Modelling the Rheology of Olive Paste for Oil Extraction Plant Automation: Effects of the Crushing Process on the Rheology of Olive Pastes

Abstract

In extra virgin olive oil production, it is essential to obtain a well-prepared olive paste which allows not only the extraction of the oil drops from the olives, but also the achievement of a high-quality oil while maintaining high yields. This work addresses the problem of determining the effect of three crushing machines on the viscosity of the olive paste: a hammer crusher, a disk crusher and a de-stoner were tested. The tests were repeated on both the paste leaving each machine and the paste to which water was added; this was done with the main aim of considering the different dilutions of the paste while entering the decanter. A power law and the Zhang and Evans model were used to analyse the rheological behaviour of the paste. The experimental results allow validation of the two models with a high (more than 0.9) coefficient of determination between experimental and numerical data. The results also show that the pastes obtained with the two classic crushing methods (hammers and disks) are almost identical, with a packing factor of about 17.9% and 18.6%, respectively. Conversely, the paste obtained with the de-stoner entails higher viscosity values and a smaller solid packing factor, of about 2.8%. At 30% dilution with water, the volume of the solid concentration dropped to about 11.6% for the hammer and disc crushers, while for the de-stoner it only reached 1.8%. This behaviour is also reflected in the evaluation of yields, which were 6% lower with the de-stoner. No significant differences regarding the legal parameters of oil quality were found using the three different crushing systems. Finally, this paper establishes some fundamental pillars in the research for an optimal model for identifying the rheological behaviour of the paste as a function of the crusher used. Indeed, since there is an increasing need for automation in the oil extraction process, these models can be of great help in optimizing this process.