Cunliang Pan¹, Shi Feng², Shengyang Tao², Hongwu Zhang¹, Yonggang Zheng^1,3, Hongfei Ye^1,*

The International Conference on Computational & Experimental Engineering and Sciences, Vol.30, No.1, pp. 1-1, 2024, DOI:10.32604/icces.2024.011132

Abstract Capillarity is prevalent in nature, daily life, and industrial processes, governed by the fundamental Young-Laplace equation. Solving this equation not only enhances our understanding of natural phenomena but also provides valuable insights into industrial advancements. To address challenges posed by conventional numerical methods in parameter identification and complex boundary condition handling, the Young-Laplace Physics-informed Neural Network (Y-L PINN) is introduced to solve the Young-Laplace equation within a tubular domain. Through computational analyses focusing on the classical capillary rise case, the proposed method's accuracy is affirmed through comparisons with Jurin's law, experimental data, and numerical results.… More >

Meng Ma^1,2,*, Liu Fu^1,2, Xu Guo³, Zhi Zhai^1,2

CMES-Computer Modeling in Engineering & Sciences, Vol.141, No.1, pp. 385-399, 2024, DOI:10.32604/cmes.2024.052585 - 20 August 2024

Abstract Partial Differential Equation (PDE) is among the most fundamental tools employed to model dynamic systems. Existing PDE modeling methods are typically derived from established knowledge and known phenomena, which are time-consuming and labor-intensive. Recently, discovering governing PDEs from collected actual data via Physics Informed Neural Networks (PINNs) provides a more efficient way to analyze fresh dynamic systems and establish PED models. This study proposes Sequentially Threshold Least Squares-Lasso (STLasso), a module constructed by incorporating Lasso regression into the Sequentially Threshold Least Squares (STLS) algorithm, which can complete sparse regression of PDE coefficients with the constraints More >

Haojie Lian¹, Jiaqi Wang¹, Leilei Chen^2,*, Shengze Li³, Ruochen Cao⁴, Qingyuan Hu⁵, Peiyun Zhao¹

CMES-Computer Modeling in Engineering & Sciences, Vol.140, No.1, pp. 1143-1163, 2024, DOI:10.32604/cmes.2024.048549 - 16 April 2024

Abstract This paper presents a novel framework aimed at quantifying uncertainties associated with the 3D reconstruction of smoke from 2D images. This approach reconstructs color and density fields from 2D images using Neural Radiance Field (NeRF) and improves image quality using frequency regularization. The NeRF model is obtained via joint training of multiple artificial neural networks, whereby the expectation and standard deviation of density fields and RGB values can be evaluated for each pixel. In addition, customized physics-informed neural network (PINN) with residual blocks and two-layer activation functions are utilized to input the density fields of More >

Jiepeng Yao^1,2, Zhanjia Peng^1,2, Jingjing Liu^1,2, Chengxiao Fan^1,2, Zhongyi Wang^1,2,3, Lan Huang^1,2,*

CMES-Computer Modeling in Engineering & Sciences, Vol.137, No.3, pp. 2243-2265, 2023, DOI:10.32604/cmes.2023.028101 - 03 August 2023

Abstract In the establishment of differential equations, the determination of time-varying parameters is a difficult problem, especially for equations related to life activities. Thus, we propose a new framework named BioE-PINN based on a physical information neural network that successfully obtains the time-varying parameters of differential equations. In the proposed framework, the learnable factors and scale parameters are used to implement adaptive activation functions, and hard constraints and loss function weights are skillfully added to the neural network output to speed up the training convergence and improve the accuracy of physical information neural networks. In this… More >

Loc Vu-Quoc^1,*, Alexander Humer²

CMES-Computer Modeling in Engineering & Sciences, Vol.137, No.2, pp. 1069-1343, 2023, DOI:10.32604/cmes.2023.028130 - 26 June 2023

Abstract Three recent breakthroughs due to AI in arts and science serve as motivation: An award winning digital image, protein folding, fast matrix multiplication. Many recent developments in artificial neural networks, particularly deep learning (DL), applied and relevant to computational mechanics (solid, fluids, finite-element technology) are reviewed in detail. Both hybrid and pure machine learning (ML) methods are discussed. Hybrid methods combine traditional PDE discretizations with ML methods either (1) to help model complex nonlinear constitutive relations, (2) to nonlinearly reduce the model order for efficient simulation (turbulence), or (3) to accelerate the simulation by predicting… More >