DETERMINATION OF JUICE STORAGE CELLS BROKEN
4 0
I
49 during each successive cycle. Further, Graph I I I indicated that a 15-second cycle is too short, a 30- second cycle giving better results.
The maximum degree of vacuum required was investigated, b u t invariably during normal operation at Illovo factory, only 50 to 56 cm Hg of vacuum could be obtained.
Table II indicates that a 30-sccon.d evacuation cycle, with the vacuum increasing from 0 to 50 cm of Hg in steps of approximately 2 cm of Hg vacuum per cycle, gives optimum results.
Tests also show t h a t a higher final degree of vacuum (70 cm of Hg) gives greater extraction on samples with higher free sucrose content (crusher and first mill bagasse), and not much difference on final, mill bagasse samples. Here again, the bagasse par- ticle size m a y be retarding extraction, or perhaps sucrose is being extracted from closed cells. This is a m a t t e r for further investigation.
Ultimate Test Procedure To summarise:
(1) Only very gentle oscillatory motion was given to the suspended apparatus during test.
(2) Apparatus was well shaken during final evacua- tion cycle.
(2) Evacuation increased in steps of approximately 2 cm of Hg per evacuation cycle, from zero.
(4) Evacuation cycle time of approximately 30 seconds taken.
(5) Maximum degree of vacuum of 50 cm of Hg applied; giving an overall time of at least I2|- minutes.
From the graphs it can ben seen t h a t the free pol so obtained may be equal to and not greater t h a n the true free pol. In this case then, the percentage of broken cells calculated from these results is less t h a n or equal to the true value, where broken cells refers to the definition previously given.
Graph I I I does not appear to support the above summarised method, but it will be appreciated t h a t these tests were made for a different purpose; samples were withdrawn at various intervals throughout the test, and before sampling, 300 to 500 ml of extract was withdrawn and returned to the vessel. This could easily have upset the conditions for maximum extraction. Also the apparatus could not be shaken to mix the contents before sub-sampling. Hence, more reliance is placed on the tests in Table II for giving optimum conditions, though this should be confirmed in further tests.
Application of Method to Bagasse from Consecutive Mill Units
Tests were then carried out on the Illovo milling train where samples of cane were followed through each mill and a sample taken at each unit. It was found that the percentage broken cells increased with, each unit; sub-samples of a particular mill bagasse did not differ by more than .1. per cent in the number of cells broken, except for the crusher sample, where sub-sampling and pneumatic extraction are in- accurate. Also, it is seen from Table I I I A t h a t the fibre content of the bagasse has very little effect on the calculated broken, cells. These results in Tables
I I I A and I I I B indicate only the difference between mill units since the sucrose per cent cane value was taken from t h e factory laboratory data, and need not necessarily represent the sucrose in the cane, which actually passed through the mill at the time of t h e test.
Mill tests were then carried out, in which the mill was run. for two hours, enabling a mill balance to be carried out. Samples of bagasse from each unit were taken throughout the run by the factory staff under
"wet-milling" conditions. These samples were then sub-sampled and tested in the factory laboratory.
The writer also received sub-samples, which were tested for free pol in all cases, and total pol and fibre in some cases. The results are given in Table IV in fuller detail.
Under "mill balance" in Table IV, all details except free pol are taken from factory laboratory data. In each case, however, the pol per cent bagasse was recalculated, taking the natural fibre content of the bagasse into account, as opposed to the so-called
"normal weight" method of determining pol in bagasse.
Under "S.M.R.I. Analysis" in Table IV, the fibre and pol content of the bagasse was determined by the high-speed extractor methods.
Broken cells per cent total cells is shown in Column 11.
In Column 12, a figure called "Imbibition Effi- ciency" is included, where—
"Imbibition efficiency"
/ p e r cent extraction by tandem up to u n i t \ . „ft
— \ per cent broken cells /
— per cent sucrose available due to broken cells extracted by the tandem up to that unit.
This then is a measure of the imbibition efficiency.
In other words, if "imbibition efficiency" is 100 per cent, i.e. all the free sucrose in bagasse was extracted, then the extraction by tandem per cent sucrose in cane is identically equal to the per cent broken cells,
51 i.e. the sucrose content of the broken cells, which was originally present in the cane per cent sucrose in cane.
In Column 13, pol in closed cells per cent pol in bagasse= pol % bagasse — free pol % bagasse
pol % bagasse 100 This figure decreases from 100 per cent for cane to about 35 per cent at the first mill, and increases to about 65 per cent for the fifth mill bagasse.
Effect of Bagasse Particle Size on Number of Cells Opened
The question arose as to how the broken cells varied between the different sized particles of bagasse.
One test was fully carried out on final bagasse in which the fibre content and total pol of the bagasse was determined by the high-speed extractor method
—see Table V. The per cent broken cells was 97.5 for the bagasse and increased from 95.2 per cent for the larger particles to 99.9 per cent for the very fine particles. This alone is evidence that the pneumatic extraction method cannot be far out on the free pol in final bagasse determination.
Conclusion
As seen in Table IV, the method given for deter- mining the number of broken juice storage cells gives a good guide to the milling efficiency, where this refers simply to the mechanical action of breaking cells. The operative word is "guide", since the ulti- m a t e object of any such test is to determine accurately t h e number of cells broken at the end of the milling train. To ascertain this figure, more work will have to be done on the determination of free sucrose in bagasse by pneumatic extraction and other methods.
However, the figures given for broken cells in Table IV may be a minimum, since the true free pol per cent bagasse is not less than, but may be equal to the value ascertained. This will be clear on examining equation (5).
It should be noted, however, t h a t not all the juice in cells broken by the last mill is available for extrac-
tion. Therefore, the number of cells broken by the tandem up to and including the last unit but one, becomes the main criterion of the overall milling efficiency.
Acknowledgments
The tests were carried out at Illovo Sugar Estates L t d factory, and the writer thanks the management for their assistance, and for all the facilities given him. Thanks are also due to Mr. Draeger, Chief Chemist at the factory, who personally supervised t h e mill tests, and Mr. K. Gilmour, who assisted with t h e laboratory tests.
References
1 Khainovsky, V. (1920): I'roc. I.S.S.C.T. 3 437.
- Kerr, [I. \V. (Uir>7): Proc. Q.S.S.C.T., 20 221.
3 "Fibro-vascular Water of Sugar Cane," Technical Reports, 0 and 23 of the Sugar Research Institute, Mackay.
4"Determination of Certain Qualities of Individual Cane Con- signments," IV, S.M.R.I. Quarterly Bulletin No. 5, p. 11.
•' "Determination of Certain Qualities of Individual Cane Con- signments," III, S.M.R.I. Quarterly Bulletin, No. 3, p. 22.
TABLE I
Pneumatic Extraction of Bagasse
JO g Bagasse -f 10,000 g Water -'r Xg 5 per cent w/w Na, Co;i solution. pH reading given after extraction
of free pol.
Bagasse Sample
Waddell Shredded cane
Crusher 1st Mill 2nd Mill 3rd Mill 4th Mill 5th Mill
W T Ka2 CO,
Soln.
(X)
20 g 20 g 20 g 20 g 20 g 20 g 20 g
p H
— 7.5 6.8 6.8 6.7 6.8 7.0
WT Na2 COa
Soln.
(X)
30 g 30 g 30 g 30 g 30 g 30 g 30 g
PH
7.8 8 . 0 , 8 . 0 6 . 9 , 7 . 1 7 . 0 , 7 . 1 7 . 0 , 7 . 5 7 . 2 , 7 . 7 7.5, 8.2 30 g 5 per cent w/w Na2 C03 solution used.
56 Mr. Dick wished to know why it was necessary to assume that all the cells were of equal size, because under a microscope they did not appear to be so.
Mr. Young replied t h a t the basic assumption was necessary to give practical value to the determin- ations. He referred to the last paragraph at the bottom of page 2, and said t h a t had the assumptions not been made, this statement would not be true.
Mr. Rault asked if this method could serve as a basis for demonstrating whether a mill was working efficiently or not.
Dr. Douwes Dekker said t h a t only in the case of very low percentage of unbroken cells could the method be used to point to inefficiency of mills, but we did also find t h a t one could get quite a low extraction for a high proportion of broken cells.
Mr. Rault said that they had found by increasing imbibition they did not necessarily get a better extraction.
Dr. Douwes Dekker said t h a t if this method was applied to various mills in Natal we would see if this method could give better results. In replying to Mr. Rault he said that from many calculations made by himself he found that a mixture of imbi- bition and residual juice was far from perfect. If complete mixing could be assured after each mill the loss in bagasse could be reduced down to one third of what it is now.
Mr. Royston said it would appear that shredding was not particularly effective in opening the cells as it seemed to shred mostly along the length of the cane.
Dr. Douwes Dekker said that when the method was applied to mills it would be possible to determine how effective the shredder actually is.
Mr. Rault inquired if the high temperature maceration tried with the Nobel carrier in Java, was now abandoned, and if so, was t h a t due to high steam requirements.
Dr. Douwes Dekker agreed t h a t the abandonment of this system was due to high steam consumption.
Mr. Boyes said t h a t a great deal of reliance had been placed on the high speed extractor. He wished to know if this could be used for determination of sucrose in normal mill bagasse.
Mr. Young said t h a t only in the case of those tests headed S.M.R.I, analysis was the high speed extractor used. It could only be used if great care were exercised.
Dr. Douwes Dekker said that at the moment he was not quite sure if this method could be recom- mended to the industry. It had obvious advantages in taking a much shorter time t h a n the boiling method, but it remained to be seen if it worked accurately enough for general application to mill bagasse. He hoped that Mr. Boyes would be in a position to compare the two methods next year.
58 of P on K is also comparatively small, no significant error is made by assuming that P is the total pressure exerted by the top roller.
3. Formula (5) shows the relationship between K, C, f, n, P and F at any moment of the crushing operation in which under normal conditions the various variables are likely to vary all the time.
We aim, however, at a constant F in spite of the normal variations in fibre throughput and formula (5) shows how this can be achieved. In the first place we must keep P constant, which is done (assuming the mill is provided with hydraulics or toggles) by not allowing the top roller to occupy either the highest or the lowest possible position, i.e. the top roller must float. More precisely for reasons to be explained under 7, we must aim at keeping the top roller continuously at a pre- determined position somewhere in between the extremes. In the third place we must aim at keeping the fibre throughput as given by Cf constant. This, however, is not very well possible and hence fluctua- tions will occur. They will show as variations in the lift of the top roller but then they can adequately be compensated for by regulating n. Formula (5) shows clearly that by regulating the rev's of the rollers, it is theoretically feasible to compensate for not too big variations in Cf in such a way that K and F remain constant, assuming the mill is pro- vided with normal hydraulics which keep P constant.
4. If we have a mill of mill constant M, expected to crush consistently at a fibre rate of Cf tons per hour and to discharge bagasse containing 100F per cent fibre; if in addition the mill engineer has decided to apply pressure P on the top roller and to rotate the t o p roller on the average at n revolutions per minute, formula (5) shows what the radial opening K between the top and discharge rollers will be under these conditions. This radial opening, usually called the work opening, is not equal to the set opening because as we have said above we try to keep the top roller at a pre-determined position between the extremes which are possible, and the set opening is the radial opening when the top roller is at its lowest position.
Usually the pre-determined position of the top roller is t h a t where its lift is 30 per cent of the total possible lift. This position is selected because—
(a) it allows a considerable further lift if a foreign object passes the mill
(b) it allows sufficient play in the position of the top roller when the regulation of the speed of the rollers does not compensate completely for all the variations in Cf which consistently occur.
5. To explain how, after we have calculated with the aid of formula (5) the work opening K commen- surate with the average conditions under which our mill will operate, K can be reduced to S, i.e. the
corresponding set opening, we will assume in the first instance t h a t the rollers are smooth cylinders. In this case the set opening S is found from K by apply- ing the formula:
S = K — 0 . 8 L (6) where L is 30 per cent of the total possible lift of
the top roller.
(N.B.—Appendix I shows how formula (6) has been developed.)
6. Normally roller surfaces are grooved and the volume of the grooves have to be taken into account when the mills are set. In Appendix II it is explained how the smallest diameter of rollers (i.e. measured in the grooves) is corrected into a hypothetical diameter corresponding to hypothetical surfaces between which the hypothetical distance should be equal to S. Appendix II also briefly describes a convenient method of using the ccnter-to-center distance of the rollers as the basis of the actual setting operation.
7. Under 3 it was said that the reasons why we should aim at keeping the top roller as much as possible at a pre-determined position was to be explained later on. We will do this now. The maxi- jum lift of the top roller in a modern mill is £ in.
to l£ in. and hence the lift when t h e mill operates
"normally" is, say, 0 . 3 X I i n . = 0 . 3 in. This means that the calculated K has to be reduced by 0 . 8 x 0 . 3 = 0 . 2 4 in. to find S, the set opening. For example, if—
C = 5 0 tons of cane/hr. D = 32 in.
f = 0 . 1 5 L = 6 0 in. and F = 0 . 5 0 P =450,000 lbs.
n =3 r.p.m.
we find K = 0 . 6 4 i n . , using formula (5).
Thus, S = 0 . 6 4 - 0 . 2 4 = 0 . 4 in.
As said in the beginning of this paper, the feed opening of a mill is found by multiplying the dis- charge opening with an empirical factor, i.e. a factor which has shown to yield optimum results if main- tained throughout the crushing operation, for example 1.9.
This means that the feed opening must be 1.9 X the discharge opening when the lift of the top roller is 30 per cent of the maximum lift b u t it means also that at any other lift the ratio is not longer 1.9, i.e.
the mill is not working under optimum conditions.
Fig. I shows how the ratio between feed and dis- charge openings changes when the lift increases from 0 to 100 per cent of its maximum value, being 1.9 at 30 per cent lift. This change is considerable and can definitely affect the milling result, hence should be restricted to a minimum. Another reason for endeavouring to keep the top roller at its pre- determined position is t h a t the trash plate has been
Gl
The actual relationship between V and F as found in 1940 for the first and respective last units of J a v a tandems is shown by two groups of dots of which the lower one represents No. 1 mills and the top one the last mills.
The curves DE and HG are representative of the average behaviour of the No. 1 and last mills respec- tively. T h e deviation of the empirical curves DE and HG from the theoretical curve AB is due to Re-absorption. It will, however, also be noticed t h a t a few last mills show higher values for V t h a n is theoretically possible. This must be explained by assuming that not only the juice but also the fibre particles in these last mills travelled faster t h a n t h e peripheral speed of the top roller. In other words: a positive slip.2
As said above the curves DE and HG indicate t h a t a higher F was not obtained when V was raised to over 35 lbs. fibre/cu. ft. e.v. in the case of No. 1 mills a n d to over 55 lbs. fibre/cu. ft. e.v. in the case of t h e last mills. We might call these values the o p t i m u m values of V for No. 1 and last mills. As such t h e y can be used for the calculation of K using formula (7a).
2 T h e possibility of the fibre travelling faster t h a n the peri- pheral speed of the top roller has already been demonstrated in the Technical Report No. 41 of the Sugar Research Institute, Mackay.
It has to be appreciated t h a t work openings calcu- lated with formula (10) and using the target F values of Table II are based on the performance of J a v a mills, crushing J a v a cane. The performance of a mill depends to a large extent on the Re-absorption phenomenon which in its t u r n should depend a.o.
on the nature of the fibre of the cane crushed. We have no means of comparing the nature of the fibre of J a v a cane with, let us say, Natal cane, and we must admit that for this reason a difference in Re- absorption between J a v a and corresponding Natal mills is a distinct possibility. The work openings calculated with formula (10) and Table II should therefore be regarded merely as an initial guidance when the object is to establish as correctly as possible the work openings of, say, a Natal tandem. They should be adhered to as long as an adequate milling control has not shown t h a t modified settings yield a superior result.
It is, however, not always the setting of the dis- charge opening which is responsible for an inferior result. The overall performance of an individual mill depends also on the setting of the feed opening and on the position of the trash plate. Obtaining maximum results is for this reason a complicated problem in which the correct setting of the discharge opening is nevertheless the most important factor.
The setting of the feed opening is usually derived
62
from the discharge opening and the position of the trash plate from the feed opening. Most mill experts seem to agree that depending on the place of the unit in the tandem and on the presence or not of a 2-roller crusher, the feed opening can be found from the discharge opening by multiplication by a factor of which the magnitude is apparently based on experience only. Table 111 gives these factors—or mill ratio's—as recommended by various authors:
Hugot
Maxwell
J a v a 1940 (average) Dr. Kerr
T 2-roller crusher
W i t h . . . Without W i t h . . . Without With Without
ABLE
1st mill
2.0 2.5 2.2 2.6 2.5 2.5 Between 1.
I I I 2nd mill
2.0 2.0 1.9 1.9 1.9 1.9 /een 1
3rd mill
2.0 2.0 1.9 1.9 1.8 1.8 .75 an
4th mill
2.0 2.0 1.9 1.9 1.9 1.7
m
mill 2.0 2.0
1.9 1.6 d 1.5%
For the setting of trash plate Hugot and also Maxwell gives a method developed by Miiller von Czemicky which is also generally used in Java. To explain this method we refer to Fig. III.
A drawing to scale is made of the rollers, the top roller being in its lowest position where the radial distance between them equals the set opening S.
The toe of the trash plate isnow found by drawing a line under an angle of 13° with the line connecting the centres of the top and feed roller, the former, however, being in its normal working position, i.e.
lifted over 30 per cent of its maximum lift. The curve of the trash plate is a circle and to allow for a gradually increasing opening between top roller and trash plate, the centre of the circle is located in point A on Fig. I l l , i.e. at a distance D from the vertical through the centre of the top roller and on the horizontal line through this center when the top roller is in its working position.
D is found from formula D=-?— (R+s), where p = per cent drop over the length of the trash
plate
R=radius of the top roller s = setting of feed roller.
Normally p is 4 or 5, or D i s — ^ - or '
25 20
Had we based the position of the trash plate either on the top roller being in its lowest, set, position or
Fig. I l l
on the top roller being lifted over the maximum distance, we would have found a not inconsiderable difference. This is illustrated in Fig. IV which is an enlarged part of Fig. III. In this figure T gives the former position of the trash plate, T1 the latter.
The correct position based on 30 per cent lift is somewhere between T and T1.3
It should be clear from the above considerations that the feed opening through the mill ratio depends on the discharge opening, further that the position of the trash plate depends on the feed opening, and the lift of the top roller. This narrow correlation between the various variables makes it the more important to do our utmost to get the best possible value of the discharge openings. The openings to start a tandem with can be found as already ex- plained, but from then on the correctness of the openings should regularly be checked.
In our opinion the best way to provide the engineer in charge with adequate data on which he can base a decision regarding whether or not to alter his settings is to initiate a system of milling control which is somewhat more complete than the one at present in force at many Natal mills.
It entails preparing Weekly Data Sheets showing the required details of each unit. These sheets have to be filed and will in due time provide an invaluable wealth of information on the performance of such a unit when working under varying conditions.
More specifically they will show what is the best factor to be substituted for factor 167 in formula K=~-—--,, or in other words they will show what
nDLF
is the best practical ratio between V and F. Java experience accepts that for that country ==110 gave V the best results but for Natal this may be different.
If the above suggested system of milling control will
3 Fig. IV is approximately 40 per cent of t h e true size.