计算机科学 ›› 2017, Vol. 44 ›› Issue (2): 239-243.doi: 10.11896/j.issn.1002-137X.2017.02.039

• 人工智能 • 上一篇    下一篇

基于大规模训练神经网络的微小故障在线检测

司文杰,杨飞飞   

  1. 华南理工大学机械与汽车工程学院 广州510641,华南理工大学材料科学与工程学院 广州 510641
  • 出版日期:2018-11-13 发布日期:2018-11-13
  • 基金资助:
    本文受国家自然科学基金重点项目(60934001)资助

Online Detection of Incipient Fault Based on Large-scale Neural Networks

SI Wen-jie and YANG Fei-fei   

  • Online:2018-11-13 Published:2018-11-13

摘要: 神经网络已经广泛应用于系统建模和模式识别领域。但为了逼近未知的参数或者系统动态,需要大量的神经元达到足够的逼近精度,因此导致了计算负荷的增大。运算量制约着大规模神经网络计算,无法使其应用到实际的在线系统中。CPU处理无法保证在线数据的同步运算,需要借助图形处理单元GPU(Graphic Processing Unit)来解决实时性同步运算问题。首先,利用RBF神经网络的持续激励PE(Persistent Excitation)特性对系统输入进行分析,减少神经元的数目且优化设计算法,从而提高逼近精度。其次,基于LabVIEW平台,利用LabVIEW的GPU高性能分析工具包实现神经网络算法和并行计算。最后,在一台航空低速轴流压气机中开发基于大规模训练神经网络的LabVIEW系统。实验结果表明,提出的方法可以实现对系统的在线实时运行,满足航空失速检测的要求。

关键词: 神经网络,持续激励,LabVIEW,GPU,大规模计算,在线实验

Abstract: Neural networks have been widely used for the system modeling and pattern recognition.However,in order to approximate the unknown parameters or system dynamics,it needs enough neurons to achieve sufficiently accurate approximation,which leads to increase of the computational cost.The computation would restrict the online application of the large-scale neural networks.Because CPU processing cannot keep pace with online data capture,the commonly available graphics processors are used for the bulk of data processing in online systems.First,the input of the system was analyzed by persistent excitation characteristics of RBF neural network,reducing the number of neurons and optimizing design optimization algorithm to improve the approximation error.Secondly,LabVIEW and LabVIEW GPU analysis toolkit were used to achieve algorithm implementation and parallel computing.Finally,online experiment of stall detection was conducted in a low speed axial compressor based on LabVIEW.Experimental results show that the proposed method can meet compressors stall detection of online operating system.

Key words: Neural network,Persistent excitation,LabVIEW,GPU,Large-scale computing,Online experiment

[1] ANDERSON J A.An introduction to neural networks[M].MIT press,1995.
[2] HAGAN M T,DEMUTH H B,BEALA M H.Neural network design[M].Boston:Pws Pub.,1996.
[3] ZENG D,GAO L,LIN L,et al.Application of LabVIEW in online monitoring and automatic control of fermentation process[J].Control & Computer,2006(22):48-50.
[4] HORNG J H.Hybrid MATLAB and LabVIEW with neural network to implement a SCADA system of AC servo motor[J].Advances in Engineering Software,2008,39(3):149-155.
[5] LIU J M,XU Z Z,SUN D H.Virtual instrument and neural network in application of grain moisture detection[C]∥Procee-dings of the 2014 International Conference on Control Enginee-ring and Information Systems.2014:155-158.
[6] LU F S,SONG J Q,YIN F K,et al.Survey of CPU/GPU synergetic parell computing[J].Computer Science,2011,8(3):5-9.(in Chinese) 卢风顺,宋君强,银福康,等.CPU-GPU协同并行计算研究综述[J].计算机科学,2011,38(3):5-9.
[7] GAO Z,CECATI C,DING S X.A survey of fault diagnosis and fault-tolerant techniques-Part I:fault diagnosis With model-based and signal-based approaches[J].IEEE Transactions on Industrial Electronics,2015,62(6):3757-3767.
[8] WU Y,JIANG B,LU N,et al.ToMFIR-based incipient fault de-tection and estimation for high-speed rail vehicle suspension system[J].Journal of the Franklin Institute,2015,352(4):1672-1692.
[9] HWANG W,HUH K.Fault Detection and Estimation for Electromechanical Brake Systems Using Parity Space Approach[J].Journal of Dynamic Systems,Measurement,and Control,2015,137(1):014504.
[10] KURDILA A J,NARCOWICH F J,WARD J D.Persistency of excitation in identification using radial basis function approxi-mants[J].SIAM Journal on Control and Optimization,1995,33(2):625-642.
[11] WANG C,HILL D J.Deterministic learning theory for identification,recognition,and control[M].CRC Press,2009.
[12] POLYCARPOU M M,TRUNOV A B.Learning approach tononlinear fault diagnosis:detectability analysis[J].IEEE Tran-sactions on Automatic Control,2000,45(4):806-812.

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!