文献类型：期刊文献

英文题名：Multimodal fusion and transfer learning enable cross-material life prediction of defect dominated fatigue in additively manufactured metals

作者：Wang, Haijie[1,2];Li, Bo[1,3];Wu, Qimin[1];Zhang, Jianrui[1];Xuan, Fuzhen[1,2,3]

机构：[1]East China Univ Sci & Technol, Sch Mech & Power Engn, Shanghai 200237, Peoples R China;[2]East China Univ Sci & Technol, Key Lab Pressure Syst & Safety, Minist Educ, Shanghai 200237, Peoples R China;[3]Shanghai Collaborat Innovat Ctr High end Equipment, Shanghai 200237, Peoples R China

年份：2026

卷号：211

外文期刊名：INTERNATIONAL JOURNAL OF FATIGUE

收录：;EI(收录号：20261920674290);WOS:【SCI-EXPANDED(收录号:WOS:001766263200001)】；

基金：This research work is sponsored by Science Fund for Creative Research Groups of the National Natural Science Foundation of China (Grant No. 52321002) , Postdoctoral Fellowship Program (Grade C) of China Postdoctoral Science Foundation (Grant No. GZC20250420) , National Key R&D Program of China (Grant No. 2024YFF0505204) , National Natural Science Foundation of China (Grant Nos. 52175140

语种：英文

外文关键词：Fatigue lifeprediction; Defect; Multimodal fusion; Transfer learning; Additive manufacturing

摘要：Accurately predicting the fatigue life of additively manufactured (AM) components remains challenging due to the complex interplay of process-induced defects, microstructural variability, and loading conditions. Here, we present a machine learning (ML) framework that combines multimodal data fusion with transfer learning (TL) to predict the fatigue life of laser powder bed fusion (LPBF)-built metals. By integrating image-based and text-based attributes, we develop a multimodal model for Hastelloy X that significantly improves prediction accuracy by leveraging the complementary information across modalities. We further employ TL to fine-tune this pretrained model using limited data from AlSi10Mg, demonstrating successful knowledge transfer across distinct AM alloys. This approach reduces data requirements for new materials while maintaining high predictive performance. Our results highlight the synergistic potential of multimodal fusion and TL for fatigue life prediction in AM metals, paving the way for more generalizable and data-efficient performance modeling in metal additive manufacturing.