面向氢能机车的加氢机器人力控算法设计

打开文本图片集

中图分类号：TP23;TP249 文献标志码：B doi：10.20214/j.cnki.zhgdjt.2025.06.012

Abstract：With the rapid development of hydrogen locomotives，hydrogen refueling robots，as the core equipment for realizing automated hydrogen refueling technology，have become a research focus. However， traditional force control methods struggle to achieve precise and anthropomorphic motion adaptation under complex environmental disturbances. To address this limitation，this paper proposes a neural network-based force control algorithm designed to enhance the accuracy and safety of robotic hydrogen refueling operations. The algorithm incorporates three key innovations：establishing a Lagrangian dynamics model to characterize robotic motion properties，utilizing neural networks for real-time model parameter identification；combining servo motor current measurements with the identified dynamics model to estimate equivalent external force disturbances during hydrogen refueling by integrating servo motor current measurements with the identified dynamic model，eliminating dependency on torque sensors；and enhancing trajectory smoothness in motion replay phases through optimized spline interpolation strategies.Simulation and physical experiments show that compared with traditional impedance control methods，the proposed algorithm reduces position error by 38.7% ，while stably controlling the contact force within a safe threshold of 45to 55N . The research results verify the effectiveness of the algorithm in achieving anthropomorphic motion customization and strong antiinterference capabilities，providing a theoretical basis for inteligent hydrogen refueling systems.

Key words：hydrogen-powered locomotive；force control algorithm；automated hydrogen refueling; industrial robot；deep neural network

在低碳环保背景下，氢能机车在铁路行业加快推广应用。（剩余6272字）