Fault detection and diagnosis of nonlinear processes based on kernel ICA-KCCA
2010
Fault detection and diagnosis based on multivariate statistical way is a hotspot in recent years. According to the nonlinear property of Continuous Annealing Line, this article developes a nonlinear ICA, which combined the predominance of ICA and reproducing kernel Hilbert space, to monitor process. This method has better statistical attribute than traditional ICA algorithm based on maximum negentropy, and it performs more robust and flexible to the variety of signal source. At last, the simulation results of practical production reveal that the kernel ICA-KCCA algorithm is more effective than traditional ICA method.
Keywords:
- Principal component analysis
- Signal processing
- Kernel (linear algebra)
- Multivariate statistics
- Reproducing kernel Hilbert space
- Independent component analysis
- Machine learning
- Fault detection and isolation
- Nonlinear system
- Computer science
- Artificial intelligence
- Pattern recognition
- Mutual information
- Negentropy
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
11
References
3
Citations
NaN
KQI