Abstract

Sunflower (Helianthus annus L.) is one of the most cultivated oil crops in the world. Seed oil of cultivated sunflower has good quality that they are important source of unsaturated oil. Various types of cleaning, grading and separation equipment are designed on the basis of the physical properties of grains or seeds. However knowing the average of seed dimensions are important in designing some apparatuses, but knowing the dimensional distribution is necessary for designing apparatuses such as sorters, separators, and planters. The aim of this study was to model the size distribution of two hybrid of sunflower (hysun33 and hysun25) dimension by three statistic models namely, Normal, Lognormal and Weibull distribution functions. For comparison the goodness of fit of the models, three statistic tests namely, Kolmogrov-Smirnov, Anderson-Darling and Chi-squared tests were used. The results showed that for hysun33 Hybrid, the length, width and thickness of seeds varied from 6.36 to 12.87, 4.02 to 7.57 and 2.65 to 5.52 mm, respectively. Corresponding values for Hysun25 variety were 7.57 to 10.99, 2.54 to 6.82 and 2.03 to 4.93 mm, respectively. According to the results, the normal function was the best model to predict the length, width and thickness distributions of the two hybrids.

Key word: Sunflower seed, physical properties, size distribution, statistic distribution.