Linear PI control of batch exothermic reactors with temperature measurement

Abstract

SUMMARYA wide variety of speciality materials and fine chemicals such as plastics, pharmaceutical and microelectronics components are produced in batch reactors. The nonlinear, transient and finite-time features of the batch reactors give rise to complex process and control design problems. In particular, the safe operation of exothermic reactors depends on the adequate functioning of a temperature tracking controller, and to a good extent, the same is true for the attainment of a suitable compromise between productivity and product quality attributes. While the stabilization problem of continuous exothermic chemical reactors has been recently addressed with rigorous asymptotic-stability methods, the same kind of studies have not yet been performed for the finite-time batch reactor case. In this paper, the problem of designing a temperature tracking controller for an exothermic batch reactor, with n species and m reactions, is addressed under the following premises: (i) only the reactor temperature is measured, (ii) the (typically uncertain) reaction rate and heat exchange nonlinear functions are unknown, (iii) the controller must be linear and easy to tune, and (iv) the closed-loop reactor motion must be stable in a suitable sense. The combination of industrial-oriented inventory control concepts in conjunction with singular perturbation results yields a linear controller with a combined feedforward-PI feedback structure, antireset windup scheme, and conventional-like tuning rules. The controller: (i) tracks, arbitrarily fast and close, a prescribed temperature trajectory, with admissibly deviated concentration motions, and (ii) quickly recovers the behaviour of an exact model-based nonlinear I/O linearizing controller. The proposed design is put in perspective with the geometric and IMC nonlinear control approaches.

Operation range extensionviahot-spot control for catalytic CO2methanation reactors

“…Other studies focused on reactor control aspects such as robustness, uncertainties, Kalman filters, state observer, and optimal control. [51][52][53][54][55] Forced-periodic operation is not the objective of this work as it is continuously stressing process units and actuators which often leads to serious doubts in its technical feasibility and scale-up potential. 13 Instead, process control is an essential and already implemented component of many technical applications, such that the realization of stabilizing control requires just little effort in comparison to its achievable process improvements.…”