|Table of Contents|

SVM implemented in fault diagnosis of liquid rocket engine(PDF)

《火箭推进》[ISSN:1672-9374/CN:CN 61-1436/V]

Issue:
2008年03期
Page:
7-12
Research Field:
研究与设计
Publishing date:

Info

Title:
SVM implemented in fault diagnosis of liquid rocket engine
Author(s):
He Hao; Hu Xiaoping; Jiang Zhijie; Liu Weiqiang
Inst．of Aerospace and Material Engineering,National Univ．of Defense Technology，Changsha 410073，China
Keywords:
SVM; liquid rocket engine; fault diagnosis; pattern recognition
PACS:
V434
DOI:
-
Abstract:
SVM which is based on machine learning algorithm is a method for pattern classifi— cation．The advantage of SVM is to solve the small samples，no-liner and pattern recognition with high dimension．In this paper．the method of SVM is used in fault diagnosis for the data of practical LRE trial run and simulated model．The SVM classifier detects the four faults of heat run complete— ly．Among the eighteen groups of simulated model data，seventeen groups of them can be detected， although incorrect warnings are happened during four groups detecting．Through learning and detecting on the four groups again，the faults are divided and detected correctly．

References:

[l]张育林．液体火箭发动机健康监控技术[M]．长沙：国防 科技大学．1998．
[2]吴建军．液体火箭发动机故障检测与诊断[D]．长沙：国防 科技大学．1995．
[3]张学工．关于统计学习理论与支持向量机[J]．自动化学 报。2000。26(1)：33_42．
[4]祁亨年．支持向量机及其应用研究综述[J]．计算机工程， 2004。30(10)：6-9．
[5]Vapnik V N．Estimation of Dependences Based on Empirical Data[J]．Berlin：pringer-Vedag，1982，13(2)：46-49．
[6]Davide Mattera，Francesco Palmieri，Simon Haykin．An Explicit Algorithm for training Support Vector Machines [J]．Signal processing letters，1999，24(5)：53-57．
[7]齐志泉，田英杰，等．支持向量机中的核参数选择问题[J]． 控制T程，2005，12(4)：379—381．
[8]李盼池，许少华．支持向量机在模式识别中的核函数特 性分析[J]．计算机工程与设计，2005，26(2)：302-304．

Memo

Memo:
-
Last Update: 1900-01-01