摘要
高速干式滚齿加工过程是一个高度非线性、强时变性的复杂动态过程,难以建立准确有效的滚削力波动特性的动力学模型。因此,本文提出一种基于机器学习的SOA-XGBoost的滚削力波动特性预测模型,通过使用海鸥优化算法SOA对XGBoost模型的超参数进行迭代训练,将优化后的超参数赋给XGBoost模型进行滚削力波动特性的预测,与已有的GABP、PSO-SVR和XGBoost算法模型进行预测对比分析,对预测模型进行优选。实验结果表明SOA-XGBoost在粗加工和精加工测试集上的平均绝对预测误差分别为1.86%和2.51%,提高了滚削力波动特性的预测精度,可为滚齿加工工艺参数优化提供指导。
High-speed dry hobbing process is a highly nonlinear and time-varying complex dynamic process,so it is difficult to establish an accurate and effective dynamic model for hobbing force fluctuation characteristics.Therefore,a prediction model for rolling force fluctuation characteristics of SOA-XGBoost(Seagull optimization algorithm-extreme gradient boosting)based on machine learning is put forward.By using the seagull optimization algorithm to iteratively train the hyperparameters of XGBoost model,and assigning the optimized hyperparameters to XGBoost model to predict the rolling force fluctuation characteristics,and comparing with the existing GABP,PSO-SVR and XGBoost model,the prediction model is optimized.The experimental results show that the average absolute prediction errors via SOA-XGBoost in the rough machining and finishing test sets are 1.86%and 2.51%respectively,which improves the prediction accuracy of hobbing force fluctuation characteristics and can provide a guidance for optimizing the hobbing processing parameters.
作者
郑荣盛
丁国龙
周明成
ZHENG Rongsheng;DING Guolong;ZHOU Mingcheng(School of Mechanical Engineering,HuBei University of Technology,Wuhan 430068,China)
出处
《机械科学与技术》
北大核心
2025年第12期2146-2160,共15页
Mechanical Science and Technology for Aerospace Engineering
基金
湖北省技术创新专项(2019AAA069)
湖北省重点研发计划(2022AAA004)
湖北省科技重大专项(2023BAA010-3)。
关键词
高速干式滚齿
机器学习
融合算法
滚削力波动特性预测
high-speed dry hobbing machining
machine learning
fusion algorithm
prediction of fluctuation characteristics of hobbing force
