插电式混合动力汽车建模仿真及离合器控制技术研究

发布时间：2018-10-13 11:56

【摘要】：随着全球气候变暖,环境污染加重,同时伴随着能源危机和能源安全的压力,以石油为主要能源的传统汽车产业面临着前所未有的挑战,开发清洁、高效、智能的新型汽车已迫在眉睫。插电式混合动力不需要构建专用充电设施,便于推广应用,与纯电动相比行驶里程大大增加,有效地解决了纯电动汽车续航里程不足的问题;与传统汽车相比,有效地减少了尾气排放造成的环境污染。论文主要研究工作如下:1.为了确保车辆仿真的精确性,本文根据汽车重量、迎风面积、轮胎尺寸等参数和整车的性能指标,结合汽车理论、汽车设计、车辆动力学等有关知识,为整车设计匹配了合适的电机、动力电池、发动机和速比等参数。2.本文选择Matlab/Simulink平台建立发动机、电机、动力电池、驾驶员、整车控制器等模型,然后按照P1+P4的混合动力系统结构搭建成整车仿真模型。采用这个平台进行建模仿真既可以对整车控制器内的控制策略模型进行MIL测试和验证,又可通过代码生成技术与底层驱动程序直接集成,形成实际控制软件。与传统C代码开发技术相比,这种技术手段在很大程度上提高了开发效率,降低了研究的成本。3.针对离合器模型,本文分别采用了PID控制和自适应模糊神经网络控制对其进行调节。PID控制的使用非常广泛,但是在复杂且非线性的离合器系统中,它的控制效果受到了很大的影响,因此引入了自适应模糊神经网络对其进行调节。仿真表明,采用自适应模糊神经网络控制的离合器性能比PID控制有了很大的提升,接合时间缩短了27%,滑磨功减小了15.8%,冲击度减小了51.4%,这将使车辆的性能和驾驶舒适性有极大的改善。4.为了使仿真的操作更加便捷高效,仿真的结果更加清晰易读,本文通过Matlab平台设计了GUI图形用户界面,该软件综合了车型选择、参数修改、模型更正、工况设定等诸多功能,可以很好的与模型和仿真兼容。此外为了便于车辆控制器的实车测试,通过Lab VIEW设计了基于车辆CAN网络的监控软件,实现了车辆信息和故障状态的实时采集、解析和保存,并可根据需求向指定节点发送报文。
[Abstract]:With the global warming and environmental pollution increasing, along with the energy crisis and the pressure of energy security, the traditional automobile industry with oil as the main energy source is facing unprecedented challenges. Intelligent new cars are imminent. The plug-in hybrid electric power does not need to build special charging facilities, which is convenient for popularization and application. Compared with the pure electric vehicle, the driving mileage is greatly increased, which effectively solves the problem of the shortage of the pure electric vehicle's endurance mileage. It can effectively reduce the environmental pollution caused by tail gas emission. The main research work is as follows: 1. In order to ensure the accuracy of vehicle simulation, according to the parameters of vehicle weight, upwind area, tire size and the performance index of the whole vehicle, this paper combines the relevant knowledge of automobile theory, automobile design, vehicle dynamics and so on. Suitable parameters such as motor, power battery, engine and speed ratio are matched for the whole vehicle design. 2. In this paper, the engine, motor, power battery, driver and vehicle controller are built on Matlab/Simulink platform, and then the vehicle simulation model is built according to the structure of P1P4 hybrid power system. This platform can not only test and verify the control strategy model in the whole vehicle controller by MIL, but also directly integrate the code generation technology with the underlying driver to form the actual control software. Compared with the traditional C code development technology, this technique improves the development efficiency and reduces the research cost to a great extent. In this paper, PID control and adaptive fuzzy neural network control are used to adjust the clutch model. The PID control is widely used, but in the complex and nonlinear clutch system, Its control effect is greatly affected, so adaptive fuzzy neural network is introduced to adjust it. Simulation results show that the performance of clutch with adaptive fuzzy neural network control is much better than that of PID control. The bonding time is reduced by 27%, the sliding power is reduced by 15.8 and the impact degree is reduced by 51.4, which will greatly improve the performance and driving comfort of the vehicle. 4. In order to make the operation of simulation more convenient and efficient, and the results of simulation more clear and easy to read, this paper designs the GUI graphical user interface through Matlab platform. The software integrates many functions, such as model selection, parameter modification, model correction, working condition setting and so on. Good compatibility with model and simulation. In addition, in order to facilitate the vehicle controller test, a monitoring software based on the vehicle CAN network is designed by Lab VIEW, which can collect, analyze and save the vehicle information and fault state in real time, and can send the message to the designated node according to the requirement.
【学位授予单位】：浙江工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：U469.7;U463.211

【参考文献】