【作者】 刘长亮；

【导师】 杨兆军；

【作者基本信息】 吉林大学 ， 机械制造及其自动化， 2011， 硕士

【摘要】 鼓式制动器是一个柔性多体系统,各个零件在制动过程中的运动规律和受力情况比较复杂,通过建立刚柔耦合模型可较真实得仿真制动过程,得到相对准确的动力学分析结果,为制动性能的预测与优化提供基础。目前,在建立鼓式制动器刚柔耦合模型进行动力学分析并进行基于仿真的优化研究上还未见成熟报道。因此,本文在借鉴国内外相关研究成果的基础上,进行了两方面的研究：探索了建立鼓式制动器刚柔耦合模型的通用方法,采用参数化设计方法开发了界面友好的建模与仿真平台,以E260鼓式制动器为实例进行了动力学仿真分析；然后,引入多岛遗传算法与代理模型技术,按照“试验设计-获取样本点-构建代理模型-误差分析-建立优化数学模型-选择优化算法-进行优化设计-验证优化结果”的路线,对E260鼓式制动器进行了基于动力学仿真的制动性能优化,加快了优化进程。主要研究如下：(1)探索了建立鼓式制动器刚柔耦合模型的通用方法。忽略一次制动中温升对制动效能的影响,提出了基于鼓式制动器刚柔耦合模型进行动力学仿真的方法。研究了有限元模态中性文件的生成与导入方法；利用迭代法建立了渐开线凸轮模型；建立了摩擦片与制动鼓之间基于哑物体的接触力；通过能量守恒推导了整车的等效转动惯量公式,建立了“四分之一”整车模型；然后在MSC.ADAMS中开发了鼓式制动器刚柔耦合模型建立与动力学仿真平台,解决了建模耗时、耗力的问题。(2)建立了E260鼓式制动器的刚柔耦合模型,并对满载时初速度60 km/h的制动工况进行了动力学仿真。结果显示：持续制动阶段制动力矩为19574.53 Nm；制动减速度5.24 m/s2,制动距离为27.89m,满足国标要求；接触力作用线与两滚轮中心连线夹角16.70。,可通过优化结构减小此夹角来提高制动效能。(3)构建了预测制动力矩的代理模型。选取滚轮中心坐标A、滚轮中心坐标P、滚轮半径、内盖板宽度的一半、摩擦片包角和摩擦片起始角为六个设计变量,选制动力矩为响应,采用拉丁超立方试验设计方法安排30个样本点,用平台计算响应后构建了二次多项式响应面模型和径向基函数模型。然后用复相关系数法和均方根误差法进行误差分析,结果表明,径向基函数模型可较好得预测制动力矩。(4)进行了基于多岛遗传算法与代理模型的制动性能优化并进行仿真验证。引入多岛遗传算法,利用建立好的径向基函数代理模型,以制动力矩最大为目标对E260鼓式制动器的六个设计变量进行了优化,优化后制动力矩提高25.60%。把修正后结果代入平台进行建模与仿真的结果表明：优化后持续制动阶段制动力矩为24698.15 Nm,制动力矩的波动程度略有增大；制动减速度增加到6.74 m/s2；制动距离为22.34 m；接触力作用线与滚轮中心连线的夹角减小了4.76°,凸轮对制动蹄的张开效果更好,制动效能更高。更多还原

【Abstract】 Drum brake is a flexible multi-body system, and the situations of the movement and force of varoius parts are complicated during the braking process. More realistic simulation of the braking process can be achieved and the relatively accurate dynamic analysis results can be also obtained through the establishment of rigid-flexible coupling model of drum brake, which provide the basis for the prediction and optimization of braking performance. Currently, mature research of rigid-flexible coupling model establishing of drum brake for dynamic analysis and simulation-based optimizing of drum brake has not been reported. Therefore, two aspects of work were completed based on referring the related research achievements home and abroad in this thesis:first, the general methods of building rigid-flexible coupling model of drum brake were explored, and a user-friendly modeling platform was developed using parametric design method, and the E260 drum brake was taked as an actual example to do dynamic simulation and analysis; second, metamodling and multi-island genetic algorithms (MIGA) were introduced to optimize braking performance of E260 drum brake based on dynamic simulation according the line of "Design of experiment (DOE)-Access to samples-Building surrogate model-Verification of error-Establishment of optimal mathematical model-Selection of optimization algorithm-Realization of optimization and design-Verification of results", which greatly accelerated the optimization process. The main research is as follows:(1) The general methods of establishing rigid-flexible coupling model of drum brake were explored. Ignoring the effect of the rise of temperature during a braking peocess on the braking efficiency, the methods of building a rigid-flexible coupling model of drum brake were presented. The generation and import methods of finite element modal neutral file (MNF) were researched. The involute cam model was built by iteration method. The contact force model between the friction lining and brake drum was established based on dummy parts. The equivalent inertia formula of the vehicle was derived according energy conservation, and the "quarter" vehicle model was established. Then a user-friendly modeling and simulation platform was developed in MSC.ADAMS, solving the time-consuming and labor-intensive problems of modeling.(2) The rigid-flexible coupling model of E260 drum brake was built and the dynamic simulation was done under the braking conditions of full loaded at the initial braking speed of 60 km/h. The results showed that:the brake torque during the continuous braking stage was 19574.53 Nm; the braking deceleration was 5.24 m/s2 and the braking distance was 27.89 m, which met the national standard; the angle between the contact force action line and the connection line of two roller centers was 16.70°, which could be reduced by optimizing the structure.(3) The surrogate model which predicted the brake torque was constructed. The roller center coordinates A and P, the roller radius, half the width of the internal cover, the friction lining wrap angle, the friction lining initial angle were selected as six design variables, and braking torqe was selected as response.30 sample points were arranged using Latin hypercube (LH) experimental design, and after the responses being calculated using the platform the quadratic polynomial response surface model (RSM) and radial basis function (RBF) model were constructed. The error analysis of multiple correlation coefficient and root mean square error (RMSE) method shows that RBF surrogate model can predict the brake torque better.(4) The braking performance was optimized based on multi-island genetic algorithm (MIGA) and surrogate model, and the results were verified by simulation. Introducing MIGA and metamodling, taking the maximum brake torque as the optimization object, the six design variables of the E260 drum brake were optimized using well-established RBF surrogate model. After the optimization the brake torque was improved by 25.60%. The results were modified and corresponding models were established on the platform, the simulation results showed that:the optimized brake torque during the continuous braking stage was 24698.15 Nm, but the fluctuation degree of the brake torque increased little; the braking deceleration was 6.74 m/s2; the braking distance was 22.34 m; the angle between the contact force action line and the connection line of two roller centers was decreased by 4.76°, and the cam could better push the brake shoes resulting in higher braking performance.更多还原