A common method of analyzing experimental data is to determine the distributional model which best describes the process under study. In this paper, theoretical statistical models discussed by Tarcolea et al. (2008) are corroborated with data for the cleaning system of a combine harvester, data obtained experimentally in laboratory conditions. The purpose of this paper is to illustrate how some of the continuous distributions can be used for describing the variation separation intensity of seeds on sieve length. The Pearson coefficients show that some curves are far from the normal distribution, and better fits can be obtained with other distributions which can describe more adequately different degrees of skewness and peakedness of the curves. The considered probability laws are: normal, gamma, Weibull and beta distributions. The best results were obtained with gamma and beta distributions, since, for example, the values of the correlation coefficient R2 are in the most of the corresponding cases close to 1.

combine harvester; cleaning system; seed separation; normal; Weibull; gamma and beta distributions

Published by University of Zagreb, Faculty of Agriculture, Zagreb, CROATIA
Sponsored by Ministry Science and Education of the Republic of Croatia