very familiar with the rock, judges its composition with fair accuracy from its appearance, and regulates the mixture on the basis of this judgment. Needless to say, the method demands constant vigilance and careful quarry supervision.

METHODS OF OVERCOMING FLUCTUATIONS.

One fairly effective method of meeting the difficulty is to run the mixed and pulverized material into a large storage bin before it enters the kilns. If a supply sufficient for eight or nine hours' run is maintained the operator may compensate for any undesirable variation. Thus, if the mixture is found to be too high in calcium, an excess of clay or shale may be added for a time sufficient to reduce the average calcium content to the desired point. This method can succeed only where proper mixing of the materials may be brought about in the bin.

A more accurate method of mixing is followed by one Oklahoma plant where the limestone is sampled before the shale is added. The limestone is crushed, dried, and ground, and then by means of an intermittent sampler a sample is taken about every 65 seconds. All the samples for a given period of time are mixed, an average analysis obtained, and the proper proportion of shale calculated and added by weight. As clay stripping is shot down and loaded with the rock in the quarry, a small amount of shale only is added. The mixture is then pulverized in tube mills. The product of five tube mills is all conducted to a single bin, from which it is distributed to the kilns. Such mixing gives greater uniformity to the product.

THE BEST METHODS OBSERVED.

An admirable method observed of making a dry mix is successfully used by a company in the Middle West. Blasting in churn-drill holes and steam-shovel loading are employed. A clay overburden of good quality is shot down and loaded with the rock, and on this account the addition of shale is unnecessary. The rock is crushed, dried, and ground in a ball mill to 14-mesh size. The ground material is then taken to storage bins, and during transit a continuous sampler takes a representative sample, which is analyzed every 3 hours. A record is kept of the rate of entry, the number of hours that rock is discharged into the bin, and the average composition for each 3-hour period. From this data a fairly accurate analysis of the entire contents of the bin may be deduced. The ratio of silica and alumina to calcium is calculated by Newberry's formula that carbonate of lime silicaX5+alumina×2. For example, if the estimated analysis of the entire bin shows 12 per cent silica, 7 per cent alumina, and 72 per cent calcium carbonate, the desired percentage of calcium car

bonate would be 74. As only 72 per cent calcium carbonate is present, the contents of this bin would be expressed as -2. Twelve such bins are maintained, and each is designated by a plus or minus figure, which indicates the excess or shortage of calcium carbonate.

In drawing off the material it is taken from not less than three bins at once, and a combination of bins is selected such that the + and quantities will give a zero result. The first run is sampled and its composition approved before the mixture is allowed to go to the tube mills for final grinding. This method has advantages over methods of mixing pure shale and limestone. In the latter methods temporary obstruction of the flow of shale or limestone decidedly affects the composition of the mixture, whereas with the method outlined, the contents of the various bins differ so little in composition that the effect of obstructing any one outlet for a brief period would be small. For plants where shale is mixed with limestone, it is, therefore, desirable to add the shale before the material enters the storage bins.

An improved type of bin, having for its object more complete mixing of the materials, was observed by the writer at an Illinois plant. Four bins are provided for storing rock after it is crushed, dried, and ground to 20 mesh. Each bin is divided into four parts by vertical partitions. The parts are filled in succession, but the material is drawn off from all four simultaneously. This tends to equalize any fluctuations in composition. Uniformity in the final mixture is further promoted by having sufficient storage room at the kilns for 3 or 4 hours' supply.

Another improvement suggested is a simple device for proportioning the amount of material from each bin. It consists of a series of sprocket-driven conveyors, each having three speeds. Thus, if material is to be taken from three bins in the proportion of 1 to 2 to 3, the conveyor for the first bin could be operated at slow speed, the second at intermediate, and the third at high speed.

Another company quarrying limestone interbedded with shale maintains a two weeks' supply of rock. The proportions of shale and limestone in the ledge are such that if the entire face is thrown down as a single bench and the mass thoroughly mixed, the resulting mixture has approximately the proper composition for cement. The large storage bin assists in obtaining a uniform mixture. The rock is drawn from the storage bin, pulverized, and conducted to 16 concrete tanks. Continuous samples are taken, the samples being analyzed every hour. From the data thus obtained the average analysis of the material in each tank is determined. The material is drawn from one, two, three, or a maximum of four openings in each tank, and a combination of tanks is determined that will give the desired mixture.

One New York plant maintains six storage bins. The material may be drawn from each bin by two worm feeds, each controlled by a 3-speed cone pulley. Thus six rates of removal from each bin are possible.

a

Meade suggests that the raw material, after it is passed through ball mills, be stored in tanks, the composition of the contents of each tank being slightly higher in lime than is desired. The material is sampled as it enters the tank and an analysis calculated for the entire mass. It is then an easy matter to calculate the small proportion of clay or shale that must be added to bring the mixture to the proper composition. At least three bins should be maintained for the mixture and one for shale.

A slight modification of this method is proposed by an Illinois cement company. The chemist of this company has found that the composition of the cinders from the boiler room is approximately the same as of the shale used in the cement mixture. It has been proposed, therefore, to run the mixture a little higher in calcium. than is desired for cement, and to correct the mixture by means of an auxiliary cinder feed. By this means any fluctuation in composition could be corrected immediately, whereas by the method used at the time of the writer's visit the correction was made about 24 hours after the defective mixture has gone by.

THE SEMIWET PROCESS.

DESCRIPTION OF THE PROCESS.

Fluctuations in the composition of raw materials may also be overcome successfully by use of the semiwet process, also called the semidry process. Both terms are misnomers, as the materials are just as wet as in the wet process. The term "semiwet" implies the subsequent wetting of materials originally dry. The advantages claimed for the semiwet process are intimate mixture of the particles and ease of obtaining a uniform mixture of the cement constituents. This latter phase of the process will be considered here. The process is best described by reference to a concrete example.

At an Iowa plant where the semiwet process is successfully employed, the raw materials consist of limestone and shale. The limestone is variable in composition, the calcium carbonate content varying from 72 to 92 per cent. The rock is shot down by blasts in churndrill holes and is loaded with two steam shovels. The quarry operators aim to keep one shovel working in the high-calcium and the other in the low-calcium stone, and thus equalize the product to

some extent.

'Meade, R. K., Portland cement, 2d ed., 1911, p. 103.

The rock is crushed and passed through a ball mill, after which ground shale and water are added. The mixture is pulverized wet. The slurry, containing about 40 per cent of water, is conveyed by screw feed through horizontal troughs to storage tanks. Seven tanks are maintained, each having a capacity of about 600 barrels of cement. Samples are taken from the troughs with a continuous sampler and analyzed every hour. The composition of the contents of each tank is determined from these analyses. It is customary in some plants to maintain high-calcium and low-calcium tanks.

The method of making the final mixture is very similar to that employed for the dry process described on page 143. Knowing the analysis of the contents of each tank, the operator can easily work out a combination from two or more tanks that will give the desired composition. The eighth tank is reserved for mixing the slurry drawn from the storage tanks. An air agitator is employed to thoroughly mix the contents of the tank, and when so mixed the composition of the entire contents may be determined from a single analysis.

ADVANTAGES OF THE SEMIWET PROCESS.

When wetted to the consistency of a slurry, a very intimate mixture of the particles may be obtained. It is claimed that grinding or pulverizing the material wet requires at least one-third less power than when ground or pulverized dry, although the counter claim is made that the wet material wears the pulverizing machinery much more rapidly than the dry. Pumping the slurry is a very simple and cheap method of moving it from place to place. Also materials mixed and conveyed in a wet condition make no dust.

The greatest advantage, however, is the ease with which variable raw materials may be mixed together and a final mixture of the desired composition obtained, as outlined above. This feature of the semiwet process recommends it as an alternative method in plants working on materials of variable composition.

DISADVANTAGES OF THE SEMIWET PROCESS.

The greatest difficulty in connection with the semiwet process is getting rid of the 30 or 40 per cent of water that has been added. The process was first employed where the water could be evaporated at small cost, because of cheap natural gas. Where powdered coal is used for fuel, drying is more costly. The difficulty has of late years been overcome with fair success by the use of long kilns. In kilns 160 feet long, so it is claimed, all the moisture is driven out during the passage of the material through the first 35 feet of the kiln. Therefore, the same output may be obtained with slurry in a 160-foot kiln as by the dry process in a 125-foot kiln. It must not

45723°-18-10

be presumed, however, that the entire disadvantage of the semiwet process is the additional first cost and maintenance of long kilns. It is unwise to assume that the heat used in evaporating the water would, with a dry mixture in shorter kilns, be entirely wasted. By means of waste-heat boilers this heat may be saved and utilized.

DESIRABILITY OF PROPER PLANT DESIGN.

The realization that cheaper and more rapid methods of quarrying than the selective hand-loading method are to be desired has led a number of operators to seek some process of mixing raw materials that will permit the use of churn drills and steam shovels. The semiwet process has apparently appealed to several operators as the best means of obtaining a uniform mixture of variable raw materials. As a consequence a number of plants that were originally designed for dry mixing have been remodeled and the necessary equipment added to adapt them to the semiwet process. Several plants of this type differ in some important respects from one properly constructed. Thus in the attempt to retain as much as possible of the old equipment, water may be added at a much later stage than in plants originally designed to operate by the semiwet process. The grinding and pulverizing may be done while the materials are dry, the water being added subsequently. As a result such plants have the disadvantages of both processes, the slower and more costly pulverizing of dry materials, and the high fuel cost of burning a wet mixture. Also, it has been pointed out that long kilns are required in operating a semiwet process to advantage. In remodeling a plant to adapt it for this process, the introduction of new and longer kilns is an expense that few cement-plant operators would care to undertake.

In planning to remodel a plant in order to modify the mixing system in such a way as to encourage more advantageous quarrying, the operator should carefully consider the original design of his plant and the extent of remodeling demanded. Considered from every angle, it is probable that the dry method should be retained in plants designed for a dry process. It is evident that less modification would be required to introduce a system of bins as described on page 143, than is necessary to convert the plant into a properly equipped semiwet plant.

THE STORAGE BIN AS A MIXING DEVICE.

It has been pointed out that variable constituents may be mixed to some extent in the quarry by intelligent deep-hole blasting, and that the quarry products may at a later stage be mixed together as desired, either by the employment of a series of bins, the contents of which may be apportioned in accordance with their composition,