Abstract
In this paper, we discuss model predictive control applied to blending processes. Blending processes are ubiquitous in the chemical process industries since reactants usually need to be mixed before entering a reactor. Many times, the blending is trivial as pure streams of reactants are mixed. We consider non-trivial blending problems in which non-pure streams with the reactants are to be mixed. The motivating example is the blending problem that occurs in cement production. The raw mix for the cement kiln must have a specified chemical composition. This composition is obtained by mixing piles with different chemical compositions and economic value such that the raw mix meets specifications in the cheapest possible way. We formulate the blending problem as a nonlinear optimization problem that can be approximated well as a convex quadratic optimization problem. We implement the corresponding nonlinear and linear model predictive controllers (NMPC, LMPC) using a continuous-time transfer function description that is realized as a discrete-time linear state space model. The controller obtains feedback by combining regularly sampled online measurements and irregularly sampled laboratory measurements using a time variant dynamic Kalman filter with memory. Numerical simulations demonstrate that the NMPC and LMPC have similar performance.
Published Version (
Free)
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.
