Data-driven process decomposition and robust online distributed modelling for large-scale processes

ABSTRACT

With the increasing attention of networked control, system decomposition and distributed models show significant importance in the implementation of model-based control strategy. In this paper, a data-driven system decomposition and online distributed subsystem modelling algorithm was proposed for large-scale chemical processes. The key controlled variables are first partitioned by affinity propagation clustering algorithm into several clusters. Each cluster can be regarded as a subsystem. Then the inputs of each subsystem are selected by offline canonical correlation analysis between all process variables and its controlled variables. Process decomposition is then realised after the screening of input and output variables. When the system decomposition is finished, the online subsystem modelling can be carried out by recursively block-wise renewing the samples. The proposed algorithm was applied in the Tennessee Eastman process and the validity was verified.

KEYWORDS:

• Canonical correlation analysis
• affinity propagation clustering
• block-wise RPLS, model reduction, model-predictive control
• process control
• parameter identification

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

National Natural Science Foundation of China [grant number 61203072], [grant number 61403190], [grant number 61773366]; Research Innovation Program for College Graduates of Jiangsu Province [grant number KYLX16_0598].

Notes on contributors

Zhang Shu

Zhang Shuis a graduate in Measurement Control Technology and Instrumentation from Nanjing Tech University in 2015 and will receive the master degree in Control Theory and Control Engineering from Nanjing Tech University in 2018. Her current research interests include large-scale system decomposition and distributed model predictive control.

Li Lijuan

Li Lijuan is an associate professor in the Industrial System and Automation Department, Nanjing Tech University, Nanjing, China. She received the BS and MS degrees from Nanjing University of Technology, Nanjing, China, in 1997 and 2004, respectively, and PhD degree from Zhejiang University, Hangzhou, China, in 2008. She worked as a visiting scholar in the University of Southern California from 2013 to 2014. His current research interests include industrial process modeling and advanced control, control performance monitoring and diagnosis.

Yao Lijuan

Yao Lijuan is an lecture in the Industrial Robot Department, Suzhou higher vocational and technical school. She received the bachelor's and the master degrees from Nanjing Tech University. Her current research interests include industry robot technology.

Yang Shipin

Yang Shipin is a lecturer in the college of Electrical Engineering and Control Science, Nanjing Tech University, Nanjing, China. He received the BS and MS degrees from Wuhan University of Science and Technology, Wuhan, China, and the PhD degree from Zhejiang University, Hangzhou, China. His research interests include bionic-based optimization computation and its application such as electronics design and parameter estimation in the chemical plant process.

Zou Tao

Zou Tao is a professor in the Industrial Control Networks and Systems Department, Shenyang Institute of Automation, Chinese Academy of Sciences, China. He received the BS and MS degrees from Shenyang Institute of Chemical Technology and PhD degree from Shanghai Jiaotong University, China. His current research interests include model predictive control and its applications.