基于BP神经网络的高速动车组网络控制系统时延研究  

Research on Time Delay of Networked Control System for high-speed EMU Based on BP Neural Network

作　　者：宋程 王元东 徐滨全 

机构地区：[1]中车青岛四方机车车辆股份有限公司总装分厂

出　　处：《科技创新与应用》2020年第9期1-5,共5页Technology Innovation and Application

摘　　要：当前,高速动车组的控制系统迅速向网络化与智能化方向发展,网络控制系统已被大规模地应用于高速动车组的控制系统之中,网络控制系统可实时地传输高速动车组中各个设备的数据并监控高速动车组的运行状态,以保证高速动车组安全地运行。但高速动车组的通信网络控制端口和变量数目非常多且各端口的长度与特征周期也各异。故高速动车组的网络控制系统在输出数据的时候,不可避免地出现网络时延现象,该现象严重地威胁着高速动车组网络控制系统的实时性和稳定性。由于高速动车组的网络控制系统是一种非线性系统,故文章针对带有时延问题的高速动车组非线性网络控制系统,提出一种有效的优化方法,即将BP神经网络递推预测与广义预测控制结合起来,先根据BP神经网络递推预测的方法对网络控制系统未来的输出进行预测,然后利用递推最小二乘法对网络控制系统未来时刻的系统参数进行辨识,将非线性网络控制系统在每一时刻进行线性化处理,最后由广义预测控制算法补偿非线性网络控制系统的时延且对所提出的方法进行实验仿真,实验结果表明:该方法可有效地消除高速动车组网络控制系统的时延,同时具有实时性高、计算速度快及鲁棒性强等优点,显著地优化了带有时延问题的高速动车组非线性网络控制系统。At present, the control system of high-speed EMU is rapidly developing in the direction of networking and intelligence, and the network control system has been widely used in the control system of high-speed EMU. The network control system can transmit the data of each equipment in the high-speed EMU in real time and monitor the running status of the high-speed EMU, so as to ensure the safe operation of the high-speed EMU. However, the communication network control ports and variables of high-speed EMU are very large, and the length and characteristic period of each port are also different. Therefore, when the networked control system of high-speed EMU outputs data, the phenomenon of network delay is inevitable, which seriously threatens the real-time performance and stability of the networked control system of high-speed EMU. Because the networked control system of high-speed EMU is a kind of nonlinear system, this paper proposes an effective optimization method for the nonlinear networked control system of high-speed EMU with time delay, that is, combining BP neural network recursive prediction with generalized predictive control, firstly, the future output of networked control system is predicted according to the recursive prediction method of BP neural network. Then the recursive least square method is used to identify the system parameters of the networked control system in the future, and the nonlinear networked control system is linearized at every moment. Finally, the time delay of the nonlinear networked control system is compensated by the generalized predictive control algorithm, and the experimental simulation of the proposed method is carried out. The experimental results show that this method can effectively eliminate the time delay of the networked control system of high-speed EMU. At the same time, it has the advantages of high real-time performance, fast computing speed and strong robustness, which significantly optimizes the nonlinear networked control system of high-speed EMU with time delay.

关 键 词：BP神经网络 广义预测控制 递推最小二乘法 高速动车组 网络时延 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]