基于转向架姿态轨迹的轨道线路曲率探测系统研究

Research on Railway Track Curvature Detection System Based on Bogie Attitude Trajectory

主动径向技术是进一步提高轨道车辆曲线通过性能、彻底解决直线运行稳定性和曲线通过之间矛盾的有效途径,而实现轨道线路曲率准确、高效的探测是其中的重要环节。为了提高轨道线路曲率探测的精度和实时性,首先提出了一种基于转向架姿态轨迹的曲率探测系统(Bogie attitude trajectory based curvature detection system,BATCDS),该系统由布置于转向架上的速度传感器、角速度传感器、倾角传感器以及算法单元组成,布局简单且易于实现。此后,提出了与上述系统相匹配的曲率探测算法,该算法利用转向架偏航角速度ω和车速v获取转向架在二维水平面上的运行轨迹,并据此初步估计线路曲率;利用转向架侧滚角α和车速v获取转向架的侧滚姿态轨迹,进而估计转向架侧滚角;综合转向架上述各项姿态信息,利用轨道线路固有几何规律融合计算线路曲率。最后,建立了BATCDS的联合仿真模型,仿真计算的结果表明,该系统在有效滤除原始信号中的高频噪声的基础上,仍能维持较高的实时性;仿真工况下,相较传统的低通滤波而言曲率探测结果的精度和实时性得到显著改善,相对误差率均降低至传统方法的50%以下,最大降幅可达65%。

Active steering technology is an effective approach to improve the curving performance of railway vehicles,and can completely solve the contradiction between straight-running stability and curve negotiation.Accurate and efficient detection of railway track curvature is an important part of active steering technology.In order to improve the accuracy and real-time performance of railway track curvature detection,a curvature detection system based on the bogie attitude trajectory(BATCDS)is proposed.This system is composed of a velocity sensor,an angular velocity sensor,an inclination sensor arranged on the bogie,and an algorithm unit,with an easy-to-implement layout.A curvature detection algorithm that matches the BATCDS system is further proposed.The bogie yaw angular velocity(ω)and the vehicle speed(v)are used to compute the running trajectory of the bogie on the two-dimensional horizontal plane,based on which the track curvature is preliminarily estimated.The bogie rolling angle(α)and v are used to obtain the rolling attitude trajectory of the bogie,which is used to further estimate the rolling angle of the bogie.By using the inherent geometric law of the railway track,the above-mentioned attitude information of the bogie is fused to calculate the line curvature.Finally,the co-simulation model of BATCDS is established.The simulation results show that this system can maintain high real-time performance and effectively filter the high-frequency noise in the original signal in the meanwhile.The accuracy and real-time performance of the curvature detection are significantly improved,with a more than 50%(up to 65%)reduction of relative error rate compared to the traditional low-pass filter.