【作者】 张珂；

【导师】 范秋霞；

【作者基本信息】 山西大学 ， 控制科学与工程， 2024， 硕士

【摘要】 近年来人们对汽车从最初的代步需求转变上升为对舒适性的追求。悬架系统作为汽车的减振装置,对人们驾乘舒适性有着重要作用。被动悬架成本较低,应用广泛,但因自身固定的弹性刚度和阻尼系数,导致其在复杂道路或者车速变化等工况下的适应性较差。因此,主动悬架已然成为各大车企抢占市场的必备因素。为解决主动悬架系统线性二次型输出调节器(Linear Quadratic Regulator,LQR)权重系数矩阵的整定问题,本文提出一种基于多策略增强型蛇优化算法(multi strategy Advanced Snake Optimizer,ASO)的LQR控制策略即ASO-LQR控制策略。首先以二自由度四分之一、四自由度半车这两个主动悬架为研究对象,建立其动力学模型并转化为状态空间表达式,作为后续线性二次型输出调节器设计的前提。其次基于谐波白噪声法推导出了随机路面模型,还有依据减速带路面的振动效果而建立了冲击路面模型。这两种路面模型作为了悬架系统的扰动激励输入。然后针对蛇优化算法(Snake Optimizer,SO)存在的收敛速度慢、寻优精度不足、稳定性不足等问题,提出3点增强型策略,从而设计出多策略增强型蛇优化算法(ASO算法),即通过佳点集和对立学习提高初始蛇群的质量;利用自适应振荡权重增强蛇群个体的移动能力;引入遗传算法(Genetic Algorithm,GA)的交叉-变异算子并行搜索解空间,以及莱维飞行扰动最优解,提高算法对局部最优的逃逸能力。基于23个基准函数对ASO算法、SO算法、GA算法、灰狼算法(Grey Wolf Optimizer,GWO)、人工蜂群算法(Artificial Bee Colony Algorithm,ABC)先后进行30维变量情况、10至50维变量变化情况下的测试对比。数据结果表明,ASO算法可对多个函数达到理论最优均值,并拥有极低的波动,相比其他优化算法具有较强的优势。最后对主动悬架系统设计LQR控制器,构造基于多个悬架性能指标的线性加权适应度函数,用来作为ASO算法与悬架系统LQR控制器的连接纽带。对二自由度四分之一悬架、四自由度半车悬架分别进行随机路面、冲击路面工况下的振动仿真。仿真结果表明:ASO-LQR-主动悬架系统可在轮胎动载荷和悬架动行程的安全范围内,较好地抑制车身加速度和俯仰角加速度,保障了驾乘人员的舒适感。更多还原

【Abstract】 In recent years,people’s demand for vehicles has changed from the initial demand for mobility to the pursuit of comfort.Suspension,as a vibration damping device in vehicles,plays an important role in people’s driving comfort.Passive suspension is low cost and widely used,but due to its own fixed elastic stiffness and damping coefficients,its adaptability is poor under the working conditions of complex roads or speed changes.Therefore,active suspension has become a necessary factor for major car companies to conquer the market.In order to solve the problem of rectifying the weight coefficient matrixs of Linear Quadratic Regulator(LQR)for active suspension system,this paper proposes an LQR control strategy based on the multi strategy Advanced Snake Optimizer(ASO),i.e.,ASOLQR control strategy.Firstly,the two active suspensions of a two-degree-of-freedom quarter and a four-degree-of-freedom half vehicle are taken as the research objects,and their dynamic models are established and transformed into state space expressions,which serve as the premise for the subsequent design of their LQR.Secondly,the random road model is derived based on the harmonic white noise method,and the bump road model is established based on the vibration effect of speed bump road.These two types of road models are used as perturbation excitation inputs to the suspension system.Then,to address the problems of slow convergence speed,insufficient search accuracy and insufficient stability of Snake Optimizer(SO),three advanced strategies are proposed to design the multi strategy Advanced Snake Optimizer(ASO),i.e.,to improve the the quality of initial snakes through the good-point set and the oppositional learning;to enhance the movement ability of the snakes by using the adaptive oscillation weights;and to introduce the crossover-mutation operator in Genetic Algorithm(GA)to search the solution space in parallel,as well as the optimal solution of Levy flight perturbation to enhance the algorithm’s ability to escape from the local optimum.Based on 23 benchmark functions,ASO,SO,GA,Grey Wolf Optimizer(GWO)and Artificial Bee Colony Algorithm(ABC)are successively tested and compared in 30-dimensional variables and10 to 50-dimensional variable changes.The data results show that the ASO can reach the theoretical optimal mean value for multiple functions and has very low fluctuation,which is a strong advantage over other optimization algorithms.Finally,the LQR is designed for the suspension system,and the linear weighted fitness functions based on several suspension system performance indexes are constructed,which is used as the connecting link between the ASO and the LQR of the suspension systems.And the vibration simulation is carried out for the two-degree-of-freedom quarter suspension and four-degree-of-freedom half vehicle suspension under the conditions of random road and bump road,respectively.The simulation results show that the ASOLQR-active suspension system can better suppress the sprung mass acceleration and pitch angle acceleration within the safety range of dynamic tyre load and suspension working space,which guarantees the comfort of the driver and passengers.更多还原