DocumentCode
1867298
Title
Multiple shooting SQP-line search algorithm for optimal control of pressure-constrained batch reactor
Author
Drag, Pawel ; Styczen, Krystyn
Author_Institution
Inst. of Comput. Eng., Control & Robot., Wroclaw Univ. of Technol., Wrocław, Poland
fYear
2013
fDate
8-11 Sept. 2013
Firstpage
307
Lastpage
313
Abstract
In the article a new approach for control of a pressure-constrained batch reactor and a new multi-step optimization algorithm were presented. The considered batch reactor was described by both differential and algebraic equations. State constraints incorporate always difficulties into a mathematical model of the reactor, so a new algorithm based on a multiple shooting SQP-line search method was proposed and tested. The multiple shooting method was used not only to ensure a stability of the solution, but to divide a system into smaller subsystems, so a large-scale problem is considered. The considerations were made for a simultaneous approach, which allows to apply this algorithm to a wide class of differential-algebraic systems. The simulations were executed in Matlab environment using Wroclaw Centre for Networking and Supercomputing.
Keywords
batch processing (industrial); chemical reactors; differential algebraic equations; optimal control; quadratic programming; search problems; stability; Matlab environment; Wroclaw Centre for Networking and Supercomputing; algebraic equation; differential equation; differential-algebraic systems; multiple shooting SQP-line search algorithm; multistep optimization algorithm; optimal control; pressure-constrained batch reactor; reactor mathematical model; stability; state constraints; Algorithm design and analysis; Equations; Inductors; Mathematical model; Optimal control; Optimization; Trajectory; DAE systems; multiple shooting method; optimal control; state constraints;
fLanguage
English
Publisher
ieee
Conference_Titel
Computer Science and Information Systems (FedCSIS), 2013 Federated Conference on
Conference_Location
Krako??w
Type
conf
Filename
6644018
Link To Document