【作者】 王朝庆；

【导师】 申铁龙； 季海波；

【作者基本信息】 中国科学技术大学 ， 控制理论与控制工程， 2013， 博士

【摘要】 所谓制动就是利用制动装置使车辆减慢或者停止的过程。就混合动力汽车而言,制动装置主要分为两种：传统的机械式和能量再生式。前者是利用摩擦原理将反相制动力矩直接施加在车轮上,后者是电机产生的反相制动力矩通过传动系统最终传递到半轴上。在制动过程中如何合理地操作两个制动系统,是车辆安全稳定运行的前提。也是目前混合动力汽车研究的关键内容之一。本文以并联混合电动卡车为主要研究对象,从分析再生制动模式为出发点,针对带有换挡操作的制动过程,鉴于以往控制策略的不足,着力改善混合电动卡车的制动性能,在深入分析各部件的特性基础上,提出适合本课题特定需求的控制方案和策略。集中表现为在制动过程中满足车辆安全性及驾驶员舒适性要求前提下同时尽可能多的回收能量。本文研究工作如下：(一)根据混合电动卡车制动控制问题的要求,主要对混合动力汽车的传统机械制动系统、再生制动系统(电机、超级电容、逆变器)传动系(电机减速器、手自一体化机械变速箱、主减速器、传动轴)和车轮建立数学模型。为混合动力汽车制动力矩控制和模拟仿真提供基础。(二)根据制动系统模型设计了电机和传统机械制动系统的力矩匹配协调控制策略,并从理论上对稳定性加以证明。具体是将制动过程分别处理,划分为制动无换挡和制动有换挡阶段两个阶段。在前一个阶段,分别对制动力矩在可测与不可测条件下,设计近似比例-积分的反馈控制器,并对其稳定性做了证明。在后一个阶段,分别采用两种控制策略：电机提前补偿策略和电机同步器、同步器齿轮压力同时补偿策略。从而保证了车辆换挡的顺利完成并减少在换挡过程中制动力矩的波动。(三)根据超级电容等价模型对实际电容两端电压利用观测器进行估计。在模型参数已知的条件下,运用类似Luenberger观测器估计实际电容端电压,并分析了估计结果关于观测器模型参数的敏感性以及观测器关于标称模型参数变化的敏感性。在模型参数未知的情况下,运用带有遗忘因子的最小二乘离线辨识方法来估计模型的参数,从而我们可以利用被辨识的参数作为上述观测器模型参数估计实际电容端电压。考虑到实际过程中扰动的影响,引入高斯白噪声作为状态过程扰动和输出测量扰动,并根据滤波理论设计了拓展的Kalman滤波器并对模型参数和实际电容端电压同时进行估计。更多还原

【Abstract】 The process that using brake devices makes the vehicle slow down or stop is called braking. As far as hybrid electric vehicles, the brake system consists of two kinds:the traditional mechanical and energy regenerative brake system. The torque generated by the former can be acted on wheels, directly; Torque generated by the latter is transmitted to wheels by transmission. During braking, how to distribute properly brake torque between the two devices is premise of stability of vehicle and also one of the important content of study for hybrid electric vehicle.In this paper, parallel hybrid electric trucks can be considered as the main object of the research. Consider the analysis of the regenerative brake mode as starting point, point to brake process with gear shifting, find shortage of preceding control scheme, make efforts to improve braking performance of hybrid electric trucks, analyze deeply the properties of brake devices, and present control schemes and strategies for the sub-ject. The research problems embody retrieving the kinetic energy of vehicle as much as possible satisfying vehicle safety and the comfort requirement of driver. The main research work is as follows:(1) According to requirements of brake control of hybrid electric truck, the pa-per will give the mathematical model of the conventional mechanical brake system, the regenerative brake system (electrical, super capacitors, inverters) driveline (motor re-ducer, automated manual mechanical gearbox, main gear, drive shaft) and wheels. This provides the basis for brake torque control and simulation for hybrid electric truck.(2) Based on the brake system model, the coordinated torque control scheme is p-resented between the conventional mechanical brake system and the regenerative brake system, and the stability of the closed system is proved. The braking process can be divided into brake with no gear shifting and brake with gear shifting. In the former stage, approximate proportional-integral feedback controller is designed and stability is proved under the condition of the measurable brake torque and unmeasurable brake torque. In the latter stage, the two strategies are presented:motor advanced compensa-tion strategy and motor-synchronizer gears pressure compensation strategy, which en-sures successful completion of gear shifting and reduces the vibration of brake torque.(3) Based on equivalent model of super-capacitor, observer to estimate voltage of real capacitor is presented. Under the condition of known model parameter, the actual capacitor voltage is estimated by using observer similar to Luenberger type, and the estimated results is analyzed on sensitivity of observer parameters and the model param-eters. Under the condition of unknown model parameter, the least squares identification method with forgetting factor is used to estimate the real model parameters, and then the estimated parameters can be used as parameters of above observer to estimate real capacitor terminal voltage. With the consideration of disturbance effect in real process, Gaussian white noise is considered as real state disturbance and output measurement disturbance. The extended Kalman filter is designed to estimate simultaneously mod-el parameters and real capacitor terminal voltage.更多还原