The invention discloses a high-speed train group distributed data driving control method under fault saturation, and the method specifically comprises the steps: building a high-speed train group dynamic model through the Newton's law, and carrying out the discretization and normalization; carrying out equivalent linearization on the nonlinear model by utilizing a pseudo partial derivative and a differential median theorem; establishing a criterion function about system data model parameters and control signals, and preliminarily obtaining a distributed data driving control scheme of the train group according to optimization conditions; establishing an intermittent actuator fault and nonlinear input saturation model existing in the running process of the high-speed train, and introducing a multi-agent theory and a parameter resetting algorithm to obtain a data-driven high-speed train group distributed anti-saturation robust control method; and performing test verification on a numerical simulation platform, and transplanting to a high-speed train automatic operation system. The method gets rid of dependence on a train model, is simple in structure and easy to implement, and can improve the utilization rate of inter-station lines and the transportation efficiency of a high-speed train.
本发明公开了一种故障饱和下的高速列车群组分布式数据驱动控制方法,具体为:利用牛顿定律建立高速列车群组动力学模型,并进行离散化和归一化;利用伪偏导数和微分中值定理将非线性模型进行等价线性化;建立关于系统数据模型参数和控制信号的准则函数,并根据优化条件初步得到列车群的分布式数据驱动控制方案;建立高速列车运行过程中存在的间歇性执行器故障以及非线性输入饱和模型,引入多智能体理论以及参数重置算法,得到数据驱动的高速列车群组分布式抗饱和鲁棒控制方法;在数值仿真平台上进行测试验证,移植到高速列车自动运行系统。本发明摆脱了对列车模型的依赖,结构简单,易于实施,能够提升站间线路的利用率和高速列车的运输效率。
High-speed train group distributed data driving control method under fault saturation
一种故障饱和下的高速列车群组分布式数据驱动控制方法
2024-01-02
Patent
Elektronische Ressource
Chinesisch
High-speed train aided driving control method and system
Europäisches Patentamt | 2023
|Implementation method for autonomous driving control of high-speed train
Europäisches Patentamt | 2021
|Distributed deep learning train emergency driving control method
Europäisches Patentamt | 2023
|