基于模型和数据驱动的智能轮胎载荷估计  

Model-based and data-driven load estimation for smart tires

作　　者：胡梦丹 李波[1,2] 贝绍轶[1] 林棻[3] 殷国栋[4] HU Mengdan;LI Bo;BEI Shaoyi;LIN Fen;YIN Guodong(School of Automobile and Transportation,Jiangsu University of Technology,Changzhou 213001,China;Suzhou Automobile Research Institute,Tsinghua University,Suzhou 215200,China;College of Energy and Power Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 210016;School of Mechanical Engineering,Southeast University,Nanjing 210096,China)

机构地区：[1]江苏理工学院汽车与交通工程学院,江苏常州213001 [2]清华大学苏州汽车研究院,江苏苏州215200 [3]南京航空航天大学能源与动力学院,南京210016 [4]东南大学机械工程学院,南京210096

出　　处：《重庆理工大学学报(自然科学)》2025年第2期55-64,共10页Journal of Chongqing University of Technology:Natural Science

基　　金：国家自然科学基金项目(52172367,51705220);江苏省高校自然科学基金重大项目(21KJA580001);常州市国际科技合作基金项目(cz20220031)。

摘　　要：轮胎载荷是车辆稳定性控制以及车辆参数估计的关键,提出一种结合物理模型和数据驱动的轮胎载荷估计算法。采用ABAQUS建立轮胎模型,搭建试验台架对模型进行验证;根据柔性环模型研究轮胎载荷与中心角的内在机理,提取轮胎不同位置的应变信号,对不同载荷下的应变进行响应分析,采用皮尔逊相关系数法选取特征点;基于中心角等模型参数和应变数据,通过BP神经网络搭建轮胎载荷预测模型。结果表明,结合应变信号与柔性环模型参数的预测效果比单一使用应变信号或物理模型的效果更好,该方法增加了数据驱动的可解释性,可为研究车辆的稳定性控制提供参考。Tire load is the key to vehicle stability control as well as parameter estimation.We propose a tire load estimation algorithm that integrates physical and data-driven models.First,ABAQUS is employed to build a tire model,and a test rig is built to verify the model.Then,the intrinsic mechanism of tire load and center angle is studied based on the flexible ring model.The strain signals at different positions of the tire are extracted,the strain curves under different loads are subjected to response analysis,and the Pearson correlation coefficient method is used to select the characteristic points.Finally,based on the center angle and other model parameters and strain data,the tire load prediction model is constructed by BP neural network.Our results show the prediction effect of combining the strain signal and the parameters of the flexible ring model is superior to the single use of the strain signal or the physical model.Moreover,the method increases the data-driven interpretability,providing some references for the study of the stability control of vehicles.

关 键 词：智能轮胎 应变 轮胎载荷 物理模型 BP神经网络 

分 类 号：U463.341[机械工程—车辆工程]