Application of mixed integer nonlinear programming for system identification

Fernandes, Natércia C.P.; Fernandes, Florbela P.; Romanenko, Andrey

http://hdl.handle.net/10198/27255

Use this identifier to reference this record.

Name:	Description:	Size:	Format:
5.0026410.pdf		421.89 KB	Adobe PDF	Download

Send Feedback

Authors

Fernandes, Natércia C.P.

Fernandes, Florbela P.

Romanenko, Andrey

Abstract(s)

This work describes a method of deadtime approximation in dynamic systems, particularly in the context of nonlinear model predictive control based on mechanistic models where the differentiability of the equations must be ensured. The resulting system identification system is solved using the BBMCSFilter (Branch and Bound based on a Multistart Coordinate Search Filter) global optimization algorithm to determine the order and the parameters of the resulting model, taking into account not only the model-plant mismatch but also the model complexity and the resulting computation time. The application of the method is illustrated with a simulated example of a chemical process unit. © 2020 American Institute of Physics Inc.. All rights reserved.

URI

http://hdl.handle.net/10198/27255

Citation

Fernandes, Natércia C.P.; Fernandes, Florbela P.; Romanenko, Andrey (2020). Application of mixed integer nonlinear programming for system identification. In International Conference on Numerical Analysis and Applied Mathematics 2019, ICNAAM 2019. Rhodes