- Tytuł:
- A modified independent component analysis-based fault detection method in plant-wide systems
- Autorzy:
-
Teimoortashloo, M.
Sedigh, A. K. - Powiązania:
- https://bibliotekanauki.pl/articles/205999.pdf
- Data publikacji:
- 2015
- Wydawca:
- Polska Akademia Nauk. Instytut Badań Systemowych PAN
- Tematy:
-
DICA
fault detection
running time
orthogonalization - Opis:
- This paper presents a modified Independent Component Analysis (ICA)-based Fault Detection Method (FDM). The proposed FDM constructs a set of matrices, revealingthe trend of the variable samples and execute ICA algorithm for each set of matrices in contrast to the FDM based on dynamic ICA (DICA) which constructs the high dimensional augmented matrix. This paper shows that the proposed FDM decreases the matrix dimensions and as a result compensates for some disadvantages of using the high dimensional matrix discussed in previous articles. Furthermore, other advantages of the proposed FDM are the decreases in the running time, computational cost of the algorithm and the orthogonalization estimation errors. Moreover, the proposed method improves the detectability for a class of faults compared to DICA-based FDM. This class of fault occurs when two or more consecutive samples of fault source signal have opposite signs and cancel out each other. Simulation results are provided to show the effectiveness of the proposed methodology.
- Źródło:
-
Control and Cybernetics; 2015, 44, 2; 287-310
0324-8569 - Pojawia się w:
- Control and Cybernetics
- Dostawca treści:
- Biblioteka Nauki
Artykuł