节点文献
轮轨/磁浮动车组虚拟样机若干关键技术研究与应用
Studies and Applications of Some Key Technics in Railway/Maglev Trainset Virtual Prototype
【作者】 闫雪冬;
【导师】 兆文忠;
【作者基本信息】 大连交通大学 , 机械制造及其自动化, 2005, 博士
【摘要】 虚拟样机(VP)技术是复杂产品设计的一种新方法。论文首先给出了虚拟样机的定义和内涵,并对国内外虚拟样机技术的研究现状及发展趋势进行了综述。根据铁路机车车辆行业中典型的复杂产品——轮轨/磁浮动车组的技术复杂性,论证虚拟样机技术可以缩短轮轨/磁浮动车组的研发周期,最大可能地降低物理样机投资风险的可行性及紧迫性。 以虚拟样机的核心技术——性能仿真为切入点,论文从理论上论证了性能仿真与CAD技术相比可以为产品创造更多技术附加值。论文并依此构建了主要研究方向。 在虚拟样机性能仿真过程中,提高仿真置信度至关重要,而提高仿真置信度的关键却在于建模。论文针对机车车辆行业性能仿真中过分依赖仿真软件的错误倾向,将多年的性能仿真建模经验予以归纳、提炼和升华,总结了四项建模原则,即:能量等效替换原则;定性力学原则;主要矛盾原则;反馈原则。同时,还讨论了边界条件模型化以及如何发现模型中错误等技术问题。论文将这些建模原则及技术具体化,并应用于结构复杂的三个典型产品:时速210公里中空挤压铝型材的高速动车组;不锈钢点焊结构的城市轻轨动车组;中空挤压铝型材/夹层蜂窝板混合结构的中低速常导磁浮动车组。在这三个工程应用中,针对“点传力”的结构特点,提出了位移主—从建模思想;针对蜂窝夹层复合材料蜂窝尺寸极小的结构特点,提出了不丧失蜂窝材料基本功能的等效“大蜂窝”建模思想。三个产品物理样机的测试结果证明了上述建模原则及建模思想非常实用,论文并给出了具体的数据对比。 在动力学仿真方面,论文开展了刚-柔混合建模技术研究,不仅讨论了刚-柔混合建模算法,还讨论了刚-柔混合建模的关键接口技术。轮轨动车组的盘形制动实例给出了一个完整的刚-柔混合建模技术路线。 考虑到轮轨/磁浮动车组多学科交叉的技术特点,论文研究了性能仿真过程中跨平台工具软件仿真的协同与集成技术。基于液-固耦合的动车组水箱起皱机理分析,表明协同与集成是虚拟样机性能仿真的重要平台技术,为此,基于PDM技术管理模式,论文提出了轮轨/磁浮动车组性能仿真中间数据文件的协同管理框架。 由于主动控制已经成为不提高线路等级而提速的核心技术及未来趋势,也由于中低速常导磁浮动车组已对城市轨道交通带来了另一个极有经济、环保潜力,且又可行的新思路,因此,论文最后又开展了轮轨/磁浮动车组主动控制探索性研究。关于轮轨动车组垂向主动控制,论文针对跨平台工具软件算法中的积分步长不协调导致的输出偏差问题;ADAMS环境中控制器代码非光滑连续导致的高频成分误差问题;以及系统动力学线性化的速度敏感性问题,提出了基于模型移植的求解策略,并用工程实例验证了此项技术的实际意义。作为探索,又基于遗传算法对PID控制器三个参数进行了优化,从而提高了控制的平稳性。 关于中低速常导磁浮动车组,论文构建了包括磁间隙在内的磁浮车单转向架动力学仿真模型,并成功地解决了磁转向架防侧滚板的机械耦合问题。 在结论与展望一章,提出了论文的三个创新点,它们是: (1) 在复杂产品结构建模方面,针对点焊结构、铆焊结构的难点,提出了基于主—从关系的点传力建模技术;针对磁浮车蜂窝复合材料建模难点,提出了“大蜂窝”建模技
【Abstract】 Virtual Prototype (VP) technology is a new design method for all complex products. In this paper, the definition and implied meaning of Virtual Prototype is firstly given, and then, the research situation and future development about Virtual Prototype technology are summarized. According to the technological complexity of rail-wheel/MagLev trainset, which is very typical products in railway vehicle industry, it is proved that shortening the design period and decreasing as large as possible the venture of physical product are total feasible and urgently needed while based on the Virtual Prototype technology.In this paper, it is theoretically addressed that performance simulation is able to create more additive value than CAD technology. As performance simulation is the key technology of Virtual Prototype, so the main research direction in this paper is therefore constructed following it.While using the Virtual Prototype technology, how to keep the high fidelity of performance simulation is key issue, and it has being pointed out how to create good simulation model is the most key issue also. In order to avoid such a wrong incline excessively depending on commercial simulation software, the many-year experiences coming from dealing with the real engineering problems are summarized, and the four principles, which could be used to create the simulation models, are established as well. The four principles are: (1) Energy Equivalent Exchange Principle; (2) Qualitative Mechanics Principle; (3) Keeping Main Contradiction Principle; (4) Feedback Principle. At same time, some useful technology for creating models and how to avoid and find some mistakes in models are suggested as well. Based on these principles and useful technology, three FEA models of newest and most complex products among all railway vehicle factories in China are created. And they are: (1) high-speed train with 210Km/h constructed with hollow-shape aluminum beams; (2) stainless steel city-car constructed with stainless steel and using spot-welding technology; (3) mid-/low-speed EMS MagLev trainset constructed with hollow-shape aluminum beams and composite material named as cellular layers structures. In these three engineering applications, according to the structure feature, the master-slave displacement constrain modeling technology is employed to define the special node which is able to keep two nodes always have same displacements such as spot nodes and rivets. Since the size of each cell is too small to create its finite element model, so keeping its main features, a new special method is employed which is named big-cell method. Comparing the results between the finite element analysis results of virtual prototypes and the measured results of corresponding physical prototypes, it is clear that the above modeling principles and technology are very useful and practical. The more details within comparisons are given in the paper in addition.In the dynamical simulation aspect, in order to obtain more information, rigid-flex hybrid modeling technology is developed. Both arithmetic and interface technologies of rigid-flex hybrid modeling are discussed. Taking the disc-brake paradigm of rail-wheel trainset as an example, the whole rigid-flex hybrid modeling technological flow chart is established.Taking into account that the coupling features of multiple-domain in rail-wheel/MagLev train set, the co-simulation and integration among many different software environments are investigated. Taking buckling-mechanism analysis of solid/liquid coupling in water-tank as an example, it is shown that co-simulation and integration are important technology underperformance simulation environment. Based on PDM technical management mode, collaborative management frame is presented in this paper for managing so many midway data files.Because the active control is another key technology and it could be used to increase the running speed when train is still running on the existing-grade railway, and also because the mid-/low-speed EMS (Electro-Magnetic Suspension) MagLev trainset comes into the city railway transportations, which has the extreme economical and ecological advantages, so at the end of this paper, the research about active control technology is focus on such three difficult problems: (1) output errors caused by the integral step harmonization; (2) the high-frequency error produced when the controller code-model being imported into ADAMS environment; (3) initial velocity non-sensitivity in linearising the dynamical model. After solving such problems, various engineering applications are given to verify these solving strategys. As a trial technical exposition, the three parameters in PID control are optimized through the genetic arithmetic and the control system stability is also improved.On mid-/low-speed EMS MagLev trainset, the single-bogie MagLev vehicle dynamical model is constructed including the electro-magnetic levitation control, and successfully solved the mechanical coupling problem solved through the anti-roll beams finally.In the last chapter, three innovations are given as follows:(1) In the field of creating simulation models of complex products, in order to overcome the difficulties such as how to create the model of the spot welds and rivet-jointed structures, based on the master-slave displacement constrain relationship, node-transferred force modeling technology is employed; in order to overcome the difficulties such as how to avoid large scale of elements, based on the static and dynamic energy equivalent exchange principle, the large-cell modeling is employed. And the comparisons between simulation results and the tested results prove the new technology mentioned above is very useful in engineering applications.(2) In the field of increasing the fidelity of performance simulation, four practical modeling principles are proposed: (1) Energy Equivalent Exchange Principle; (2) Qualitative Mechanics Principle; (3) Keeping Main Contradiction Principle; (4) Feedback Principle. These modeling principles have presented their engineering values.(3) In the field of active control, a strategy named as ’seamless solution’ is presented based on model immigration, and three very difficult problems have been solved : (1) output errors caused by the integral step harmonization; (2) the high-frequency error produced when the controller code-model being imported into ADAMS environment; (3) initial velocity non-sensitivity in linearising the dynamical model. The consistence between the dynamical control system simulation and the expected theoretical values shows the effectivity of model immigration.
【Key words】 Virtual Prototype; railway-wheel trainset; MagLev trainset; performance simulation; active control;