基于机器学习的航空发动机智能本机平衡研究
Research on Machine Learning-based Aero-engine Local Balancing Technology
王跃 1葛向东1
作者信息
- 1. 中国航发沈阳发动机研究所,沈阳 110015
- 折叠
摘要
针对传统航空发动机本机平衡技术多次试重导致的平衡效率低且成本高的现状,通过搜集历次成功平衡的试验数据,分别采用单一随机森林、多方法融合的人工智能与机器学习方法,深度挖掘蕴含于试验数据中的相关规律,建立航空发动机本机智能平衡模型,实现平衡加重角度的智能预测.基于多方法融合的本机智能平衡模型,所有样本进行训练预测误差仅为3.06°,表明了多方法融合的优势.该智能动平衡技术为航空发动机本机平衡的使用与推广提供了重要的技术途径.
Abstract
Aiming at the low equilibrium efficiency and high cost caused by the traditional aero-engine local balancing technology,by collecting experimental data that had been successfully balanced,the artificial intelligence and machine learning methods of single random forest and multi-method fusion were adopted respectively.The relevant laws embedded in experimental data was deeply mined.The intelligent local balancing model of the aero-engine was established,and the intelligent prediction of the balanced weighting angles was realized.Based on the multi-method fusion of the local balancing technology model,all samples training prediction error was only 3.06 °,indicating the advantages of multi-method fusion.The intelligent dynamic equilibrium technology provided an important technical approach for the use and popularization of aero-engine equilibrium.
关键词
航空发动机/本机平衡/随机森林/多方法融合Key words
aero-engine/local balancing technology/random forest/multi-method fusion引用本文复制引用
出版年
2024
NETL
NSTL
万方数据