Predictive Multiscale Modeling Research Articles

Multi-scale modelling and statistical analysis of heterogeneous characteristics effect on chloride transport properties in concrete

The transport properties of concrete are closely related to its heterogeneous features. In this study, by considering the multi-scale microstructural characteristics, the chloride transport properties of concrete are comprehensively analysed. Based on the selected representative elementary volumes from bulk cement paste to concrete, a multi-scale predictive model was first proposed for diffusivity prediction of bulk cement paste, mortar and concrete. By taking the hydration process, the presence of sand/interfacial transition zones (ITZs) and the aggregate shape into account, the predicted diffusivities of each scale were validated against the experimental results. To further analyse the weights of different heterogeneous characteristics on chloride transport properties of cementitious materials, a statistical analysis method (principal component analysis) was then adopted based on the modelled results. The analysed results indicated that the capillary pore and C-S-H parts dominate the diffusivity estimation of bulk cement paste, and the consideration of ITZ and multi-species ions interaction can enhance the prediction accuracy for diffusivities of mortars and chloride penetration depths in concrete, respectively. The proposed multi-scale framework provides a novel perspective to study the chloride penetration in concrete by considering the effects of both the multi-scale microstructural characteristics and multi-species interactions, which can also enhance the understanding of deterioration mechanisms for reinforced concrete structures serving in a complex or marine environment.

Atomistic mechanisms underlying plastic flow at ultralow yield stress in ductile carbon aerogels.

We investigated carbon aerogel samples with super low densities of 0.013 g cm-3 (graphite is 2.5) and conducted compression experiments showing a very low yield stress of 5-8 kPa. To understand the atomistic mechanisms operating in these super low density aerogels, we present a computational study of the mechanical response of very low-density amorphous carbonaceous materials. We start from our previously derived atomistic models (based on the DynReaxMas method) with a density of 0.16 g cm-3 representing the core regions of carbon aerogels. We considered three different phases exhibiting either a fiber-like clump morphology interconnected with string-like units or a more reticulated framework. We subjected these phases to compression and shear deformations and analyzed the resulting plastic response via an inherent-structure protocol. Strikingly, we find that these materials possess shear plastic relaxation modes with extremely low values of yield stress, negligible with respect to the finite values predicted outside this "zero-stress" region. This is followed by a succession of two additional regimes with increasing yield stress values. Our analysis of the atomistic relaxation mechanisms finds that these modes have a collective and cooperative character, taking the form of nanoscopic shear bands within the clumps. These findings rationalize our experimental observations of very low-stress plastic deformation modes in carbon aerogels, providing the first steps for developing a predictive multi-scale modeling of the mechanical properties of aerogel materials.

Understanding the thermomechanical behavior of graphene-reinforced conjugated polymer nanocomposites via coarse-grained modeling.

Graphene-reinforced conjugated polymer (CP) nanocomposites are attractive for flexible and electronic devices, but their mechanical properties have been less explored at a fundamental level. Here, we present a predictive multiscale modeling framework for graphene-reinforced poly(3-alkylthiophene) (P3AT) nanocomposites via atomistically informed coarse-grained molecular dynamics simulations to investigate temperature-dependent thermomechanical properties at a molecular level. Our results reveal reduced graphene dispersion with increasing graphene loading. Nanocomposites with shorter P3AT side chains, lower temperatures, and higher graphene content exhibit stronger mechanical responses, which correlates with polymer dynamics. The elastic modulus increases linearly with the graphene content, which slightly deviates from the "Halpin-Tsai" micromechanical model prediction. Local stiffness analysis shows that graphene possesses the highest stiffness, followed by the P3AT backbone and side chains. Deformation-induced stronger chain alignment of the P3AT backbone compared to graphene may further promote conductive behavior. Our findings provide insights into the dynamical heterogeneity of nanocomposites, paving the way for understanding and predicting their thermomechanical properties.

Progressive failure prediction of three-dimensional woven composites using a generic multi-scale analytical model

Three-dimensional (3D) woven composites with higher interlaminar fracture toughness, better damage tolerance and improved mechanical stability are being increasingly used in aviation and automotive industries. To expedite their analysis and design efficiency, developing simple and accurate modeling tools capable of quickly evaluating their mechanical properties and progressive failure is necessary. In this paper, a generic multi-scale analytical model for mechanical response prediction and failure behavior characterization of 3D woven composites is presented and validated by using the existing experimental and numerical results. Subsequently, the analytical model is adopted to evaluate the predictive capability of different unit cell model schemes and to investigate the effects of geometric parameters on stiffness and strength properties of composites. The results indicate that the interior-surface integrated cell model scheme has the overall best performance in mechanical response prediction of composites. In addition, an exponential model and a linear model are proposed to quantify the relationships between the number of weft layer and yarn density and mechanical properties, respectively. The present results demonstrate the analytical model can be a useful tool to provide an insight for the analysis and design of other 3D woven composites.

Linear viscoelastic shear and bulk relaxation moduli in poly(tetramethylene oxide) (PTMO) using united-atom molecular dynamics

A quantitative understanding and prediction of the viscoelastic relaxation behavior of polymer melts is required to build up predictive multiscale models, which can be utilized for the practical design of viscoelastic materials. We report the application of the Green–Kubo method to compute the common shear relaxation modulus and the rarely reported bulk relaxation modulus for a realistic united atom model of poly(tetramethylene oxide) (PTMO), which is one key components of the segmented polyurethane (PU) elastomers. The temperature and molecular weight dependent viscoelastic behavior is investigated in detail by computing the bulk relaxation modulus K(t) along with shear relaxation modulus G(t). Our results provide new rheological data for PTMO melt based on the united atom model, which incorporates the chemical details of the polymer backbone. The predicted stress relaxation are mapped onto master curves using the time–temperature superposition principle with horizontal and vertical shift factors. The computed shift factors agree with the Williams–Landel–Ferry equation and appear to be similar for the shear and bulk relaxation modulus. The underlying dynamics for the shear and bulk relaxation modulus appear to stem from the same mechanistic origin as both processes seem to sample similar kinetic signature of the reference time scale of the underlying thermally activated processes in the liquid. Our results demonstrate furthermore, that MD simulations with the chemical details beyond coarse-grained models can sample the Rouse dynamics and the transitions towards entanglement, which is evident by the emergence of a plateau in shear relaxation modulus and scaling relations in the diffusion behavior of the monomer at larger molecular weights.

Multi-scale modeling of shock initiation of a pressed energetic material. II. Effect of void–void interactions on energy localization

Heterogeneous energetic materials (EMs) contain microstructural defects such as voids, cracks, interfaces, and delaminated zones. Under shock loading, these defects offer potential sites for energy localization, i.e., hotspot formation. In a porous EM, the collapse of one void can generate propagating blast waves and hotspots that can influence the hotspot phenomena at neighboring voids. Such void–void interactions must be accounted for in predictive multi-scale models for the reactive response of a porous EM. To infuse such meso-scale phenomena into a multi-scale framework, a meso-informed ignition and growth model (MES-IG) has been developed, where the influence of void–void interactions is incorporated into the overall reaction rate through a function, fv−v. Previously, MES-IG was applied to predict the sensitivity and reactive response of EM, where fv−v was assumed to be a function of the overall sample porosity alone. This paper performs a deeper analysis to model the strong dependency of fv−v on other factors, such as void size and shock strength. The improved model for void–void interactions produces good agreement with direct numerical simulations of the HE microstructures and, thus, advances the predictive capability of multi-scale models of the shock response and sensitivity of EM.

Modeling Polyzwitterion-Based Drug Delivery Platforms: A Perspective of the Current State-of-the-Art and Beyond.

Drug delivery platforms are anticipated to have biocompatible and bioinert surfaces. PEGylation of drug carriers is the most approved method since it improves water solubility and colloid stability and decreases the drug vehicles’ interactions with blood components. Although this approach extends their biocompatibility, biorecognition mechanisms prevent them from biodistribution and thus efficient drug transfer. Recent studies have shown (poly)zwitterions to be alternatives for PEG with superior biocompatibility. (Poly)zwitterions are super hydrophilic, mainly stimuli-responsive, easy to functionalize and they display an extremely low protein adsorption and long biodistribution time. These unique characteristics make them already promising candidates as drug delivery carriers. Furthermore, since they have highly dense charged groups with opposite signs, (poly)zwitterions are intensely hydrated under physiological conditions. This exceptional hydration potential makes them ideal for the design of therapeutic vehicles with antifouling capability, i.e., preventing undesired sorption of biologics from the human body in the drug delivery vehicle. Therefore, (poly)zwitterionic materials have been broadly applied in stimuli-responsive “intelligent” drug delivery systems as well as tumor-targeting carriers because of their excellent biocompatibility, low cytotoxicity, insignificant immunogenicity, high stability, and long circulation time. To tailor (poly)zwitterionic drug vehicles, an interpretation of the structural and stimuli-responsive behavior of this type of polymer is essential. To this end, a direct study of molecular-level interactions, orientations, configurations, and physicochemical properties of (poly)zwitterions is required, which can be achieved via molecular modeling, which has become an influential tool for discovering new materials and understanding diverse material phenomena. As the essential bridge between science and engineering, molecular simulations enable the fundamental understanding of the encapsulation and release behavior of intelligent drug-loaded (poly)zwitterion nanoparticles and can help us to systematically design their next generations. When combined with experiments, modeling can make quantitative predictions. This perspective article aims to illustrate key recent developments in (poly)zwitterion-based drug delivery systems. We summarize how to use predictive multiscale molecular modeling techniques to successfully boost the development of intelligent multifunctional (poly)zwitterions-based systems.

A review of microwave-assisted process intensified multiphase reactors

Microwaves provide alternative heating and allow process intensification due to their rapid, volumetric, and selective nature. Recognizing the central role of multiphase reactors in chemical industry, a recent surge in employing microwaves is observed. Here, we review the recent experimental and modeling investigations of microwave heating of multiphase reactors with emphasis on chemical engineering applications. We demonstrate that there is accumulated evidence for improved performance via microwave heating and a clear opportunity for further process intensification. In most of the cases, this improved performance stems from a temperature gradient between two phases. We discuss the ongoing debate on the mechanism by which microwaves affect chemical processes exacerbated by the inability of measuring the temperature distribution in a microwave cavity. We outline recent progress in this direction in monolith reactors and needs for future work. We underscore the lack of detailed modeling and simulation tools even for single-phase systems and emphasize the imperative for multiscale predictive modeling to bridge the experimental-modeling gap. Promising results are shown by a few recently published modeling studies that can predict the experimental measurements in complex multiphase reactors. A combination of experimental and modeling tools can provide a comprehensive picture of the microwave multiphase reactors as well as a means toward scale-up and optimization.

Effect of Microstructural Damage on the Thermomechanical Properties of Electrodes in Proton Exchange Membrane Fuel Cells.

Advanced functional materials composed of multiple nanoscale phases, including pores and interfaces, have been extensively applied in the fields of new energy, architecture, and aerospace. However, insufficient knowledge of the thermomechanical properties resulting from material failures, such as interfacial delamination and porosity deformations, which limit the durability and lifetime of these materials, has hindered their further application, demanding a deeper understanding of microstructural changes. Based on the fuel cell electrode, we explore a multiscale prediction model that correlates the atomic interactions between interfaces with a microscopic thermomechanical model to illuminate the effects of interface binding characteristics on the materials' mechanical response and heat conduction mechanisms. Compared with experimental measurements and theoretical calculations at the macroscopic scale, our model excels in predicting the initiation and propagation of interfacial debonding and the thermal conductivity of the electrode, with the resistance factors for the interface, pores, and cracks taken into consideration. This work provides guidance for designing robust electrodes resistant to thermomechanical failure and serves as a reference method for predicting damage in heterogeneous porous materials.

A Practical Guide to Integrating Multimodal Machine Learning and Metabolic Modeling.

Complex, distributed, and dynamic sets of clinical biomedical data are collectively referred to as multimodal clinical data. In order to accommodate the volume and heterogeneity of such diverse data types and aid in their interpretation when they are combined with a multi-scale predictive model, machine learning is a useful tool that can be wielded to deconstruct biological complexity and extract relevant outputs. Additionally, genome-scale metabolic models (GSMMs) are one of the main frameworks striving to bridge the gap between genotype and phenotype by incorporating prior biological knowledge into mechanistic models. Consequently, the utilization of GSMMs as a foundation for the integration of multi-omic data originating from different domains is a valuable pursuit towards refining predictions. In this chapter, we show how cancer multi-omic data can be analyzed via multimodal machine learning and metabolic modeling. Firstly, we focus on the merits of adopting an integrative systems biology led approach to biomedical data mining. Following this, we propose how constraint-based metabolic models can provide a stable yet adaptable foundation for the integration of multimodal data with machine learning. Finally, we provide a step-by-step tutorial for the combination of machine learning and GSMMs, which includes: (i) tissue-specific constraint-based modeling; (ii) survival analysis using time-to-event prediction for cancer; and (iii) classification and regression approaches for multimodal machine learning. The code associated with the tutorial can be found at https://github.com/Angione-Lab/Tutorials_Combining_ML_and_GSMM .

Segmentation-based multi-scale attention model for KRAS mutation prediction in rectal cancer

Segmentation-based multi-scale attention model for KRAS mutation prediction in rectal cancer

(Invited) Hydrogen Materials Advanced Research Consortium R&D Activities

The Hydrogen Materials—Advanced Research Consortium (HyMARC) is a National Lab-led consortium working to advance and develop hydrogen storage materials that meet DOE targets for energy density, kinetics and cost for a wide range of applications. The goal of HyMARC is to enable a doubling of the hydrogen storage density compared with state-of-the-art compressed storage systems. DOE’s Hydrogen and Fuel Cell Technologies Office launched HyMARC in 2015, bringing together world-class National Lab capabilities in materials modelling, characterization and synthesis; and fostering collaboration with industry and academia to overcome R&D challenges in materials-based hydrogen storage. The consortium is led by teams at Sandia National Laboratories and the National Renewable Energy Laboratory and includes groups at Lawrence Livermore, Lawrence Berkeley, and Pacific Northwest National Laboratories. HyMARC is working to develop foundational understanding of phenomena governing thermodynamics and kinetics of hydrogen release and uptake in all classes of hydrogen storage materials, including metal hydrides, adsorbents, and hydrogen carriers. It seeks to accelerate materials development by addressing critical R&D gaps in hydrogen storage—leveraging recent advances in predictive multiscale modeling, machine learning, high-resolution in situ characterization, and novel materials synthesis techniques to realize their goal to enable twice the energy density for hydrogen storage in the coming years. This presentation will describe HyMARC’s ongoing R&D activities and past accomplishments which have pushed the bar forward for materials-based hydrogen storage.

(Invited) Multiscale Modeling and Theory Experiment Integration for Understanding Complex Interfaces in Materials for Hydrogen Production and Storage

Chemical processes occurring at solid-gas, solid-liquid, and solid-solid interfaces critically determine the performance and durability of hydrogen production and storage technologies. While probing behavior of these interfaces under operating conditions remains challenging, recent progress in predictive multiscale modeling methods together with high-fidelity experimental characterization has led to significant advancements in understanding how microscopic interfacial processes translate to macroscopic device behavior. In this talk, we will provide an overview of our materials modeling efforts within the DOE HydroGEN, HyMARC, and H2NEW hydrogen production and storage consortia, which span ab-initio atomistic modeling to macroscopic device modeling while leveraging state-of-art experimental characterization to ensure fidelity of interpretation. Concrete examples will be provided from our recent activities focused on identifying factors that dictate performance and durability of materials under operating conditions. In particular, we will discuss how computational models have helped to elucidate mechanisms of interface chemistry, formation of new phases, and the impact of interfaces on key reaction pathways that are intimately related to the device performance and durability. We will also illustrate how experimental characterization data has been incorporated to constrain the models and obtain new understanding of interfacial phenomena. Finally, we will discuss how the obtained knowledge may be used to develop strategies for improving materials used in hydrogen production and storage technologies. This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.

A multi-scale prediction model of elastic modulus for ceramic matrix composites considering oxidation damage

The oxidation damage evolution law of ceramic matrix composites prepared by chemical vapor infiltration (CVI) process was analyzed. Considering the effects of temperature and oxidation time, prediction models of elastic modulus about fiber and cell were established based on the distribution of microcracks in matrix and the oxidation process of interface, fiber and matrix. The prediction results show that the tensile elastic moduli of carbon fiber (Cf)/SiC and SiC fiber (SiCf)/SiC composites decrease more obviously with the increase of temperature and oxidation time. The elastic prediction model was verified by the tensile tests of the composites after high temperature oxidation. The error between the prediction results and the test results of SiCf/SiC composites with BN interphase after oxidation at 1000 ℃ for different time is no more than 2%, and the error between the prediction results and the test results of Cf/SiC composites with PyC interphase after oxidation at 700℃ for different time is no more than 7%.

Heat Transfer Efficiency Prediction of Coal-Fired Power Plant Boiler Based on CEEMDAN-NAR Considering Ash Fouling

Ash fouling has been an important factor in reducing the heat transfer efficiency and safety of the coal-fired power plant boilers. Scientific and accurate prediction of ash fouling of heat transfer surfaces is the basis of formulating a reasonable soot blowing strategy to improve energy efficiency. This study presented a comprehensive approach of dynamic prediction of the ash fouling of heat transfer surfaces in coal-fired power plant boilers. At first, the cleanliness factor is used to reflect the fouling level of the heat transfer surfaces. Then, a dynamic model is proposed to predict ash deposits in the coal-fired boilers by combining complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and nonlinear autoregressive neural networks (NARNN). To construct a reasonable network model, the minimum information criterion and trial-and-error method are used to determine the delay orders and hidden layers. Finally, the experimental object is established on the 300 MV economizer clearness factor dataset of the power station, and the root mean square error and mean absolute percentage error of the proposed method are the smallest. In addition, the experimental results show that this multiscale prediction model is more competitive than the Elman model.

Calibration and Validation of Multiscale Model for Ultimate Strength Prediction of Composite Laminates Under Uncertainty

Abstract A methodology to account for the effect of epistemic uncertainty (regarding model parameters) on the strength prediction of carbon fiber reinforced polymer (CFRP) composite laminates is presented. A three-dimensional concurrent multiscale physics modeling framework is considered. A continuum damage mechanics-based constitutive relation is used for multiscale analysis. The parameters for the constitutive model are unknown and need to be calibrated. A least squares-based approach is employed for the calibration of model parameters and a model discrepancy term. The calibrated constitutive model is validated quantitatively using experimental data for both unnotched and open-hole specimens with different composite layups. The quantitative validation results are used to indicate further steps for model improvement.

Experimental study and production model of multiscale stress sensitivity in tight oil reservoirs

ABSTRACT Stress sensitivity is an inevitable issue in the depletion exploitation of tight oil reservoirs, especially in the reformed multiscale region. Through a series of improved laboratory experiments, the stress sensitivity characteristics of multiscale media were thoroughly evaluated based on the primitive effective stress. Furthermore, based on the nonlinear seepage theory and equivalent seepage resistance method, a corresponding multiscale production prediction model of vertical wells in tight oil reservoirs with stress sensitivity established. The results showed that the decline in permeability of main fracture rock (12.86%) was more than twice that of the microfracture rock (5.25%), which was approximately four times that of the matrix rock (3.32%) in the selected pressure range. In addition, there was a fine power function relationship between permeability variation and effective stress in multiscale representative media. Compared with the actual daily production and cumulative production of the coring well, the production curve fitting of the proposed multiscale model was 91.25%, which was reliable to guide the development scheme design in the target block of tight oil reservoirs.

FraGAT: a fragment-oriented multi-scale graph attention model for molecular property prediction.

MotivationMolecular property prediction is a hot topic in recent years. Existing graph-based models ignore the hierarchical structures of molecules. According to the knowledge of chemistry and pharmacy, the functional groups of molecules are closely related to its physio-chemical properties and binding affinities. So, it should be helpful to represent molecular graphs by fragments that contain functional groups for molecular property prediction.ResultsIn this article, to boost the performance of molecule property prediction, we first propose a definition of molecule graph fragments that may be or contain functional groups, which are relevant to molecular properties, then develop a fragment-oriented multi-scale graph attention network for molecular property prediction, which is called FraGAT. Experiments on several widely used benchmarks are conducted to evaluate FraGAT. Experimental results show that FraGAT achieves state-of-the-art predictive performance in most cases. Furthermore, our case studies show that when the fragments used to represent the molecule graphs contain functional groups, the model can make better predictions. This conforms to our expectation and demonstrates the interpretability of the proposed model.Availability and implementationThe code and data underlying this work are available in GitHub, at https://github.com/ZiqiaoZhang/FraGAT.Supplementary informationSupplementary data are available at Bioinformatics online.

Molecular Modelling Guided Modulation of Molecular Shape and Charge for Design of Smart Self-Assembled Polymeric Drug Transporters.

Nanomedicine employs molecular materials for prevention and treatment of disease. Recently, smart nanoparticle (NP)-based drug delivery systems were developed for the advanced transport of drug molecules. Rationally engineered organic and inorganic NP platforms hold the promise of improving drug targeting, solubility, prolonged circulation, and tissue penetration. However, despite great progress in the synthesis of NP building blocks, more interdisciplinary research is needed to understand their self-assembly and optimize their performance as smart nanocarriers. Multi-scale modeling and simulations provide a valuable ally to experiment by mapping the potential energy landscape of self-assembly, translocation, and delivery of smart drug-loaded NPs. Here, we highlight key recent advances to illustrate the concepts, methods, and applications of smart polymer-based NP drug delivery. We summarize the key design principles emerging for advanced multifunctional polymer topologies, illustrating how the unusual architecture and chemistry of dendritic polymers, self-assembling polyelectrolytes and cyclic polymers can provide exceptional drug delivery platforms. We provide a roadmap outlining the opportunities and challenges for the effective use of predictive multiscale molecular modeling techniques to accelerate the development of smart polymer-based drug delivery systems.

Perspective for Future Medicine: Multidisciplinary Computational Anatomy-Based Medicine with Artificial Intelligence.

Multidisciplinary computational anatomy (MCA) is a new frontier of science that provides a mathematical analysis basis for the comprehensive and useful understanding of “dynamic living human anatomy.” It defines a new mathematical modeling method for the early detection and highly intelligent diagnosis and treatment of incurable or intractable diseases. The MCA is a method of scientific research on innovative areas based on the medical images that are integrated with the information related to: (1) the spatial axis, extending from a cell size to an organ size; (2) the time series axis, extending from an embryo to post mortem body; (3) the functional axis on physiology or metabolism which is reflected in a variety of medical image modalities; and (4) the pathological axis, extending from a healthy physical condition to a diseased condition. It aims to integrate multiple prediction models such as multiscale prediction model, temporal prediction model, anatomy function prediction model, and anatomy-pathology prediction model. Artificial intelligence has been introduced to accelerate the calculation of statistic mathematical analysis. The future perspective is expected to promote the development of human resources as well as a new MCA-based scientific interdisciplinary field composed of mathematical statistics, information sciences, computing data science, robotics, and biomedical engineering and clinical applications. The MCA-based medicine might be one of the solutions to overcome the difficulties in the current medicine.