Royal Society Publishing has recently published Part II of a special double issue of Philosophical Transactions A: Physics-informed Machine Learning and Its Structural Integrity Applications (Part 2) compiled and edited by Shun-Peng Zhu (ESIA17–ISSI2023 Session Chair for Probabilistic Failure Assessment), Abílio M P De Jesus, Filippo Berto, John G Michopoulos, Francesco Iacoviello, and Qingyuan Wang.
This theme issue explores the advances in physics-informed machine learning (PIML) and its structural integrity applications through accurate failure mechanism modelling, combining either deterministic or probabilistic analyses by using Artificial intelligence (AI) methods. Specifically, this collection discusses several critical issues related to learning from massive amounts of data, and highlights current research endeavours and the challenges to data science in structural integrity and safety, especially incorporating physics into machine learning models. Note that PIML improves consistency with prior knowledge, extrapolation performance, prediction accuracy, interpretability and computational efficiency and reduces dependence on training data, which provides an excellent opportunity to discover new physics under small samples and ambiguous physical mechanisms.
The articles can be accessed directly at www.bit.ly/TransA-2264.
A print version is also available at the special price of £40.00 per issue (usual price £60.00) by contacting the Royal Society sales team at sales@royalsociety.org and quoting the promotional code TA 2264.
Contents
Preface
Physics-informed Machine Learning and Its Structural Integrity Applications (Part 2)
Shun-Peng Zhu, Abílio M P De Jesus, Filippo Berto, John G Michopoulos, Francesco Iacoviello and Qingyuan Wang
Articles
Automatic machine learning versus human knowledge-based models, property-based models and the fatigue problem
Enrique Castillo, Alfonso Fernández Canteli and Miguel Muñiz Calvente
High-cycle fatigue life prediction of L-PBF AlSi10Mg alloys: a domain knowledge-guided symbolic regression approach
Huan Yu, Yanan Hu, Guozheng Kang, Xin Peng, Bingqing Chen, and Shengchuan Wu
Interpretable fusion methodology of health indices with an application to industrial turbine cavitation condition monitoring
Yichu Fu, Yikai Chen, Dong Wang, and Zhike Peng
Physics-based and machine-learning models for accurate scour depth prediction
Ajay Jatoliya, Debayan Bhattacharya, Bappaditya Manna, Ana Margarida Bento, and Tiago Fazeres Ferradosa
Physics-informed deep learning for structural vibration identification and its application on a benchmark structure
Minte Zhang, Tong Guo, Guodong Zhang, Zhongxiang Liu, and Weijie Xu
Semantic segmentation of defects based on DCNN and its application on fatigue lifetime prediction for SLM Ti-6Al-4V alloy
Jinchao Pan, Dianyin Hu, Liucheng Zhou, Di Huang, Ying Wang, and Rongqiao Wang
A novel learning function for adaptive surrogate-model-based reliability evaluation
Shiyuan Yang, Debiao Meng, HongtaoWang, and Chang Yang
Magnetic flux leakage defect size estimation method based on physics-informed neural network
Yi Xiong, Shuai Liu, Litao Hou, and Taotao Zhou
Data-driven based fracture prediction of notched components
Hossein Talebi, Bahador Bahrami, Mohammad Daneshfar, Sara Bagherifard, and Majid R Ayatollahi
Contact
The Royal Society
6–9 Carlton House Terrace
London SW1Y 5AG
Telephone: +44 20 7451 2500
Web: royalsociety.org
Email: philtransa@royalsociety.org