英文翻译

毕业设计/论文外文文献翻译Optimization of high speed EMU suspension parameters forvibration reduction高速动车组为了减振而对悬挂参数的优化院系机电工程学院专业班级学号姓名指导教师高速动车组为了减振而对悬挂参数的优化Seung-Guk Baek , Bumsik Shin , Sang Won Lee , Yeon-Sun Choi , Jinhyun Kim and Ja Choon Koo摘要：在许多国家积极努力下，我们设计出多种最大运行速度400km/h以上的高速动车组。

在这种极高速驾驶的情况下，配备多个动力单元和非铰接式转向架的动车组经常承受着严重的振动。

注意到动车组垂向振动的这一设计问题，目前的工作就是试图利用动车组全自由度分析模型优化车辆悬挂系统。

目标函数是由铁路功率谱密度（PSD），车身的传递函数以及可表示人体乘坐舒适性的频率加权函数组成的。

传递函数使用频域的自相关模型来制定。

为提高优化工作的可靠性，美国FRA（联邦铁路局）轨道PSD和实际轨道PSD收购了韩国Kyeongbu高速铁路，特别是将运用动车组的区段。

这充分证明了BFGS（Broyden-Fletcher-Goldfarb-Shanno）算法是适用于该优化的。

该优化设计已在不同的工作条件下使用商业动态系统仿真代码进行了验证。

关键词：优化；高速列车；悬浮液；频率域；BFGS算法；动车组1.引言许多国家正在积极开发研究时速在400km/h或更高速的列车。

不同于传统的高速动车组，在韩国新开发的列车配备由非铰接式转向架和分布式电气设备。

作为一个实际问题，设计工程师们感兴趣的是分布式电气系统造成过度振动的可能性，它们分别联接在每节旅客列车的地板下面。

很少有出版物或先前的研究活动可以用于此高速列车的设计和相关的设计问题。

此外，和其他传统高速动车组一样，轨道不平顺和轮轨之间的复杂接触关系是产生激励的主要原因，因此一篇对车辆动态的全面研究应该把动力分散式列车的减振也考虑在内。

考虑到时速400km/h的列车在极端操作条件下的车辆动态特性以及可能对其稳定性造成的影响，本文试图解决车身悬挂系统优化设计的技术与设计问题。

悬挂系统的设计与敏感性分析有关。

但是不论是提供最优值还是考虑系统的稳定性都很难做到。

所以车辆悬挂系统的优化设计问题已经作为动态稳定裕度在汽车设计和制造业进行了深入地研究，并且其乘坐舒适性也成为行业的主题。

通过限制在时域中的主观功能而尽量减小车体的垂向加速度。

Haug编写出了实现这一功能的目标函数。

但是由于时域系统中的信息很难得到，因此，许多研究都是在频域中进行的。

振动对乘坐舒适性的影响是根据其方向和频率而改变的，基于UIC 513R 和ISO 2631里的信息。

根据不同的道路情况而在时域和频域进行分析的公式是由Bruce.Kim编写而成的。

他通过使用梯度优化算法在频域中优化车辆的垂直悬挂。

Nishimura则在频域中对时速200公里每小时的高速列车悬挂进行了优化。

在之前的称述中悬挂的优化设计是在频域中完成的。

但是由于很难推导出理论上的传递函数，所以Myers将响应面模型（RSM）的方法应用到了复杂打的优化问题中。

Park则试图用实验和神经网络设计来优化韩国高速列车的各悬挂变量。

但这些方法都需要一个真实的实验平台。

因此，悬挂装置的优化设计是通过减小等效模型的自由度实现的。

Hedrick模拟了一列在轨道不平顺状态下运行的列车的功率谱密度（PSD）。

Garge则用多种方法实现了列车的建模。

Dukkipati运用数字化模拟对线性悬挂给乘坐舒适度带来的影响进行了测试。

相比被动悬挂装置，现有的研究对象主要都是主动悬挂装置。

然而，实际上高速列车都配备有一套被动悬挂装置来预防主动悬挂装置出故障的情况。

虽然在振动隔离的研究领域中主动悬挂装置是主流，但是为了安全考虑高速列车还是主要使用被动悬挂装置。

因此在现在的运用中应当考虑使用被动悬挂装置。

在意识到垂向加速度组件在整体乘坐舒适度上的优势并且考虑到UIC 513R 和ISO 2631中的规格后，我们目前的工作重点是垂向振动力的目标函数公式。

在目前的工作中，最高运行时速400公里每小时的高速列车的最优悬挂设计是通过最小化垂向振动力来实现的。

这个速度是超过韩国现有的研究高度的。

也正因为400km/h这一速度，使得悬挂系统的优化一直没有在韩国施行。

优化的目标函数是由钢轨的功率谱密度，车体的传递函数和由乘坐舒适度定义的频率加权函数构成。

第一系和第二系悬挂的最优值是通过一系列的数字化模拟来导出并验证得到的。

2.车辆模型为了对动车组车辆的悬挂进行优化，在本文 中我们对某车辆车身进行了建模。

在图1所示的韩国高速列车的动车中，采用了非铰接式转向架和动力单元组，它可以被认为是一种多功率单元型车辆。

我们大多数的设计参数都是是基于韩国的高速列车系统。

图2描绘出了这种10个自由度的系统模型，而相应的运动方程见公式（1）。

提到典型的悬挂优化设计的过程，我们将在这一节确定一个传递函数，它描述了激励源和车身加速度之间的关系。

要确定输入和输出之间的关系，就得导出传递函数。

输入的是车轮与钢轨之间的相互接触作用。

如图3所示，二系悬挂中上部的垂直加速度就是输出。

为了导出传递函数，我们在式1中把车辆模型的运动方程表示为一个向量矩阵。

而对影响单元功能的速度和加速度我们用傅里叶变换将其转化，见公式（2）~（4）。

一位轮对钢轨位移和其他位移之间的关系可以用公式（5）来表达。

其他导出公式一并如下：图1.韩国高速列车（HSR-350x ）图2. 10自由度系统示意图图3.车辆的三个关键点3.悬挂优化乘坐舒适性由车身和转向架的垂直、横向和纵向加速度来评估。

目标函数由频率加权函数、传递函数和输入功率谱密度组成，见公式（10）。

公式（11）则表示了三点位移和车身中间点旋转位移之间的关系。

点P1和P3之间的传递函数用公式（12）表示，最后的传递函数在公式（13）中给出。

公式（14）用来表示Hessian矩阵，它是该算法的关键点。

从优化算法中获得的最优值在表1中给出。

线路悬挂 优化前 优化后 △ Kyeong-Bu 线一系刚度[N/m] 741000 759000 18000 二系刚度[N/m]61600067100055000一系阻尼[Ns/m] 10200 11500 1300 二系阻尼[Ns/m] 19400 23000 3600 FRA 线一系刚度[N/m] 741000 710000 31000 二系刚度[N/m] 616000 616000 0 一系阻尼[Ns/m] 10200 11500 1300 二系阻尼[Ns/m]194002300036004.优化模型仿真为了验证由BFGS 导出的悬挂最优值，我们在ADAMS/Rail 软件中设计了高速动车组的模型。

车辆整体的模型如图7所示，其转向架模型如图8所示。

我们把悬挂优化得到的最优值运用在了ADAMS/Rail 模型中，并对其进行了仿真，以此来验证优化的悬挂。

输入的轨道不平顺功率谱密度是来自FRA 和Kyeong-Bu 高速线路上的。

其校验结果如图10所示。

它向我们展示了起始和最佳振动值，我们可以看到车身振动的减少。

表1.最优值图7.铁道车辆的ADAMS/Rail 模型图8.转向架的ADAMS/Rail 模型垂向悬挂RMS值速度（km/h ）(a) Kyeong-Bu 高速线路轨道速度（km/h ） 垂向悬挂RMS值(b) FRA 高速线路轨道 图10.车身垂向加速度5.结论为了提高最高运行时速400km/h的高速列车的乘坐舒适度并减小其振动，我们建立了一个十自由度的垂向模型，并且在频域中定义了有关乘坐舒适度和车体振动的目标函数。

转向架上半部分的振动也被考虑在内，所以相当于给目标函数施加了一个旋转的影响。

悬挂装置的最优值是由优化算法推导出来的。

被推导出的最优值被运用在ADAMS/Rail模型中，该模型模拟写入的是FRA 和Kyeong-Bu高速线路上的数据，并以此来研究振动的减小。

最优设计的一系和二系悬挂的刚度分别是759000N/m和671000N/m，最优设计的一系和二系悬挂的阻尼分别是11500Ns/m和23000Ns/m。

在未来的研究中，横向自由度、非线性和车体的灵活性也将被考虑在内。

6.参考文献（略）原文：Optimization of high speed EMU suspension parameters for vibrationreductionSeung-Guk Baek , Bumsik Shin , Sang Won Lee , Yeon-Sun Choi , Jinhyun Kim and Ja Choon Koo AbstractVarious designs of high speed EMU (electric multiple units) operating at maximum speed of 400 km/h or above are under active development in many countries. Driving at such extreme high speed, the EMUs that are equipped with multiple power units and non articulated bogies normally experience severe level of vibration. Noting one of the primary design issues of the vertical EMU vibration, the present work tries an optimization of vehicle suspension system using ten degree of freedom analytical EMU model. The objective function for the work is consisted of the rail power spectral density (PSD), transfer function of car body, and frequency dependent weighting function which may represent human riding comfort. The transfer function is formulated using an autocorrelation of the model in frequency domain. For the enhanced reliability of the optimization work, both the US FRA (Federal Railway Administration) rail PSD and the actual rail PSD acquired from Kyeong-bu high speed rail in Korea on which the particular EMU will be driven are used. The well proven BFGS (Broyden-Fletcher-Goldfarb-Shanno) method is adopted for the optimization. The optimized design is verified using a commercial dynamic system simulation code at various operating conditions.Keywords: Optimization; High speed train; Suspension; Frequency domain; BFGS algorithm; Electricmultiple units1. IntroductionSome of the high-speed trains which are under active development in many countries operate at the maximum speed of 400km/h or higher. Unlike the traditional high-speed train power car, the new version of vehicles developed in Korea is to be equipped with non-articulated bogies and distributed power plant. As a matter of fact, design engineers are interested in the study for the possibility of excessive vibrations caused by the distributed power generation systems, which are attached underneath the floor of each passenger car. Few of publications or previous research activities are available for this high-speed train designs and related design concerns. In addition, similar to other traditional high speed trains since rail irregularity and complicate contact situation between the rail and the wheel of a train are the major source of excitation, a comprehensive study on the overall vehicle dynamics should be carried for the reduction of the vibration of the distributed power system trains. Recognizing the vehicle dynamic characteristics under the extreme operating condition such as speed of 400Km/h which might affect the stability of dynamics system, the present paper tries to address technical and design issues of the optimization of car body suspension system.Design of suspension related to the sensitivity analysis. But it is hard to provide optimum value and consider stability of system. So vehicle suspension design optimization problems have been heavily investigated in automobile design and manufacturing industry as the dynamic stability margin and the comfort of rider are major interests. Minimization of vertical acceleration of a car body is carried out by defining a objective function in time domain. An objective function was formulated to minimize the vertical maximum acceleration of the car body in time domain by Haug.However, it is difficult to know much information of system in time domain. Therefore many researches were performed in frequency domain. Effects of vibration on ride comfort differs according to direction and frequency, based on information in UIC 513R and ISO 2631. Analysis method in time and frequency domain was represented with various road about a car by Bruce. Kim obtained a transfer function using a numerical method. And he optimized the vertical suspension of a car by using gradient optimal algorithm in the frequency domain. Nishimura optimized the suspension of a high speed train for speeds below 300km/h in the frequency domain .The optimization of the suspension is done in frequency domain as previous statement. But it is difficult to derive the transfer function theoretically. So Myers applied the method of a response surface model (RSM) to problems of complex optimization. Park tried to optimize the suspension of the Korean High Speed Train by using design of experiments and neural network to consider the various design variables of a railway vehicle. But these methods need real plant.Therefore, the optimization of the suspension is carried out by decreasing the degrees of freedom of an equivalent model. Motion of a train was simulated with power spectral density (PSD) of rail irregularities by Hedrick. Garge represented diverse methods of the modeling of a train.The effect of linear suspension to ride comfort of a train was tested using numerical simulation by Dukkipati .Other existing research is one about active suspension which has been the main currents than that aboutpassive suspension. In reality, however, a high speed train has a passive suspension for safety in relation with the malfunction of an active suspension. Although the active suspension system for a vehicle is the main trend in current vibration isolation research, the passive suspensions are still dominantly used for the high speed trains mainly due to the safety concerns. Hence a passive suspension system is to be considered in the present work. Noting the dominancy of vertical acceleration component in the overall ride comfort and considering the information given by the UIC 513R and ISO 2631 specifications, the present work focuses on the vertical vibrations for the formulation of objective function.In the present work the optimum suspension design of a high speed train which operates at maximum speed of 400 km/h is acquired through the minimization of vertical vibration. The speed is faster than one of existing research in Korea.In 400 km/h speed, optimization of suspension has not been performed in Korea. The objective function for the optimization target consists of rail PSD, transfer function of the car body, and frequency weighting function which is defined by the definition of ride comfort. The optimal values of the primary and the secondary suspensions are derived and validated through a series of numerical simulations.2.Modeling of a vehicleIn order to perform the EMU vehicle suspension optimization,one car body has been modeled in this article. Since the powercar of the Korean High Speed Train which is shown in Fig. 1has a non-articulated bogie and power units, it can be consideredto represent a multiple power unit type vehicle. Most of designparameter sets are based from the Korean High Speed TrainFig 1.system. A model of 10 DOF system is sketched in Fig. 2 and the corresponding equations of motion are provided Eq. (1) [8].Noting a typical process for suspension design optimization, a transfer function which describes a relationship between the excitation source and the acceleration of the car body is to be determined in this section.To determine the relationship between input and output, a transfer function was derived. The input is the contact interaction between wheel and rail. The outputs are the vertical accelerations of the middle and upper secondary suspensions at the car body in Fig. 3. To obtain the transfer function, the equation of motion of the vehicle model is represented as Eq. (1) as a matrix and vector.The velocity and acceleration of the impact unit function can be transformed using Fourier frequency transformation in Eqs. (2)-(4).So the relationship between the rail displacement of the first wheelset and the rail displacement of the others can be expressed in Eq. (5).3. Optimization of suspensionRide comfort is evaluated with respect to the vertical, lateral and longitudinal accelerations of the car body and bogie. The objective function consists of the frequency weighting function, transfer function and input兰州交通大学毕业设计（论文）PSD in Eq. (10). Eq. (11) is the relationship among the displacements of the three points and the rotation of the middle points of the car body. The transfer function of points P1 and P3 can be found using Eq. (12), a relation among transfer functions. The final transfer function is given by Eq. (13). Eq. (14) represents the modification of the Hessian matrix which is the key point of this algorithm.The optimal values obtained from the optimization algorithm are given in Table 1.4.Simulation of optimization modelTo verify the optimal value of the suspension derived by BFGS, an ADAMS/Rail model of the high speed EMU was designed. One volume of the vehicle is shown in Fig. 7. A model of its bogie is shown in Fig. 8. The optimal value of suspension derived by optimization was applied to the ADAMS/Rail model. And a simulation was performed to verify the optimized suspension. The input is the track irregularity PSD of FRA and Kyeong-Bu high speed line. Results are shown in Fig. 10. It represents original and optimal vibration. We can see the reduction of vibration of the car body.5. ConclusionTo improve the ride comfort and reduce the vibration of a high speed EMU with maximum speed 400 km/h, a vertical model of 10 DOF was designed and an objective function related to the ride comfort and vibration of the car body was defined in the frequency domain. The vibration of the upper bogie was considered, so there were rotation effects on the objective function. The optimal value of the suspension was derived by using an optimization algorithm. The derived value of suspension was applied to the ADAMS/Rail model, which was simulated with the track input data of FRA and Kyeong-Bu high speed line, to investigate the reduction of vibration. The optimal primary and secondary stiffnesses were 759000N/m and 671000 N/m respectively. The optimal primary and secondary damping were 11500 Ns/m and 23000 Ns/m, respectively. In the future, the lateral degree of freedom, nonlinearity and flexibility of the car body will need to be considered.10。