Report 4: DSIR 16/227: 1/10/1920 to 31/12/1920

ROTARY KILN RESEARCH

In accordance with the programme for the reduction of coal used by rotary kilns, further investigations into the working of kilns and coal dryers have been made during the past three months, and drawings for the improvement of various kiln plants have been proceeded with.

CONVENTIONAL TERMS

Certain conventions are used in recording the results of experiments on rotary kilns. They are as follows:-

Unit Output - The unit output of a kiln is the clinker produced in cwt/hour, per 1,000 cubic feet of shell capacity, measured inside the lining. Thus a kiln of 5,000 cubic feet capacity which produced 85 cwt of clinker per hour, would have a unit output of 85/5 equals 17.0.

Standard Coal - All kiln test results are reduced to percentage (on clinker) of standard coal used. This is dry coal of 7,000 calories, or 12,600 B.T.U. per pound. Chimney Draught is expressed in "cents". A cent is equal to 1/100 inch of water gauge.

ENQUIRY No. 14 - ROTARY KILN 130'0" × 7'4⅝"

This enquiry relates to a kiln 130'4" long. It was referred to in the previous quarterly report.

This is APCM's Vectis plant. The Ernest Newells rotary kiln had been installed in 1913. The coal mills, referred to as pendulum mills, were Griffin mills.

General description of plant

The wet process is used.
Open type cooler under natural draught.
Two pendulum mills are used for coal grinding.
The coal dryer is of the Ruggles Coles type. It is heated by a separate furnace, fired with coke.

Kiln shell

The diameter inside the shell plates is 7'10¾" for a length of 47'6" at the hot end, the diameter of the remaining portion being 7'4⅝".
Overall length = 130'4"
Slope of Shell = 4.1%
Capacity inside lining = 4,312 ft³
Slurry lifters area = 144 ft²
The slurry lifters are of the cast iron bracket type, and are in good condition. The area is reckoned on the leading surface only.

Cooler shell

Inside diameter = 6'4"
Length = 40'1"
Slope = 7.3%
RPM = 0.72

Air passages cooler to kiln

The firebrick shoot is of ample area, this being the first kiln tested in which the chimney draught has been sufficient to draw the proper quantity of air through the cooler. The temperature of the air entering the kiln from the cooler was 537°F.

Coal Firing Apparatus

The fine coal hopper for the kiln holds 21 tons. The coal feed is provided with an overflow arrangement as indicated below. Conveyor A brings more coal from the hopper than is required by the kiln, the unused portion being returned to the hopper by the elevator indicated. Conveyor B measures out the coal under the constant head D.E. Conveyor C is not essential to the arrangement.

The measuring screw B, 2⅝" dia., 2" pitch, spiral ¾" thick. It makes about 500 rpm and should deliver, when flowing full, 16.5 lb of coal per 100 revolutions. Actually the delivery is considerably less.

Kiln speed, coal screw speed, slurry feed speed

The drive is by a variable speed motor in each case. The coal screw diagram indicated that the coal feed was fairly regular. The slurry feed is perhaps not sufficiently uniform and it would be an improvement to drive this feed from the kiln gear.

Kiln flues and chimney

The flues are of compact design, being not deeper than is necessary; they are carried on a reinforced concrete support. There is an iron chimney 4' dia. inside the firebrick lining, and 80 ft. in height, above the top of the flues. This gives a satisfactory draught.

Particulars of measurements made

Raw Coal: This was weighed out in barrow loads of 2 cwt.

Clinker: The weight of the cement produced, as bagged for shipment was accepted, suitable allowance being made for the gypsum and moisture added.

This probably resulted in an over-estimate of the output.

Draught Measurements: Draught recorders were used as follows: -
(1) In the flue, opposite the kiln exit end.
(2) In the chimney base, below the damper.
Owing to the large airway through the cooler, the draught in the kiln hood only averaged 1 cent, and this was too small to be taken on a recorder.

Temperature measurements: A continuous record was taken of the temperature of the kiln exit gases. The other temperature measurements were observed at frequent intervals.

Speed measurements: Speed recorders were driven from the kiln gear countershaft, from the coal feed screw B, and from the worm shaft of the rotary slurry feed. The diagrams obtained have already been alluded to.

Kiln Test Sheet: The more important observations are recorded on the kiln test sheet - sheet No.3.

Only the averages were used from this data sheet, which is too large to show. Available in Excel form on application: enq14_kiln_data

The chief results are as follows:-
(a) Duration of Test: 8 days
(b) Unit Output: 13.29
(c) Slurry moisture: 42.98%
(d) Fine coal, residue on 180#: 28.43%
(e) Standard coal consumption: 34.40%
(f) SCC reduced to 40% slurry and allowing for water run on to clinker rings: 32.6%
These figures do not include coal drying.

Temperature measurements, °F
(1) Clinker leaving cooler: 256
(2) Clinker leaving kiln: 2042
(3) Clinkering zone, flame temperature: 2692
(4) Clinkering zone, clinker temperature: 2447
(5) Air entering kiln from cooler: 537
(6) Waste gases at kiln exit end: 637
The waste gas temperature is the average for the test. It was considerably higher during the intervals that the kiln was clear of clinker rings. The temperatures in lines (2), (3) and (4) were taken by an optical pyrometer. These temperatures were also observed by a Féry radiation pyrometer, with the following result:-
For line (2) the temperature measured was 144°F lower.
For line (3) the temperature measured was 189°F lower.
For line (4) the temperature measured was180°F lower.
The optical pyrometer measurements are thought to be more nearly correct.

All the values are probably seriously low. Pyrometric measurements were (and are) compromised by inability to distinguish between clinker, lining, gas and dust, and by a lack of knowledge of the relevant emissivities.

Draughts, Cents
(1) In kiln hood: 1
(2) In flue, opposite kiln exit end: 20
(3) In chimney below damper: 28
The draught in line (2) varied from 9 to 36 cents according to the condition of the kiln as regards clinker rings, similarly the draught in line (3) varied from 11 to 47 cents.

Air measurements: As calculated from gas analysis observations, the air supply to the kiln (lb/min) works out as follows:-
(1) Coal: 39.2 lb/min
(2) Air required: 8.89 lb per pound of coal
(3) Excess air used: 10.00%
Hence air used = 39.2 × 8.89 × 1.10 = 384 lb/min

Of this quantity 98 lb entered the kiln from the coal firing pipe leaving 286 lb per minute to come through the cooler, and up the clinker shoot. Actually the quantity coming through the cooler, as measured by an anemometer was 448 lbs per minute, and the amount of air gauged as entering the kiln from the clinker shoot was 423 lbs per minute. Apparently the figures can only be reconciled by assuming a leakage outwards, at the joint where the kiln enters the hood, but further air measurements, and gas analysis observations are desirable, under steady running conditions.

Percentage CaCO₃ (as recorded) in slurry entering kiln: This was observed hourly. The fluctuations are usually well under 1%.

The hour-to-hour CaCO₃ 1-s variability during the test was 0.26, corresponding to an LSF variability of (at least) 1.55 and a loss-free C₃S variability of (at least) 3.43. This is apparent variability - the CaCO₃ testing 1-s precision was probably close to 0.2.

Kiln Radiation: The surface temperature of the kiln shell, at regular intervals, is given in diagram form below. The position of a clinker ring will be noted.

Chemical Measurements: The chemical department again carried out tests having for their object the determination of how much of the coal ash was mixed in with the clinker. The test of 8 days was divided into four 48 hour periods, and average samples were obtained for each of the periods. The results are as follows:-

Slurry analysis (loss-free)

Clinker analysis (loss-free)

Average analysis

It will be seen that the clinker contains too little alumina and too much iron to justify the assumption that the whole of the coal ash has combined with the clinker.

One would have expected that they would have thought twice about doing this after the stupid results of the Shoreham test. What the results do show is that (1) the lack of a mass balance including dust loss makes logical deductions impossible and (2) the analyses are probably not internally consistent.

Clinker rings and the use of sand in the coal

The experiment was tried of mixing in 5% of sand, and later 7% of sand with the raw coal. This had the effect of raising the S/R in the coal ash, from 1.15 to 1.81, the latter value being a little less than that obtaining in the clinker. The works reported that the effect of the sand was to prevent a lining being obtained in the burning zone, but it did not prevent the formation of clinker rings. They attributed this to the fact that the sand being coarse and heavier than particles of coal of equal size, was falling in and about the burning zone, and was not being carried far enough into the kiln to be effective on the clinker rings.

It was noticed from the bypass measurements that the coal came much more freely from the hopper when mixed with sand, and also irregularly at times, hence it is possible that the kiln lining was burnt off, or damaged by an excess of sand, due to uneven mixing.

An experiment might be made by grinding the sand separately, and to a greater degree of fineness.

The whole idea was misconceived and doomed to failure. The plant was using the local Hamstead Clay, which had a low silica ratio, since they lost their supply of alluvium from the Hamble which was more satisfactory chemically. The obvious remedy was to add sand (suitably ground) to the rawmix.

Heat balance for cooler - % of total heat

Heat balance for kiln, % standard coal

Summary

(1) A short cooler, large in diameter, and a large firebrick shoot, enables all the air required by the kiln to be drawn through the cooler by natural draught, the output being relatively small.
(2) The kiln coal feed is provided with a bypass arrangement which is satisfactory.
(3) There is rather a large speed variation of the motor driven slurry feed, and it is considered that a rotary feed driven from the kiln gear countershaft would be preferable.
(4) The rather poor result obtained on the test was mainly due to the formation of clinker rings, and in this connection the use of coal much more finely ground might be tried.
(5) The quantity of air entering the kiln hood (as measured) was considerably in excess of the amount deduced from the gas analysis measurement, and verification is required.
(6) The percentage of CaCO₃ in the slurry was reasonably uniform.
(7) The chemical measurements made to determine the percentage of the coal ash which combined with the clinker are not altogether consistent, but probably most of the coal ash did so combine.
(8) To obviate clinker rings 5%, and later 7% of sand, was mixed with the raw coal; apparently the sand fell short in the kiln, for reasons which are not altogether clear.

ENQUIRY No. 17: ROTARY KILN 228'6" × 8'10"

This was BPCM's Penarth plant, which had installed this FLS kiln in 1914. It was similar to Wouldham Kiln 9 with similar, but slightly smaller coal dryer and ball-and-tube coal mill system.

This enquiry relates to a rotary kiln 8'10⅝" dia and 228'6" long, working on the wet process. There is an enlarged clinkering zone 10'0¾" dia × 32'1⅝" long. The cooler is of the return-tube type, and is operated under forced draught.

The kiln dimensions are questionable. The kiln as supplied was metric, 2.7 m and 3.0 m diameter inside the shell. The burning zone diameter given here (and repeated in the 1924 plant schedule) is not an FLS dimension, but the design is as original with a flange at the hot end. It is difficult to see how a replacement shell of different diameter could have been attached to the flange, so one must assume that, although the main diameter is correct, the burning zone diameter must be external.

The coal is dried, in a rotary dryer, using hot air from the cooler.
The coal grinding plant is of the ball and tube mill type.
Previous to the test, a large leakage area where the kiln shell entered the kiln hood, was made good by a packing of the loose ring type.

Method of conducting test

The test was at first made under usual works condition. It was found however that too much excess air was being used (about 20%) and far too much air was being passed through the cooler, resulting in the low hot air chamber temperature of 375°F. The research staff then reduced the excess air to 7% and reduced the quantity of air passing through the cooler until its temperature rose to 510°F. The general working of the kiln was thereby considerably improved, and the output was increased by about 16 cwt per hour.

Summary of Results
(1) Slurry moisture: 36.87%
(2) Slurry residue on 180#: 14.4%
(3) Raw coal moisture: 10.2%
(4) Moisture in coal leaving dryer: 2.6%
(5) Fine coal moisture: 1.9%
(6) Fine coal residue on 180#: 5.4%
(7) Fine coal calorific value: 6677 (no units given, but it is gross kCal/kg, dry basis)
(8) Clinker, average output: 7.47 ton/hr
(9) Clinker, unit output: 13.3
(10) Standard coal used: 30.5%
(11) Standard coal, referred to a slurry moisture of 40%: 31.9%

Temperatures, °F
(12) In hot air chamber: 414
(13) In kiln clinkering zone, flame: 2536
(14) In kiln clinkering zone, material: 2298
(15) Kiln exit gases: 796
(16) Clinker leaving cooler: 222

Clinker cooler: It was found by trial that the cooler holds at least 4 tons of clinker, so that the time occupied in passing through is not less than 30 minutes.

Slurry feed: This of the "orifice and constant head" type. In practice the orifice area is not varied; when the kiln is on full speed it receives all the slurry leaving the feed tank by the orifice, but when the kiln goes on half speed the delivery from the orifice is deflected, by a rope operated by the burner, and about half of the slurry is diverted to the overflow pipe. Considerably too much slurry is fed to the kiln when it is running at full speed, consequently it is frequently put on to half speed, and the working of the kiln is therefore not as regular as it should be.

"Orifice and constant head" systems were installed on most early kilns, and relied upon the slurry being thin and having a constant viscosity - an unlikely prospect. There was no logical way of ensuring a constant slurry loading at different kiln speeds, and operators had to resort to desperate measures to cope with the system's inadequacies. Variations in loading led to a vicious cycle of ever-more-frequent speed changes.

Coal feed: An inspection of the coal feed screw diagrams show that the burners do not vary the rate of coal feed. The variations in the excess air used, as shown by the percentage of oxygen in the exit gases, are much greater than the variations in the coal screw speeds, indicating that the supply of coal from the hopper to the screws is not very uniform. This supposition is confirmed by an inspection of the kiln exit gas temperature chart. A continuous variation in temperature is indicated which is somewhat unusual. A full report of this test with diagrams will shortly be issued.

STANDARDISATION OF INSTRUMENTS

Pyrometer standardisation

The iron constantan thermo-electric pyrometers, and the temperature indicators and recorders have been in use about 6 years. Each pyrometer, indicator and recorder has been tested (throughout a range of temperatures) at Ingress Abbey, and a table of the corrections to be applied to each instrument has been prepared.

The pyrometers (10 in number) have an average error of 1.5° at 212°F. At 900°F, the errors range from -15°F to 2°F. The method of testing the pyrometers is as follows:- The iron-constantan pyrometer under test is immersed side by side with a standard platinum-rhodium pyrometer in a bath of molten tin, the temperature of which can be varied as required. The EMF produced by each couple is measured on a works potentiometer, and a reference to suitable tables enables the corresponding temperature to be found in each case. The readings being compared, the error of the iron-constantan pyrometer is then known.

Tests of Indicators and Recorders. These instruments are essentially millivoltmeters; they register the electromotive force produced by the thermocouples, but they are graduated to read the temperature directly, instead of the equivalent millivolts. To test them it was found most convenient to apply a series of known electromotive forces to the terminals of each instrument (the exact value in each case being determined by a works potentiometer), and the reading of the instrument was compared with the temperature equivalent to the EMF applied, as taken from the tables. Iron-Eureka thermocouples immersed in boiling water, were employed as a source of EMF, the number of couples in series being varied to give the corresponding temperatures desired.

A table of corrections has been drawn up for each instrument, generally speaking the errors are within 1% of the temperatures indicated.

ENQUIRY No.18: COAL DRYER RESEARCH

This dryer which is of the Ruggles Coles type is shown on Sheet No.5. There is a coke-fired furnace - see figs. (2) and (3) - with provision for admitting auxiliary air, at the back of the fire bridge. The hot gases are led through firebrick channels so that they do not impinge on the dryer shell, immediately they enter the outer chamber. Passing through the outer chamber, the gases traverse the two horizontal elbow pipes C, to the chamber D; they next travel through the annular portion of the dryer shell, and return by the inner tube to the chamber E, whence they pass to the dust silo, and the exhaust fan. The cascading irons on the inner and outer tube are shown on Fig.4.

Sheet 5. View HD image in a new window.

Method of Working. The dryer is used during the day only, from 6 a.m. to 9 or 10 p.m., hence the brickwork and ironwork cool down each night. There was not sufficient assistance available to test the dryer for eight days with the kiln, hence a test of 9 hours was made subsequent to the kiln test.

TABLE OF TEST RESULTS

(1) Duration of Test: 9 hours
(2) Moisture in coal before drying: 12.40%
(3) Moisture in sand before drying: 3.90%
(4) Moisture in coal and sand after drying: 6.40%
(5) Moisture evaporated: 4.69 lb/min
(6) Coal entering dryer: 40.0 raw cwt/hr
(7) Sand entering dryer: 2.0
Size after Drying:
(8) Residue on ½": 18.7%
(9) Residue on ¼": 31.7%
Temperatures:
(10) gases at back of furnace, at point A: 926°F
(11) gases impinging on dryer shell at B: 638°F
(12) gases leaving outer chamber, in pipe C: 382°F
(13) gases leaving inner tube at F: 149°F
(14) gases entering exhaust fan: 131°F
(15) coal entering dryer (partly on fire) : 103°F
(16) coal leaving dryer: 141°F
Air quantities:
(17) Quantity required to burn coke, net: 14.7 lb/min
(18) Entering ash pit: 69 lb/min
(19) At back of furnace, point A: 71 lb/min
(20) Under dryer shell, point B: 104 lb/min
(21) Leaving outer chamber, in pipe C: 136 lb/min
(22) Entering dust silo at F: 219 lb/min
(23) Entering exhaust fan: 233 lb/min
Furnace of Dryer:
(24) Moisture in coke, as received: 4.39%
(25) Dry coke used on furnace: 1.51 lb/min
(26) Calorific value of coke: 12,000 BTU/lb
Dust Silo:
(27) Coal intercepted per hour: 3.1%
(28) Time occupied by air in passing through dust silo: 27 s

Heat balance, BTU/min:

Leakages &c. The increase in air quantity, as the hot gases pass through the dryer, will be apparent from lines 18 to 23. The air quantity at A, line 19, was deduced from the coke burnt per minute, and the gas analyses at A, viz. CO₂: 4.29%, O₂: 16.30, the excess air being 350%. Similarly, at hole B, the gas analysis observations were CO₂: 2.84, O₂: 18.01, excess air: 578%.

Loss unaccounted for. The heat balance shows a large loss unaccounted for, a portion is no doubt due to closing down the dryer each night. To aid in tracing out the losses, the total sensible heat in the gases (BTU/min), in various positions, is given in the table below:

Total Stage loss

It will be seen that before reaching hole B, where the heat becomes effective, 3730 BTU/min has been lost. Also the heat given up between hole B, and the point where the gases leave the inner tube is 4120 plus 5870 = 9990 BTU/min. It appears from the heat balance that only 4950 plus 1042 = 5992 BTU/min, have been usefully expended in evaporation, and in warming the coal, leaving the balance (9993 - 5992) = 3998 BTU/min as loss by conduction and radiation.

Suggestions for Improvement

(1) The use of force draught on the furnace, and a suitable speed reduction of the induced draught fan, would enable the interior of the dryer to be worked at or near atmospheric pressure and thus reduce leakage. The air supply could also be better controlled. At present the air enters too freely at the back of the furnace owing to the brickwork bring broken away round the regulating damper.
(2) A pyrometer at point B, to enable the temperature at which the hot gases impinge on the dryer shell, to be kept at say 900°F would be useful.
(3) The use of a water spray, to reduce the loss of dust from the fan chimney should be considered.
(4) A coating of pitch and tar on the furnace brickwork would assist in preventing leakage.
(5) The elbow pipes leading from the outer chamber to the dryer interior should be lagged.

GRINDING PLANT RESEARCH

ENQUIRY No.16: Test of Pendulum Mills.

This returns to the summer's visit to Vectis. The delay was due to the tedious process of estimating surface production by elutriation. It gave useful information on the performance of Griffin mills. The Vectis mills were later replaced with a Newells combination ball mill.

This investigation was referred to in the previous Quarterly report, and a short summary of the results was given. A full report is now made. The coal grinding plant consists of two pendulum mills. They are belt driven from a common countershaft, which is in turn rope driven by a DC motor. The driving pulley of each mill can be disconnected from the countershaft, by means of a friction clutch.

Mill Dimensions, Speed &c:-
(a) Diameter of roll head 18 1/16" average
(b) Depth of roll head: 6⅛"
(c) Diameter of grinding ring: 29¾"
(d) Screen used: 50#
(e) Mill pulley speed: 209 rpm

Ventilating arrangement. A fan is used to draw off the dust from the mills, the delivery being to a reinforced concrete settling chamber of 12500 ft³ capacity which is located under the turning platform.

The coal firing fan draws its air supply from the same chamber, at approximately the same rate, viz. 100 to 120 lb/min of air. An overflow chimney is provided to the settling chamber, and any air drawn from the coal mills, in excess of the requirement of the coal firing fan, escapes through it. The actual air speed through the dust chamber, assumed uniform, is only 3.7 feet per minute; the dust deposited is 8.5 lb/min, or 12.6% of the total coal ground. The ventilating apparatus is efficient, and the use of two fans has the advantage that the coal mill can be ventilated when the kiln is not running. Also the settling chamber can be kept under a slight air pressure and in-leakage of cold air is thus prevented.

Method of Testing. The coal grinding plant was tested for 8 days with the kiln, the weight of coal ground is therefore that used on the kiln test, and the running time of each mill being observed, the quality ground per hour can be calculated. There was no method of separating out the rate of feed to each mill, so the average is taken.

Hourly samples were obtained of the coal entering the fine coal hopper, and a sieve test was made every 6 hours, on an average sample.

Running Times & Output &c
Date of Test: 20 to 28/8/1920
(1) Mill No.1 running time: 62.89 hr
(2) Mill No.1 power: 32.0 BHP
(3) Mill No.2 running time: 38.50 hr
(4) Mill No.2 power: 23.3 BHP
(5) Average output per mill: 1.77 ton/hr
(6) Residue on 180# 28.4%
(7) Residue on 100# 8.5%
(8) Moisture in fine coal: 4.7%

Power measurements. A detail of the method employed to obtain the BHP used by the motor, the countershaft, and each mill is given below:-
(1) Motor, armature resistance as measured: 0.114 Ω
(2) Motor, field resistance as measured 115.7 Ω

Motor running light:
(3) Current: 8.5 A
(4) EMF: 244.0 V
(5) Motor speed: 476.0 rpm
(6) Field current: 2.1 A
(7) Armature C²R loss: 4.7 W
Field C²R loss: 510.2 W
Total C²R loss: 514.9 W
(8) Power supplied (8.5 × 244.0): 2074.0 W
(9) Constant losses due to friction, hysteresis, windage &c. = 2074.0 - 514.9 = 1559.1 W

Motor and Countershaft:
(10) Current: 40.9 A
(11) EMF: 241.2 V
(12) Motor speed: 472 rpm
(13) Countershaft speed: 131 rpm
(14) Armature C²R loss: 171.6 W
Field C²R loss: 504.3 W
Constant loss (line 9): 1559.1 W
(15) Total losses: 2235.1 W
(16) Power supplied (40.9 × 241.2): 9865.1 W
(17) Power to drive countershaft = 9865.1 - 2235.1 = 7630.0 W = 10.23 HP

Mill No.1 with countershaft:
(18) Current: 149.2 A
(19) EMF: 241.6 V
(20) Countershaft speed: 132 rpm
(21) Mill speed: 210 rpm
(22) Armature C²R loss: 2466.8 W
Field C²R loss: 505.2 W
Constant loss (line 9): 1559.1 W
Total losses: 4531.1 W
(23) Power supplied (149.2 × 241.6): 36046.7 W
(24) Power to drive mill & countershaft = 36046.7 - 4531.3 = 31515.4 W = 42.25 HP
(25) Power Mill No.1 only = 42.25 - 10.23 = 32.02 HP

Mill No.2 with countershaft:
(18) Current: 118.15 A
(19) EMF: 242.2 V
(20) Countershaft speed: 131 rpm
(21) Mill speed: 208 rpm
(22) Armature C²R loss: 1535.3 W
Field C²R loss: 506.5 W
Constant loss (line 9): 1559.3 W
Total losses: 3600.9 W
(23) Power supplied (118.15 × 242.2): 28615.9 W
(24) Power to drive mill & countershaft = 28615.9 - 3600.9 = 25015.1 W = 33.53 HP
(25) Power Mill No.1 only = 33.53 - 10.23 = 23.30 HP

The subdivision of the BHP expended is as follows:-

Calculation of New Surface Produced

The method of calculating the nets surface produced has been previously described, and it was shown that:
Surface of cubes in grade = 0.825 × W / S
where W = weight of coal in grade
S = average particle width in grade

The new surface produced per 100 lb of coal, is 179,751 ft2. The average quantity ground per hour was 1.77 tons for an expenditure of 28.73 BHP hence the new surface produced per BHP hour is 179751 × 22.40 × 1.77 / 28.73 = 248,000 ft².

Comparison of Tests. The results so far obtained from test of coal grinding mills may be compared as follows:-

The figures show the pendulum mill to be very efficient as a preliminary grinder.

The tedious tables showing the surface calculation have been left out. The finished fine coal had 1.88% on 76#, 6.62% on 100# and 19.9% on 180#.

TESTS ON 18" EXPERIMENTAL MILL

At the request of Mr. Chas. Blyth, comparative tests on 1" dia. steel balls, cylpebs, and holpebs, have recently been made. The charge used in each case occupied approximately 40% of the mill volume. l" steel balls = 2.35 oz: Cylpebs = 2.15 oz: Holpebs = 1.57.

Charles Edward Blyth of Stockton appears to be the only person "at whose request" work was done. He was the token non-Blue Circle person on the Association's board. Of course, all the other work was at the anonymous request of Blue Circle.

The mill lining plates were fitted with lifter bars.

From the above figures additional results are calculated, all figures below relate to grinding to 5% on 180#.

In these tests the Holpebs are somewhat at a disadvantage as a 300 lb charge cannot be got into the mill.

Tests on 18" mill using smooth lining plates. In tests previously reported, the lining plates were fitted with lifter bars; at this point smooth lining plates were fitted to the mill. Generally speaking, when using sand as the material to be ground, it is found that the smooth plates are preferable.

In the tests about to be described the charge occupied in each case approximately 40% of the mill volume, or 42% of the diameter. The material put into the mill at the beginning of each test was 35 lb of Standard Sand - 20# + 30#.

The grinding media used were l¾" and l" steel balls, Cylpebs, large flints, medium flints, and small flints.

A summary of all the results obtained is given on Sheet No.6.

I have not bothered to digitise this table, but here it is, if anyone wants to do it.

Selection of best result in each grade

It will be seen that the 1" dia. steel balls gave the best result; also that the small flints are considerably more efficient than the large flints. The small flints ground more sand per hour, with a smaller expenditure of BHP hr per ton ground. An inspection of Sheet No.7 also shows that the small flints were preferable even for the coarser stages of the grinding.

Sheet 7 - life's too short to bother showing this boring graph. Suffice it to say, as the grinding continues, the sand gets finer. The salient results are in the above table.

Col. (1) shows that a relatively low speed is preferable cylpebs.

As seen from Col. (3), the BHP hr per ton ground required by the flints is relatively high. This is partly due to the fact that the frictional losses of the mill gearing, belt, bearings etc., are nearly the same with flints as with steel balls, but much less useful work is done by the flints, owing to the lighter charge, hence the frictional losses are large in proportion to the useful work, and the efficiency of grinding is reduced.

With a view of placing the flints on the same footing as steel balls, as regards grinding efficiency, measurements were made of the BHP required to run the 18" mill when empty, at various speeds. The additional power required to run the mill under load is made up of the actual work done in grinding, plus the extra bearing friction due to the weight of the charge, the latter being very small. Col. 4 above gives the BHP hr per ton ground after the HP expended in mill friction has been deducted.

Looked at in this manner it will be seen that the small flints are somewhat better than cylpebs.

Sedimentation tests of sand as ground by steel balls, Cylpebs and flints. A large number of tests have been made using standard sand in the 18" mill, and it was considered desirable to examine a few samples of the ground material by Hall's sedimentation test, and to work out the new surface produced per BHP hour.

A typical sedimentation test is given on page 25. It will be noticed that as compared with coal there is much less fine material in the ground sand.

The formula used is surface in grade = W × 4393 / S
W = weight of sand in grade in lb
S = side of cubes in grade in units of 0.0001".

The sides of the cubes in each grade are the averages of about 600 microscopic measurements, the size so determined being accepted as standard for all tests made with sand.

Taking the figures overleaf, which are given for tests Nos. 30, 45 and 50, it is found that the relative efficiency for 1" steel balls : medium flints : small flints, is 100 : 57.7 : 84,7, when the calculation is based on the BHP hr per ton ground to 5% on 180#, and 100 : 57.5 : 75.5 when the calculation is based on the new surface produced per BHP hour. These results are with mill friction deducted.

Hence, the results given by the two methods are fairly consistent.

The result of six sedimentation tests on sand, and the grinding efficiency calculations based on them, are given in the table following:-

Tests Nos.25, 26 and 27 were made with lifter bars on the lining plates; tests Nos. 30, 45 and 50 were made with smooth lining plates. Columns (3) and (4) show that the proportion of "flour" is approximately the same, whether the grinding is done by steel balls, cylpebs, or flints. The tests were partly undertaken to find out if flints produced more "flour" than steel grinding media. Apparently this is not the case.

There followed an example sedimentation calculation of no interest included probably to acknowledge the inordinate amount of laboratory time required to produce these data.

Tests on 18" experimental mill with partial charges of 1" steel balls and Cylpebs

The tests hereafter described were made to determine the best speed, and the grinding efficiency of a tube mill when loaded with a partial charge only of 1" steel balls, or cylpebs. Such conditions would probably obtain when a flint stone charge is removed from a tube mill, and iron grinding media substituted.

Early tube mills, particularly those of FLS, used flint media. Their conversion to steel media, which have three times the density, was still in progress at this time.

The normal charge in the 18" mill for steel balls and Cylpebs is 300 lb, the partial charges used were of 225 lb and 150 lb respectively, the quantity of sand used being proportionately reduced in each case.

Charge 225 lb steel balls. Refer Sheet No.8.

Comparing lines 8 and 12 it will be seen that a partial charge admits of a wide speed variation without much loss of efficiency, a speed of 58 rpm corresponding to a speed constant of 245 giving slightly the best result.

Charge 150 lb steel balls. This charge is too small for maximum efficiency. The speed may vary through a fairly wide range. There was a sudden drop of efficiency at 60.5 rpm but this is close to the critical speed which is 63 rpm. At the latter speed the outer ring of balls would adhere to the inner circumference of the mill during the complete revolution.

Tests of Cylpebs. The results are shown on Sheet No.8. They are generally the same as for steel bells, but the efficiency is less.

I have not bothered to digitise this table, but here it is, if anyone wants to do it.

SUMMARY

The table on Sheet 8 does not give the results when using a normal charge of 300 lb, as these figures have been reported previously. For the sake of comparison the best results both for a charge weight of 300 lbs, and for smaller charges, are given in the following table:-

It appears that a charge volume of 40.5% is somewhat too large for general practice, on account of the speed limitation imposed. A charge of 30.4% gives nearly the same output, at relatively less BHP and the speed may vary between wide limits.

In this case it will be seen that a charge volume of 30.4% gives the best result. This would occupy 34% of the mill diameter. When flints are replaced in a tubemill by steel balls or cylpebs, it is apparently preferable to contract the charge volume 1ongitudinally, in order that 34% of the mill diameter or thereabouts may be covered by the grinding media.

New Surface Calculations. It is convenient to put the formula Previously given for calculating particle surface in such a manner that variations in the specific gravity of the materials can be readily allowed for. The formula then becomes:-

Surface of cubes in grade = 1.154 W/Sg sq. feet
where W = weight of material in grade, in lb
S = average particle width in inches
g = specific gravity of material, referred to water.

For some coals, g = 1.40 and the formula becomes:-
Surface = 0.824 W/S sq. feet

For standard sand, g = 2.628 and the formula becomes:-
Surface = 0.439 W/S sq. feet

The relative density of acid-washed Leighton Buzzard sand at 15°C does not vary outside the range 2.647-2.651. Raw sand is somewhat denser. The value given gives an indication of the quality of density determination. Clearly, air had not been eliminated.

Chemical Department

Setting Time Research

The work introduced in Report 2 continued. The remarks on the work given there continued to hold. The precision was too poor to glean any valuable information. I have left out the extensive data, and just give the conclusions, such as they were.

Three samples: No.4 (rotary), No.5 (shaft) and No.6 (chamber kiln) have been subjected to the action of dry and moist carbon dioxide and dry oxygen. No.4 cement has also been subjected to dry and moist ozone, to ascertain the effect both on setting time and soundness.

To ascertain the effect of ozone on the soundness of cement, a sample was specially obtained, which was stated. to expand 30 mm in boiling water. On its arrival, however, it was found to be not beyond the permissible limits for soundness. Portions of this cement were exposed to dry and moist ozone respectively and tested by the Le Chatelier method.

This bizarre investigation was conducted, presumably, because someone wanted to buy an ozoniser.

In the gas exposure tests, by "constant water" test, the initial and final setting times were all accelerated by from 40% to 71%. The normal consistency tests in two cases show a retardation of setting time, but all the cements, after treatment, required an increased percentage of water to gauge them, and previous experiments have shown that increase of gauging water gives a slower setting time.

On exposure to moist carbon dioxide the cements absorbed carbon dioxide and water in the following proportions:-

The setting times were accelerated to a remarkable degree, the slow-setting No.4 cement being rendered very quick. Nos.5 and 6 cements were probably rendered quicker than the figures indicate, as it was strongly suspected that the initial set was worked into during the process of rapid gauging. No figures are given for No.5 cement (normal consistency) as in this case there is no doubt that the cement initially set during the gauging.

Exposure to dry oxygen had no material effect upon the setting time, in this respect agreeing with the effect of pure dry air. Some differences in the actual number of minutes are recorded but this is due to the imperfections of the standard test for setting time. As great, or greater differences have been repeatedly observed in the same sample of cement, tested on different days under standard conditions of temperature, amount of gauging water and humidity. Hence it is necessary to interpret the figures for setting time rather broadly.

Exposure to dry ozone had no effect on the setting time of the one sample treated. Moist ozone retarded the setting time, with an absorption of .17% of water. Neither dry nor moist ozone had any appreciable effect on the soundness of the cement.

Effect on setting time of small additions of soluble salts

Experiments have been made with a number of soluble salts of the alkali metals, using the 6 cements of Class (b) (regulated by steam and gypsum).

Where salts containing water of crystallization were employed, the combined water was allowed for, and the amount of addition stated represents anhydrous material.

In the first instance, the salt was dissolved in the water used for gauging the cement, but it was later considered better to grind the salt with the cement, and thoroughly mix before testing. In all cases the cement alone was tested side by side, under identical conditions, with the samples containing additions of salts.

The carbonates of sodium, potassium and ammonium, when dissolved in the gauging water, have a strong accelerating action on the setting time of cement. The figures indicate a slightly lengthened setting time with the higher percentages of salt, but it must be noted that with these percentages an increased quantity of water was generally required for gauging to a plastic consistency, and the slight retardation is therefore only apparent, not real.

Ammonium hydroxide, up to 1% of ammonia (NH₃) had little or no influence on the setting time of cement. Beyond 1%, there was a slight acceleration.

The sulphates of sodium and magnesium (containing water of crystallization) had, broadly speaking, no material effect on the character of the setting time of cement. The tendency of sodium sulphate, however, was to accelerate the setting time, and of magnesium sulphate to retard it. The sulphates of potassium and ammonium (containing no water of crystallization) had an accelerating effect, very slight in the case of ammonium sulphate, but distinctly marked in the case of potassium sulphate.

In the experiments described below, the various salts were all ground into the dry cement, instead of being dissolved in the water employed for gauging.

Both anhydrous and crystallised sodium carbonate accelerated the setting time of all cements to a marked degree. Ammonium sulphate had no definite effect except that the final setting time of the chamber kiln cement was retarded. Potassium sulphate accelerated the setting times of all the cements, particularly marked in the rotary and chamber kiln samples.

Anhydrous sodium sulphate, on the other hand, had no very definite effect on either of the cements; but crystallized sodium sulphate (containing 10 molecules of combined water) had a decided retarding effect on the Schneider and chamber kiln cements although its action on the rotary cement was indefinite.

Anhydrous magnesium sulphate, like potassium sulphate, accelerated all cements. Crystallized magnesium sulphate, however, decidedly retarded the setting times of the Schneider and chamber kiln cements. The figures for the rotary cement are rather erratic.

The action of various soluble salts in accelerating or retarding the setting time of cement, has hitherto been ascribed to "catalysis". No means have yet been discovered of determining exactly what takes place, but it appears that in the case of ammonium salts, at any rate, a definite chemical reaction occurs, inasmuch as all the ammonium salts so far experimented upon are decomposed in the act of grinding or mixing with the cement, ammonia (NH₃) being evolved in quantity. The experiments are being continued, with the carbonates, chlorides and nitrates of the alkali metals, before passing on to the soluble salts of other groups.

This is beginning to look like a job for life. As any entry-level chemistry student knows, one releases an amine from its salts by treating it with an alkali, such as calcium hydroxide.