FeynKrack: A continuum model for quasi-brittle damage through Feynman-Kac killed diffusion
Ved Prakash, Upadhyayula M. M. A. Sai Gopal, Sanhita Das, Ananth Ramaswamy, Debasish Roy · 2024
Read FeynKrack: A continuum model for quasi-brittle damage through Feynman-Kac killed diffusion by Ved Prakash, Upadhyayula M. M. A. Sai Gopal, Sanhita Das, Ananth Ramaswamy, Debasish Roy (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-00791.txt.
Introduction
Ved Prakash and four co-authors model cracking in brittle materials with a tool borrowed from quantum physics, in this 2024 paper. Continuum damage mechanics tracks a damage parameter as cracks grow, but the equations for that parameter rarely match its supposed physical meaning. The team re-reads damage as a change of probability measure, with the damage variable acting as the conversion factor.
In their model, FeynKrack, the density of microcracks evolves by a diffusion that can be killed off, the Feynman-Kac picture. Damage growth is then the shrinking of that measure in a region, counting bonds that no longer transmit force between neighboring points. A well-chosen killing rate builds in irreversibility, so no ad-hoc history rules are needed. The evolution even admits an approximate closed-form solution, making simulations fast. Tests on quasi-brittle damage supported the approach.
Catalog record
| CREATOR | Ved Prakash, Upadhyayula M. M. A. Sai Gopal, Sanhita Das, Ananth Ramaswamy, Debasish Roy |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 68 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-00791
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
FeynKrack: A continuum model for quasi-brittle damage through Feynman-Kac killed diffusion Ved Prakash, Upadhyayula M. M. A. Sai Gopal, Sanhita Das, Ananth Ramaswamy, Debasish Roy Abstract Continuum damage mechanics (CDM) is a popular framework for modelling crack propagation in solids. The CDM uses a damage parameter to quantitatively assess what one loosely calls ‘material degradation’. While this parameter is sometimes given a physical meaning, the mathematical equations for its evolution are generally not consistent with such physical interpretations. Curiously, degradation in the CDM may be viewed as a change of measures, wherein the damage variable appears as the Radon-Nikodym derivative. We adopt this point of view and use a probabilistic measure-valued description for the random microcracks underlying quasi-brittle damage. We show that the evolution of the underlying density may be described via killed diffusion as in the Feynman-Kac theory. Damage growth is then interpreted as the reduction in this measure over a region, which in turn quantifies the disruption of bonds through a loss of force-transmitting mechanisms between nearby material points. Remarkably, the evolution of damage admits an approximate closed-form solution. This brings forth substantive computational ease, facilitating fast yet accurate simulations of large dimensional problems. By selecting an appropriate killing rate, one accounts for the irreversibility of damage and thus eliminates the need for ad-hoc history-dependent routes typically employed, say, in phase field modelling of damage. Our proposal FeynKrack (a short form for Feynman-Kac crack propagator) is validated and demonstrated for its efficacy through several simulations on quasi-brittle damage. It also offers a promising stochastic route for future explorations of non-equilibrium thermodynamic aspects of damage. 1 Introduction Quasi-brittle failure in materials such as concrete, rocks, ceramics, etc. is typically preceded by an intricate web of random microcracks. A faithful prediction of such complex crack patterns at the mesoscale is generally infeasible and one is thus more interested in models that can reproduce the propagation of quasi-brittle damage at the macroscale. However, except for a few attempts [ 1 ] , most available models for quasi-brittle damage do not have the physical motivation or mathematical grounding necessary for high predictive fidelity to replace repetitive and costly experiments. Given the inherent randomness in the distribution of microcracks, a model that exploits a probabilistic measure-valued description of microcracks seems to be a possibility. In this context, we note that the degradation function used with continuum damage mechanics (CDM) [ 2 ] may be looked upon as a Radon Nikodym derivative associated with a change of probability measures. We plan to follow this thread further in this article. Initially, linear elastic fracture mechanics (LEFM) and elasto-plastic fracture mechanics (EPFM) formed the basis for many continuum models on fracture and damage. But in the absence of proper length scales, they could not model the geometric and material non-linearities in the process-zone mechanics, an aspect considered crucial to quasi-brittle fracture [ 3 ] . Moreover, fracture preceded by degradation processes not necessarily arising due to mechanical loads cannot be captured by LEFM or EPFM. In contrast, damage mechanics [ 2 ] equipped with proper variables and length scales not only modelled the damage phenomena, but also the continuous material degradation as a precursor to crack propagation. Damage models may be discrete or continuous. While a discrete approach allows for a jump in the deformation field due to the presence of a crack, the jump is smeared out in a continuous approach. Discrete approaches require prior knowledge of the crack path and thus pose challenges in numerical implementation. Several enhancements in discrete approaches have been proposed, e.g. FEM with remeshing, XFEM, and cohesive- zone [ 4 , 5 ] requiring local enrichment and complex crack-tracking algorithms. But they still remain highly mesh-dependent and computationally intensive. Further developments such as discontinuous FEM [ 6 ] , inter-element separation model [ 7 ] , cracking particle method [ 8 ] , though significant, fall short of expectations in modelling complex crack trajectories, especially in three-dimensional fracture. Continuum damage models are superior to discrete approaches as they do not require a prior knowledge of the crack path and are numerically more stable. Most of these models are phenomenological. For example, the size effect model and the two-parameter fracture model [ 9 ] rely on crack-width, stress intensity factors with no reference to underlying bond-breakage or stiffness degradation. Microplane and crack band models [ 10 , 11 ] are two approaches widely used in modelling quasi-brittle fracture. Though they accurately capture damage orientation, damage evolution laws often violate the laws of thermodynamics. In modelling damage orientation, however, they are superior to second and fourth-order tensor-based continuum approaches as constitutive equations are defined using vectors representing stress and strain on randomly oriented planes. Of most local continuum models, approaches based on continuum damage mechanics (CDM) invoke the concept of a damage variable describing the physics of underlying degradation. These models are thus physically more meaningful. Damage variables with different interpretations have been used, such as measuring effective load-carrying areas [ 12 ] , variations in elastic modulus [ 2 ] , void volume fractions in ductile materials [ 13 ] , microcrack density [ 14 ] , post-rupture bond-density, nucleation-growth-coalescence of microvoids etc. We find scalar, vector, or higher-order tensor representations of these damage variables [ 15 ] carrying information on degradation and its orientation. The key to successfully capture complex crack paths with the CDM is by constructing evolution equations consistent with the adopted interpretation of the damage variable. This important feature is however conspicuous by its absence in most models. CDM models also lack numerical stability unless length scales are incorporated. Using length scales, non-local and hence regularized damage models with length scales have been prescribed [ 16 ] so as to prevent localisation of damage to zero-volume, mesh dependence and unphysical softening [ 17 , 18 ] . Even from a physical standpoint, non-locality makes sense as damage at a material point might indeed depend on its neighbourhood [ 19 ] , especially via interactions among microcracks at the mesoscale. Both gradient and integral-based non-local models have been used for quasi-brittle fracture; but they have their drawbacks. Gradient-based models essentially originate from integral approaches [ 16 ] . Among these, phase-field damage models are widely used for quasi-brittle fracture [ 20 ] . Despite several advantages, it suffers from crucial drawbacks such as opaqueness in a physical interpretation of phase field, unphysical energy functionals, high computational cost, mesh-sensitivity [ 21 ] etc. The governing equation is ill-conditioned for soft materials, dynamic fracture, multi-modal fracture and material and geometric nonlinearities. To address these challenges, elaborate measures are typically required [ 22 ] , possibly at an additional computational cost. In some cases, numerical issues arise from the monolithic solving of the balance of linear momentum and damage evolution. To prevent this, an arbitrary history variable is introduced [ 23 ] . The high sensitivity of gradient-type models to algorithmic parameters is, to an extent, eliminated in integral-type models. Peridynamics [ 24 ] is a widely used integral model, but with questionable accuracy [ 25 ] . Efforts to enhance it have increased its complexity [ 26 ] . In the context of quasi-brittle fracture, various attempts to improve the