【作者】 景立新；

【导师】 郭孔辉；

【作者基本信息】 吉林大学 ， 车辆工程， 2011， 博士

【摘要】 悬架系统的研究是车辆底盘设计开发的重要内容,是汽车自主研发的重要核心技术。在悬架开发过程中,设计人员需要做的就是如何匹配悬架以满足人们对车辆性能“好”的要求。这包括两方面研究内容:首先需要研究什么样的车辆性能可以称之为“好”;然后再研究如何设计悬架以实现车辆性能的“好”。第一方面是车辆性能的评价问题,车辆性能包括:操纵稳定性能、平顺性能及NVH性能,本文主要考察操纵稳定性能。操纵稳定性评价可分为主观评价及客观评价,主观评价是对真实样车的感受评价,无法在开发初期为悬架设计提供帮助;客观评价则是通过测试与车辆性能相关的客观量并与相应的标准比较进行评量的方法,可借助虚拟样机技术在开发初期完成悬架设计匹配,本文采用客观评价来完成车辆性能优化。同时,“好”还包含另外一种含义:稳健性。由于车辆使用工况包含众多不确定性因素(载荷状态变化、行驶速度变化及路面摩擦系数变化等),基于单一确定工况优化得到的悬架结构在其它工况下可能表现很差,这就需要在设计时充分考虑不确定因素的影响,利用稳健性设计使车辆操稳性能在各种工况下均具有较好特性,为此本文采用了田口稳健性设计及双响应面稳健性设计。第二方面是悬架结构设计问题。KnC是悬架运动学及弹性运动学的简称,综合体现了悬架的稳态特性。悬架KnC具有多目标特性,各目标相互影响,互相耦合,有些甚至是相互矛盾,很难找到能使各目标均最优的解,本文引用多目标优化问题中的Pareto解集概念,得到一系列优化结果,并由决策者折衷选择最终解。在悬架结构设计阶段同样需要考虑稳健性问题,此时不确定性主要来自于悬架硬点位置的偏差及衬套刚度误差,这些因素将导致悬架性能的不一致。利用稳健性设计可使悬架性能指标在不确定性因素变化的情况下较好,且波动最小,最稳定,为此本文采用了多目标稳健性设计。悬架结构设计目标(即KnC特性)与车辆操纵稳定性能之间是通过基于特性的整车模型(CarSim)来联系的,由整车模型对车辆的操纵稳定性进行稳健性优化,得到悬架的目标特性;再由基于结构的悬架模型(Adams)来稳健性优化悬架的硬点位置及衬套刚度。本文主要内容包括:1.介绍了悬架KnC特性概念、台架试验及虚拟仿真方法,并对其进行了定性分析;2.根据参考车型的悬架及整车参数,对其进行基于结构的悬架系统建模和基于特性的整车建模,利用实车悬架KnC特性试验及整车操纵稳定性试验数据,对悬架模型及整车模型的建模精度做了验证,为后续悬架系统特性的优化匹配做了模型上的准备;3.介绍稳健性设计理论及主要研究方法;4.利用基于特性的整车模型稳健性优化悬架KnC目标特性:优化中以车辆的使用工况为不确定性因素,以车辆操纵稳定性客观评价指标为优化目标,采用了田口稳健性设计及双响应面稳健性设计;5.利用基于结构的悬架模型稳健性设计悬架硬点位置及衬套刚度:优化中以悬架硬点位置偏差及衬套刚度误差为不确定性因素,以KnC特性与设计要求一致为优化目标,并引入蒙特卡罗采样方法计算不确定性因素引起的均值及方差,采用了基于改进NSGAⅡ遗传算法的多目标稳健性设计;6.由于优化参数多且每次仿真计算耗时使得优化过程效率较低,为此提出用近似模型来代替原有的真实模型,并分别对整车模型及悬架模型进行近似模拟,在保证精度的前提下大大提高了优化效率。本文主要创新点如下:1.对悬架KnC特性做了深入分析,特别是扭转梁式半独立悬架设计开发中的关键因素研究,总结得到横梁纵向位置、横梁开口方向及衬套安装角度对悬架KnC特性的影响规律,为扭转梁式半独立悬架设计开发及后期调校提供参考。2.将稳健性设计引入到悬架系统开发流程:首先针对本文的车辆操纵稳定性优化为单目标(闭环操纵稳定性综合评价指标)优化问题,利用田口方法及双响应面法以信噪比最大稳健性优化悬架KnC目标特性,优化中充分考虑了不同使用工况对车辆操纵稳定性的影响;之后针对悬架结构优化为多目标问题,利用多目标优化方法以KnC特性均值与目标值偏差最小及离散度最小稳健性优化悬架硬点坐标及衬套刚度,尽量降低加工制造误差对悬架性能影响。3.鉴于稳健性优化计算量大,采用复杂模型优化效率低下,应用近似模型代替真实复杂模型进行稳健性优化计算,在保证精度的同时大大提高优化效率。更多还原

【Abstract】 Suspension system research is an important part of the vehicle chassis design and development, and also is the core technology of the automobile self-developed.In suspension development process, designers hope the vehicle have good performance. This includes two studies: first, we need to study what kind of vehicle performance is "good"; and then we should study how to design the suspension structure to achieve the "good" performance.The first aspect is about the evaluation of vehicle performance. Vehicle performance, including: handling stability, ride comfort and NVH, this paper mainly focuses on the handing stability. Handling stability evaluation can be subjective and objective. Subjective evaluation is the feeling of the real vehicle, and it can’t provide help in the early days of the suspension design; Objective evaluation is testing the objective parameters which are relative to the vehicle performance and comparing to the relevant standards. Objective evaluation can be used in virtual prototyping technology in the early days of the design. In this paper, objective evaluation is used to complete the optimization of the vehicle performance. Meanwhile, the "good" also includes other meanings: robustness. As the working conditions include uncertainties (load state changes, speed changes and the changes in road surface friction coefficient, etc), the suspension structure which is optimized based on a single condition may have poor performance in other conditions. Including full consideration of the uncertain factors in the design, the robust design has good performance in most conditions.This paper used the Taguchi robust design and the dual response surface robust design to optimize the vehicle handing stability.The second aspect is about the suspension structure design. KnC means the kinematics and compliance, and is the steady-state characteristics of the suspension system. KnC includes multi-objective characteristics which may influence each other, and even conflict with each other, this make it difficult to find a solution which is suitable for all characteristics. This paper used Pareto solution, and got a series of optimization results, from which the final solution was selected by a decision-makers. In the same time, the robust design also needs to be considered. Here, we mainly study the uncertainty of the hard points and the bushing stiffness, which can lead to inconsistent performance of the suspension. Using multi-objective robust design can help us get good suspension performance even there are uncertain factors.Suspension characteristics (KnC) were associated with vehicle handling stability through CarSim model,and by optimizing vehicle handling stability,we can get the KnC targets; Suspension hard points and bushing stiffness were associated with Suspension characteristics through Adams model, and by Using multi-objective robust design, we can get hard points and bushing stiffness.The main content of this paper:1. The introduction of the concept of suspension KnC,laboratory tests,virtual simulation and the qualitative analysis of KnC;2. Using the benchmarking vehicle parameters to build the structure-based suspension model and the property-based vehicle model, and these models were verified by KnC laboratory test and vehicle handling stability experimental test,which laid the foundation of future work;3. The introduction of the robust design theory and research methods;4. Robust optimization of suspension KnC characteristics based on property-based vehicle model: vehicle handling stability was taken as the objective, working conditions was taken as the uncertainties, and Taguchi robust design and dual response surface robust design were taken as optimization method.5. Robust optimization of hard points and bushing stiffness based on structure-based suspension model: Being consistent with the KnC targets was taken as the objective, hard points error and bushing stiffness error were taken as the uncertainties, Monte Carlo was used to calculate the mean and variance, and NSGAⅡwas used in the multi-objective robust design.6. Because there are too many parameters and the calculation is time-consuming, approximate model was used to replace the real model, which greatly improved the efficiency of the optimization without reducing the accuracy.Major Innovations of the Dissertation:(1) In-depth study on suspension characteristics (KnC), especially the study on semi-independent torsion beam suspension;(2) The robust design was introduced into Suspension system development process;(3) The approximate model instead of the real model was applied to the robust design.更多还原