INFLUENCE OF THE GRANULAR MIXTURE PARTICLES SIZES AND OSCILLATION FREQUENCY ON THE SEPARATION PROCESS AT THE CONICAL SIEVES WITH OSCILLATING MOVEMENT

The paper presents the results of experimental researches on the working process and his efficiency for a sieve with external conical surface with alternative periodic motion. The sieve is suspended at the top and bottom with three elastic threads (of silk), with thread diameter 1.5 mm and is provided with circular apertures of 4.2 mm. Vibratory motion was obtained with an eccentric mechanism, placed horizontally and acting on tangential direction to the sieve at adjustable distance. . Was thus obtained amplitude oscillation of 3.58, 3.74, 3.91 and 4.10 mm. The oscillation frequencies used were 250, 520 and 790 osc/min. The sieve was used for sorting of canola seeds having sizes between 1.25 mm and 2.5 mm, in percent of 95 %. It was found that seeds are separated closer to the feeding point of sieve the more so as the oscillation frequency is smaller. Seeds of a size less than 1.5 mm is separated close to the axis of rotation. Also, the sieve orifices can properly calibrate a seeds mix for all oscillation frequencies analyzed.


INTRODUCTION
Based on the data in the table and the determinations performed, were plotted the distribution curves of material on conical sieve generatrix (according to the number of seeds collected under the sieve and beyond the edge thereof).In order for a granular mixture can be sifted and sorted into fractions using sieves is required to exist a state of sifting of the material, so that is a relative movement between the particles and the separation surface.In another order of ideas for the particles may be sieved, two conditions must be met: first, the particles must reach on the aperture, which is achieved by the relative movement of these on the separation surface and, second, particles have sufficient time to be involved in passing through apertures [1][2][3][4][5].
A mixture of seeds (grains, technical plants, oilseeds, etc.), after being cleansed of impurities, is required to be sorted by size, because each fraction should be directed to a particular technological processing route or may be given a different destination (the more developed seeds can be used as seeding material, seeds with a lower degree of development can be routed for industrial processing, poorly developed seeds are used for animal feed or disposed of as waste, etc.).Before the processing of seeds and the use of it, either in milling industry, either as seeding material, they go through a complex process of cleaning and conditioning to ensure that all foreign bodies are removed from their mass [6].
Sorting of seeds by the particle size is carried out, generally, into separation blocks with oscillating movement.Sifting and sorting on sieve requires relative movement of the material particles on the separation surface, which is performed either by a corresponding inclination of the sieves, either by the oscillation movement of the sieve blocks [3].Separation surfaces may have different shapes (plane, conical, cylindrical, parabolic), and oscillatory movement is obtained using mechanisms of different construction, either crank drive (with eccentric), or with the aid of unbalanced rotating masses, or with other mechanisms that ensure the required oscillating movement.Also, separation surfaces can be realized in the form of metal or textile braids or in the form of perforated sheets with apertures, ordinarily, elongated or circular.
Experimental research on the process of separation and sorting on sieve with oscillatory motion have been made by numerous researchers worldwide.Thus, the separation process of material on the sieve is influenced by the amplitude and frequency of oscillation, which provide the relative movement of material on separation surfaces.Other parameters that influence the separation are: sieve slope, angles of internal and external friction of the material, sieve apertures, average particle size [4,7,8].Clogging of sieve apertures depends on the particle size, phenomenon investigated by R. Feller in papers [6,9].Mechanical classification of seeds with sieves is not a complete classification, each time in a fraction being found 10-20% from particles of fraction that should have separate [10].
In our paper, are presented the results of experimental research on a sieve with perforated sheet with a external conical separation surface, with angle of inclination of 8 o , provided with circular apertures with a diameter of 4.2 mm, used to study the separation and sorting of the canola seeds, the sieve may be used to separate impurities from the seeds mass.

MATERIAL AND METHOD
Conical sieve is suspended with steel cables with a diameter 1.5 mm, and actuating in oscillating movement (circular alternative) was made by an electric motor through a worm adjustment and an oscillating slide.Actuation was done at a distance d (adjustable) from sieve center (in the rest position), being rigidly connected by the edge of sieve, radially, and a link arm 2 (Figure 1).Fig. 1.Scheme of conical sieve used in experiments: 1-sieve with circular apertures; 2-actuating arm; 3-material collecting box; 4-actuating mechanism with worm wheel and oscillating slide, 5-elastic cables.
These has been linked to the lever of the oscillating slide through a coupler with helical spring in order to reduce the oscillation of sieve in radial direction of the lever 2, at the ends of the stroke of slide and.Thus, the amplitude of oscillation could be adjusted, in four steps, by changing the distance d, to obtain sieve amplitude oscillation of 3.58, 3.74, 3.91 and 4.10 mm.Also, sieve oscillation frequency could be adjusted with actuating mechanism, being used three frequencies of oscillation: 250, 520 and 790 osc/min.Diameter of conical sieve, at the bottom thereof, is of 430 mm, and the density of the circular aperture on the separation surface is 2.25 aperture/cm 2 .Length of suspension wire is l 1 = 240 mm, and l 2 = 180 mm.A detailed description of the construction of the conical sieve is shown in the papers [2,3,11].For experiments were used canola seeds (harvested in the south of Romania) after a storage period of about six months.The material sample mass was always the same 0.5 kg seeds.Canola seeds humidity was about 7.65-8.05%, determined using a thermobalance Partner MAC 110, at a drying temperature of 105 o C. Seeds had sizes between 1.25-2.5 mm, (in a proportion of more than 95 %).Seeds had sizes between 1.25-2.5 mm, (in a proportion of more than 95 %) determination which has been performed using a sieve shaker for granulometric analysis VAPO model, with sieve aperture 125 μm, and 250 μm.Were carried out determinations at three oscillation frequency, four sieve oscillation amplitudes and three different feed rates, obtained by the position above or below of outflow opening of feed funnel, (Figure 1).The three feed rates used were: 0.20, 0.33 and 0.42 kg/s canola seeds.The outflow opening of the feeding funnel has a diameter of 25 mm.
Collecting of canola seeds under the sieve was performed in a box with concentric circular compartments, having diameters: 80, 140, 200, 260, 320, 410 and 460 mm.Notification of beginning and end of material feeding was made visual, starting and stopping each time a stopwatch with digital display.
It was followed, in the main, distribution of canola seeds collected under the sieve and loss of seeds that has passed the edge of the sieve.To estimate the coefficient of sorting seeds were performed experimental determinations by sorting on fractions of the material collected in each of the boxes under the sieve to identify the fraction with the highest percentage in the box (sorting was done in four size classes: 1.0-1.25 mm; 1.25-1.6mm; 1.6-2 mm; 2-2.5 mm).To sort were used sieves from accessories of the sieve shaker.

RESULTS AND DISCUSSION
Some of the results obtained in experiments (only for A=3.91 mm, feeding rate Q=0.033 kg/s and those three oscillation frequency) are presented in Table 1.
With values from Table 1 were drawn the distribution curves of seeds separated on sieve generatrix for the three values of oscillation frequency.
Estimation of the influence of oscillation frequency on the sorting process of seeds on sieve was performed by regression analysis of the experimental data (in %) with Gaussian distribution function, for which have been determined the correlation coefficient χ 2 and R 2 values: where: p x represents the weight percentage of separated material over a range (radius) of sieve length; e.g. between 0-r 1 , r 1 -r 2 , r 2 -r 3 etc.); y 0 , A, x c and w -experimental coefficients.
In relation (1) A represents the maximum percentage of material collected in the boxes under the sieve, x c radius of the sieve base corresponding to the maximum percentage of separated seeds (or average values determined), and w is compared to the maximum position.These coefficients depend on the working regime parameters and are presented in Table 2. Charts of the regression function, for the four size class of the seeds and the three oscillation frequencies, are shown in Figure 2. First, the graphs show the distribution of percentage of seeds on the four classes of sizes.It is observed so that the largest percentage share in the mixture of sorting seeds has class of sizes 1.6 -2 mm, followed by sizes class above 2 mm.The lowest percentage shares have sizes classes 1.25 -1.6 mm and 1.25 mm below.Placing of the peak position of curves on horizontal is but different (not on the same vertical).
From the analysis of graphs it is found approximately the same allure of distribution curves for each of the four size classes with that proper to general mixture of seeds.Parameter values x c shows that at higher oscillation frequency separation is more quickly, while at oscillation frequency F 1 =250 osc/min (small frequency) all four fractions is separated away from the center of the sieve.
Still, for a better sorting by size would require a lower frequency as shown in the charts presented, at which the maximum position of the distribution curve increases with decreasing of seeds sizes.At none of feeding rate and amplitude of oscillations, for the three oscillation frequencies used in experiments, was not found a good sorting on fractions of seeds with different sizes.Although the maximum of the separation curve is modified less for the fractions mentioned, in each compartment of the collecting box were found each time seeds of other fractions.It follows that, in the actual form of conical sieve (with apertures of the same dimensions on the entire surface), its use as equipment for sorting on fractions of different sizes is not effective.Is, so, necessary to perform apertures of different sizes on the cone generatrix, from larger to smaller sizes, from top to bottom.In Figure 3 are graphically represented variations of peak positions of the separations curves for each size class at those three oscillation frequency of sieve.
It is found that for lower oscillation frequencies, seeds of small size classes are separated before the larger size classes of seeds.At higher oscillation frequency (790 osc/min) separation of seeds becomes random possible due to entry into resonance of sieve and of elastic system for clamping of actuating mechanism lever.For size class 2.2 -2.5 mm seeds are separated closer to the feeding point of sieve as the oscillation frequency is smaller.This it is found otherwise for all size classes with a more pronounced character for the first two size classes (1 -1.25 mm; 1.25 -1.6 mm).
If at oscillation frequency of 250 osc/min seeds from the third size class separates closer to the vertical axis of the sieve compared to other size classes at the other two oscillation frequency (520, 790 osc/min) separation of seeds begin to depart from the vertical axis of oscillation.
Therefore we can say that for seeds with sizes under 1.5 mm separation is made closer to the axis of rotation for used sieve and that the size of sieve aperture can properly calibrate the seeds mixture for all oscillation frequency analyzed.If it is desired an appropriate calibration for larger classes is necessary either increasing of aperture diameter, either use of sieves with apertures of stepwise increasing, either reduce of oscillation frequency.It should also be said that the inclination angle has a great influence on the movement of seeds, them rolling on the inclined surface of the sieve as the angle is greater.Perhaps if the angle of sieve with the horizontal would be smaller, the sorting degree of seeds would improve.

CONCLUSIONS
Conical sieve with oscillatory movement around the axis of the cone can be used successfully for separation of impurities from the mixture of cereals crops, technical or legumes if the separation surface is appropriately selected.Regarding the use of conical sieves for sorting by size, researches from this paper were not very conclusive, but if working regime parameters (amplitude, frequency and feeding rate) are chosen properly can be obtained good results.
It is necessary so, further research for lower oscillation frequencies and lower inclination angles of sieve in the case of sorting canola seeds on experimented sieve with apertures and also seeds of other crops.

Fig. 2 .
Fig. 2. The regression curves of experimental data with normal distribution function.

Fig. 3 .
Fig. 3. Variations in peak positions of separation curves for each size class at those three oscillation frequency.

Table 1 .
The amounts of separated seeds on the sieve radius base for A=3.91mm, Q =0.033 kg/s at three different oscillation frequencies.

Table 2 .
Regression coefficients, respectively of correlation, obtained from regression analysis.