您好，欢迎访问江苏省农业科学院 机构知识库！
江苏省农业科学院机构知识库

全部

  • 全部
  • 标题
  • 学者
  • 机构
  • 关键词

Wettability prediction and feature importance analysis for laser-chemical induced superwetting surface based on machine learning

文献类型： 外文期刊

作者： Fu, Jiajun 1 ; Li, Yide 3 ; Song, Xinrong 1 ; Liu, Chao 1 ; Zhao, Runhan 1 ; Wang, Huixin 4 ; Wang, Qinghua 1 ;

作者机构： 1.Southeast Univ, Sch Mech Engn, Nanjing 211189, Peoples R China

2.Jiangsu Key Lab Design & Mfg Precis Med Equipment, Nanjing 211189, Peoples R China

3.Shandong Special Equipment Inspect Inst Grp Co Ltd, Jinan 250101, Peoples R China

4.Jiangsu Acad Agr Sci, Inst Agr Facil & Equipment, Nanjing 210014, Peoples R China

5.Minist Agr & Rural Affairs, Key Lab Protected Agr Engn Middle & Lower Reaches, Nanjing 210014, Peoples R China

关键词： Laser surface texturing; structure and chemistry; superwetting surface; machine learning; wettability prediction

期刊名称：JOURNAL OF MANUFACTURING PROCESSES （ 影响因子：6.8； 五年影响因子：7.0 ）

ISSN： 1526-6125

年卷期： 2025 年 152 卷

页码：

收录情况： SCI

摘要： Surface with extreme wettability prepared by laser-chemical composite process has been widely used in the fields of anti-icing, self-cleaning, oil/water separation and biological detection. However, it was difficult to accurately predict whether a group of surfaces possessed extreme wettability based on the traditional physical models due to the complex coupling between the influencing factors of superwetting surface including structure and chemistry. In this study, a set of surface morphology descriptors in micro/nano scale and surface chemistry descriptors in molecular-scale were designed as the inputs for the establishment of a machine learning (ML)-based prediction model in terms of surface wettability. The XGBoost model was utilized due to its superior performance and the parameters of the model were optimized. The SHapley Additive exPlanations (SHAP) method was used to obtain the most relevant features of the surface wetting characteristics. The contact angle of the newly prepared superwetting surface was predicted. By experimental validation, the results showed that the Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) values of the XGBoost model on the validation set were 6.55 and 4.61 degrees, respectively. In addition, the R2 value of the prediction results reached 0.99, indicating that the model was able to adequately capture the connection between the selected features and surface wettability. This proved that the proposed ML model exhibited excellent generalization ability and could perform well on new data sets, even though these data were obtained from different treatment methods. In addition, the experimental data required by this method can be easily obtained, and thus the subsequent data set can be quickly supplemented. Therefore, it could provide good reference and guidance for the design and fabrication of different types of superwetting surfaces.

  • 相关文献
作者其他论文 更多>>
置顶
购物车
反馈

© 京ICP备09089781号-29 主办单位：江苏省农业科学院

学术资源： 中国农业科学院 农业专业知识服务系统 农科联盟资源共建共享平台 中国农业科技文献与信息服务平台 中国农业期刊集成服务平台