Abstract A novel multi-objective optimization of planetary gearbox is presented using a discrete version of Non-Dominated Sorting Genetic Algorithm (NSGA-II). Minimization of weight and total power loss of planetary gearbox are two objective functions. Number of teeth in the sun, planet and ring gears, module, face width, input shaft and planet pin diameter are design variables. Regular mechanical, bearing selection and scuffing are various design constraints considered. Investigations were done for three different gear profiles on an industrial planetary gearbox. The results were compared with an industrial gearbox provided in the AGMA standard, which showed not only significant reduction of weight and power loss, but were also found safer in scuffing. Pareto fronts of different ISO grade oils were compared for all gear profiles, and ISO VG 460 was found to be the best oil. Further analysis without a scuffing constraint showed moderate to high levels of scuffing risk for lower-grade ISO oils and lower level risk for higher-grade ISO oils. Ultimately, this study helps to reduce weight, enhance the efficiency and prevent early failures in planetary gearbox.

中文翻译：

---

考虑划伤约束的行星齿轮箱宏观几何多目标优化

摘要 使用离散版本的非支配排序遗传算法 (NSGA-II) 提出了一种新颖的行星齿轮箱多目标优化方法。行星齿轮箱的重量和总功率损失的最小化是两个目标函数。太阳轮齿数、行星齿轮和齿圈、模数、面宽、输入轴和行星销直径都是设计变量。常规机械、轴承选择和磨损是考虑的各种设计约束。对工业行星齿轮箱上的三种不同齿轮齿廓进行了调查。结果与 AGMA 标准中提供的工业齿轮箱进行了比较，这表明不仅重量和功率损失显着减少，而且磨损更安全。比较了所有齿轮轮廓的不同 ISO 等级油的帕累托前沿，并且发现 ISO VG 460 是最好的油。没有划伤限制的进一步分析表明，低等级 ISO 油的划伤风险为中等至高水平，而高等级 ISO 油的风险水平较低。最终，这项研究有助于减轻重量、提高效率并防止行星齿轮箱的早期故障。