The crystallization is a unitary operation that consists on organic and inorganic chemical purifying by liquid - solid separation, where mass is transferred from a dissolved solute in a liquid phase to a solid phase (crystal pure); One of the key attributes of particulate systems is the co-presence of a continuous phase and a dispersed phase, which leads to the occurrence of physiochemical phenomena such as nucleation, growth and birth-death rates of the crystals. The driving force in order to occur the nucleation and the growth rates of the crystals is the supersaturation, therefore these phenomena has not any possibility to occur in a saturated or unsaturated solution.

In this work, an operation alternative was studied by simulation of a stirred tank continuous isothermic crystallizer through of its dynamic characterization. The process model is based on the moments technique, which has a less complex solution than to the population balance model (PBM); since the last one regularly presents complex distributed parameters models of solving mathematically. This technique has the advantage of being used by many investigators in crystallization in the last years. The methodological process consisted on the implementation of a SISO (single input-single output) feedback servocontrol algorithm and a PID (proportional-integral-derivative) controller. The tuning was carried out with the strategies of internal model control (IMC). The process variable to control was the average diameter in % volume D(4,3) = M4/M3, derivate of moments of the crystal size distribution (CSD) because is a useful variable in both commercial and industrial environment; the manipulate variable was the solute concentration in the feeder.

From obtained results, it is outstanding the stationary state analysis of the response from the process to different residence times (tao), where is observed that to tao < 0.6 hr., the crystallizer that generally has an oscillatory behavior (cycles limits); it has a trend toward a stable and controllable state. Simulation runs of the close-loop system clearly demonstrate that the implemented controller reached in smaller time (2 hrs) the set point value of (D(3,4) = 1,000 micron) than the open-loop process (approximately 10 hrs). With this control action, both the offset and the initial oscillations were eliminated; a good response was presented in the pursuit from the positive and negative step changes on set point. It is concluded that the system is functional; however it presents two inflection points, what makes it unstable in these zones.