Improvement of sunflower oil extraction by modelling and simulation

Dorina Brǎtfǎlean, Vasile Mircea Cristea, Paul Şerban Agachi, Dan Florin Irimie

Research output: Contribution to journalArticlepeer-review

9 Citations (Scopus)


The paper describes the results obtained by modelling and simulation of sunflower oil extraction. The methods used for lipid extraction have been the conventional ISO extraction Soxhlet method (SO) and the mechanical oil press method (PO). These methods have been investigated on 14 Roumanian genotypes of sunflower seeds (10 inbred lines and 4 sunflower hybrids). Genetic material was obtained from the Agricultural Research and Development Institute Fundulea, Roumania. The sunflower seeds variability proved to be responsible for the quantity of the extracted oil. A difference in the oil extraction yield has been detected between Soxhlet and mechanical extraction. Compared to the mechanical press extraction, the Soxhlet oil extraction yield is higher. The diffusion coefficient through the seed, mainly depending on particle pores, has been fitted with experimental observations in the SO extraction. The yield of the oil extraction by oil press method has been investigated as depending on a set of parameters such as: feed stream, mechanical energy, speed of rotation, moisture and particle size. The oil press extraction yield has a large variability among the 14 investigated sunflower genotypes. It ranges from the minimum value of 66.3% for the LC-L3 type to the maximum value of 94.2% for the LC-L10 type. Based on the quantitative analysis of the extracted fat, the study reveals the fact that assessment of the sunflower seed genotype depends on the extraction methods, PO or SO, used in the investigation.

Original languageEnglish
Pages (from-to)881-888
Number of pages8
JournalRevue Roumaine de Chimie
Issue number9
Publication statusPublished - Sept 1 2008

All Science Journal Classification (ASJC) codes

  • Chemistry(all)


Dive into the research topics of 'Improvement of sunflower oil extraction by modelling and simulation'. Together they form a unique fingerprint.

Cite this