基于改进XGBoost的注塑机液压系统故障诊断研究  

Research on fault diagnosis of hydraulic system ofinjection molding machine based on improved XGBoost

作　　者：欧阳琦 张倍培 刘晋康 袁栋 黄昌隆 Ouyang Qi;Zhang Beipei;Liu Jinkang;Yuan Dong;Huang Changlong(College of Mechanical and Electrical Engineering,Hohai University,Jiangsu Changzhou,213022,China)

机构地区：[1]河海大学机电工程学院,江苏常州213022

出　　处：《机械设计与制造工程》2023年第7期78-83,共6页Machine Design and Manufacturing Engineering

基　　金：江苏省高等学校大学生创新创业训练计划项目(202210294144Y)。

摘　　要：针对注塑机液压系统故障隐蔽且难以准确诊断的问题,提出一种基于改进XGBoost的液压系统故障诊断方法。通过改进Tent混沌映射改善麻雀搜索算法(SSA)的全局寻优能力,并采用改进SSA对XGBoost中的超参数进行协调优化,弥补传统XGBoost凭经验设定超参数的不足。实验表明,改进XGBoost故障诊断模型具有更快的诊断速度和更高的诊断精度,能够实现对注塑机液压系统故障的准确诊断。Aiming at the problem that the hydraulic system fault of injection molding machine is hidden and difficult to diagnose accurately,a fault diagnosis method of hydraulic system based on improved XGBoost is proposed.By improving Tent chaotic mapping,the global optimization ability of sparrow search algorithm(SSA)is improved,and the superparameter in XGBoost is coordinated and optimized by improved SSA,which changes the deficiency of traditional XGBoost in setting super-parameter by experience.The experiment shows that the improved XGBoost fault diagnosis model has faster diagnosis speed and higher diagnosis accuracy,and can accurately locate the valve fault of hydraulic system.

关 键 词：注塑机 麻雀搜索算法 故障诊断 XGBoost Tent混沌映射 

分 类 号：TH133.33[机械工程—机械制造及自动化]