【作者】 张永林；

【导师】 钟毅芳；

【作者基本信息】 华中科技大学 ， 机械设计及理论， 2010， 博士

【摘要】 论文从标准等级道路的频域模型——随机道路功率谱密度函数(PSD)和非标准道路的实测样本数据出发,针对现代车辆具有的非线性系统特征,重构了用于车辆时程分析所需的道路激励时域模型。论文从随机道路基本描述,道路时域建模的基本理论与方法,车辆单点激励的时域模型,车辆单轮辙多点激励的时域模型,车辆相关双轮辙激励的时域模型,任意道路(标准和非标准)的时间序列AR和ARMA模型,2维路面的时域模型,非平稳随机道路模型及其小波分析等方面进行了较为全面、系统而深入的研究,并应用所提方法实现了对车辆道路和路面的仿真。所建立的道路和路面时域模型成为道路或路面频域模型的等价模型,成为道路或路面数据库建构的基础,成为虚拟样车技术和虚拟试车场技术所必需的环境模型。论文首先介绍了随机道路的基本描述方法——道路频域模型,给出了道路频域模型——随机道路功率谱密度函数的2种典型表达式,指出了道路频域模型在车辆动力学分析中的作用。基于对车辆道路随机不平顺性态的分析,结合其在频域内的PSD统计表示方法,以载重车六轮激励输入为例,导出了多点激励的道路高程的功率谱密度函数矩阵,该矩阵可用于线性或线性化的车辆动力学分析中。现代车辆系统的非线性特征规定了其动力学分析必须置于时域进行。论文描述了道路时域模型建构的4种基本方法：谐波叠加法、白噪声过滤法、基于逆Fourier变换的方法和时间序列建模方法。并采用前3种方法对标准等级道路进行了时域建模,尤其是,应用3种方法实现了对于时空相关的双轮辙道路激励输入的建模,其中,基于逆Fourier变换方法被首次应用于道路建模中。在建模过程中,针对伪白噪声生成算法中不精确的缺陷,改进了伪白噪声的生成算法,提高了仿真道路的精度。除了常规的由PSD表示的标准等级道路外,道路时域建模的内涵可以向2个方向拓展,一是向建立任意道路(包括非平稳道路)的时域模型方向,一是向建立2维路面时域模型方向。前者丰富了车辆动力学分析所用的激励模型素材,后者更直接服务于虚拟试车场技术。利用时间序列模型方法,以标准等级道路的时序模型为对照测试,建立了任意非标准实测道路的时域模型,并实现了非标准道路的时域仿真。在路面建模方面,利用谐波叠加法、逆FFT方法和白噪声过滤法建立了与标准道路频域统计特征一致的道路路面模型；进一步,基于任意子样测量数据,用时间序列方法建立了任意道路路面模型,实现了路面仿真。非平稳道路路面是对车辆进行深层次动力学分析所必需的基础激励模型。对非平稳道路,既能够基于平稳道路模型数据方便快捷地实现其非平稳化重构,而且对于所重构的非平稳道路,采用小波分析,把道路激励信号中的平稳成分与非平稳成分、低频成分与高频成分分开,从而揭示了非平稳道路激励的特征,同时,非平稳随机道路小波分析也涵盖了车辆变速行驶、启动、制动等工况下道路激励的特性。通过全面深入的道路及路面建模研究,可以方便地建成随机道路及路面的时域模型库,该模型库与其他确定性道路及路面模型函数库一起,在虚拟样车及虚拟试车场技术中可实现任意的集成、组装,为车辆数字化设计、试验提供实时道路及路面激励数据。更多还原

【Abstract】 From the frequency domain model, the power spectrum density (PSD) function, of the standard grade roads and the measuring sample data of the nonstandard road, the time domain model of roads or road surfaces was regenerated to meet the requirement on the time history analysis of the advanced nonlinear vehicle system in the paper. The complete, systemic and deep study was made on several aspects including the basic description of stochastic road, the time domain modeling theories and methods of roads, the time domain model of one input excitation and multi-point excitation from road to vehicles, the domain model of the temporal-spatial correlation from the bilateral correlative track excitation of a road, the time series AR or ARMA model of arbitrary roads, the time domain model of 2 dimensional road surface, and the wavelet analysis on the nonstationary stochastic road process. The numerical road simulation was realized by using the time domain methods. The simulated road or road surface in time domain, equal to the model in frequency domain, was established as the basic road data and used as enviroment model in the virtual prototyping of vehicles (VP) and the virtual proving ground of vehicles (VPG).Firstly, the frequency domain model was introduced as a fundamental description to the stochastic roads. Two different expressions of road PSD were presented and their applications to the vehicle dynamics illustrated. Based on the property of the road random irregularities and combining with the expression of PSD in statistics, taking a six-wheel vehicle as an example, a matrix of PSD function with respect to multi-point excitation from road irregularities was deduced for the puopose of the analysis on a linear or linearized vehicle system.Then, in terms of the requirements of dynamics analysis in time domain on the nonlinear vehicle system, the basic pinciples of the four methods, harmony supperposition method, white noise filtration method, inverse fast Fourier transform (IFFT) method and the time series AR or ARMA modeling method, available for time domain modeling to road irregularities were illustrated. The first three methods were used to model the standard grade road and the domain model of the temporal-spatial correlation from the bilateral correlative track excitation of a road reached. Among these, the IFFT method was firstly successfully implemented in time domain model of vehicle road excitation. Morever, aiming at the low accuracy of psuedo-white noise generation, a improved algorithm to generat quality white noise was designed to rise the accuracy of the simulated roads. Followed was the two developments, one orienting to the modeling of an arbitrary road besides the standard grade road and another orienting to the modeling of 2 dimensional road surface. The former made the model data of vehicle dynamics richer and the latter provided direct model data for VPG.Using time series AR or ARMA modeling method, through the model test in time domain to the standard grade road, the time series AR or ARMA model of an arbitrary nonstandard measuring road was established and the simulation realization implemented in time domain.Based on the development of time domain model of road irregularities, the road surface model in time domain was constructed with the consistence to the frequency domain statistics of the standard grade road using harmony supperposition method, IFFT method and white noise filtration method respectively. Morever, based on the sample data from arbitrary measuring road, the road surface model of arbitrary road was established in time domain and the simulation realized.Finally, the model of nonstationary road was necessary for the deeper dynamics of vehicles. The numerical rengeneration of the nonstationary road was expediently reached based on the data from the numerical models of the stationary road or surface. In turn, the regenerated nonstationary road was analyzed using wavelet method in order to separate the nonstationary signal from the stationary signal and the high frequency signal from the low frequency signal. By doing so, the excitation feature of the nonstationary road or surface was revealed. On the other hand, nonstationary analysis on road was applied to study the excitation characteristics of roads under the condition of speed changing, starting and stoping of vehicles.In a word, through the complete and deep study on the time domain model of stochastic road or road surface, a data-base about the time domain model of stochastic road and surface was expedientlly constructed. In combining with some deterministic functions of roads or road surfaces, the excitation data from road irregularities were obtained in real time by means of integration, assembly or supperposition, according to the requirements of VP and VPG technologies.更多还原