【作者】 林颖；

【导师】 张卫华； 王长林；

【作者基本信息】 西南交通大学 ， 交通信息工程及控制， 2013， 博士

【摘要】 车载ATO (Automatic Train Operation)是实现列车自动运行的核心设备之一,属于现代的列控设备,是轨道交通智能化的表现。自动调速功能是车载ATO最为核心的功能之一,通常由运行模式曲线计算模块和速度自动控制模块共同实现列车的自动运行功能。运行模式曲线计算模块根据计划运行时间和计划运行路径,综合考虑准点、精确停车、节能、舒适性、线路状况、列车性能等条件,计算列车的最佳运行轨迹。速度自动控制模块以该运行轨迹为目标,控制列车运行。运行模式曲线计算方式可分为在线计算方式和离线计算方式。在线计算方式能实时获取列车运行过程中的各种信息,及时做出有利调整,提高了车载ATO的适应性能力。基于在线计算方式的ATO运行模式曲线优化模型是一个高实时性的多目标优化问题。本文分析和总结了现有的多目标优化算法,在动态规划最优性原理的基础上,以时间为依据进行阶段划分,将运行曲线划分为若干相互关联的阶段,在每个阶段采用多目标层次决策的思路计算最优解,实现ATO运行模式曲线的优化。论文的主要研究内容如下：1.根据列车运行控制和优化目标的特点,建立了列车运行控制的数学模型体系。该数学模型体系包括了列车运动学优化模型、控制目标计算模型、控制目标评价模型、列车状态转移模型等。2.针对ATO运行模式曲线优化问题中,主、客观权值存在的不足,提出基于主观权值、局部客观权值以及全局客观权值的ATO控制目标权值计算和权值组合计算模型。将ATO控制目标的客观权值分为局部客观权值和全局客观权值,实现了对不同阶段客观因素对阶段性能指标和整体性能指标影响的描述。同时,可通过改变主观权值、局部客观权值和全局客观权值的权值组合方案,计算满足不同决策目的的决策权值,从而直接影响多目标决策的结果,达到预定的优化目的。3.针对准点、精确停车、节能以及舒适性目标优化的特点,提出ATO多目标层次决策思路,将多目标决策分为两层。第一层多目标决策只对节能和舒适性目标进行优化,使得节能和舒适性指标达到最优；第二层多目标决策对准点、精确停车、节能和舒适性目标进行优化,通过牺牲节能和舒适性指标,使得准点和精确停车指标满足约束条件。论文提出的多目标层次决策中的两层多目标决策之间彼此相互独立,其本质是在进行所有目标优化之前先对节能和舒适性目标进行优化,判断优化结果是否仍能满足准点和精确停车约束条件,从而在不影响准点和精确停车指标的前提下,实现最大化提升节能和舒适性指标的目的。4.基于在线计算方式的ATO运行模式曲线计算是一个具有高实时性、多目标、控制对象非线性、且复杂多变的外界条件对控制目标的影响显著等特点的多目标优化问题。本文以列车运行控制的数学模型体系为基础,采用多目标层次决策优化思路,提出了基于在线计算方式的ATO运行模式曲线计算模型。通过实验仿真证明,本文建立的ATO运行模式曲线计算模型在实时性和多目标优化方面都能满足了车载ATO自动调速功能的需求,对于该领域的技术进步具有一定的积极意义。更多还原

【Abstract】 On-board ATO (Automatic Train Operation) is the core parts equipment of automatic train control. It is an advanced train control equipment and demonstrates the intelligence of rail transport. Automatic speed control is one core functions of ATO, usually, it is realized through the combination of the operation type profile calculation module and the module of automatic speed control. After taking comprehensive consideration on time, accurate parking, energy saving, comfort, track condition, train performance etc, operation type profile calculation module will calculate the best train operation track based on the programmed operation time and operation paths. Automatic speed control module control train operation according to the operation truck. The operation type profile calculation module has two different models:on-line and off-line. The online calculation module can improve the adaptive capacity of ATO, which is able to obtain real-time information and adjust the operation in time during the operating process.The online calculation of operation type profile is a high real time capability of multi-objective optimization problem. After analyzes and summarizes the existing multi-objective optimization algorithm and on the basis of the dynamic programming principle of optimality, the method of the thesis is to divide the profile into a number of interrelated phases according to the operation time. At each stage, using multi-objective hierarchical decision to calculate the optimal solution and achieve the optimizing of ATO operational mode profile.The main works of this thesis can be surmised as the followings:1. According to the characteristic of train operation and control target, the thesis establishes the mathematical model system of train operational control. The mathematical model system includes the optimized model of train kinematics, the calculation model of control parameters, the calculation model of control target, the evaluation model of control target and the state transition model of train operation, etc.2. To overcome the shortcomings of single subjective and objective weight existing in solving the optimization problem, this thesis proposed the model of ATO controlling target weight calculation and weight combination calculation based on subjective weights, local objective weights and objective weights.By dividing ATO control objective weights into local objective weights and global objective weights can realize the description of different stages of objective factors impacts on the stage performance and the overall performance. Through changing the combination plan of subjective weights of objective weights, local and global weights of objective weights, the decision will directly influence the outcome of multiple objective decision making, and finally achieve the expected purpose of optimization3. The thesis proposes the idea of ATO multi-objective hierarchical decision method and divided Multi-objective decision-making into two layers according to features of the Punctual, precise Parking, energy-saving and comfort-objective optimization. The function of first layer of multi-objective decision is only optimizing the energy efficiency and comfort goals and make energy efficiency and comfort s be the optimal; The second layer is used to Optimize punctuality, precision Parking, energy-saving and comfort index and makes punctuality and precision Parking s satisfy the constraints through the expense of energy efficiency and comfort index.Multi-objective hierarchical decision between the two layers of multi-objective decision-making proposed in the thesis is independent from each other. It optimizes energy efficiency and comfort index before optimizes all the index and judge whether the optimization results meet punctuality and precision parking constraints and realize the purpose of maximizing enhance energy saving and comfort index under the premise that punctuality and precision parking index is not affected.4. The on-line calculation of ATO best operational type profile is a multiple optimization problem, which has the characteristic of high real-time, multiple target, nonlinearity of control system and various and complex environment obviously effect on control targets.Based on the mathematical model system of train operation, using the idea of multi-objectives comprehensive optimization based on multi-level decision, this thesis proposes the on-line calculation model of ATO best operational type profile. Through the experimental simulation, the on-line calculation model of ATO best operational type profile can satisfy the real-time and optimization requirements, which has certain positive significance in the field of technological progress.更多还原