Abstract: According to the structure feature and application background of dual-cavity half toroidal CVT, an optimization model with objective functions of the split-torque ratio of input torque and the ratio of output torque to input torque is established. Pareto-optimal solutions of this model are obtained by means of non-dominated sorting genetic algorithm with the elite strategy(NSGA-II). To determine the best Pareto-optimal solution, firstly the decision matrix is constructed by the Pareto-optimal solutions, then the weights of attributes are calculated by objective information entropy weight(IEW) approach, and finally the Pareto-optimal solutions are ranked through technique for order preference by similarity to ideal solution(TOPSIS). The values of the design variables in the best Pareto-optimal solution are substituted for the expressions for the split-torque ratio of input torque, and the ratio of output torque to input torque and transmission efficiency. The characteristic curves on the three performance indicators are drawn by MATLAB. The result shows that the best Pareto-optimal solution decided by TOPSIS can make dual-cavity half toroidal CVT high-efficiently transmit, output high torque and have continuously variable unit bear smaller input torque.

Key words: CVT, performance indicators, multi-objective optimization, NSGA-II, TOPSIS

摘要： 针对双腔半环面型CVT的结构特点和应用背景,建立以输入扭矩分矩比和输出输入扭矩比为目标函数的优化模型。采用带有精英策略的非支配排序遗传算法(non-dominated sorting genetic algorithm with the elite strategy, NSGA-II)求解该模型的Pareto最优解。为了确定最佳的Pareto最优解,首先将获得的Pareto最优解构成决策矩阵,其次利用客观赋权的信息熵(information entropy weight, IEW)法计算各属性的权值,最后运用逼近理想解的排序法(technique for order preference by similarity to ideal solution, TOPSIS)对Pareto最优解排序。将最佳Pareto最优解对应的设计变量值代入输入扭矩分矩比、输出输入扭矩比以及传动效率的表达式,使用MATLAB软件绘制相应的特性曲线。结果表明,通过TOPSIS确定的最佳Pareto最优解能够实现双腔半环面型CVT高效率传动、大扭矩输出和无级变速单元承受较小输入扭矩的功能。

关键词: CVT, 性能指标, 多目标优化, NSGA-II, TOPSIS法