arXiv subjects

Joyjit Chattoraj

Publications and source records attributed to Joyjit Chattoraj.

16 recordsLinked to original sources

arXiv2025

AceWGS: An LLM-Aided Framework to Accelerate Catalyst Design for Water-Gas Shift Reactions

While the Water-Gas Shift (WGS) reaction plays a crucial role in hydrogen production for fuel cells, finding suitable catalysts to achieve high yields for low-temperature WGS reactions remains a persistent challenge. Artificial Intelligence (AI) has shown promise in accelerating catalyst design by exploring vast candidate spaces, however, two key gaps limit its effectiveness. First, AI models primarily train on numerical data, which fail to capture essential text-based information, such as catalyst synthesis methods. Second, the cross-disciplinary nature of catalyst design requires seamless collaboration between AI, theory, experiments, and numerical simulations, often leading to communication barriers. To address these gaps, we present AceWGS, a Large Language Models (LLMs)-aided framework to streamline WGS catalyst design. AceWGS interacts with researchers through natural language, answering queries based on four features: (i) answering general queries, (ii) extracting information about the database comprising WGS-related journal articles, (iii) comprehending the context described in these articles, and (iv) identifying catalyst candidates using our proposed AI inverse model. We presented a practical case study demonstrating how AceWGS can accelerate the catalyst design process. AceWGS, built with open-source tools, offers an adjustable framework that researchers can readily adapt for a range of AI-accelerated catalyst design applications, supporting seamless integration across cross-disciplinary studies.

cs.CL↗

arXiv2024

Tailoring Generative Adversarial Networks for Smooth Airfoil Design

In the realm of aerospace design, achieving smooth curves is paramount, particularly when crafting objects such as airfoils. Generative Adversarial Network (GAN), a widely employed generative AI technique, has proven instrumental in synthesizing airfoil designs. However, a common limitation of GAN is the inherent lack of smoothness in the generated airfoil surfaces. To address this issue, we present a GAN model featuring a customized loss function built to produce seamlessly contoured airfoil designs. Additionally, our model demonstrates a substantial increase in design diversity compared to a conventional GAN augmented with a post-processing smoothing filter.

cs.LG↗

arXiv2023

A Surrogate-Assisted Extended Generative Adversarial Network for Parameter Optimization in Free-Form Metasurface Design

Metasurfaces have widespread applications in fifth-generation (5G) microwave communication. Among the metasurface family, free-form metasurfaces excel in achieving intricate spectral responses compared to regular-shape counterparts. However, conventional numerical methods for free-form metasurfaces are time-consuming and demand specialized expertise. Alternatively, recent studies demonstrate that deep learning has great potential to accelerate and refine metasurface designs. Here, we present XGAN, an extended generative adversarial network (GAN) with a surrogate for high-quality free-form metasurface designs. The proposed surrogate provides a physical constraint to XGAN so that XGAN can accurately generate metasurfaces monolithically from input spectral responses. In comparative experiments involving 20000 free-form metasurface designs, XGAN achieves 0.9734 average accuracy and is 500 times faster than the conventional methodology. This method facilitates the metasurface library building for specific spectral responses and can be extended to various inverse design problems, including optical metamaterials, nanophotonic devices, and drug discovery.

cs.LG↗

arXiv2023

Graph Neural Network Based Surrogate Model of Physics Simulations for Geometry Design

Computational Intelligence (CI) techniques have shown great potential as a surrogate model of expensive physics simulation, with demonstrated ability to make fast predictions, albeit at the expense of accuracy in some cases. For many scientific and engineering problems involving geometrical design, it is desirable for the surrogate models to precisely describe the change in geometry and predict the consequences. In that context, we develop graph neural networks (GNNs) as fast surrogate models for physics simulation, which allow us to directly train the models on 2/3D geometry designs that are represented by an unstructured mesh or point cloud, without the need for any explicit or hand-crafted parameterization. We utilize an encoder-processor-decoder-type architecture which can flexibly make prediction at both node level and graph level. The performance of our proposed GNN-based surrogate model is demonstrated on 2 example applications: feature designs in the domain of additive engineering and airfoil design in the domain of aerodynamics. The models show good accuracy in their predictions on a separate set of test geometries after training, with almost instant prediction speeds, as compared to O(hour) for the high-fidelity simulations required otherwise.

cs.LG↗

arXiv2022

Shear-induced mixing of granular materials featuring broad granule size distributions

Granular flows during a shear-induced mixing process are studied using Discrete Element Methods. The aim is to understand the underlying elementary mechanisms of transition from unmixed to mixed phases for a granular material featuring a broad distribution of particles, which we investigate systematically by varying the strain rate and system size. Here the strain rate varies over four orders of magnitude and the system size varies from ten thousand to more than a million granules. A strain rate-dependent transition from quasistatic to purely inertial flow is observed. At the macroscopic scale, the contact stresses drop due to the formation of shear-induced instabilities that serves as an onset of granular flows and initiates mixing between the granules. The stress-drop displays a profound system size dependence. At the granular scale, mixing dynamics are correlated with the formation of shear bands, which result in significantly different timescales of mixing, especially for those regions that are close to the system walls and the bulk. Overall, our results reveal that although the transient dynamics display a generic behavior these have a significant finite-size effect. In contrast, macroscopic behaviors at steady states have negligible system size dependence.

cond-mat.soft↗

arXiv2021

Emergence of linear isotropic elasticity in amorphous and polycrystalline materials

We investigate the emergence of isotropic linear elasticity in amorphous and polycrystalline solids, via extensive numerical simulations. We show that the elastic properties are correlated over a finite length scale $ξ_E$, so that the central limit theorem dictates the emergence of continuum linear isotropic elasticity on increasing the specimen size. The stiffness matrix of systems of finite size $L > ξ_E$ is obtained adding to that predicted by linear isotropic elasticity a random one of spectral norm $(L/ξ_E)^{-3/2}$, in three spatial dimensions. We further demonstrate that the elastic length scale corresponds to that of structural correlations, which in polycrystals reflect the typical size of the grain boundaries and length scales characterizing correlations in the stress field. We finally demonstrate that the elastic length scale affects the decay of the anisotropic long-ranged correlations of locally defined shear modulus and shear stress.

cond-mat.soft↗

arXiv2021

Designing Phononic Band Gaps with Sticky Potentials

Spectral gaps in the vibrational modes of disordered solids are key design elements in the synthesis and control of phononic metamaterials that exhibit a plethora of novel elastic and mechanical properties. However, reliably producing these gaps often require a high degree of network specificity through complex control optimization procedures. In this work, we present as an additional tool to the existing repertoire, a numerical scheme that rapidly generates sizeable spectral gaps in absence of any fine tuning of the network structure or elastic parameters. These gaps occur even in disordered polydisperse systems consisting of relatively few particles ($N \sim 10^2-10^3$). Our proposed procedure exploits sticky potentials that have recently been shown to suppress the formation of soft modes, thus effectively recovering the linear elastic regime where band structures appear, at much shorter length scales than in conventional models of disordered solids. Our approach is relevant to design and realization of gapped spectra in a variety of physical setups ranging from colloidal suspensions to 3D-printed elastic networks.

cond-mat.soft↗

arXiv2020

Frictional Active Brownian Particles

Frictional forces affect the rheology of hard-sphere colloids, at high shear rate. Here we demonstrate, via numerical simulations, that they also affect the dynamics of active Brownian particles, and their motility induced phase separation. Frictional forces increase the angular diffusivity of the particles, in the dilute phase, and prevent colliding particles from resolving their collision by sliding one past to the other. This leads to qualitatively changes of motility-induced phase diagram in the volume-fraction motility plane. While frictionless systems become unstable towards phase separation as the motility increases only if their volume fraction overcomes a threshold, frictional system become unstable regardless of their volume fraction. These results suggest the possibility of controlling the motility induced phase diagram by tuning the roughness of the particles.

cond-mat.soft↗

arXiv2020

The stability phase diagram of active Brownian particles

Phase separation in a low-density gas-like phase and a high-density liquid-like one is a common trait of biological and synthetic self-propelling particles' systems. The competition between motility and stochastic forces is assumed to fix the boundary between the homogeneous and the phase-separated phase. Here we demonstrate that motility does also promote the homogeneous phase allowing particles to resolve their collisions. This new understanding allows quantitatively predicting the spinodal-line of hard self-propelling Brownian particles, the prototypical model exhibiting a motility induced phase separation. Furthermore, we demonstrate that frictional forces control the physical process by which motility promotes the homogeneous phase. Hence, friction emerges as an experimentally variable parameter to control the motility induced phase diagram.

cond-mat.soft↗

arXiv2020

Role of attractive forces in the relaxation dynamics of supercooled liquids

The attractive tail of the intermolecular interaction affects very weakly the structural properties of liquids, while it affects dramatically their dynamical ones. Via the numerical simulations of model systems not prone to crystallization, both in three and in two spatial dimensions, here we demonstrate that the non-perturbative dynamical effects of the attractive forces are tantamount to a rescaling of the activation energy by the glass transition temperature $T_g$: systems only differing in their attractive interaction have the same structural and dynamical properties if compared at the same value of $T/T_g$.

cond-mat.soft↗

arXiv2019

Oscillatory Instabilities in 3-Dimensional Frictional Granular Matter

The dynamics of amorphous granular matter with frictional interactions cannot be derived in general from a Hamiltonian and therefore displays oscillatory instabilities stemming from the onset of complex eigenvalues in the stability matrix. These instabilities were discovered in the context of one and two dimensional systems, while the three dimensional case was never studied in detail. Here we fill this gap by deriving and demonstrating the presence of oscillatory instabilities in a three dimensional granular packing. We study binary assemblies of spheres of two sizes interacting via classical Hertz and Mindlin force laws for the longitudinal and tangent interactions, respectively. We formulate analytically the stability matrix in 3D and observe that a couple of complex eigenvalues emerges at the onset of the instability as in the case of frictional disks in two-dimensions. The dynamics then shows oscillatory exponential growth in the Mean-Square-Displacement, followed by a catastrophic event. The generality of these results for any choice of forces that break the symplectic Hamiltonian symmetry is discussed.

cond-mat.stat-mech↗

arXiv2019

Transition from Static to Dynamic Friction in an Array of Frictional Disks

The nature of an instability that controls the transition from static to dynamical friction is studied in the the context of an array of frictional disks that are pressed from above on a substrate. In this case the forces are all explicit and Newtonian dynamics can be employed without any phenomenological assumptions. We show that an oscillatory instability that had been discovered recently is responsible for the transition, allowing individual disks to spontaneously reach the Coulomb limit and slide with dynamic friction. The transparency of the model allows a full understanding of the phenomenon, including the speeds of the waves that travel from the trailing to the leading edge and vice versa.

cond-mat.stat-mech↗

arXiv2019

Noise Amplification in Frictional Systems: Oscillatory Instabilities

It was discovered recently that frictional granular materials can exhibit an important mechanism for instabilities, i.e the appearance of pairs of complex eigenvalues in their stability matrix. The consequence is an oscillatory exponential growth of small perturbations which are tamed by dynamical nonlinearities. The amplification can be giant, many orders of magnitude, and it ends with a catastrophic system-spanning plastic event. Here we follow up on this discovery, explore the scaling laws characterizing the onset of the instability, the scenarios of the development of the instability with and without damping, and the nature of the eventual system spanning events. The possible relevance to earthquake physics and to the transition from static to dynamic friction is discussed.

cond-mat.soft↗

arXiv2019

Oscillatory Instabilities in Frictional Granular Matter

Frictional granular matter is shown to be fundamentally different in its plastic responses to external strains from generic glasses and amorphous solids without friction. While regular glasses exhibit plastic instabilities due to a vanishing of a real eigenvalue of the Hessian matrix, frictional granular materials can exhibit a previously unnoticed additional mechanism for instabilities, i.e. the appearance of a pair of complex eigenvalues leading to oscillatory exponential growth of perturbations which are tamed by dynamical nonlinearities. This fundamental difference appears crucial for the understanding of plasticity and failure in frictional granular materials. The possible relevance to earthquake physics is discussed.

cond-mat.soft↗

arXiv2011

Robustness of avalanche dynamics in sheared amorphous solids as probed by transverse diffusion

Using numerical simulations, we perform an extensive finite-size analysis of the transverse diffusion coefficient in a sheared 2D amorphous solid, over a broad range of strain rates, at temperatures up to the supercooled liquid regime. We thus obtain direct qualitative evidence for the persistence of correlations between elementary plastic events up to the vicinity of the glass transition temperature $T_g$. A quantitative analysis of the data, combined with a previous study of the $T$- and $\dotγ$-dependence of the macroscopic stress \cite{ChattorajCaroliLemaitre2010}, leads us to conclude that the average avalanche size remains essentially unaffected by temperature up to $T\sim0.75 T_g$.

cond-mat.mtrl-sci↗

arXiv2010

Universal, additive effect of temperature on the rheology of amorphous solids

Extensive measurements of macroscopic stress in a 2D LJ glass, over a broad range of temperatures ($T$) and strain rates ($\dotγ$), demonstrate a very significant decrease of the flowing stress with $T$, even much below the glass transition. A detailed analysis of the interplay between loading, thermal activation, and mechanical noise leads us to propose that over a broad ($\dotγ,T$) region, the effect of temperature amounts to a mere lowering of the strains at which plastic events occur, while the athermal avalanche dynamics remains essentially unperturbed. Temperature is then shown to correct the athermal stress by a (negative) additive contribution which presents a universal form, thus bringing support to and extending an expression proposed by Johnson and Samwer [1]. Our prediction is shown to match strikingly well numerical data up to the vicinity of $T_g$.

cond-mat.mtrl-sci↗