上海交通大学学报(英文版) ›› 2015, Vol. 20 ›› Issue (3): 358-362.doi: 10.1007/s12204-015-1637-4

• • 上一篇    下一篇

Primary Research on Real-Time Fault Diagnosis Platform for Fuel Tank System of an Aircraft

BAO Yong-lin (鲍泳林)   

  1. (Institute of Systems Engineering, China Academy of Engineering Physics, Mianyang 621999, Sichuan, China)
  • 发布日期:2015-06-11
  • 通讯作者: BAO Yong-lin (鲍泳林) E-mail:yonglin bao@hotmail.com

Primary Research on Real-Time Fault Diagnosis Platform for Fuel Tank System of an Aircraft

BAO Yong-lin (鲍泳林)   

  1. (Institute of Systems Engineering, China Academy of Engineering Physics, Mianyang 621999, Sichuan, China)
  • Published:2015-06-11
  • Contact: BAO Yong-lin (鲍泳林) E-mail:yonglin bao@hotmail.com

摘要: Sub-tanks in fuel tank systems of aircrafts transfer fuel to engines in certain order. These sub-tanks and attached tank-accessories affect each other, and make fault diagnosis in such systems rather difficult. Without real measured data, this paper analyzes fault modes and fault effects of the fuel tank system, including its tankaccessories, of a given aircraft. Fault model of the system is built theoretically, and fault diagnosis criteria are deduced. Such criteria are then quantified to train a back propagation neural network (BPNN) as fault diagnosis model. To realize fault diagnosis of the real fuel tank system, a real-time fault diagnosis platform based on LabView and VxWorks to perform this diagnosis method is discussed. This platform is a technical groundwork for fault diagnosis in real fuel tank systems.

关键词: fuel tank systems, fault diagnosis, real-time platform, neural network

Abstract: Sub-tanks in fuel tank systems of aircrafts transfer fuel to engines in certain order. These sub-tanks and attached tank-accessories affect each other, and make fault diagnosis in such systems rather difficult. Without real measured data, this paper analyzes fault modes and fault effects of the fuel tank system, including its tankaccessories, of a given aircraft. Fault model of the system is built theoretically, and fault diagnosis criteria are deduced. Such criteria are then quantified to train a back propagation neural network (BPNN) as fault diagnosis model. To realize fault diagnosis of the real fuel tank system, a real-time fault diagnosis platform based on LabView and VxWorks to perform this diagnosis method is discussed. This platform is a technical groundwork for fault diagnosis in real fuel tank systems.

Key words: fuel tank systems, fault diagnosis, real-time platform, neural network

中图分类号: