摘要
随着能源问题和环境问题的日益突出，新能源汽车受到了越来越多的关注，我国也已将新能源汽车列入七大战略性新兴产业之一。轮毂电机电动汽车作为新能源汽车的重要一员，直接通过安装在轮毂内的轮毂电机进行驱动，具有机械损失低、动力响应快、可实现线控独立驱动等诸多优点，备受青睐。本文以四轮独立驱动的轮毂电机电动汽车为研究对象，主要研究在保证车辆动力性的前提下，通过驱动转矩的合理分配，达到提高整车经济性和稳定性的目标。具体内容如下：
（1）基于Matlab/Simulink和CarSim软件搭建联合仿真平台。利用CarSim软件建立轮毂电机电动汽车整车动力模型，通过Matlab/Simulink软件建立轮毂电机模型、轮毂电机矢量控制策略及驾驶员意图识别模型，用于后续控制策略的仿真验证。
（2）基于卷积神经网络进行路面条件识别。基于卷积神经网络设计路面条件识别模型，通过路面图像的有效识别，获得相应的路面附着系数，用于驱动转矩优化控制。并在此基础上利用Burckhardt轮胎模型识别出不同路面条件下的最优滑转率，用于轮毂电机电动汽车的驱动轮防滑控制。
长安星卡图片
（3）轮毂电机电动汽车驱动转矩控制策略设计。对轮毂电机进行能量消耗分析，并根据车辆行驶状态，设计不同的控制策略。在常规工况下，考虑节能和载荷转移问题，进行驱动转矩多目标协调控制，提高整车的动力性和经济性。在紧急工况下，进行驱动轮防滑控制，解决驱动轮过度滑转问题，提高整车的
行驶稳定性。
（4）轮毂电机电动汽车驱动转矩控制仿真试验验证。基于Matlab/Simulink与CarSim联合仿真平台，通过NEDC循环工况对驱动转矩优化控制策略进行试验验证，并利用轮毂电机效率工作点和整车能耗指标验证控制策略的有效性。设计高附着干沥青路面和低附着冰路面两种典型路面下的加速行驶工况，对驱动轮防滑控制策略进行试验验证，结果表明所设计的控制策略能够较好的抑制车轮滑转。
关键词：轮毂电机电动汽车卷积神经网络路面条件识别驱动转矩控制稳定性控制
Abstract
吉利gc9报价With the increasingly prominent energy and environmental issues, new energy vehicles have attracted more and more attention , and China has also listed new energy vehicles as one of the seven strategic emerging industries. As an important member of new energy vehicles, the in-wheel motor electric vehicle is directly driven by the in-wheel motor installed in the hub, which has many advantages, such as low mechanical loss, fast power response, independent drive by wire and so on. In this paper, four-wheel independently driven in-wheel motor electric vehicles as the research object, and mainly studies to achieve the goal of improving the economy and stability of the whole vehicle through the reasonable distribution of driving torque under the premise of ensuring the vehicl
e dynamics. The details are as follows:
(1) A joint simulation platform is built based on Matlab / Simulink and CarSim software. The whole vehicle dynamic model of the in-wheel motor electric vehicle is established by using CarSim software, and the in-wheel motor model, the vector control strategy of the in-wheel motor and the driver intention recognition model are established by using Matlab / Simulink software, which are used for the simulation verification of the subsequent control strategy.
(2) Pavement condition recognition based on convolutional neural network. Based on the convolution neural network, pavement condition recognition is designed. Through the effective recognition of the road image, the corresponding road adhesion coefficient is obtained, which is used for the optimal control of driving torque. On this basis, Burckhardt tire model is used to identify the optimal slip ratio under different road conditions, which can be used for the anti-skid control of the driving wheel of the in-wheel motor electric vehicle.
(3) Design of driving torque control strategy for in-wheel motor electric vehicles. The energy consumption of in-wheel motor is analyzed, and different control strategies are designed according to the driving state of the vehicle. Under normal working conditions, considering energy saving and l
oad transfer, multi-objective coordinated control of driving torque is adopted to improve the power and economy of the vehicle. Under the emergency condition, the anti-skid control of the driving wheel is carried out to solve the problem of excessive sliding of the driving wheel and improve the driving stability of the whole
vehicle.
(4) Simulation test and verification of the drive torque control of the in-wheel motor electric vehicle. Based on the Matlab / Simulink and CarSim joint simulation platform, the drive torque optimization control strategy is tested and verified through the NEDC cycle conditions, and the effectiveness of the control strategy is verified using the wheel hub motor efficiency operating point and the vehicle energy consumption index. Accelerating driving conditions on two typical roads with high adhesion dry asphalt pavement and low adhesion ice pavement were designed, and the drive wheel anti-skid control strategy was tested and verified. The results show that the designed control strategy can effectively restrain the wheel slip.
Key words:In-wheel motor electric vehicle Convolutional neural network Pavement condition recognition Driving torque control Stability control
目录
摘要................................................................ I Abstract ............................................................. II 第1章绪论.. (1)
1.1课题研究背景和意义 (1)
1.2轮毂电机电动汽车国内外发展现状 (2)
1.3轮毂电机电动汽车纵向动力学控制国内外研究现状 (7)
1.4本文主要研究内容 (10)
第2章轮毂电机电动汽车模型建立 (11)
2.1基于CarSim的车辆动力模型建立 (11)
2.1.1 CarSim软件介绍 (11)
2.1.2 CarSim车辆动力模型建立 (12)
2.2轮毂电机模型建立 (14)
2.2.1 轮毂电机数学模型建立 (15)
2.2.2 轮毂电机矢量控制策略 (17)
2.2.3 轮毂电机模型仿真与分析 (19)
2.3基于神经网络的驾驶员意图识别模型建立 (20)
2.3.1 驾驶员意图识别控制策略 (20)
2.3.2 驾驶员意图识别模型建立与仿真 (21)
2.4本章小结 (23)
第3章基于卷积神经网络的路面条件识别研究 (24)
3.1路面条件识别理论 (24)
国产c级3.2基于卷积神经网络的路面条件识别 (25)
3.2.1 数据预处理与数据集建立 (26)
3.2.2 基于卷积神经网络的路面条件识别模型建立 (27)
3.2.3 路面条件识别模型的训练与预测 (32)
3.3基于路面附着条件的最优滑转率识别 (34)
3.4本章小结 (37)
第4章轮毂电机电动汽车驱动转矩控制研究 (38)
4.1轮毂电机能耗分析 (38)
4.2轮毂电机电动汽车驱动转矩优化控制研究 (39)
4.2.1 基于能耗最优的驱动转矩优化控制 (40)桑塔纳停产
4.2.2 考虑轴荷转移的驱动转矩优化控制 (42)
4.2.3 轮毂电机驱动转矩多目标协调控制 (44)
4.3轮毂电机电动汽车驱动轮防滑控制研究 (45)
a8制造4.4本章小结 (49)
第5章轮毂电机电动汽车联合仿真与分析 (50)
5.1轮毂电机电动汽车联合仿真平台建立 (50)
5.2轮毂电机电动汽车驱动转矩优化控制仿真与分析 (51)
5.3轮毂电机电动汽车驱动轮防滑控制仿真与分析 (54)
5.3.1 高附着路面仿真与分析 (54)
5.3.2 低附着路面仿真与分析 (55)
5.4本章小结 (56)
第6章总结与展望 (57)
6.1总结 (57)
6.2展望 (57)
宇通
致谢 (58)
参考文献 (59)
作者简介 (63)
攻读硕士学位期间研究成果 (64)