中国民航大学学报

• 民用航空 • 上一篇    下一篇

基于AdaBoost 的飞机部件DMC预计方法研究

徐建新,孙发东   

  1. (中国民航大学航空工程学院,天津300300)
  • 收稿日期:2015-12-10 修回日期:2016-03-08 出版日期:2016-10-19 发布日期:2016-12-06
  • 作者简介:徐建新(1967—),男,江苏苏州人,教授,博士,研究方向为结构细节应力分析、复合材料结构力学

Study on DMC estimation of aircraft components based on AdaBoost

XU Jianxin, SUN Fadong   

  1. (College of Aeronautical Engineering, CAUC, Tianjin 300300, China)
  • Received:2015-12-10 Revised:2016-03-08 Online:2016-10-19 Published:2016-12-06

摘要:

飞机部件的直接维修成本(DMC)预计是控制整机维修成本的关键步骤。鉴于现有方法预计精度不高,波动性大,并针对经验数据匮乏的问题,引入了一种基于AdaBoost(adaptive boosting)算法的飞机部件直接维修成本预计方法。对比分析结果表明,此方法不依赖于经验数据袁可较准确地预测飞机部件的实际直接维修成本,比以往的方法在精度和稳定性上有显著提高,适用于设计阶段及维修过程中的部件维修成本预计。

关键词: 飞机部件, 直接维修成本, 预计模型, 偏最小二乘法, 极端学习机, AdaBoost

Abstract:

Component DMC estimation is a vital procedure in maintenance cost control. Aiming at the disadvantages of existing estimation methods such as low accuracy and large fluctuation, a new method based on AdaBoost (adaptive boosting) is introduced. Experiments show that this method does not rely on empirical data, it is obviously more accurate and stable than former ones. It is applicable to estimate the component DMC during design and maintenance process.

Key words: aircraft components, DMC, prediction model, PLS, ELM, AdaBoost

中图分类号: