高速机车轮轨动力学计算机视觉实验研究

The wheel/rail dynamic performance of high speed locomotive not only affects the increase of speed, but also the ride comfort and running safety. In view of the faults in the current locomotive dynamics performance detection methods, such as big detection error, poor performance of anti-noise, high difficulty of the wheel/rail force measurement and multi-source vibration signal separation and etc, an analysis instrument based on computer vision will be designed and developed in this project. In the scheme of the analysis instrument, high-speed cameras and laser sources are assembled together to gain the image. Then, a serial of image segmentation algorithms are used to abstract the geometrical state of the wheel-rail contact, diamond deformation degree and move status of the train bogie. As a result, many dynamic performance parameters, such as critical hunting speed, derailment coefficient, reduction ratio of wheel-load, overturning moment coefficient and etc will be measured directly. Further more, the inherent link between the geometrical state of the wheel-rail contact and the high speed locomotive's nonlinear dynamics behaviors, such as bifurcations and chaotic motions, will be revealed. The research outcomes of this project can not only supply much more scientific and effective technical means for the the wheel/rail dynamic performance detection of high speed locomotive but olso have extremely important theoretical significance and engineering application value in the experiment study field of the high speed locomotive's nonlinear dynamics behaviors.

高速运行时机车轮轨间的剧烈作用不仅会降低机车蛇行运动稳定性，还会影响到机车的运行安全性和乘坐舒适性。实时动态监测轮轨关系、提高轮轨动力学性能至关重要。目前，采用机车动力学性能常规振动检测方法测量高速时的轮轨关系遇到了很多技术瓶颈，如多源振动信号分离困难、容易受噪声干扰、信噪比低、检测误差大等。因此，迫切需要研制一套基于计算机视觉技术的高速机车轮轨动力学性能检测分析仪。本项目拟采用高速摄像机和激光源组合作为信息采集手段，通过先进的图像分割方法来获取机车轮轨接触几何关系、转向架菱形形变程度和运动状态，实现失稳临界速度、脱轨系数、减载率、倾覆系数等关键动力学参量的动态、直接测量，揭示轮轨非线性接触状态与分叉、混沌等非线性动力学行为的内在关系。该项目的研究成果可以为高速机车动力学性能检测提供更加科学、有效的技术手段，对推动高速机车非线性动力学行为的实验研究具有十分重要的理论意义和应用价值。

近年来，我国高速铁路技术发展迅速，高速机车车辆（动车组）的运行安全是高铁运营的最关键的问题之一。随着车速的提高，轮轨间作用逐渐变得剧烈，更容易引起轮轨的磨损，疲劳严重时还会出现脱轨事故。机车轮轨间复杂运动是由外界环境、轨道和机车共同作用的结果，到目前为止轮轨接触状态监测及运行安全评价一直是各国相关领域专家关注和研究的热点。.本项目所提出的轮轨动力学计算机视觉测量方法采用高速摄像机与激光源相组合的方案，将轮轨动态位移测量巧秒地转换为激光源在车轮表面斑点的图像提取，对轮轨接触的每一个参数如轮轨冲角、摇头角、轮轨位移等，从系统设计、方法原理、处理算法、实验过程、数据分析及结论等方面分别进行了系统研究，验证了轮轨接触状态参数视频检测方案的可行性及有效性。然后，基于VI-Rail软件建立了车辆动力学模型，通过仿真获取了脱轨系数和轮轨减载率等参数，为列车运行安全评价提供必要的数据支持。最后通过车辆动力学建模与仿真，给出了基于轮轨位移的列车运行安全状态评价方法。.该项目所研究的轮轨间动力学问题是影响高速机车安全、稳定运行的根本因素，它直接影响机车车辆的运动稳定性、运行平稳性和运行安全性，其研究成果可以为高速机车轮轨动力学性能检测提供更加科学、有效的的技术手段，对推动高速机车在稳定性演化和动态脱轨过程中的非线性动力学行为研究具有非常重要的理论意义和应用价值。