Abstract
Many industrial processes are operated in multiple modes due to different manufacturing strategies. Multimodality of process data is often accompanied with nonlinear and non-Gaussian characteristics, which makes data-driven monitoring more complicated. In this paper, statistics pattern analysis (SPA) is introduced to extract low- and high-order statistics from raw process data. Support vector data description (SVDD), which can deal with nonlinear and non-Gaussian problems, is applied to monitor multimode process in this paper. To improve detection performance of SVDD for training multimode data with outliers, modified local reachability density ratio (mLRDR) is proposed as a weight factor to be embedded in the weighted-SVDD (wSVDD) model, in which the local neighbors in terms of both space and time are considered. Finally, the effectiveness and superiority of our proposed method are demonstrated by the Tennessee-Eastman (TE) process and wastewater treatment process (WWTP).
Highlights
MethodologySmLRDR-wSVDD method is proposed to monitor nonlinear multimode process, in which statistics pattern dataset is formed by Statistics pattern analysis (SPA), and a new weight factor named modified local reachability density ratio (mLRDR) is proposed based on LRD for wSVDD monitoring method
Compared with the multimodel scheme, single-modelbased method simplifies the modeling and monitoring procedure
The spatial information is still not fully mined, and the temporal information is not employed in the construction of the weight for wSVDD, which may lead to unsatisfactory performance of the existing methods
Summary
SmLRDR-wSVDD method is proposed to monitor nonlinear multimode process, in which statistics pattern dataset is formed by SPA, and a new weight factor named mLRDR is proposed based on LRD for wSVDD monitoring method. To weaken the multimodality feature and widen the gap between outlier and the normal data, a modified local reachability density ratio (mLRDR) is developed to be the weight factor of wSVDD as w si mLRDR si. Compared with the existing weight factors such as LDR [26], mLRDR contains more comprehensive spatial and additional temporal information, and it can distinguish normal data with lower density and the outliers around the data points with higher density data; it will improve the monitoring performance of wSVDD. The proposed method adopts SPA to mine more features from different-order statistics of the original data compared with other existing SVDD-based methods. A new weight factor is proposed for wSVDD modeling, in which the additional neighbors’ local spatial information and temporal information are considered, which makes the proposed method more sensitive to outliers and density compared with other methods. Tennessee-Eastman process (TE process) was put forward in 1993 by Downs and Vogel [35], which has been widely adopted for scientific research [36]. ere are five units, Statistics matrix S M
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
