Non linear receding horizon observer and model based predictive control of evaporators in cane sugar industry

The control of the evaporating process is of the highest importance in the sugar industry, on one hand because of the influence of its behaviour on the regularity of the sugar quality and on the other hand because of the large energy involvement. Based on the thought of the concrete industrial problem, and in order to improve the existing control law performances, we have been using some advanced tools, which should be well adapted to industrial processes. This paper deals with the design of an observer of the unmeasured state variables which are needed for the control law, and its insertion in an advanced control structure. A nonlinear predictive control law, based on a knowledge model which has been validated with data collected from a specific industrial plant, has been established. Some simulation results illustrate the performances of the global control system.