taptap下载安装安卓学报

• 民用航空 • 上一篇    下一篇

基于改进SDAE 算法的复杂系统故障识别

史永胜,衡慧娟   

  1. (taptap下载安装安卓航空工程学院,天津300300)
  • 收稿日期:2018-03-22 修回日期:2018-05-04 出版日期:2019-04-26 发布日期:2019-05-10
  • 作者简介:史永胜(1965—),男,辽宁锦州人,教授,博士,研究方向为结构智能化设计、故障诊断与结构修理、专家系统及知识表示.

Fault recognition of complex system based on improved SDAE algorithm

SHI Yongsheng, HENG Huijuan   

  1. (College of Aeronautical Engineering, CAUC, Tianjin 300300, China)
  • Received:2018-03-22 Revised:2018-05-04 Online:2019-04-26 Published:2019-05-10

摘要: 针对传统稀疏去噪自编码网络算法对复杂系统故障识别效率低、误差损失值大等问题,提出了一种改进算法,引入数据量化和故障码匹配的概念,从而达到降低输入与输出的误差损失值,提高故障识别精度的目的。结果表明:改进后算法的重构误差平均降低1.5%,故障识别率平均提高9%,复杂系统故障识别精度得到提高。

关键词: 复杂系统, 故障码匹配, 数据量化, 故障识别, 精度分析

Abstract: Aiming at the low fault recognition rate and large error loss of traditional SDAE algorithm, an improved algorithm is proposed. Data quantification and fault code matching are introduced in order to reduce the input/output error loss and raise the fault recognition accuracy. Results show that the reconstruction error of the improved SDAE algorithm is decreased by 1.5% on average, and the fault recognition rate is increased by 9% on average, the fault recognition accuracy of complex system has been improved.

Key words: complex system, fault code matching, data quantization, fault identification, accuracy analysis

中图分类号: 

Baidu
map