Seid Koric

• PhD
• Technical Associate Director and Research Professor at University of Illinois Urbana-Champaign

Real-time monitoring is a foundation of nuclear digital twin technology, crucial for detecting material degradation and maintaining nuclear system integrity. Traditional physical sensor systems face limitations, particularly in measuring critical parameters in hard-to-reach or harsh environments, often resulting in incomplete data coverage. Machine...

Under some initial and boundary conditions, the rapid reaction-thermal diffusion process taking place during frontal polymerization (FP) destabilizes the planar mode of front propagation, leading to spatially varying, complex hierarchical patterns in thermoset polymeric materials. Although modern reaction-diffusion models can predict the patterns r...

Effective real-time monitoring technique is crucial for detecting material degradation and maintaining the structural integrity of nuclear systems to ensure both safety and operational efficiency. Traditional physical sensor systems face limitations such as installation challenges, high costs, and difficulties in measuring critical parameters in ha...

Metamaterials, synthetic materials with customized properties, have emerged as a promising field due to advancements in additive manufacturing. These materials derive unique mechanical properties from their internal lattice structures, which are often composed of multiple materials that repeat geometric patterns. While traditional inverse design ap...

Data-driven models that act as surrogates for computationally costly 3D topology optimization techniques are very popular because they help alleviate multiple time-consuming 3D finite element analyses during optimization. In this study, one such 3D CNN-based surrogate model for the topology optimization of Schoen’s gyroid triply periodic minimal su...

The deep operator network (DeepONet) structure has shown great potential in approximating complex solution operators with low generalization errors. Recently, a sequential DeepONet (S-DeepONet) was proposed to use sequential learning models in the branch of DeepONet to predict final solutions given time-dependent inputs. In the current work, the S-...

Crystal plasticity (CP) model is a vital tool for understanding structure–property relations, but it is computationally expensive. Hence, data-driven models have been used as surrogate. We proposed a Deep Operator Network (DeepONet) to predict polycrystal stress–strain response. It employs a convolutional network to encode microstructure. To accoun...

Unlike classical artificial neural networks, which require retraining for each new set of parametric inputs, the Deep Operator Network (DeepONet), a lately introduced deep learning framework, approximates linear and nonlinear solution operators by taking parametric functions (infinite-dimensional objects) as inputs and mapping them to complete solu...

A state-of-the-art large eddy simulation code has been developed to solve compressible flows in turbomachinery. The code has been engineered with a high degree of scalability, enabling it to effectively leverage the many-core architecture of the new Sunway system. A consistent performance of 115.8 DP-PFLOPs has been achieved on a high-pressure turb...

The deep energy method (DEM), a type of physics-informed neural network, is evolving as an alternative to finite element analysis. It employs the principle of minimum potential energy to predict an object’s behavior under various boundary conditions. However, the model’s accuracy is contingent upon choosing the appropriate architecture for the mode...

A state-of-the-art large eddy simulation code has been developed to solve compressible flows in turbomachinery. The code has been engineered with a high degree of scalability, enabling it to effectively leverage the many-core architecture of the new Sunway system. A consistent performance of 115.8 DP-PFLOPs has been achieved on a high-pressure turb...

Deep Operator Network (DeepONet), a recently introduced deep learning operator network, approximates linear and nonlinear solution operators by taking parametric functions (infinite-dimensional objects) as inputs and mapping them to solution functions in contrast to classical neural networks (NNs) that need re-training for every new set of parametr...

A novel deep operator network (DeepONet) with a residual U-Net (ResUNet) as the trunk network is devised to predict full-field highly nonlinear elastic-plastic stress response for complex geometries obtained from topology optimization under variable loads. The proposed DeepONet uses a ResUNet in the trunk to encode complex input geometries, and a f...

The advent of data-driven and physics-informed neural networks has sparked interest in deep neural networks as universal approximators of solutions in various scientific and engineering communities. However, in most existing approaches, neural networks can only provide solutions for a fixed set of input parameters such as material properties, sourc...

Deep neural networks as universal approximators of partial differential equations (PDEs) have attracted attention in numerous scientific and technical circles with the introduction of Physics-informed Neural Networks (PINNs). However, in most existing approaches, PINN can only provide solutions for defined input parameters, such as source terms, lo...

The deep energy method (DEM), a type of physics-informed neural network, is evolving as an alternative to finite element analysis. This method employs the principle of minimum potential energy to predict deformations under static loading conditions. However, the model’s accuracy is contingent upon choosing the appropriate architecture for the model...

Triply periodic minimal surface (TPMS) metamaterials characterized by mathematically-controlled topologies exhibit better mechanical properties compared to uniform structures. The unit cell topology of such metamaterials can be further optimized to improve a desired mechanical property for a specific application. However, such inverse design involv...

The deep energy method (DEM) has been used to solve the elastic deformation of structures with linear elasticity, hyperelasticity, and strain-gradient elasticity material models based on the principle of minimum potential energy. In this work, we extend DEM to elastoplasticity problems involving path dependence and irreversibility. A loss function...

The cover image is based on the Research Article On the use of graph neural networks and shape‐function‐based gradient computation in the deep energy method by Junyan He et al., https://doi.org/10.1002/nme.7146.

Metal additive manufacturing (AM) involves complex multiscale and multiphysics processes. Physics-based modeling approaches to simulate such processes face challenges in their predictions due to the several time and length scales involved in the thermomechanical effects that are inherent in AM. Deep learning-based approaches have been recently expl...

This paper explores the possibilities of applying physics-informed neural networks (PINNs) in topology optimization (TO) by introducing a fully self-supervised TO framework based on PINNs. This framework solves the forward elasticity problem by the deep energy method (DEM). Instead of training a separate neural network to update the density distrib...

The paper explores the possibility of using the novel Deep Operator Networks (DeepONet) for forward analysis of numerically intensive and challenging multiphysics designs and optimizations of advanced materials and processes. As an important step towards that goal, DeepONet networks were devised and trained on GPUs to solve the Poisson equation (he...

Physics‐informed neural networks have gained growing interest. Specifically, they are used to solve partial differential equations governing several physical phenomena. However, physics‐informed neural network models suffer from several issues and can fail to provide accurate solutions in many scenarios. We discuss a few of these challenges and the...

A graph neural network (GCN) is employed in the deep energy method (DEM) model to solve the momentum balance equation in 3D for the deformation of linear elastic and hyperelastic materials due to its ability to handle irregular domains over the traditional DEM method based on a multilayer perceptron (MLP) network. Its accuracy and solution time are...

In hydrology, projected climate change impact assessment studies typically rely on ensembles of downscaled climate model outputs. Due to large modeling uncertainties, the ensembles are often averaged to provide a basis for studying the effects of climate change. A key issue when analyzing averages of a climate model ensemble is whether to weight al...

Surrogate neural network-based models have been lately trained and used in a variety of science and engineering applications where the number of evaluations of a target function is limited by execution time. In cell phone camera systems, various errors, such as interferences at the lens-barrel and lens-lens interfaces and axial, radial, and tilt mi...

In this work, we extend the deep energy method (DEM), which has been used to solve elastic deformation of structures, to problems involving classical elastoplasticity. A loss function for elastoplastic DEM is proposed, inspired by the discrete variational formulation of plasticity. The radial return algorithm is coupled with DEM to update the plast...

A graph neural network (GCN) is employed in the deep energy method (DEM) model to solve the momentum balance equation in 3D for the deformation of linear elastic and hyperelastic materials due to its ability to handle irregular domains over the traditional DEM method based on a multilayer perceptron (MLP) network. Its accuracy and solution time are...

This paper explores the possibilities of applying physics-informed neural networks (PINNs) in topology optimization (TO) by introducing a fully self-supervised TO framework that is based on PINNs. This framework solves the forward elasticity problem by the deep energy method (DEM). Instead of training a separate neural network to update the density...

The deep energy method (DEM) employs the principle of minimum potential energy to train neural network models to predict displacement at a state of equilibrium under given boundary conditions. The accuracy of the model is contingent upon choosing appropriate hyperparameters. The hyperparameters have traditionally been chosen based on literature or...

Using recent advancements in high-performance computing data assimilation to combine satellite InSAR data with numerical models, the prolonged unrest of the Sierra Negra volcano in the Galápagos was tracked to provide a fortuitous, but successful, forecast 5 months in advance of the 26 June 2018 eruption. Subsequent numerical simulations reveal tha...

Physics-informed neural networks have gained growing interest. Specifically, they are used to solve partial differential equations governing several physical phenomena. However, physics-informed neural network models suffer from several issues and can fail to provide accurate solutions in many scenarios. We discuss a few of these challenges and the...

The potential energy formulation and deep learning are merged to solve partial differential equations governing the deformation in hyperelastic and viscoelastic materials. The presented deep energy method (DEM) is self-contained and meshfree. It can accurately capture the three-dimensional (3D) mechanical response without requiring any time-consumi...

The potential energy formulation and deep learning are merged to solve partial differential equations governing the deformation in hyperelastic and viscoelastic materials. The presented deep energy method (DEM) is self-contained and meshfree. It can accurately capture the three-dimensional (3D) mechanical response without requiring any time-consumi...

Surrogate neural network-based models have been lately trained and used in a variety of science and engineering applications where the number of evaluations of a target function is limited by execution time. In cell phone camera systems, various errors, such as interferences at the lens-barrel and lens-lens interfaces and axial, radial, and tilt mi...

Recent work has shown that deep learning provides an alternative solution as an efficient function approximation technique for airfoil surrogate modeling. In this paper we present the feasibility of convolutional neural network (CNN) techniques for aerodynamic performance evaluation. CNN approach will enable designer to fully utilize the ability of...

Iterative methods are widely used for solving sparse linear systems of equations and eigenvalue problems. Their performances are relevant to the conditioning of the linear systems. This work explores factors which affects the conditioning of the discretized system, including material heterogeneity, different constitutive characteristics and element...

Deep learning and the collocation method are merged and used to solve partial differential equations describing structures' deformation. We have considered different types of materials: linear elasticity, hyperelasticity (neo‐Hookean) with large deformation, and von Mises plasticity with isotropic and kinematic hardening. The performance of this de...

Recent work has shown that deep learning provides an alternative solution as an efficient function approximation technique for airfoil surrogate modeling. In this paper we present the feasibility of convolutional neural network (CNN) techniques for aerodynamic performance evaluation. CNN approach will enable designer to fully utilize the ability of...

This paper introduces a computational design framework for obtaining 3D periodic elastoplastic architected materials with enhanced performance, subject to uniaxial or shear strain. A nonlinear finite element model accounting for plastic deformation is developed, where a Lagrange multiplier approach is utilized to impose periodicity constraints. The...

The solidifying steel follows highly nonlinear thermo-mechanical behavior depending on the loading history, temperature, and metallurgical phase fraction calculations (liquid, ferrite, and austenite). Numerical modeling with a computationally challenging multiphysics approach is used on high-performance computing to generate sufficient training and...

Predicting history-dependent materials’ responses is crucial, as path-dependent behavior appears while characterizing or geometrically designing many materials (e.g., metallic and polymeric cellular materials), and it takes place in manufacturing and processing of many materials (e.g., metal solidification). Such phenomena can be computationally in...

Deep learning and the collocation method are merged and used to solve partial differential equations describing structures' deformation. We have considered different types of materials: linear elasticity, hyperelasticity (neo-Hookean) with large deformation, and von Mises plasticity with isotropic and kinematic hardening. The performance of this de...

Fluid-Structure Interaction (FSI) simulations have applications to a wide range of engineering areas. One popular technique to solve FSI problems is the Arbitrary Lagrangian-Eulerian (ALE) method. Both academic and industry communities developed codes to implement the ALE method. One of them is Alya, a Finite Element Method (FEM) based code develop...

Data-driven models are rising as an auspicious method for the geometrical design of materials and structural systems. Nevertheless, existing data-driven models customarily address the optimization of structural designs rather than metamaterial designs. Metamaterials are emerging as promising materials exhibiting tailorable and unprecedented propert...

Background and objective Finite element models built from micro-computed tomography scans have become a powerful tool to investigate the mechanical properties of trabecular bone. There are two types of solving algorithms in the finite element method: implicit and explicit. Both of these methods have been utilized to study the trabecular bone. Howev...

Identifying the appropriate parameters of a turbulence model for a class of flow usually requires extensive experimentation and numerical simulations. Therefore even a modest improvement of the turbulence model can significantly reduce the overall cost of a three-dimensional, time-dependent simulation. In this paper we demonstrate a novel method to...

Significant investments to upgrade and construct large-scale scientific facilities demand commensurate investments in R&D to design algorithms and computing approaches to enable scientific and engineering breakthroughs in the big data era. Innovative Artificial Intelligence (AI) applications have powered transformational solutions for big data chal...

Among advanced manufacturing techniques for Fiber-Reinforced Polymer-matrix Composites (FRPCs) which are critical for aerospace, marine, automotive, and energy industries, Frontal Polymerization (FP) has been recently proposed to save orders of magnitude time and energy. However, the cure kinetics of the matrix phase, usually a thermosetting polyme...

The field of optimal design of linear elastic structures has seen many exciting successes that resulted in new architected materials and structural designs. With the availability of cloud computing, including high-performance computing, machine learning, and simulation, searching for optimal nonlinear structures is now within reach. In this study,...

LS-DYNA is a well-known multiphysics code with both explicit and implicit time stepping capabilities. Implicit simulations rely heavily on sparse matrix computations, in particular direct solvers, and are notoriously much harder to scale than explicit simulations. In this paper, we investigate the scalability challenges of the implicit structural m...

Significant investments to upgrade or construct large-scale scientific facilities demand commensurate investments in R&D to design algorithms and computing approaches to enable scientific and engineering breakthroughs in the big data era. The remarkable success of Artificial Intelligence (AI) algorithms to turn big-data challenges in industry and t...

The field of optimal design of linear elastic structures has seen many exciting successes that resulted in new architected materials and structural designs. With the availability of cloud computing, including high-performance computing, machine learning, and simulation, searching for optimal nonlinear structures is now within reach. In this study,...

Model extrapolation to unseen flow is one of the biggest challenges facing data-driven turbulence modeling, especially for models with high dimensional inputs that involve many flow features. In this study we review previous efforts on data-driven Reynolds-Averaged Naiver Stokes (RANS) turbulence modeling and model extrapolation, with main focus on...

During the solidification of stainless steel, the mechanical behavior of the solidifying shell follows nonlinear elastic-viscoplastic constitutive laws depending on metallurgical phase fraction calculations (liquid, ferrite and austenite). A multiphysics model that couples thermal and mechanical behavior in a Lagrangian reference frame, including b...

High-resolution spatiotemporal data is crucial for characterizing, modeling, and monitoring the space–time dynamics of complex systems in manufacturing. However, the acquisition of such data is generally expensive and time-consuming. Spatiotemporal interpolation aims to predict the values at unmeasured locations using measured data, and emerges as...

This manuscript discusses the role of high performance computing (HPC) in direct numerical simulation (DNS) of the nonlinear mechanical behavior of 3D architectured materials for which there has been recently an impressive renewal of interest predominantly because of the development of advanced manufacturing methods. The complex mechanical behavior...

Geopolymers are X-ray amorphous materials with appealing properties such as high flexural strength and high compressive strength. Yet, the influence of the heterogeneity and porosity on the constitutive behavior is not fully understood. We formulate a nonlinear multiscale mathematical model to describe the strength behavior of geopolymer composites...

As a type of architectured material, knitted textiles exhibit global mechanical behavior which is affected by their microstructure defined at the scale at which yarns are arranged topologically given the type of textile manufactured. To relate local geometrical, interfacial, material, kinematic and kinetic properties to global mechanical behavior,...

Direct numerical simulations (DNS) of knitted textile mechanical behavior are for the first time conducted on high performance computing (HPC) using both the explicit and implicit finite element analysis (FEA) to directly assess effective ways to model the behavior of such complex material systems. Yarn-level models including interyarn interactions...

Fracture analysis of a cortical bone sample from a tibia of a 70 years-old human male donor is conducted computationally using an extended finite element method. The cortical bone microstructure is represented by several osteons arranged based on bone microscopy image. The accuracy of results is examined by comparing a linear elastic fracture mecha...

The discretization of partial differential equations of complex physical problems involves solving linear systems of equations with a great number of unknowns. The resultant matrix obtained from this discretization is often sparse and ill-conditioned. In many cases problems are solved in fine structured meshes with irregular geometries yielding ill...

With the rapid development of sensing, communication, and computing technologies and infrastructure, today’s manufacturing industry is marching towards a big data era and a new generation of digitalization and intelligence. The availability of big data provides us with a golden opportunity to promote smart manufacturing. Nevertheless, the deploymen...

Larger supercomputers allow the simulation of more complex phenomena with increased accuracy. Eventually this requires finer and thus also larger geometric discretizations. In this context, and extrapolating to the Exascale paradigm, meshing operations such as generation, deformation, adaptation/regeneration or partition/load balance, become a crit...

Bone consists of cortical (compact) bone and trabecular (spongy) bone. Cortical bone forms the outer shell, and trabecular bone is present within or at the ends of long bones. Trabecular bone plays a significant role in load carrying and fracture resistance of bone because of its spongy structure. There is a high interest in tissue level and overal...

The flow characteristics of spherical bodies, arising in a variety of important engineering and environmental problems, range from laminar to turbulent flow. Turbulent flows are predominantly studied using the models based on Reynolds-averaged Navier-Stokes (RANS) equations. Especially, in case of flows around bluff bodies RANS models have limitati...

In this study, the sensitivity of the apparent response of trabecular bone to different constitutive models at the tissue-level was investigated using finite element modeling based on micro-computed tomography. Trabecular bone specimens from porcine femurs were loaded under a uniaxial compression experimentally and computationally. The apparent beh...

Emerging peta-scale computing is already a strategic enabler of large- scale simulations in many scientific areas such as astronomy, biology, material science and chemistry. In 2014 it was also proven that engi- neering simulation codes scaled on supercomputers, scaling BSC’s Alya multiphysics code up to 100,000 cores on NCSA’s Blue Waters. Recentl...

The research objective is to elucidate the microstructure-strength relationships in geopolymer composites. Geopolymer is an inorganic polymeric material, which consists primarily of alumina, silica and alkali metal oxides. Geopolymer composites exhibit excellent adhesive, thermal, and chemical properties making them relevant in a diverse range of a...

Geopolymer is a novel cementitious material that was first developed by Davidovits in 1978. It serves as a good alternative for ordinary Portland cement (OPC). Geopolymers are a class of inorganic polymeric amorphous materials that can be synthesized at low temperature. The synthesis of geopolymer involves the mixing of an aluminosilicate source wi...

Geopolymer is an inorganic polymeric material, which consists primarily of alumina, silica and alkali metal oxides. Geopolymer exhibits excellent adhesive, thermal, and chemical properties making them relevant in a diverse range of applications such as fire resistant materials, environment-friendly binders, composites for infrastructure repair/stre...

With prolific advances in science and technology, there is a constant need for state-of-the-art materials that are strong, tough, light weight, yet thermally and chemically stable. In the recent past, geopolymer has attained much consideration due to numerous benefits over common matrix materials. Geopolymer is an inorganic polymeric material with...

Highly nonlinear finite element (FE) model of trabecular bone, using its actual complex micro-scale geometry imaged with micro-computed tomography (micro-CT), was used to study the apparent trabecular bone mechanical behavior and compared with experimental data. Sensitivity of the apparent-level response to different plasticity constitutive models...

Sparse matrix factorization, a critical algorithm in many science and engineering applications, has had difficulty leveraging the additional computational power afforded by the infusion of heterogeneous accelerators in HPC clusters. We present a minimally invasive approach to the GPU acceleration of a hybrid multifrontal solver, the Watson Sparse M...

We study computationally, using a finite element method, plastic deformations and strength of trabecular bone. We represent the trabecular bone as a material with hierarchical structure and consider four structural scales: the nanoscale (mineralized collagen fibril), the sub-microscale (single lamella consisting of mineralized collagen fibrils), th...

The depth-dependent strain partitioning across the interfaces in the growth direction of the NiAl/Cr(Mo) nanocomposite between the Cr and NiAl lamellae was directly measured experimentally and simulated using a finite element method (FEM). Depth-resolved X-ray microdiffraction demonstrated that in the as-grown state both Cr and NiAl lamellae grow a...

Parallel linear equation solvers are one of the most important components determining the scalability and efficiency of many supercomputing applications. Several groups and companies are leading the development of linear system solver libraries for HPC applications. In this paper, we present an objective performance test study for the solvers avail...

Alya is a multi-physics simulation code developed at Barcelona Supercomputing Center (BSC). From its inception Alya code is designed using advanced High Performance Computing programming techniques to solve coupled problems on supercomputers efficiently. The target domain is engineering, with all its particular features: complex geometries and unst...

High performance computing is absolutely necessary for large-scale geophysical simulations. In order to obtain a realistic image of a geologically complex area, industrial surveys collect vast amounts of data making the computational cost extremely high for the subsequent simulations. A major computational bottleneck of modeling and inversion algor...

The progress in high performance computing and parallel algorithms now allows the modelling of trabecular (spongy) bone using its actual complex geometry. In this paper, the nonlinear computational model of trabecular bone, that includes plasticity and contact mechanics, is applied to trabecular bone architectures imaged using micro-computed tomogr...

Modern numerical algorithms for computational electromagnetics lead to many large sparse systems of linear equations. Their solution takes up to 90% of the total computational time in the geophysical inversion process. This paper provides evaluation and comparison of several state-of-the-art direct solvers in a massively parallel environment. We de...

Turbulent flow at Reynolds number 50,000 around a sphere with trip wire placed 75 from the frontal stagnation point is investigated using detached eddy simulation (DES). The time-averaged results of the velocity field agree considerably well with the measurements from literature and reveal good reproduction of the separation zone, as well as the e...

The spring lattice models offer a powerful way of modeling damage evolution in disordered materials by explicitly representing the disorder, microcrack nucleation and coalescence processes. The evolution of anisotropic damage tensor is studied using spring lattice models in 2D and 3D, where presence of disorder leads to the size effects in stre...

In this study, a planar spring lattice model is used to study the evolution of damage variable d L in disordered media. An elastoplastic softening damage constitutive law is implemented which introduces a cohesive length scale in addition to the disorder-induced one. The cohesive length scale affects the macroscopic response of the lattice with the...

Industrial Applications of High-Performance Computing: Best Global Practices offers a global overview of high-performance computing (HPC) for industrial applications, along with a discussion of software challenges, business models, access models (e.g., cloud computing), public-private partnerships, simulation and modeling, visualization, big data a...

In this study 2d two phase microstructures closely resembling the experimentally captured micrographs of the interpenetrating phase composites are generated using a Gaussian correlation function based method. The scale dependent bounds on the effective thermal conductivity of such microstructures are then studied using Hill-Mandel boundary conditio...