Representation Of Microstructure Research Articles

A deep learning convolutional neural network and multi-layer perceptron hybrid fusion model for predicting the mechanical properties of carbon fiber

Recently, deep learning methods have become one of the hottest topics in predicting material properties, however, one bottleneck in current research is the simultaneous analysis of heterogeneous data. In this study, a deep learning fusion model is developed for the first time to predict the material properties of carbon fiber monofilament using textual (macroscopic properties of composites and matrix) and visual (two-point statistics of microstructures) data. For this, 1200 stochastic microstructures are generated using the greedy-based generation (GBG) algorithm. Then, the statistical representations of microstructures are determined using two-point statistics and the macroscopic properties are calculated based on a micro-scale finite element (FE) simulation. Finally, the visual and textual data are fed into the convolutional neural network (CNN) and multi-layer perceptron (MLP) fusion model for predicting the mechanical properties of carbon fibers. The developed hybrid CNN-MLP fusion model achieves encouraging average testing R2 of longitudinal modulus, transverse modulus, in-plane shear modulus, major Poisson’s ratio, and out-of-plane shear modulus of carbon fibers with values of 0.991, 0.969, 0.984, 0.903, and 0.955, respectively. Thus, the proposed strategy provides a promising framework for predicting material properties via multisource heterogeneous data and is expected to accelerate the smart design and optimization of materials.

Effects of lead and lean in multi-axis directed energy deposition

The present study examines the effect of varying laser incidence angles on textural, microstructural, and geometric characteristics of directed energy deposition (DED) processed materials, providing a more comprehensive outlook on participating laser-matter interaction phenomena and ultimately devising strategies to ameliorate print performance. In this study, single-layer, single-/multi-track specimens were processed to examine the effect of non-orthogonal angular configurations on bead morphology, microstructure, phase composition, and textural representation of DED-processed 316L stainless steel materials. It was observed that bead size decreased at increasing lead and lean angles. Asymmetry in the distribution of the bead morphology as a function of lead angle indicates better catchment for acute lead angle configurations over obtuse configurations. No significant differences in phase composition, texture, and microstructure were observed in moderate off-axis configurations. When the penetration depth for the deposits was below 20 μm, columnar structures dominated the microstructure of the deposited material. At deeper penetration depths, columnar and equiaxed structures were observed at the bead-substrate interface and center of the bead, respectively. Compared to powder-blown DED, wire-DED dilution profiles were found to be asymmetric in both orthogonal and non-orthogonal wire DED samples.

A microstructure-based fatigue model for additively manufactured Ti-6Al-4V, including the role of prior [formula omitted] boundaries

Microstructure-based models of additive manufactured (AM) Ti-6Al-4V should faithfully represent the unique microstructural features of these materials to provide a more thorough understanding of their role in the mechanical performance of the material. For AM Ti-6Al-4V, the prior β boundaries are likely sites of microscopic plastic strain localization, often leading to fatigue crack initiation. Within the context of crystal plasticity finite element methods, there is an existing gap in the current literature for creating synthetic microstructures of Ti-6Al-4V that capture both the prior β boundaries and α laths. This work focuses on generating such statistically equivalent microstructures, where the prior β boundaries and α laths are explicitly modeled. The framework can generate synthetic microstructures that consider one prior β grain or multiple β grains (and thus prior β grain boundaries) and their associated α laths forming a Widmanstätten microstructure, which are used as inputs for fatigue simulations. Within the fatigue model, the critical accumulated plastic strain energy density (APSED) is a metric used to predict the location and number of cycles of crack initiation. From a statistical average perspective, the presence of prior β grain boundaries was detrimental to the fatigue performance. The effectiveness of the prior β grain boundary in impeding slip, and therefore localizing APSED, is dependent on the loading conditions, prior β orientations, and α variants present in neighboring prior β grains. A higher percentage of failures occurred at the prior β grain boundary in the case of lower applied stress amplitudes, where the applied loads were below the percolation limit of the material and microplasticity was localized to a few locations relative to the microstructure. Additionally, for the case of crack initiation associated with the α variants within a given prior β grain, a combination of soft-hard α lath orientations resulted in sites of APSED localization and predicted fatigue crack initiation. The framework generated in this paper to create representative synthetic microstructures of the Widmanstätten microstructures associated with AM Ti-6Al-4V and the associated fatigue modeling enables trade-off studies between microstructural features and fatigue performance.

Polarization multiplexing multichannel high-Q terahertz sensing system

Terahertz functional devices with high-Q factor play an important role in spectral sensing, security imaging, and wireless communication. The reported terahertz devices based on the electromagnetic induction transparency (EIT) effect cannot meet the needs of high-Q in practical applications due to the low-Q factor. Therefore, to increase the Q-factor of resonance, researchers introduced the concept of bound state in the continuum (BIC). In the quasi-BIC state, the metasurface can be excited by the incident wave and provide resonance with a high-Q factor because the condition that the resonant state of the BIC state is orthogonal is not satisfied. The split ring resonator (SRR) is one of the most representative artificial microstructures in the metasurface field, and it shows great potential in BIC. In this paper, based on the classical single-SRR array structure, we combine the large and small SRR and change the resonance mode of the inner and outer SRR by changing the outer radius of the inner SRR. The metasurface based on parameter-tuned BIC verified that the continuous modulation of parameters in a system could make a pair of resonant states strongly coupled, and the coherent cancellation of the resonant states will cause the linewidth of one of the resonant states to disappear, thus forming BIC. Compared with the single-SRR array metasurface based on symmetry-protected BIC, the dual-SRR array metasurface designed in this paper has multiple accidental BICs and realizes multichannel multiplexing of X-polarization and Y-polarization. It provides a brilliant platform for high-sensitivity optical sensor array, low threshold laser and efficient optical harmonic generation.

A 3-D micromechanical framework to study damage propagation and failure of printed aligned discontinuous fiber-reinforced composites

This paper presents a novel and efficient micromechanical framework for finite element simulation of damage and failure in 3-D printed aligned discontinuous fiber-reinforced composites. The framework can predict the initiation and propagation of different types of damage in the composite under tensile loading along the fibers’ axis. The micromechanical framework includes the microstructural representation of the composite with explicit fibers and matrix in addition to linear expansions at the ends. Accurate constitutive equations are utilized for fibers, matrix, and fiber/matrix interfaces in the microstructural representation. Fibers’ locations and lengths are generated randomly; based on a target distribution for the fibers’ aspect ratios measured experimentally; within the microstructural representation. Optimal microstructural representative dimensions for reliable investigation of the mechanical response are computed by conducting sensitivity analyses. The accuracy of the micromechanical framework is validated versus the experimental results of a 3-D printed aligned discontinuous fiber-reinforced composite as the composite of interest. It is shown that the proposed framework can simulate various aspects of the mechanical response, including the failure pattern and stress-strain behavior. Subsequently, the sensitivity of the mechanical response of the composite to a few constitutive equation-related parameters, including the strength and fracture toughness of the fiber/matrix interfaces and the matrix strength, is investigated. The correlation between the studied parameters and the composite’s strength and failure pattern is also analyzed and discussed. This paper delivers valuable guidance on the mechanical response of printed aligned discontinuous fiber-reinforced composites, eventually resulting in the paradigm-shifting design, manufacturing, and analysis of such advanced composites.

Effect of microstructure heterogeneity shapes on constitutive behaviour of encapsulated self-healing cementitious materials

Self-healing cementitious materials with microcapsules are complex multiscale and multiphase materials. The random microstructure of these materials governs their mechanical and transport behaviour. The actual microstructure can be represented accurately with a discrete lattice model, but computational restrictions mean that the size of domain that can be considered with this approach is limited. By contrast, a smeared approach, based on a micromechanical formulation, provides an approximate representation of the material microstructure with low computational costs. The aim of this paper is to compare simulations of a microcapsule-based self-healing cementitious system with discrete-lattice and smeared-micromechanical models, and to assess the relative strengths and weaknesses of these models for simulating distributed fracture and healing in this type of self-healing material. A novel random field generation technique is used to represent the microstructure of a cementitious mortar specimen. The meshes and elements are created by the triangulation method and used to determine the input required for the lattice model. The paper also describes the enhancement of the TUDelft lattice model to include self-healing behaviour. The extended micromechanical model considers both microcracking and healing. The findings from the study provide insight into the relative merits of these two modelling approaches.

On the anisotropy of thick-walled wire arc additively manufactured stainless steel parts

Wire Arc Additive Manufacturing (WAAM) is an emerging group of methods for producing large parts with complex geometries and varying wall thicknesses. These parts usually exhibit anisotropic material behavior due to their intrinsic heterogeneous microstructure. To fully exploit the versatility of WAAM, a rigorous understanding of the relationship between processing conditions, microstructure, and mechanical response of WAAM parts is necessary. To this end, this paper investigates the structure-property relationship for thick-walled austenitic stainless steel WAAM parts experimentally and numerically using a mean-field crystal plasticity model. The major microstructural features are studied using optical microscopy and electron backscattered diffraction. A representative microstructure volume element is obtained with averaged features to study spatial variations in the microstructure across the WAAM part. Uniaxial tensile tests assisted with Digital Image Correlation along the transverse direction, diagonal (45o from the transverse direction), and building direction within the transverse direction-building direction plane are used to study the mechanical properties and associated deformation fields. The resulting heterogeneous microstructure with periodically alternating microstructural features reveals a clear anisotropic material behavior. Furthermore, distinct plastic deformation patterns for different loading directions arise from the spatially varying microstructure. The proposed crystal plasticity model adequately describes the crystallographic texture-induced orientation-dependent yield strength.

Detection of defects in cellular solids using highly nonlinear solitary waves: a numerical study of the proximal femur.

This study investigates the suitability of a relatively new non-destructive evaluation (NDE) technique for the detection of non-visible defects in cellular solids using highly nonlinear solitary waves (HNSWs) in a one-dimensional granular chain. Specifically, the HNSW-based NDE approach is employed to identify the existence of micro-fractures in trabecular bone within the femoral neck (FN) and the intertrochanteric (IT) region of the proximal femur which are fracture-prone sites due to their relatively low bone density, particularly in osteoporosis patients. The availability of a HNSW-based bone quality assessment tool could not only help in early diagnosis of osteoporosis but also affect surgical decisions and improve clinical outcomes in joint replacement surgeries which motivated this study. To obtain a realistic representation of the trabecular microstructure, high-resolution finite-element (FE) models of the FN and the IT region are first constructed using a topology optimization-based bone reconstruction scheme. Then, artificial defects in the form of fractured ligaments are generated in the FN and IT models by selectively disconnecting various struts within the trabecular network. Using the FE models as the inspection medium, hybrid discrete-element/finite-element (DE/FE) simulations are performed to examine the interaction of the HNSWs with the cellular bone samples through two different inspection modes, i.e., inspection via direct contact with the sample and indirect contact through an adequately chosen face sheet inserted between the cellular sample and the granular chain. The delays and amplitudes of the HNSWs are used to estimate the effective elastic moduli of the cellular samples and these estimates were found to be reasonably accurate only in case the face sheet was applied. For the latter case, it was shown that the HNSW-based modulus estimates can be used as indicators for defect detection, allowing to discern between pristine and damaged cellular solids. These results suggest that HNSW-based NDE is a reliable and cost-effective technique for the identification of defects in cellular solids, and is expected to find applications in various fields, such as non-invasive screening of bone diseases and fractures, or damage detection in additively manufactured cellular structures.

Identification of microscale fracture models for mortar with in-situ tests

The microscale fracture modeling of concrete requires explicit descriptions of the microstructure and the fracture properties of its constituents. The identification of fracture properties of mortar was carried out by numerical analyses of two in situ (via X-ray tomography) mesoflexural tests on small-scale specimens. Realistic meshes were created for mechanical simulations to be run. The experimental kinematic fields were measured by processing the reconstructed volumes via Digital Volume Correlation. The measured displacements were used as kinematic boundary conditions to simulate three-point flexural tests. The full dataset consisting of realistic geometry and experimental boundary conditions, allowed for full simulations of the performed tests. A phase-field model for brittle materials was selected to describe damage of the cement paste, and debonding at the matrix-aggregate interfaces was modeled with a cohesive zone model. The fracture paths and ultimate forces simulated with the phase-field model were consistent with the experiment thanks to representative microstructure and boundary conditions. In both tests, the predicted crack path changed because of interface debonding. The introduction of cohesive interface elements did not significantly improve the faithfulness of the path prediction. This study demonstrates the potential of microscale simulations of in situ tests for model identification, validation and development at the microscale of mortar.

Physics-based model of irradiation creep for ferritic materials under fusion energy operation conditions

Irradiation creep is known to be an important process for structural materials in nuclear environments, potentially leading to creep failure at temperatures where thermal creep is generally negligible. While there is a great deal of data for irradiation creep in steels and zirconium alloys in light water reactor conditions, much less is known for first wall materials under fusion energy conditions. Lacking suitable fusion neutron sources for detailed experimentation, modeling, and simulation can help bridge the dose-rate and spectral-effects gap and produce quantifiable expectations for creep deformation of first wall materials under standard fusion conditions. In this paper, we develop a comprehensive model for irradiation creep created from merging a crystal plasticity representation of the dislocation microstructure and a defect evolution simulator that accounts for the entire cluster dimensionality space. Both approaches are linked by way of a climb velocity that captures dislocation-biased defect absorption and a dislocation strengthening term that reflects the accumulation of defect clusters in the system. We carry out our study in Fe under first wall fusion reactor conditions, characterized by a fusion neutron spectrum with average recoil energies of 20 keV and a damage dose rate of ≈3×10−7 dpa/s at temperatures between 300 and 800 K.

On the simulation of nematic liquid crystalline flows in a 4:1 planar contraction using the Leslie–Ericksen and Beris–Edwards models

Extrusion is an essential element in the production of nematic soft solids, ensuring that the final formulation is sufficiently refined and compacted. The interplay between complex microstructure and processing conditions affects the final product performance and may even lead to material degradation. This paper investigates the flow of nematic liquid crystals in a 4:1 planar contraction as a precursor for material extrusion. We compare two different frameworks for modelling the evolution of the nematic microstructure. The Leslie-Ericksen model represents the liquid crystal microstructure through a lower order (vector) field, where nematic elements are treated as headless rods. We find that the polar nature of the description leads to ambiguity in microstructure orientation on the boundary. As a result, boundary conditions with the same notional interpretation are found to produce significantly varying flow and microstructure distributions. The orientability issue is avoided in the higher-order Beris-Edwards model, where the local orientation and degree of molecular alignment are captured in a single tensor. We find that matching flow and microstructure predictions obtained from both approaches is possible, provided that the boundary orientation is correctly chosen in the Leslie-Ericksen theory. In the lower order framework, a poor choice of boundary condition is found to produce poorer pressure loss predictions in flows through the contraction. The effect is particularly pronounced at low Ericksen numbers, where significant over-prediction of pressure losses can arise. • An investigation of the flow of the Leslie-Ericksen and Beris-Edwards model through a 4:1 planar contraction. • The effect of elasticity and microstructure (director) boundary conditions on the flow is investigated. • Vectorial and tensorial frameworks of microstructure representation are compared. • Effects affecting the size of corner vortices are analysed.

Personalization of biomechanical simulations of the left ventricle by in-vivo cardiac DTI data: Impact of fiber interpolation methods.

Simulations of cardiac electrophysiology and mechanics have been reported to be sensitive to the microstructural anisotropy of the myocardium. Consequently, a personalized representation of cardiac microstructure is a crucial component of accurate, personalized cardiac biomechanical models. In-vivo cardiac Diffusion Tensor Imaging (cDTI) is a non-invasive magnetic resonance imaging technique capable of probing the heart's microstructure. Being a rather novel technique, issues such as low resolution, signal-to noise ratio, and spatial coverage are currently limiting factors. We outline four interpolation techniques with varying degrees of data fidelity, different amounts of smoothing strength, and varying representation error to bridge the gap between the sparse in-vivo data and the model, requiring a 3D representation of microstructure across the myocardium. We provide a workflow to incorporate in-vivo myofiber orientation into a left ventricular model and demonstrate that personalized modelling based on fiber orientations from in-vivo cDTI data is feasible. The interpolation error is correlated with a trend in personalized parameters and simulated physiological parameters, strains, and ventricular twist. This trend in simulation results is consistent across material parameter settings and therefore corresponds to a bias introduced by the interpolation method. This study suggests that using a tensor interpolation approach to personalize microstructure with in-vivo cDTI data, reduces the fiber uncertainty and thereby the bias in the simulation results.

Prediction of bulk mechanical properties of PVC foam based on microscopic model:Part I-Microstructure characterization and generation algorithm

The mechanical properties of porous materials greatly depend on their microstructures, so the generation of realistic microstructures is the premise for a sound numerical simulation and analytical prediction. This paper presented the microstructure generation of three kinds of transversely isotropic closed-cell polyvinyl chloride (PVC) foams with different densities based on the geometric characteristics obtained from the X-ray computed tomography (CT) technology. Advancing front method was used to densely pack a set of spheres whose volume distribution is proportion to the cell volume distribution and the microstructure was formed by Laguerre tessellation. Then the geometric characteristics of the numerical models were compared with those from the CT. The good consistence proves that the proposed method is accurate and efficient to produce the representative microstructure of the PVC foams. Finally, a method was proposed to optimize the microstructure of small size numerical specimens, by which the small specimens could give similar geometric characteristics to those of the large specimens.

Machine learning and materials informatics approaches for predicting transverse mechanical properties of unidirectional CFRP composites with microvoids

The mechanical properties of composites are traditionally measured using numerical and experimental approaches, which impede the innovation of materials due to the cost, time, or effort involved. This study for the first time develops a machine learning-assisted model, which aims at predicting the transverse mechanical properties of unidirectional (UD) carbon fiber reinforced polymer (CFRP) composites with microvoids. To this end, the stochastic microstructures are generated by random sequential expansion (RSE) and hard-core model. And the transverse elastic modulus, transverse tensile strength and transverse compressive strength are computed by micromechanics-based finite element (FE) method. Then, the reduced order representation of microstructures is determined using 2-point spatial correlations and principal component analysis (PCA). Finally, a genetic algorithm (GA) optimized back propagation (BP) neural network is implemented to capture the potential nonlinear relationship between microstructure and transverse mechanical properties. The presented data-driven techniques can reproduce the FE simulation results with an R-value of 0.89 or greater. The excellent agreement between the predicted results and test datasets verifies the successful application of data science methodologies in elucidating the microstructure-property linkages of UD-CFRP composites with microvoids, and thus provides a promising tool for accelerating the smart design and optimization of composites.

Carbon-Binder Optimization for Lithium-Ion Battery Extreme Fast Charge

Battery performance is strongly correlated with electrode microstructure and weight loading of the electrode components. Among them are the carbon-black and binder additives that enhance effective conductivity and provide mechanical integrity. However, these both reduce effective ionic transport in the electrolyte phase and reduce energy density. Therefore, an optimal additive loading is required to maximize performance, especially for fast charging where ionic transport is essential. Such optimization analysis is however challenging due to the nanoscale imaging limitations that prevent characterizing this additive phase and thus quantifying its impact on performance.In this work [1], an additive-phase generation algorithm [2] has been developed to remedy this limitation and identify percolation threshold used to define a minimal additive loading (cf. figure, left). Impact of carbon-black binder (CBD) loading on electrochemical performance has been investigated through a wide array of numerical methods and experiments on NMC/graphite cells. These range from the representation, characterization, and homogenization of electrode microstructures, battery macroscale modeling, and impedance and rate capabilities measurements on various cell formats. The combined work provides information on connectivity/percolation of the solid network, effective solid conductivity and ionic diffusivity, interfacial area, cell capacity, lithium plating, impedance, and transport polarization at the beginning of life (BOL) [1]. Rate capability test demonstrates capacity improvement at fast charge at BOL (cf. figure, right), from 37% to 55%, respectively for high (10%wt) and low (4%wt) additive loading during 6C CC charging (and from 80% to 86% at the end of 10 min 6 CC-CV), in agreement with macroscale model. Improvements are attributed to a combination of lower cathode impedance, reduced electrode tortuosity and cathode thickness [1].[1] F. Usseglio-Viretta et. al., Carbon-binder Weight Loading Optimization for Improved Lithium-ion Battery Rate Capability, submitted to Journal of the Electrochemical Society [2] F. Usseglio-Viretta et. al., SoftwareX, 17, 100915 (2022), https://doi.org/10.1016/j.softx.2021.100915 Figure 1

Modeling the two-way coupling of stress, diffusion, and oxidation in heterogeneous CMC microstructures

A multiphysics methodology and a corresponding numerical scheme are proposed for modeling the complex interactions between oxygen diffusion, matrix cracking, oxidation, and the state of stress in carbon silicon carbide (C/SiC) ceramic matrix composites (CMCs) at the microstructural scale. The model is derived from the governing equations for force equilibrium and conservation of mass for oxygen and carbon, which are coupled through reaction terms, oxygen diffusivities and solubilities, and damage in the matrix. The Galerkin method of weighted residuals is used to derive the finite element method (FEM) equations, and the model is demonstrated on a representative stochastic heterogeneous microstructure to investigate the creep-like strain acceleration of stressed oxidation experiments and analyze the fundamental differences between the reaction-limited and diffusion-limited temperature regimes.

Multiscale Recursive Micromechanics of Three-Dimensional Woven Composite Thermal Protection Materials Thermal Conductivities

Multiscale micromechanics predictions have been made for effective thermal conductivities and local thermal fields for a novel three-dimensional woven composite thermal protection system material. The multiscale recursive micromechanics approach, which enables micromechanics models to call other micromechanics models (or themselves recursively) to consider finer and finer length scales, has been employed. The multiscale model uses a recently developed version of the high-fidelity generalized method of cells micromechanics theory at each of three length scales. The results focus on the impact of the microstructural geometry representation at each length scale on the material’s effective thermal conductivity, along with the local thermal flux and temperature fields induced in the microstructures.

Numerical modeling of nanofibrous filter media and performance characteristics

Numerical methodology utilizing a two-dimensional random approach to model fluid flow and particle filtration through a nanofibrous filter media is presented. An algorithm for creating the representative micro-structure of filter media is developed where the statistical parameters of the filter media are given as input. The nanolayer properties are identified based on the proposition that pressure drop across nanofibrous filter media can be the sum of pressure drop across the nanolayer, and base media, which helps simplify the modeling. The simulations were performed in two approaches. In the first one, nanolayer and base media simulations are carried out separately, and in the second approach, simulation was conducted as a single composite. Simulation results infer that it is possible to predict the performance of nanofibrous filter media based on the pressure drop and collection efficiency results of the two layers, which agrees well with the experimental data.

Skeletal-based microstructure representation and featurization through descriptors

Modeling process–structure–property relationships using machine learning methods has become a valuable enabler for materials design and discovery. However, the machine learning models rely heavily on the featurization of the materials’ structure. This paper introduces a microstructure featurization framework to compute generic topological and morphological descriptors. The framework relies on our skeletal microstructure representation. The representation allows for the seamless calculation of topological descriptors, which is the main focus of this paper.To demonstrate the efficacy of our featurization framework, we couple it with a feature selection method to establish the structure–property model for organic photovoltaics (OPV). For this goal, we identify a salient set of descriptors and construct the structure–property map with high accuracy.

A multi-scale framework to predict damage initiation at martensite/ferrite interface

Martensite/ferrite (M/F) interface damage largely controls failure of dual-phase (DP) steels. In order to predict the failure and assess the ductility of DP steels, accurate models for the M/F interfacial zones are needed. Several M/F interface models have been proposed in the literature, which however do not incorporate the underlying microphysics. It has been recently suggested that (lath) martensite substructure boundary sliding dominates the M/F interface damage initiation and therefore should be taken into account. Considering the computationally infeasibility of direct numerical simulations of statistically representative DP steel microstructures, while explicitly resolving the interface microstructures and the sliding activity, a novel multi-scale approach is developed in this work. Two scales are considered: the DP steel mesostructure consisting of multiple lath martensite islands embedded in a ferrite matrix, and the microscopic M/F interfacial zone unit cell resolving the martensite substructure. Based on the emerging microscopic damage initiation pattern, an effective indicator for the M/F interface damage initiation is determined from the interface microstructural unit cell response, along with the effective sliding in this unit cell. Relating these two effective quantities for different interface microstructural configurations leads to an effective mesoscale model relating the interface damage indicator to the sliding activity of the martensite island in terms of the mesoscopic kinematics. This microphysics-based M/F interface damage indicator model, which could not be envisioned a-priori, is fully identified from a set of interfacial unit cell simulations, thus enabling the efficient prediction of interface damage initiation at the mesoscale. The capability of the developed effective model to predict the mesoscopic M/F interface damage initiation is demonstrated on an example of a realistic DP steel mesostructure.