Correlation between microstructures and macroscopic properties of nickel/yttria-stabilized zirconia (Ni-YSZ) anodes: Meso-scale modeling and deep learning with convolutional neural networks

Liu, Xuhao; Zhou, Shihao; Yan, Zhimiao; Zhong, Zheng; Shikazono, Naoki; Hara, Shotaro

doi:10.1016/j.egyai.2021.100122

Cited by 32 publications

(9 citation statements)

References 73 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As shown in Figure 5, when mean absolute percentage error (MAPE) is used as the evaluation index, there are some abnormal values with great influence, which may result in instability of the CNN training. Therefore, (Liu et al, 2022) in contrast to the MAPE metric which was applied in our previous research, in this study weighted mean absolute percentage error (WMAPE) was adopted as a new metric to evaluate the quality of results because of its better robustness. Note that for a single sample, these two metrics are equivalent.…”

Section: Resultsmentioning

confidence: 99%

“…In this study, the optimized CNN structures (Liu et al ., 2022) were adopted to link the 3D input images of 64 × 64 × 64 voxels to five macroscopic material properties. The specifics of the structures are listed in Table 3 and the structure of the CNN is visually illustrated in Figure 4.…”

Section: Methodsmentioning

confidence: 99%

“…Cecen et al (2018) successfully extracted the effective elastic properties of high contrast composites by using CNN. Liu et al (2022) proposed CNN based deep learning model to correlate the Ni-YSZ anodes in solid oxide fuel cells (SOFCs) with the effective elastic modulus, Poisson's ratio, shear modulus, coefficient of thermal expansion, and triple-phase boundary length density. EC 40,7/8 However, it should be noted that in image-based deep learning, the size of the dataset is still a concern.…”

Section: Introductionmentioning

confidence: 99%

“…In this paper, based on the dataset in our previous work (Liu et al ., 2022), the influence of data augmentation on the prediction performance of the CNN models was investigated. The dataset was augmented through image flipping and rotation.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Data augmentation and data mining towards microstructure and property relationship for composites

Guo

Liu

Pan

et al. 2023

View full text Add to dashboard Cite

PurposeIn recent years, the convolutional neural network (CNN) based deep learning approach has succeeded in data-mining the relationship between microstructures and macroscopic properties of materials. However, such CNN models usually rely heavily on a large set of labeled images to ensure the accuracy and generalization ability of the predictive models. Unfortunately, in many fields, acquiring image data is expensive and inconvenient. This study aims to propose a data augmentation technique to enhance the performance of the CNN models for linking microstructural images to the macroscopic properties of composites.Design/methodology/approachMicrostructures of composites are synthesized using discrete element simulations and Potts kinetic Monte Carlo simulations. Macroscopic properties such as the elastic modulus, Poisson's ratio, shear modulus, coefficient of thermal expansion, and triple-phase boundary length density are extracted on representative volume elements. The CNN model is trained using the 3D microstructural images as inputs and corresponding macroscopic properties as the labels. The comparison of the predictive performance of the CNN models with and without data augmentation treatment are compared.FindingsThe comparison between the prediction performance of CNN models with and without data augmentation showed that the former reduced the weighted mean absolute percentage error (WMAPE) for the prediction from 5.1627% to 1.7014%. This significant reduction signifies that the proposed data augmentation method can effectively enhance the generalization ability and robustness of CNN models.Originality/valueThis study demonstrates that data augmentation is beneficial for solving the problems of model overfitting, data scarcity, and sample imbalance for CNN-based deep learning tasks at a low cost. By developing more and advanced data augmentation techniques, deep learning accelerated homogenization will boost the multi-scale computational mechanics and materials.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Data augmentation and data mining towards microstructure and property relationship for composites

Guo

Liu

Pan

et al. 2023

View full text Add to dashboard Cite

show abstract

“…Moreover, data-driven surrogate models for property evaluation, for example, convolutional neural networks (CNNs), often directly take input electrode microstructures without any physical feature engineering and can only predict low-dimensional structural properties such as elastic properties and effective diffusivities. [35][36][37][38] Generally, the physical nature of electrochemical processes involves multiple underlying phenomena that critically determine the electrochemical performance. Therefore, it is vital to feed identified electrochemical knowledge into ML techniques as prior information for electrode candidate generation and performance evaluation to finally achieve rational data-driven electrode microstructure design.…”

Section: Introductionmentioning

confidence: 99%

π Learning: A Performance‐Informed Framework for Microstructural Electrode Design

Niu

Zhao

et al. 2023

Advanced Energy Materials

View full text Add to dashboard Cite

batteries, electrolyzers, and electrochemical CO 2 reduction reactors all depend heavily on the nature of the meso-scale reaction sites and charge conductivity in their electrodes. [2][3][4] The meso-scale electrode structure is often determined by the complex interactions of various phases under different thermal, mechanical, and chemical conditions. Therefore, how to design the next generation of high-performance electrodes remains a critical challenge.Data-driven design based on machine learning (ML) has been increasingly recognized as a revolutionary technology in structure design and has been demonstrated to successfully identify various excellent structures in protein sequences, [5] drug molecules, [6] and nanomaterials. [7] This paradigm opens the door for innovative design of high-performance catalyst materials and electrodes in electrochemical engineering. [8][9][10] A typical example is to accelerate the discovery of optimal atomic structures and compositions driven by density functional theory (DFT) calculations. [11][12][13] For instance, Ma et al. [14] proposed two deep learning algorithms to screen emerging electrode materials for sodium-ion and potassium-ion batteries. This data-driven approach for material screening is promising to accelerate the discovery of high-performance electrode materials. Benayad [15] et al. reviewed recent ML applications in Designing high-performance porous electrodes is the key to next-generation electrochemical energy devices. Current machine-learning-based electrode design strategies are mainly orientated toward physical properties; however, the electrochemical performance is the ultimate design objective. Performance-orientated electrode design is challenging because the current data driven approaches do not accurately extract high-dimensional features in complex multiphase microstructures. Herein, this work reports a novel performance-informed deep learning framework, termed π learning, which enables performance-informed microstructure generation, toward overall performance prediction of candidate electrodes by adding most relevant physical features into the learning process. This is achieved by integrating physicsinformed generative adversarial neural networks (GANs) with convolutional neural networks (CNNs) and with advanced multi-physics, multi-scale modeling of 3D porous electrodes. This work demonstrates the advantages of π learning by employing two popular design philosophies: forward and inverse designs, for the design of solid oxide fuel cells electrodes. π learning thus has the potential to unlock performance-driven learning in the design of next generation porous electrodes for advanced electrochemical energy devices such as fuel cells and batteries.

show abstract

Accessing Elastic Properties of Porous Solid Oxide Fuel Cell Electrodes Using 2D Image-Based Discrete Element Modeling and Deep Learning

Zhou,

Zeng,

Liu

et al. 2024

Acta Mech. Solida Sin.

View full text Add to dashboard Cite

Correlation between microstructures and macroscopic properties of nickel/yttria-stabilized zirconia (Ni-YSZ) anodes: Meso-scale modeling and deep learning with convolutional neural networks

Cited by 32 publications

References 73 publications

Data augmentation and data mining towards microstructure and property relationship for composites

Data augmentation and data mining towards microstructure and property relationship for composites

π Learning: A Performance‐Informed Framework for Microstructural Electrode Design

Accessing Elastic Properties of Porous Solid Oxide Fuel Cell Electrodes Using 2D Image-Based Discrete Element Modeling and Deep Learning

Contact Info

Product

Resources

About