The 13th International Conference on Fracture Fatigue and Wear (FFW 2025)

Abstract: Numerical methods have been extensively applied to the fracture mechanics, while they cannot simulate the problems without the mechanical models or constitutive equations. Artificial neural networks (ANNs) can be utilized to predict the complex fracture problems, but these approaches require large amounts of data for the training. Therefore, in this paper, to combine the advantages of the numerical methods and the ANNs, an improved back propagation neural network (BPNN) is proposed through introducing the enrichment used in the numerical methods into the activation function utilized in the neural networks. The enrichment is able to represent the crack tip field which can accelerate the convergence. At the field near the crack tip, the improved BP solution can converge to the analytical solutions which validate the high accuracy of the proposed method. Without sufficient data, especially the data are missing in the near field of the crack tip, the improved BP method can also achieve high accuracy and convergence, while the conventional BP method may not converge to the predetermined error bound. Numerical simulations of the quasi-static and fatigue crack problems demonstrate that the improved BP method can accurately predict the crack propagation and its growth rate with relatively little data.
Subsea carbon sequestration technology plays a crucial role in addressing global climate change, but CO2 leakage can harm the subsea ecosystem. Therefore, long-term monitoring and prediction of subsea carbon storage are essential. In this paper, forward and inverse Data-Assisted Physics-Informed Neural Networks (DA-PINNs) are established for subsea CO2 leakage prediction and detection. Firstly, the forward DA-PINN model integrates numerical simulation data and physical constraints including initial conditions, boundary conditions, and governing equations. This model is utilized to predict the CO2 velocity and pressure fields under different leakage widths and initial velocities. The results show that the proposed algorithm outperforms conventional Artificial Neural Networks (ANNs) in accuracy and exceeds the efficiency of conventional numerical simulations. Subsequently, the inverse model incorporates known initial and boundary conditions of leakage as training data, while the governing equations and pressure boundary conditions serve as physical inputs. The inverse DA-PINN model is then used to detect leakage widths and initial velocities under different velocity and pressure fields, achieving a prediction accuracy of over 97%. Compared to conventional ANNs and numerical simulations, the proposed DA-PINNs not only predict CO2 leakage with high accuracy and efficiency but also solve inverse problem with the same high precision and effectiveness.

Biography: Dr. Lihua Wang is a professor at School of Aerospace Engineering and Applied Mechanics in Tongji University, Shanghai, China. She is currently a General Council Member of the International Association for Computational Mechanics (IACM) and the International Chinese Association for Computational Mechanics (ICACM). She is the recipient of several awards, including the APACM Award for Young Investigators in Computational Mechanics, the Qian Linxi Computational Mechanics Award (Young Investigators), the ICACM Young Investigator Award, and the Du Qing-Hua Medal & Young Researcher Award of Computational Methods in Engineering. She has authored more than 120 peer-reviewed journal articles, including CMAME, IJNME, JCP etc., and has been invited to deliver more than 10 plenary and invited lectures at international conferences. She served as associate editor of Chinese Quarterly of Mechanics and as an editorial board member for four international/Chinese journals. Her research interests include development of meshfree methods and machine learning, fluid-structure interactions, high-speed impact, fracture mechanics.