Fault diagnosis based on signed directed graph and support vector machine

Xiaoming Han; Qing Lv; Gang Xie; Jianxia Zheng

doi:10.1117/12.920384

13 January 2012 Fault diagnosis based on signed directed graph and support vector machine

Xiaoming Han, Qing Lv, Gang Xie, Jianxia Zheng

Proceedings Volume 8349, Fourth International Conference on Machine Vision (ICMV 2011): Machine Vision, Image Processing, and Pattern Analysis; 83491H (2012) https://doi.org/10.1117/12.920384
Event: Fourth International Conference on Machine Vision (ICMV 11), 2011, Singapore, Singapore

Abstract

Support Vector Machine (SVM) based on Structural Risk Minimization (SRM) of Statistical Learning Theory has excellent performance in fault diagnosis. However, its training speed and diagnosis speed are relatively slow. Signed Directed Graph (SDG) based on deep knowledge model has better completeness that is knowledge representation ability. However, much quantitative information is not utilized in qualitative SDG model which often produces a false solution. In order to speed up the training and diagnosis of SVM and improve the diagnostic resolution of SDG, SDG and SVM are combined in this paper. Training samples' dimension of SVM is reduced to improve training speed and diagnosis speed by the consistent path of SDG; the resolution of SDG is improved by good classification performance of SVM. The Matlab simulation by Tennessee-Eastman Process (TEP) simulation system demonstrates the feasibility of the fault diagnosis algorithm proposed in this paper.

Citation Download Citation

Xiaoming Han, Qing Lv, Gang Xie, and Jianxia Zheng "Fault diagnosis based on signed directed graph and support vector machine", Proc. SPIE 8349, Fourth International Conference on Machine Vision (ICMV 2011): Machine Vision, Image Processing, and Pattern Analysis, 83491H (13 January 2012); https://doi.org/10.1117/12.920384

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available