The investigation of lamella classification process

Author: J. Marciniak-Kowalska

The paper provides a mathematical model of grain classification processes targeted at prognostication of grain size distribution on the basis of grain distribution in the feed.

Predicted (theoretical) values of the cut size and separation sharpness indices (imperfection and coefficients k sub 1 and k sub 2 are presented. Results of grain size classification in a coal suspension in a single lamella unit are duly compiled. Measurement results are compared with predicted values.

The mathematical description of the grain size distribution in the feed and final product uses Weibull's distribution (Rosin-Rammlcr--Bcnnett's distribution) as a singular case of the generalised gamma distribution. Grain sizes obtained experimentally are compared with predicted values. The experimental grade efficiency curves are approximated with several functions.

The cut sizes are derived from curves approximating the experimental data set. On that basis experimental cut sizes are obtained and then compared with predicted values. Distribution functions of grain size in classification products prognosticated on the basis of size distribution in the feed agree pretty well with the values obtained from mathematical modelling. The comparison of predicted and experimental values lead us to the conclusion that cut size obtained in experimental classification processes was nearly two times bigger than the predicted value.

Separation sharpness indices obtained experimentally arc similar to predicted values. In conclusion, the results of experimental verification of the developed model are regarded as satisfactory. The mathematical model of classification processes in lamella settling tanks agrees sufficiently well with experimental results.

While compared to other classification apparatuses (centrifugal separators, hydrocyclones), the lamella tank seems to provide for better separation sharpness.