東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Principal component analysis based f...

Cao, Jin.

FindBook

Google Book

Amazon

博客來

Principal component analysis based fault detection and isolation.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Principal component analysis based fault detection and isolation./
作者:	Cao, Jin.
面頁冊數:	175 p.
附註:	Source: Dissertation Abstracts International, Volume: 65-02, Section: B, page: 0913.
Contained By:	Dissertation Abstracts International65-02B.
標題:	Engineering, Electronics and Electrical. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3123116
ISBN:	0496704281

Principal component analysis based fault detection and isolation.
Cao, Jin.

Principal component analysis based fault detection and isolation. - 175 p.

Source: Dissertation Abstracts International, Volume: 65-02, Section: B, page: 0913.

Thesis (Ph.D.)--George Mason University, 2004.

Principal component analysis (PCA), considered by some as a "model-free" method, has recently been revealed to have a close link with parity relations through PCA transformation. As a result, PCA can also serve as a basis for analytical redundancy type fault detection and isolation (FDI).

ISBN: 0496704281Subjects--Topical Terms:

626636
Engineering, Electronics and Electrical.

Principal component analysis based fault detection and isolation.
LDR:03491nmm 2200337 4500 001 1849676
005 20051207081939.5
008 130614s2004 eng d
020 $a 0496704281
035 $a (UnM)AAI3123116
035 $a AAI3123116
040 $a UnM $c UnM
100 1 $a Cao, Jin. $3 1937624
245 1 0 $a Principal component analysis based fault detection and isolation.
300 $a 175 p.
500 $a Source: Dissertation Abstracts International, Volume: 65-02, Section: B, page: 0913.
500 $a Director: Janos Gertler.
502 $a Thesis (Ph.D.)--George Mason University, 2004.
520 $a Principal component analysis (PCA), considered by some as a "model-free" method, has recently been revealed to have a close link with parity relations through PCA transformation. As a result, PCA can also serve as a basis for analytical redundancy type fault detection and isolation (FDI).
520 $a In the modeling phase of the model-based FDI, we need to understand the mathematical expression of the model and its characteristics. The last principal component (LPC) modeling utilizes the eigenvectors associated with the last principal components and it may be considered an alternative to least squares estimation of model parameters. This modeling approach shares some basic characteristics with other estimation methods. The results can be extended to linear dynamic and polynomial static models.
520 $a The presence of noise may induce bias in modeling. It would appear as a parametric discrepancy when the model is used in model-based FDI schemes. Our research derives explicit expressions for noise-induced bias in LPC modeling. We show that LPC-based estimates are biased even when LS-based ones are not, and when the LS estimate is also biased, the LPC estimate has the LS bias plus an additional term.
520 $a PCA-based FDI mainly uses two important isolation enhancement techniques: structured residuals and directional residuals. Both types of residuals may be obtained from a full model, but structured residuals may also be generated from directly identified partial models. Accordingly, there are two approaches in the PCA framework: full PCA models and partial PCA models. An optimal structured residual design, generated from partial PCA models, is proposed for sensitivity optimization, using a max-min criterion applied to the fault-to-noise response ratio in the residual.
520 $a In PCA-based FDI, mis-isolation of some sensor and actuator faults may arise if the training data is collected under constant control. We provide a detailed analysis of how control actions affect PCA-based diagnosis. We investigate ratio and feedback control in linear static and discrete dynamic systems, with full and partial PCA models.
520 $a A non-isothermal Continuous Stirred Tank Reactor (CSTR) model and a Space Shuttle External Fuel Tank (EFT) model serve as the application studies in the various aspects throughout this dissertation research.
520 $a The main contributions in this dissertation research include investigations into LPC modeling, noise-induced bias in LPC modeling, PCA-based diagnosis in the presence of control, and partial PCA-based optimal residual design.
590 $a School code: 0883.
650 4 $a Engineering, Electronics and Electrical. $3 626636
690 $a 0544
710 2 0 $a George Mason University. $3 1019450
773 0 $t Dissertation Abstracts International $g 65-02B.
790 1 0 $a Gertler, Janos, $e advisor
790 $a 0883
791 $a Ph.D.
792 $a 2004
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3123116