Browsing by Subject "regression analysis"

Now showing 1 - 1 of 1

Access status: Open Access ,
Regresyjny model procesu klasyfikacji ziaren skrajnie drobnych
(Wydawnictwa AGH, 2009) Tajchman, Zbigniew
The paper presents the issues relied to the description and the basic technological factors of flowing classification process, which main medium were suspensions of chosen model materials ultra fine particles. After conduction of basic process factors evaluation, which originated mainly from the Tromp curve, the ones seemed to be the most appropriate to evaluate the process were chosen for the further research. Three model materials were selected: quartz glass; quartzite; barite. Furthermore, three changeable parameters of flowing classification process conduction were chosen, which were: temperature; volumetric concentration of the suspension; concentration of hydrogen ions of process conduction environment. The results of chosen process parameters significant influence on elementary phenomena occurring during particles (of sized up to 60 mim) separation process were presented, as well the results of viscosity and suspension stability. The described state of elementary phenomena recognition being the part of classification process determines unequivocally the state of models construction of this process. Their general feature is deterministic approach joined with introduction of significant simplifications. This is caused mainly by lack of many elementary phenomena measuring possibilities during the process conduction. The constant coefficients introduced to the equation, which, however, do not allow the unequivocal interpretation of all of the phenomena occurring during the classification process. This do not solve the complexity of the problem because the random interactions should be considered in process description. The randomness of the process causes many difficulties in determination of correlations between individual elementary phenomena and also in precise identification of full group of factors influencing on course and results of the process. Only in recent years, the random interactions are introduced in various range to existed deterministic process models.