扭振特性约束的混合动力汽车换挡过程转速、转矩双同步机理及控制方法

Research on the improvement of gear-shifting quality for automatic transmission, especially for hybrid electric system， is becoming spotlights focused by researchers. But the driveline’s coupling mechanism constrained by the multi-power and the torsional vibration is yet not clear. In this program, the following aspects will be carried out：Firstly，the rigid-flexible coupled driveline model considering the torsional and vibrating characteristic is built to reveal the generating and propagating mechanisms of driveline’s vibration; Secondly, the least squares method is used to identify the unknown model parameters and Discrete Kalman filter is used to estimate internal transmission torque,and then a method of vibration active elimination is proposed with the target that the torque transmitted in transmission is zero.In the following, synchronizer’s meshing principle of hybrid electric system is analyzed in detail and new control concept for double synchronization of speed and torque is proposed to improve shifting quality. Then according to phase plane method, sliding mode controller characterized with multi-variable and composite sliding surfaces is designed to realize the above concept. Finally, the effectiveness of the control algorithm is verified on the multi-domain modeling platform with AMEsim and MATLAB/Simulink.

在自动变速器的研究当中，特别对于混合动力系统来说，换挡品质的提高一直是研究人员关注的热点。但考虑混合动力传动系统多动力源以及扭振特性约束的传动系统耦合机理尚未明确。本项目对此开展以下研究：首先，建立考虑扭振特性的刚柔耦合混合动力传动系统模型，揭示传动系扭转振动产生和传播机理；然后，利用最小二乘法原理对未知模型参数进行辨识并采用离散卡尔曼滤波器估计传动系统内部传递转矩，提出以传动系统内部传递零转矩为目标的扭转振动的主动抑制方法；接着对混合动力传动系统的同步器啮合机理进行详细分析并提出同步器主从动端转速、转矩双同步的控制思想来提高换挡品质，进而基于相平面法设计具有多变量复合滑面的滑模变结构控制器来保证转速、转矩双同步的实现；最后，采用依托AMEsim与MATLAB/Simulink建立的多领域建模仿真平台验证控制算法的有效性。

单轴并联式混合动力系统架构由于动力传递效率高、承载扭矩大等优点而被广泛应用于商用车。然而，单轴并联式混合动力汽车不分离离合器的换挡方式增加了变速器输入端惯量，对换挡过程造成了不利影响，从而使得其换挡控制有别于常规车辆。本项目对此开展以下研究：首先，采用机理法和试验法相结合的方式建立了单轴并联混合动力传动系统刚-柔耦合动力学模型及解耦模型；然后，设计了基于扩展卡尔曼滤波算法的动力传动系统转矩估计器，对传动系统传递转矩进行有效估计，提出以传动系统内部传递零转矩为目标的扭转振动的主动抑制方法；接着探讨了挂挡时接合套与同步环、接合齿圈可能的作用形式，设计了转速转矩双同步的换挡控制策略，确定了电机的控制律，并通过一种模糊径向基函数(Radial Basis Function, RBF)神经网络PID控制器用来改善电机的转矩跟踪性能；最后，采用依托AMEsim与MATLAB/Simulink建立的多领域建模仿真平台验证控制算法的有效性，并通过台架、实车试验证明转速转矩双同步换挡不仅能减少换挡时间，还能大幅度减小换挡冲击。本项目的开展解决了转速转矩两个耦合变量的同步控制问题，对各领域的耦合变量同步控制具有重要的参考意义。