An integrated output space partition and optimal control method of multiple-model for nonlinear systems. (8th May 2018)
- Record Type:
- Journal Article
- Title:
- An integrated output space partition and optimal control method of multiple-model for nonlinear systems. (8th May 2018)
- Main Title:
- An integrated output space partition and optimal control method of multiple-model for nonlinear systems
- Authors:
- Song, Chunyue
Wu, Bing
Zhao, Jun
Xu, Zuhua - Abstract:
- Highlights: An integrated output space partition and optimal control method of multiple-model for nonlinear systems is initially addressed. The decomposition of multiple-model is obtained from closed-loop performance, not come from the characteristic of the open-loop systems. By constructing new states underlying each state having the same physical meaning, the input-output identification model is transferred into state space form. A projection operator is designed to project the partition in the constructed state space into output space. Abstract: A systematic method for optimally partitioning operating range of each linear subsystem in output space under the criterion of closed-loop performance is initiated, when a multiple model approach is applied to nonlinear systems. As a result, an integrated output space partition and optimal control method is proposed. Firstly, linear input-output models are identified at given operating points and then reformulated as a hybrid model underlying each state having the same physical meaning. Secondly, the optimal state space partition is obtained according to a closed-loop control index. Finally, based on the obtained optimal state space partition, an optimal output space partition is achieved with the projection technique. Furthermore, a hybrid model-MPC strategy is designed according to the obtained multiple-model associated with its optimal output space partition. The integrated output space partition and optimal control method canHighlights: An integrated output space partition and optimal control method of multiple-model for nonlinear systems is initially addressed. The decomposition of multiple-model is obtained from closed-loop performance, not come from the characteristic of the open-loop systems. By constructing new states underlying each state having the same physical meaning, the input-output identification model is transferred into state space form. A projection operator is designed to project the partition in the constructed state space into output space. Abstract: A systematic method for optimally partitioning operating range of each linear subsystem in output space under the criterion of closed-loop performance is initiated, when a multiple model approach is applied to nonlinear systems. As a result, an integrated output space partition and optimal control method is proposed. Firstly, linear input-output models are identified at given operating points and then reformulated as a hybrid model underlying each state having the same physical meaning. Secondly, the optimal state space partition is obtained according to a closed-loop control index. Finally, based on the obtained optimal state space partition, an optimal output space partition is achieved with the projection technique. Furthermore, a hybrid model-MPC strategy is designed according to the obtained multiple-model associated with its optimal output space partition. The integrated output space partition and optimal control method can improve the nonlinear system overall control performance and results of numerical simulation are provided. … (more)
- Is Part Of:
- Computers & chemical engineering. Volume 113(2018)
- Journal:
- Computers & chemical engineering
- Issue:
- Volume 113(2018)
- Issue Display:
- Volume 113, Issue 2018 (2018)
- Year:
- 2018
- Volume:
- 113
- Issue:
- 2018
- Issue Sort Value:
- 2018-0113-2018-0000
- Page Start:
- 32
- Page End:
- 43
- Publication Date:
- 2018-05-08
- Subjects:
- Nonlinear systems -- Output-space partition -- Optimal control -- Multilinear model -- Hybrid systems
Chemical engineering -- Data processing -- Periodicals
660.0285 - Journal URLs:
- http://www.sciencedirect.com/science/journal/00981354 ↗
http://www.elsevier.com/journals ↗ - DOI:
- 10.1016/j.compchemeng.2018.02.018 ↗
- Languages:
- English
- ISSNs:
- 0098-1354
- Deposit Type:
- Legaldeposit
- View Content:
- Available online (eLD content is only available in our Reading Rooms) ↗
- Physical Locations:
- British Library DSC - 3394.664000
British Library DSC - BLDSS-3PM
British Library HMNTS - ELD Digital store - Ingest File:
- 20976.xml