ISO 9276-4 pdf download – Representation of results of particle sizeanalysis — Part 4: Characterization of a classification process

ISO 9276-4 pdf download – Representation of results of particle sizeanalysis — Part 4: Characterization of a classification process.
4 The influence of systematic errors on the determination of grade efficiency curve
4.1 General
Systematic deviations from the ideal course of a grade efficiency curve can be caused by:
a) systematic analytical errors of sampling and sample splitting;
b) superposition of a classification and a splitting process within the classifier;
c) undispersed agglomerated fine partictes which are transferred to the coarse product;
d) comminution of the feed material in the classifier.
Assuming that the sampling from and the sample splitlrng of the samples of the teed material and the fine and coarse fractions have been performed with the greatest of care, the first systematic error listed above may disregarded
4.2 Systematic error due to a splitting process in the classifier
If the grade efficiency curve, as shown in Figure 7. does not &op to zero at small particle sizes, but ends parallel to the abscissa at a certain grade efficiency T r. ii is most likely that the classification pocess is supenmposed by a splitting process.
Applying the correction to the grade efficiency curves of Figure A. I • one obtains the solid lines of Figure A.3. if
aparl from the error dist,dutions, .,2(4, only one set of results of particle size analysis has been used.
The correction smoothes the grade efficiency curves especially well in the coarse size range. The results can be improved further, if the corrected grade efficiency curve is calculated from multiple analyses of the feed material and the tine and the coarse fractions.
Figure A.4 shows the final result, based on 21 analyses of each product.
The dotted line represents the average grade efficiency curve calculated from the basic data without the correction applied. While the tine end appears to be correct, the coarse end neither roaches one. nor follows the expected course. The full line shows the grade efficiency curve obtained wTth multiple analyses and an application of the suggested correction. It could be shown, that the confidence interval of this curve is of the order of the thickness of the full line shown in Figure A.4.