Plant‐wide control of a hybrid process

Abstract

This paper deals with the model predictive control (MPC) of an industrial hybrid process where continuous and batch units operate jointly: the crystallization section of a sugar factory. The paper describes a plantwide predictive controller that takes into account, both, continuous objectives and manipulated variables, as well as those related to the discrete operation and logic of the batch units. The MPC is formulated with time events, so that a more efficient NLP optimization technique, instead of MINLP, could be applied. Adaptation is provided by model updating and error estimation. Results of the controller operation in an industrial simulator are provided.

Hierarchically coordinated economic MPC plantwide control of mixed continuous‐batch units in process industries with application to a beet sugar plant

“…The scheduling problem would suggest the need to solve, at each iteration of the coordinator, a mixed-integer nonlinear problem (MINLP) that could be very demanding in the computational sense. In this particular case, this could be avoided by means of the parametrization described in the work of de Prada et al, 39 where the Load i binary command for starting subordinate cycles is replaced by the continuous time T loadi to wait from the considered time to the instant of batch initiation as shown in Figure 2B.…”

“…The optimization problem of the coordinator, as applied to the particular case at hand, is represented by Equations (30)(31)(32)(33)(34)(35)(36)(37)(38)(39). The performance function is considered to be the aggregate value of the consumption of steam by all the crystallizers in the whole exercise.…”

“…Finally, the inequality constraints (38) and (39) are required to make that all the decisions taken are compatible with keeping the level in the syrup buffer tank in the range [L min , L max ] and the pressure of steam in the evaporator effect above the value P st_min . The election of p st_min should be above the higher value of the set point of the heating steam pressure PID loop proposed by any of the crystallizer's program, taking into consideration the higher possible multiplying acceleration factor (P Fi ).…”

“…The election of p st_min should be above the higher value of the set point of the heating steam pressure PID loop proposed by any of the crystallizer's program, taking into consideration the higher possible multiplying acceleration factor (P Fi ). If constraints (38)(39) are met, the subordinate problems can be solved independently.…”

Applications of Hybrid Predictive Control

“…de Prada et al [2] describe an adaptive hybrid predictive control strategy for the crystallization section of a sugar factory where continuous and batch units operate jointly. A more efficient NLP optimization technique is proposed, instead of MINLP, and is applied to an industrial sugar factory simulator.…”