基于机器学习的耐蚀低合金钢跨尺度数据挖掘研究  

Study of Data Mining of Corrosion-resistant Low Alloy Steel Based on Machine Learning

作　　者：杨小佳 李清 朱仁政 杨国威 卫达 王昕煜 程学群 李晓刚 YANG Xiaojia;LI Qing;ZHU Renzheng;YANG Guowei;WEI Da;WANG Xinyu;CHENG Xuequn;LI Xiaogang(Institute of Advanced Materials Technology,University of Science and Technology Beijing,Beijing 100083,China)

机构地区：[1]北京科技大学新材料技术研究院,北京100083

出　　处：《鞍钢技术》2024年第6期70-83,共14页Angang Technology

基　　金：十四五科技部重点研发青年科学家项目(2023YFB3813200);国家自然科学基金青年项目(52203376)。

摘　　要：利用机器学习方法,以户外积累的环境腐蚀大数据及实验室加速试验获取的微观组织结构的腐蚀数据作为数据源,通过训练学习,获取环境因素中引起低合金结构钢腐蚀的关键因素,并从合金成分出发,分析合金元素对耐蚀性影响的权重因子;同时,结合材料微观结构数据,分析材料微观组织结构差异对耐蚀性影响的原因。基于以上学习训练模型,建立合金成分及组织结构预测低合金钢腐蚀规律的试验方法。By machine learning method,taking the environmental corrosion data accumulated outdoors and the corrosion data of microstructure obtained by accelerated tests at laboratory as data sources,the key factors causing the corrosion of low alloy structural steel were obtained through training and learning.And the weight factors affecting the corrosion resistance of alloy elements were analyzed from the alloy composition.At the same time,combined with the material microstructure data,the reasons for the influence of the material microstructure difference on the corrosion resistance were analyzed.Based on the above learning and training model,a test method for predicting the corrosion law of low alloy steel based on the alloy composition and microstructure was established.

关 键 词：机器学习 耐蚀钢 合金成分 组织结构 

分 类 号：TG142[一般工业技术—材料科学与工程] TP181[金属学及工艺—金属材料]