论文部分内容阅读
为研究摆式列车倾摆控制信号预测方法,建立“动车+拖车+拖车”3辆车编组的摆式列车机电耦合系统动力学模型。建模中考虑了列车系统中存在的轮轨蠕滑力非线性、钩缓作用力非线性和悬挂力非线性。摆式列车通过安装于头车前转向架的陀螺仪在线检测曲线,对测出的横向加速度信号进行滤波和实时生成倾摆控制信号。为了补偿加速度信号的滤波延时,对倾摆控制信号的预测分别采用线性插值法和线性BP神经网络预测,并仿真研究摆式列车曲线通过性能。数值仿真结果表明:线性插值法预测和神经网络预测均能有效补偿加速度信号的滤波延时,使头车及时倾摆,大幅度降低未平衡横向加速度;在输入信号波动较大和预测时间较长时,神经网络预测效果更好;倾摆控制信号的预测方法对车辆动力性能影响不大。
In order to study the prediction method of tilting control signal of tilting train, a dynamical model of electromechanical coupling system of tilting train was set up based on “car, trailer and trailer”. In the modeling, the nonlinearity of wheel-rail creep, nonlinearity of suspension force and non-linear suspension force are considered in the train system. The tilting train detects the curve online by the gyro installed in the front bogie, filters the measured lateral acceleration signal and generates the tilting control signal in real time. In order to compensate the filter delay of the acceleration signal, the prediction of the tilting control signal is respectively predicted by linear interpolation and linear BP neural network, and the curve passing performance of the tilting train is simulated. Numerical simulation results show that linear interpolation method and neural network prediction can both effectively compensate the filtering delay of the acceleration signal and cause the head vehicle to tilt in time, greatly reducing the unbalanced lateral acceleration. When the input signal fluctuates greatly and the prediction time is long , Neural network prediction is better; prediction method of tilting control signal has little effect on vehicle dynamic performance.