APPLICATION OF ARTIFICIAL NEURAL NETWORKS TO ESTIMATE PERMEATED FLOW IN BIPHASIC MIXTURES OF WATER AND OLEIC ACID

• DOI: 10.22533/at.ed.3172222203109

• Palavras-chave: Artificial neural networks. Tangential microfiltration. Oleic acid. Learning algorithm.

• Keywords: Artificial neural networks. Tangential microfiltration. Oleic acid. Learning algorithm.

• Abstract:

  The açaí (Euterpe oleracea) is a natural fruit from the Amazon, which has high nutritional values, being a food that is gaining national and international attention, because from it, it is possible to extract açaí juice. To guarantee the quality of the juice during storage, it is necessary to clarify it, thus facilitating other conservation processes and avoiding the generation of turbidity. The clarification process can be carried out by tangential microfiltration using ceramic membranes. The use of this method for clarification has many benefits, among them are less energy use, in addition to not having to do thermal and chemical treatment. However, in addition to the benefits, this process also has the disadvantage of decreasing the permeate flux as a function of time, due to the obstruction of the membrane pores. In view of the above, this work presents a neural model capable of estimating the permeate flux of biphasic mixtures of water and oleic acid, from açaí, which can obstruct the pores of the membranes. The neural model results were satisfactory with a mean percentage error of 9.4%.