刚性接触网-受电弓系统是城市轨道交通的驱动力来源。目前我国的刚性接触网发展存在两大问题：一是以刚性接触网为主要供电装置与以柔性接触网为主要供电装置的列车在速度方面差距较大；二是刚性接触网的检测/监测技术需要提升。为了弥补这两方面的不足，提高弓网受流质量，本文以优化弓网耦合动力学特性为目标开展了弓网系统的整体力学优化设计，作为列车提速的一个基本措施；同时探讨了基于模态实验识别刚性接触网的重要结构参数-悬挂刚度的可行性。主要研究内容和结果如下：

一、弓网结构系统的力学优化设计

1.建立刚性接触网弓网仿真模型。其中受电弓模型采用经典的三质量块模型，刚性接触网采用多跨梁模型，弓网接触采用罚函数法建立。探究单因素如单双弓、双弓间距、静态抬升力、受电弓和刚性接触网的结构参数、定位点偏差等因素对弓网受流质量的影响。结果表明，当单元内双弓间距为4m时，双弓系统的力学性能和电气性能可同时得到保证，此外适当降低双弓系统前弓的静态抬升力可以优化弓网受流质量。

2.多因素灵敏度分析。设计正交试验方案，并对正交试验方案的结果进行极差分析和方差分析，对弓网系统的结构参数进行评级，得到的结构参数敏感性由高到低排序依次为刚性接触网跨距、受电弓弓头质量、刚性接触网线密度、受电弓下框架阻尼、刚性接触网悬挂结构刚度和受电弓上框架阻尼。

3.弓网联合优化设计。设计拉丁超立方试验样本，利用样本仿真结果建立多项式响应面模型、kriging模型、神经网络模型等数学模型。利用决定系数、均方根误差对模型择优。基于自适应模拟退火法对模型的结构参数进行了优化求解。基于结构动力学分析表明：上述优化结果的本质是分别提高了刚性接触网和受电弓的结构基频，可将列车最高时速提高17%，这有助于以后的弓网系统总体方案设计。

二、刚性接触网悬挂结构等效刚度的识别方法

针对灵敏度分析过程中得出的刚性接触网悬挂结构对弓网受流质量影响较大的结论，本文提出了一种新的利用模态实验测量的刚性接触网的固有频率识别悬挂结构等效刚度的方法。此方法将悬挂刚度识别的结构动力学反问题转化为优化问题，并基于遗传算法和计算机仿真验证了该方法的可行性。研究表明常规工况下该方法的平均识别误差小于5%。该刚度识别方法既可用于新刚性接触网和维修后刚性接触网结构的仿真模型验证和修正，也可用于服役结构的损伤识别，作为实际弓网运维的一项新的检测/监测技术。

Pantograph-Overhead Conductor Rail (OCR) system is the driving force of urban rail transportation. At present, there are two problems in the development of OCR in China. One of them is that the speed of the train with OCR as the main power supply device is slower that with flexible catenary; the other is that the detection/monitoring technology of OCR needs to be improved. In order to overcome these two shortcomings, this paper carries out the overall dynamical optimization design of pantograph-OCR system with the objective of optimizing the dynamic characteristics of the systems, which is a basic measure to increase the train speed. At the same time, this paper investigates the feasibility of identifying the suspension stiffness based on modal experiment, which is one of the most important structural parameters of OCR. The main research contents and results are as follows：

Ⅰ. Dynamical optimization design of pantograph-OCR system as whole.

1.The simulation model of pantograph and OCR is established. The pantograph model adopts the classic three-mass block model and OCR adopts multi-span beam model. The contact model of pantograph and OCR is established by penalty function method. The influences of some single factors on the current collection quality of pantograph–OCR system are studied, including single or double pantographs, distance between two pantographs, static lifting force, each structural parameters of pantograph–OCR system, positioning points deviation on OCR. It is shown that when the distance between two pantographs in single unit is 4m, both the dynamical and electrical properties can be satisfied. In addition, reducing the static lifting force of the front pantograph can optimize the current collection quality of pantograph-OCR system.

2. Multi-factor sensitivity analysis is performed. The results of range analysis and variance analysis are used to grade the structural parameters of the pantograph -OCR system. The order of sensitivity of structural parameters is obtained, from high to low, as OCR span, head mass, line density of OCR, damping of lower frame, stiffness of suspension structure of OCR and damping of upper frame.

3. Joint optimization pantograph-OCR system. According to the sensitivity analysis, the factors which have the remarkable influence on the current collection quality of pantograph-OCR system are selected for joint optimization. The Latin hypercube test samples are designed, and the polynomial response surface model, Kriging model and neural network model are established based on the simulation results. The decision coefficient and root mean square error are used to select the best model. The structural parameters of the model are optimized based on the simulated annealing method. The structural dynamics analysis shows that the essence of the optimization results is to increase the fundamental frequencies of the OCR and the pantograph, respectively, and the maximum speed of the train can be increased by 17%. The present analysis is helpful to the overall design of the pantograph-OCR system.

Ⅱ. Development of the identification method of suspension structure equivalent stiffness of OCR

According to the present finding from the sensitivity analysis that the OCR suspension structure has greater influence on the current collection quality of pantograph-OCR, this paper proposes a new method to identify the equivalent stiffness of suspension structure from the measured natural frequencies of OCR based on the modal experiment. The method transforms the problem of suspension stiffness recognition, a dynamical inverse problem, into an optimization problem that is resolved by genetic algorithm. The feasibility of the method is verified by computer simulation. The result shows that the average recognition error of the method is less than 5% under normal working conditions. The stiffness identification method can be used not only to verify and modify the simulation models of the new OCR and the repaired OCR, but also to identify the damage of the structure in operation, which provides with a new detection / monitoring technology for the pantograph - OCR systems in practice.