In underwater pelletizer, different factors in the process can impact the pellet quality and the good operability. Coming from a previous stage, under very high temperature and pressure conditions, molten thermoplastic is forced to pass through a die with numerous holes of circular profile. When the molten thermoplastic passes through the die holes it is cut by a set of knives attached to a rotatory piece (a.k.a. knives-holder). Immediately the pellets get in contact with water and solidify. After, water and pellets are conveyed to a drier, where the pellets are separated from the water stream. The water is pumped back to the pelletizer chamber, in a close circuit. Poor water hydrodynamics in the interface between the die and the knives-holder front face (where the set of knives are attached) leads to an insufficient temperature drop to ensure the pellet solidification and its proper formation, causing it to stick to the knives, to the knives-holder or to each other’s, forming agglomerates and/or die holes plug. Such issues, consequently, lead to maintenance stops and/or equipment breakdown.
Using Computational Fluid Dynamics (CFD), the present work provides a consistent analysis of the implementation of an apparatus, called Diffuser, to improve the hydrodynamics within the underwater pelletizer. The Diffuser main function, from the process perspective, is to drive a large amount of process cooling water against the interface between the die and the knives-holder front face. More water in this region tends to cool the pellet more quickly. In addition, the Diffuser promotes a shorter water residence time in the cutting and rotational regions. Consequently, the pellets, driving by the water, leave those regions faster, minimizing the probability of touch each other’s while are with its surface cooling down.
Using a transient single-phase Eulerian approach, the work focused in evaluate the most suitable Diffuser geometry align with the best process conditions possible to use. To evaluate the performance of each design, the amount of water passing through the pumping holes and the water residence time within the cutting and rotational regions were defined as variables of interest. The results showed that the amount of water, when compared the case with and without the Diffuser, jumps from 59% to 77%. In terms of local residence time, at the rotational and cutting regions, the numbers are good as well. The local mean water residence time is reduced by two seconds, meaning that the water and pellets leave those regions faster than the case without the use of Diffuser.
Also, the study shows that optimal conditions can be reached for the cases evaluated, such as increase the rotation in 16% and reducing the number of knives in 17%. Playing with those two variables we can increase the so called ‘pumping effect’, which is the effect that leads more water to the interface of the die-plate and the front face of knives-holder.