently limited. Additionally, effective numerical tools require proper regularization to overcome the challenges that singularity induces. We focus on the numerical treatment of the governing equation using a nonlinear approach building on a recent adaptive stabilized finite element method. This method provides a residual representation to drive adaptive mesh refinement, a particularly useful feature for the problem at hand. We compare against analytical and standard finite element solutions to demonstrate the performance of our approach. We then investigate the sensitivity of the diffusivity ratio, main parameter of the problem, and identify multiple possible solutions, with multiple stress peaks. Finally, we show the evolution of the spacing between peaks for all solutions as a function of that parameter.

physics.comp-ph↗

arXiv2019

Contact phase-field modeling for chemo-mechanical degradation processes. Part I: Theoretical foundations

As phase-field modeling (PFM) is booming across various disciplines and has been proven fitted for numerically modeling interfacial problems, we aim at taking a step back to revisit its fundamental validity, in the light of non-equilibrium thermodynamics. For that, a general contact thermodynamics (CT) framework is derived from contact geometry, based on the maximum dissipation principle (MaxDP), thus extending Gibbs' seminal geometrical representation of thermostatics. Combining CT and micro-force balance, the gradient flow equation usually derived for PFM from the variational formulation can be written as generalized relaxation equations. The obtained viscous Allen-Cahn equation allows both the PFM kinematic degrees of freedom, the order parameter and its gradient, to be fully dissipative. The model is also extended to a double PFM, in order to include chemo-mechanical coupling, corresponding respectively to endothermic and exothermic processes and thus leading to a phase change bidirectionality. This contact PFM (CPFM) will be applied in the second part of this work to irregular microstructures like geomaterials, valid for porous media in general, with a focus on pressure solution.

cond-mat.stat-mech↗