Table of Contents

Many attempts have been made during the last forty years to evaluate the performance of gravity separation equipment, that is, the effectiveness with which light and heavy particles are separated. The most comprehensive treatment of the subject was made at the 1st International Conference on Coal Preparation held in Paris (1950) by Cerchar.. The float and sink analysis of the product is presented on a Gaussian distribution curve, resulting in an easier visualization of the inherent difficulties of separation. The gives of the distribution curve are then plotted, giving a quantitative measure of the deviation from perfect separation as an error distance instead of an error area.

**Heavy Media Separation (HMS) Performance Criteria**

In the ideal HMS process, all material lower in density than the specific gravity of separation (S.G.S.) would be recovered as floats and all material of higher density would appear as sinks.

In order to evaluate the misplaced material, the washery products are tested at the density at which the washing unit is operated. The original type of plot is shown and was developed primarily for coal cleaning units. The curve for raw coal represents the cumulative percentages of sink material. The refuse curve is also plotted as cumulative sink, the percentages being expressed in terms of raw coal. This diagrammatic representation of the results of washing units has the merit of easy visual observance of the degree of separation obtained. The error areas (cross hatched) are a measure of the amount of misplaced material and therefore they can be used to characterize the quality of the separation.

The ideal and actual separating performance between floats and sinks can be best seen from the partition curve developed by Tromp, where the ordinate is the percentage recovery of the sinks,, and the abscissa is the specific gravity. It can be seen from the shape of the curve that as the S.G.S. is approached, the proportion of material reporting to the improper product increases rapidly. In fact, the S.G.S. can be defined as the density of the material in the feed that is distributed equally between float and sink products.

**Theory of Evaluation and Representation of Washery Performance**

The proposed theory may be applied to determine the specific gravity of separation and the ideal separator efficiency with a certain feed, or to evaluate the actual washery performance by an analysis of the float and sink products. The discussion is related mainly to good (sinks) and deleterious (floats) of a gravel feed to an HMS plant, but it appears that the application can be extended to coal or any other minerals which are separated by gravity methods.

The area under the intersections of the two curves (A B C D) is due in part to the inherent specific gravity composition of the feed and also to the imperfect separation process. Point C may be defined as the specific gravity of separation, and for minimum quantities of misplaced materials, this can be the same as the specific gravity of the medium in the separator itself.

E is defined as the minimum vertical distance between curves f (g) and s(g), and it can be shown that when E is a minimum, the error area between the two weight distribution curves is also a minimum. E is designated as the index of separation and is given in percentage units:

E = f (g) – s (g)…………………………………………….(1)

The separation index (I) expressed as a percentage deviation from the maximum value is:

I = 100 – E

The separation index (IF) obtained now is the theoretical optimum which could only be achieved in the perfect heavy liquid test and it might best be called the theoretical perfect separation. The relationship between the different separation indices is represented by equation.

machine index = I – IF

IM as a percentage = (100 – IF))……………………………………………….(5)

**Separation Index for an HMS Process with Gravel Feed**

Before the float-and-sink analysis, the samples were handpicked into good and deleterious fractions. Figure 7 shows the distribution and cumulative distribution curves by specific gravity for the feed. The yields of float and sink fractions were 17.24 percent and 82.75 percent, respectively. The cumulative distribution curves show that the best possible separation index of 96 percent (100 – 4 = 96 percent). The inherent wide range of specific gravity of both types of particles in the feed does not permit a better separation than that obtained in this test.

Here the minimum distance, E, between the two lines at S.G.S. of 2.52 is 8; therefore, the index of the separation can be calculated by equation (5).

Machine Index = 100 – (8 – 4) = 96 percent

Here it was taken into account that the separator could not separate the gravel feed better than the float-and-sink test; therefore, the index is not 92 percent but 96 percent. How it would be interesting to compare the indices of separation as accomplished in two different vessels. This comparison is useful for evaluating the performance of different vessels with the same feed material.