摘要
磁悬浮列车管道复压系统需在列车行驶时保持真空状态,在停靠与检修状态时恢复至常压状态。复压过程中,管道内压力变化会影响流入气体的流量和流速,并且气体流入会对设施造成气流冲击,持续或瞬时的强冲击可能导致设备损坏。为降低气动冲击并提高复压效率,针对超高速磁悬浮列车真空管道的工作特点,研究了管道内流场特性,分析了多参数耦合关系,提出一种磁悬浮列车管道真空复压智能控制方法。该方法基于管道内监测传感器的实时数据,预测气体流入流量补偿值,以调整后的压力变化值和复压工作用时作为预测标签,采用自监督深度强化学习训练控制系统,使其自主调整气体流量,以确保复压过程的安全并提升复压效率。试验结果表明:该方法能够保障复压过程中运行设施的安全,复压效率相比PLC控制系统提升13.85%。
The recovery pressure system of the maglev train pipeline must maintain a vacuum state during operation and restore to ambient pressure during parking or maintenance.During the process of recovery pressure,variations in pipeline pressure alter the flow rate and flow velocity of the airflow inflow,whose influx may generate aerodynamic impacts on the operating facilities.Continuous or instantaneous high-intensity impacts can damage equipment.To reduce such impacts and improve the efficiency of recovery pressure,the flow field characteristics in the pipeline were studied according to the working characteristics of the ultra-high-speed maglev train vacuum pipeline.The influence coupling relationship among multiple parameters was analyzed,and an intelligent control method for vacuum recovery pressure of maglev train pipelines was proposed.This method can predict the compensation value of airflow inflow based on the real-time sensor data,using adjusted pressure variations and recovery duration as prediction labels.The self-supervised deep reinforcement learning training method was adopted to control system to autonomously adjust the flow rate of the airflow after training,ensuring the safety of the process of the recovery pressure for the vacuum pipeline and improving the efficiency of the recovery pressure.The experimental results show that this method can ensure the safety of operating facilities during the process of the recovery pressure,the recovery pressure efficiency is 13.85%higher than that of the PLC control system.
作者
王永志
李强
吴利平
吕梦璐
郑鹏达
吴佳宝
孙钦翰
WANG Yongzhi;LI Qiang;WU Liping;LYU Menglu;ZHENG Pengda;WU Jiabao;SUN Qinhan(Research Laboratory of Intelligent Fluid Control,Shenyang Aerospace Xinguang Group Co.,Ltd.,Shenyang Liaoning 110000,China;School of Mechanical Engineering,Shenyang University of Technology,Shenyang Liaoning 110870,China)
出处
《机床与液压》
北大核心
2025年第8期108-113,共6页
Machine Tool & Hydraulics
关键词
磁悬浮列车
真空管道
复压效率
深度强化学习
自监督
maglev train
vacuum pipeline
recovery pressure efficiency
deep reinforcement learning
self-supervised
作者简介
王永志(1987-),男,博士,工程师,主要研究方向为智能流体控制、机器人与智能制造系统应用技术。E-mail:wyzmach@126.com。
