Multiscale Computational Techniques Research Articles

Prediction of Electrochemical Characteristics for Yttrium Doped Barium Zirconate Based Solid Oxide Fuel and Electrolyzer Cells Using a Multiscale Framework

Yttrium doped Barium Zirconate (BZY) is one of the best-known proton-conducting solid oxides and understanding its behavior is the key to designing better proton conducting electrolytes for Solid Oxide Fuel/Electrolyzer Cells (SOFC/SOEC). Functioning of an SOFC/SOEC involves a complex interplay of factors like temperature, chemistry, microstructure, etc., at different length and time scales which necessitates the use of multiscale computational techniques. A multiscale framework involving Density Functional Theory at atomic, Transition State Theory at meso and Electrochemical Theory at continuum scales has, therefore, been developed that could accurately predict Current-Voltage Characteristics for direct current and Impedance Spectra for alternating current in BZY based SOFC/SOECs. The effect of operating conditions on charge carriers and, hence, the performance of the cell has been analyzed. Temperature and the environment (reducing/oxidizing) have a significant effect on proton conductivity of the electrolyte. The activation energy for oxidizing atmosphere is found to be lower facilitating use at lower temperatures in these environments. Using Neural Networks as a surrogate model, a multi-dimensional uncertainty quantification and sensitivity analysis has also been performed to point out the important parameters determining the performance of the SOFC/SOEC. Hydration energy, proton diffusivity, and permittivity are found to influence the efficiency of the cell significantly. These parameters serve to delineate design guidelines for the development of new chemistries and compositions of proton conducting solid oxides through a Random Forest model, correlating material property descriptors for these parameters to the proton conductivity of various proton conducting solid oxides studied in the literature.

Hydrogen storage on graphitic carbon nitride and its palladium nanocomposites: A multiscale computational approach

Hydrogen storage capacity (HSC) of multilayer graphitic carbon nitride, d-g-C3N4 (d is interlayer spacing), and its palladium nanocomposite, d-Pd@g-C3N4, were investigated using multiscale computational techniques including quantum mechanics calculations and grand canonical Monte Carlo (GCMC) simulation. According to the results, the volumetric HSC of 8-g-C3N4 and 8-Pd@g-C3N4 can reach to DOE target of 30 gH2/L at 177 K, 5.7 MPa, and 177 K, 4.0 MPa, respectively. The gravimetric HSC of 10-g-C3N4 and 12-Pd@g-C3N4 meet the DOE target of 4.5 wt% at 150 K, 3.5 MPa, and 125 K, 4.0 MPa, respectively. The incorporation of Pd atoms enhances the delivery volumetric HSC of 6-, 8-, 10-, and 12-g-C3N4 by 49, 55, 129, and 146%, respectively at 177 K and 0.5 MPa. On the other hand, the incorporation of Pd atoms has a negative effect on the delivery gravimetric HSC of 6- and 8-g-C3N4 and positive effect for 10- and 12-g-C3N4. The estimated isostric heat, Qst, of adsorption is 5.5–8.5 kJ/mol. The maximum value of Qst for both nanoadsorbents belong to those with d = 8 Å. The structure of adsorbates and possibility of multilayer adsorption occurrence were also investigated using pair correlation functions and density profiles.

Titanium-decorated boron nitride nanotubes for hydrogen storage: a multiscale theoretical investigation.

A multiscale computational technique from available quantum mechanical and Molecular Dynamics (MD) codes to user-subroutine computational fluid dynamics (CFD) files was applied to the H2 adsorption of Ti-decorated (10,0) single-walled BN nanotubes (BNNTs) with B-N defects. According to density functional theory (DFT), the Ti atom adsorbed in the defect protrudes outward from the surface and does not agglomerate. The Ti/BNNT system possesses excellent affinity towards H2 molecules with thermodynamically favorable binding energies. Up to seven H2 molecules can partially attach near Ti in quasi-molecular form due to the cationic functionalized Ti and heteropolar B-N bonds at the surface. MD simulation further reveals that the seven H2 molecules are indeed physisorbed in two distinct regions near to (at ∼2 Å) and far from (at ∼4 Å) the Ti atom. The thermal properties obtained from MD simulations were utilized in the CFD code to quantify thermal energy transfer. According to CFD, the equilibrium pressure distribution for a type III H2 cylinder is initially low within its locality. The rate of change of pressure increase is quite steep as it rises due to a sudden increase in temperature upon uptake and it eventually slows down when the local temperature reaches its saturation value.

Multiscale Finite Element Analysis of Linear Magnetic Actuators Using Asymptotic Homogenization Method

This work presents the multiscale finite element analysis of linear magnetic actuators. Here, the actuators includes unidirectional fiber reinforced magnetic composites as a back-iron core component. The composite is employed in actuators to enhance the magnetic force. However, the direct computation of actuators including heterogeneous composite structures requires high computation cost. To overcome this problem, multiscale computational technique for the analysis of a magnetic actuator is proposed in this work. First, the effective magnetic permeability of the composite is calculated at various fiber volume fractions and orientation angles. For this, the asymptotic homogenization method is applied to the composite unit cell model in microscopic coordinate system. Next, the obtained homogenized effective permeability is utilized for the macroscopic magnetostatic analysis of actuator model. Subsequently, the magnetic force acting on a actuator plunger is calculated using the Maxwell stress tensor method. To validate the accuracy and computational benefit of the proposed multiscale approach, a actuator numerical example is provided. For the accuracy validation, the magnetic force and magnetic field distribution obtained from the proposed multiscale approach are compared with those from the direct calculation. In addition, the computation time consumed for the mutiscale approach and direct calculation is compared to validate the benefit of the proposed analysis process.

Multiscale simulation of fullerene reinforced composite structures: From molecular dynamics to finite element continuum mechanics

The aim of the present study is to propose a multiscale computational technique for the prediction of the elastic mechanical properties of nanoreinforced composites. The proposed method utilizes a molecular dynamics (MD) based numerical scheme to capture the mechanical behaviour of the nanocomposite at nanoscale and then a classical continuum mechanics (CM) analysis based on the finite element method (FEM) to characterise the microscopic performance of the nanofilled composite material. The material under investigation is polyamide 12 (PA 12) randomly reinforced with fullerenes C60. At the first stage of the analysis, in order to capture the atomistic interfacial effects between C60 and PA 12, a very small cubic unit cell containing a C60 molecule, centrally positioned and surrounded by PA 12 molecular chains, is simulated via MD. Inter- and intra-molecular atomic interactions are described by using the Condensed Phase Optimized Molecular Potential for Atomistic Simulation Studies (COMPASS). According to the elastic properties data arisen by the MD simulations, an equivalent finite element volume with the same size as the unit cell is developed. At the second stage, a CM micromechanical representative volume element (RVE) of the C60 reinforced PA 12 is developed via FEM. The matrix phase of the RVE is discretised by using solid finite elements which represent the PA 12 mechanical behaviour while each C60 location is meshed with equivalent solid finite elements. Several multiscale simulations are performed to study the effect of the nanofiller volume fraction on the mechanical properties of the C60 reinforced PA 12 composite. Comparisons with other corresponding experimental results are attempted, where possible, to test the performance of the proposed method.

Challenges in Multiscale Modeling of Polymer Dynamics

The mechanical and physical properties of polymeric materials originate from the interplay of phenomena at different spatial and temporal scales. As such, it is necessary to adopt multiscale techniques when modeling polymeric materials in order to account for all important mechanisms. Over the past two decades, a number of different multiscale computational techniques have been developed that can be divided into three categories: (i) coarse-graining methods for generic polymers; (ii) systematic coarse-graining methods and (iii) multiple-scale-bridging methods. In this work, we discuss and compare eleven different multiscale computational techniques falling under these categories and assess them critically according to their ability to provide a rigorous link between polymer chemistry and rheological material properties. For each technique, the fundamental ideas and equations are introduced, and the most important results or predictions are shown and discussed. On the one hand, this review provides a comprehensive tutorial on multiscale computational techniques, which will be of interest to readers newly entering this field; on the other, it presents a critical discussion of the future opportunities and key challenges in the multiscale modeling of polymeric materials and how these methods can help us to optimize and design new polymeric materials.

Multiscale Modeling of Liquids with Molecular Specificity

The separation between molecular and mesoscopic length and time scales poses a severe limit to molecular simulations of mesoscale phenomena. We describe a hybrid multiscale computational technique which addresses this problem by keeping the full molecular nature of the system where it is of interest and coarse graining it elsewhere. This is made possible by coupling molecular dynamics with a mesoscopic description of realistic liquids based on Landau's fluctuating hydrodynamics. We show that our scheme correctly couples hydrodynamics and that fluctuations, at both the molecular and continuum levels, are thermodynamically consistent. Hybrid simulations of sound waves in bulk water and reflected by a lipid monolayer are presented as illustrations of the scheme.

Diffusion filtering in image processing based on wavelet transform

The nonlinear diffusion filtering in image processing bases on the heat diffusion equations. Its key is the control of diffusion amount. In the previous models, the diffusivity depends on the gradients of images. So it is easily affected by noises. This paper first gives a new multiscale computational technique for diffusivity. Then we proposed a class of nonlinear wavelet diffusion (NWD) models that are used to restore images. The NWD model has strong ability to resist noise. But it, like the previous models, requires higher computational effort. Thus, by simplifying the NWD, we establish linear wavelet diffusion (LWD) models that consist of advection and diffusion. Since there exists the advection, the LWD filter is anisotropic, and hence can well preserve edges although the diffusion at edges is isotropic. The advantage is that the LWD model is easy to be analyzed and has lesser computational load. Finally, a variety of numerical experiments compared with the previous model are shown.

Space-time Multiscale Laminated Theory

Multiscale computational techniques in space and time are developed to study the impact response of thin, elastic, laminated composites. The displacement field is approximated using asymptotic expansion in space and time. Using the homogenization procedure in space and time, nonlocal membrane and bending equations of motion are derived. The nonlocal equations are stabilized to filter out the higher frequency content. The multiscale model is verified for membrane and bending problems.