Enhancing forecasting of current-carrying performance through spatial frequency analysis of interface morphology  

作　　者：Nian YIN Zishuai WU Zhangli HOU Yiwei ZHANG Zhinan ZHANG 

机构地区：[1]State Key Laboratory of Mechanical System and Vibration,Shanghai Jiao Tong University,Shanghai 200240,China [2]School of Mechanical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China

出　　处：《Science China(Technological Sciences)》2025年第2期167-177,共11页中国科学(技术科学英文版)

基　　金：supported by the National Natural Science Foundation of China(Grant Nos.12072191,51875343);Science and Technology Commission of Shanghai Municipality(Grant No.24DZ2307500).

摘　　要：Electrical contact interfaces,integral to devices such as connectors,switches,and conductive slip rings,play a crucial role in determining current-carrying capabilities.This study adopts spatial frequency analysis to assess the currentcarrying performance of interfaces based on their morphological characteristics.Initially,rough surfaces with various parameters,including the cutoff frequency(N),the spectral index(β),and the gain factor(A),were generated using spatial frequency methods,followed by current-carrying simulations.The findings reveal that the characteristic parameters derived from spatial frequency methods exhibit a significantly stronger correlation with current-carrying performance than traditional roughness metrics.To enhance the practical applicability of this research,a physical-informed machine learning approach for rough surface features extraction was developed,demonstrating high predictive parameter identification accuracy for the three parameters.Overall,this paper introduces a tribo-informatics approach for analyzing and extracting rough surface features that considers the curvature characteristics of contact points.This approach holds significant potential for evaluating the processing state of current-carrying interfaces,selecting high-quality surfaces,and forecasting current-carrying performances.

关 键 词：tribo-informatics current-carrying interfaces spatial frequency analysis machine learning 

分 类 号：TM501.3[电气工程—电器] TP181[自动化与计算机技术—控制理论与控制工程]