The Iron Age 1920-01-08: Vol 105 Iss 2

New York, January 8, 1920 i i ih. 2 og”: PD ot iy Ra le VOL. 105; No.’2 Failure of Blast Furnace Stock Line Briek Carbon Deposition Was the Immediate Cause and That in Turn Was Due to the Use of War Scrap in the Charge BY L. T is generally conceded by all furnace managers that the stock line is one of the most imporcant parts of the blast furnace. It is essential, first, to have the diameter at this point correct in its relation to the other dimensions of the furnace; second, to maintain it to obtain uniform and satis- factory results. Much thought has been given to proper” protection of this part of the furnace. Various forms of protection have been developed, and the problem of protection against abrasion from the stock has been quite generally solved. Stock Line Plates As the result of the experience of the writer more than ten years ago at the Cranberry Furnace, Johnson City, Tenn., the Northern Iron Co. since 1908 has been using at its Standish furnace, Stan- dish, N. Y., and later at its Port Henry furnace, angle plates to protect its stock lines. The plates are 6 in. wide, 34 in. thick, 19 in. long, bent to form an angle of 1314 in. x 51% in. These plates are built into the brick work the 1314-in. way, and the 5%-in. way they extend down over two courses of brick; and when laid in rows around the circle of the furnace they form a completely armored stock line. Being small in size they afford a minimum of ex- pansion and contraction. This style of protection has given most excellent results against abrasion. In point of fact, blasts have been made three years on a set of these plates, and a large number of those were removed and used in relining the furnace. Following the outbreak of the war, and es- pecially after the entry of the United States, there was suddenly an abnormal demand for low phos- phorus iron, such as was produced at the Standish and Port Henry furnaces. Owing to the sudden- ness of thé demand for increased production and to scarcity of labor, the normal supply of low phos- phorus ores was inadequate for the larger tonnage required. This necessitated the use of supplemefital low phosphorus material, which fortunately, owing to war conditions, was obtainable in the shape of low phosphorus turnings produced by manufac- turers of shells. Repeated Stock Line Trouble There had been no trouble with the stock lines prior to the use of shell turnings. But in three months after they were put on the furnace, one of the stock lines had to be renewed. This was followed by repeated stock line trouble at both *General manager Northern Iron Co., Standish, N. Y. P. ROSS* - furnaces, and various explanations as to the cause were offered, such as poor brick, poor bujek lay- ing, etc. . On one occasion the brick on one sid¢ of the stock line were destroyed in about three.months after relining. In tearing out to make repairs, the condition of the brick was carefully observed. It was found that on one side the brick were com- pletely disintegrated, while on the other side and above the level of the stock they were frye from disintegration and in good condition. This seemed to disprove the theory that the quality of the brick was responsible for the trouble. At any ‘ate, it was natural to assume that if the brick were poor they would “have given way all around ingtead of on only one side of the furnace. Following this experience and in just a year after completely relining the other furnace, stock line trouble developed. Upon carefully examining the brick as they were being removed (such as were left), it was found that they were in the same condition as the ones just mentioned—-disin- tegrated and thoroughly permeated with fing, solid carbon. But in this case it was found that the brick were disintegrated uniformly around t¥xe fur- nace and to about the same depth as in th? case before mentioned, from a point where th@ stock struck the stock line to a point 15. ft. bel¥w the closed bell. . Brick Not Alone Under Suspicion An investigation was. begun to ascerta i the cause of the trouble. As will be noted fre 3 the above, the brick were disintegrated and con ed quantities of solid carbon. The protecting plates showed carbon absorption, as well as being cqvered with a thin film of carbon. Investigation was begun to ascertain why, with practically the same operating conditions (e cept the introduction of shell scrap), this trouble s* ould have developed so suddenly and continued over “uch a long period. It was evident that it was no due to the brick alone, for, as will be noted abows. in one case where it was found that most of the fas was passing up one side of the furnace, the b ick were disintegrated on that side only. The b.’ used when the trouble first developed were same as had previously given good results and were made prior to the war and under pre-war conditions, All furnace records were carefully analyzed, and 117 Ts Ps kote ears Me cw Nh i mg naling i - ee . IR SRE ARR ae 118 THE IRON AGE the only change in furnace practice or raw mater- ials was the introduction of shell scrap into the mixture. Having made this observation and feeling cer- tain that the trouble was not caused by the physical action of the shell scrap, an analysis was made to determine the effect of it upon the gases. This disclosed the fact that, due to the reduction of ore burden (and stone) to supply the heat to melt the scrap, the gas was greatly enriched in carbonic oxide (the ratio of CO, to CO having been raised from 1.83 to 2.87). Carbon Deposition An analysis of the fire brick used showed FE,O., 1.57; FeO, 0.90; Fe,O, + FeO = 2.47. It is well understood that Fe,O, exposed to car- bonic oxide, under certain conditions, causes de- position of solid carbon, the action of carbonic oxide on the Fe,O, being marked by two phenomena, namely, the removal of oxygen from the iron oxide and the deposition of carbon. On the other hand, according to Bell, who demonstrated experimentally that considerable quantities of carbon may be de- posited without the formation of metallic iron, due presumably to a catalytic splitting up of carbonic oxide into carbon and carbonic acid, the reaction being 2 CO= C+ CO.,, The most favorable conditions for carbon de- position are: (a) At temperatures from 450 to 770 deg. F. (b) In a rapid flow of carbonic oxide. (c) Where the gas carries a high percentage of carbonic oxide. (d) Where the iron oxide is exposed under pressure. In comparing these conditions with the con- ditions existing during the period under discussion, it was found: First, that the fire brick contained Fe,0O.,. Second, that the gas contained an excessive amount of carbonic oxide. Third, that the brick were affected in a zone Builds Hotel for Employees The New Departure Mfg. Co., Bristol, Conn., maker of New Departure ball bearings and coaster brakes, is going further than most manufacturers in the solving of housing difficulties. The company has already built over 100 houses on one tract, furnishing homes for 176 families, and has a majority interest in another Bristol enterprise which has constructed about 200 houses and will erect that many more before its full program is ‘ompleted. These houses take care of present needs so far as employees with families are concerned. In this factory, however, there is a very consider- able proportion of its employees who are single men, and it is for the accommodation of these in a city whose rooming capacity has been exhausted that the New Departure company has undertaken the building of a hotel. While this project is unique in that section of New England, it is by no means an experiment so far as the company is concerned. Several years ago the com- pany purchased the leading hotel in Bristol and made of it just such an institution as is now being developed and constructed on a larger scale. Both sleeping and eating accommodations are provided practically at cost, taking into consideration, of course, the investment and overhead as a factor. The new hotel will be upon the same basis. It will contain something like 300 rooms with a maximum ¢ca- pacity for 600 men, placing two men in a room. Each room will be well heated, ventilated and lighted, with windows opening to the outside. In the basement will be bowling alleys, pool room and barber shop. The first floor will be devoted to a large lobby or lounge with a January 8, 1920 where the range of temperature was probably be- tween 400 and 800 deg. Fourth, that owing to the fine magnetic ores which were used, the gas pressure and velocity were high. Fifth, that the brick above the level of the stock were not affected. From these comparisons the following conclu- sions were reached: 1. Under the then existing conditions, carbonic oxide penetrated the brick through the pores, joints and around the protecting plates. 2. The iron oxide in the brick split up the CO in the gas into iron, carbon and carbonic acid, or through a catalytic action, the carbonic oxide was split up into carbon and carbonic acid, the re- action being very likely, 2 CO = C+ CO,,. 3. The carbon thus deposited caused the brick to disintegrate. 4. The shell scrap was the primary cause of the failures of the stock lines. As is noted above, the brick were intact above the level of the stock. This was attributed to the slow velocity of the gas and low top pressure, thus allowing the gas to escape through the downcomers rather than to penetrate the brick. As there is not likely to be another world war, resulting in large supplies of munition scrap, at least during the present generation, this article may not be of special value, but it is believed that it will at least be of interest, particularly to those who may have had similar experience during the period of the war. Furthermore, the facts given would seem to emphasize the importance of using brick in the vicinity of the stock line as nearly as possible free from oxide of iron. Where the stock line protec- tion is built into the brick work, it should be so designed and so built in as to preclude, as far as possible, any joints or cracks through which the gas might enter the brick work. The above observations may also be of interest to blast furnaces which still use considerable quan- tities of turnings and borings in regular practice. cafeteria that will accommodate 600 persons. The sec- ond, third, and fourth stories will be utilized for sleep- ing rooms. On the fifth and top floor will be a club room for the foremen, with assembly hall accommodat- ing 300 persons, kitchenette and dining facilities, to- gether with a few sleeping rooms where guests of the company can be entertained. The floors will be con- nected by two elevators and the equipment throughout will be thoroughly modern. The company already has under way extensive addi- tions to its plants at Bristol, Hartford, and Meriden that will mean the employment of 3000 additional em- ployees before summer and the housing and boarding accommodations provided by the present program will be quickly absorbed. Webbing for Box Handles To conserve shipping space, box and crate handles made of webbing are recommended by the National Foreign Trade Council as the result of experiments by the United States Forest Products Laboratory. Boxes loaded with 200 to 300 lb. usually have handles made of rope, which increasés the displacement and export rates. The laboratory suggests using webbing 1% in. thick and 1% in. wide, which has a breaking strength of 800 lb. This may be inserted through saw-cuts in the box ends, turned down flat inside, and nailed securely with large-headed roofing nails. The Standard Tank Oar Co., Sharon, Pa., has prac- tically closed with the French government for 500 tank cars. Sea aa ot os ees — Design of Open-Hearth Furnaces Conditions Governing the Flow of the Burning Gases Along the Hearth of the Furnace BY A. D. NE feature of the open-hearth and similar fur- () naces that has caused sorrow and tribulation is the bottom of the furnace in that it forms a val- ley below the level of the port sills. One of the basic principles of furnace design is: The flame must lick the hearth or bottom of the furnace. The gas pressure in the heating chamber must be equal to the atmospheric pressure. This last means that a nice balance must be maintained between the volume of gas and air enter- ing the chamber and the gases removed from the chamber. When the pressure in the furnace is per- mitted to drop below the atmospheric pressure cold air will tend to flow into the furnace around the doors 0 k------ asd- He --- -05d-----> _V*Sinta h= 2 9g Fig. 6—The Parabolic Curve of an Inclined Jet and through all cracks and openings. This chills the bath and causes excessive oxidization to occur. It is well known that when patching bottom it is necessary to keep the doors closed as much as possible in order to avoid chills, which are liab:e to develop into cracks and leaks in the bottom. Many beautiful octa- hedral steel crystals have been discovered in tearing out old furnace bottoms. When the pressure in the furnace gets higher than the atmosphere a sharp sting of flame is developed and fuel must be burned to maintain it. Some melters and many heaters keep the doors of their furnaces deco- rated with a halo of flame. In some cases it is neces- sary to maintain this sting owing to the defective de- sign of the furnace. There is no question whatever that this method of working will prevent cold air being drawn into the furnace. It is “the easiest way”; the company pays for the excess fuel, and the me:ter or heater has plenty of time to sit down and “watch ‘er burn.” In foreign plants the technical control of the furnaces is more closely maintained than in this country because fuel is expensive. In order to cinter the bottom of the furnace in place it is necessary to direct the jet of flame so that it will lick the hearth. In order to do this the air and gas ports must be given a suitable inclination toward the hearth and the velocity of the flame must be sufficient to carry it down onto the bottom. The velocity which will be impressed upon the jet of air entering the fur- nace will be fixed by the height from the bottom of the regenerators to the port and by the area of the port and the flues leading to it. In the case of the gas a slightly greater head is available as it enters the re- generator under pressure from the producer. The gas pressure can be increased and usually is toward the *Copyrighted, 1920, by A. D. Williams. Second article, the first covering hearth dimensions, appearing in issue of Jan. 1. 119 WILLIAMS end of a campaign when the regenerators are partially blocked. The inclination of the ports is fixed when building the heads. It frequently happens that these heads are changed several times before they work in a satisfac- tory manner. In fact, a number of furnace drawings which have been published show heads which must have been altered considerably in order to make the bottom. The current of flame in the furnace can be consid- ered as a jet of a light fluid in motion within a heavier liquid (the air). In the case of a jet of water pro- jected vertically, neglecting the friction of the pipe and the air, the height to which the jet will ascend will be equal to the vertical head causing the hydro- static pressure. That is: v? H h - 29 In which: H head of water: h height jet rises; v = nitial velocity of jet; g = gravitational constant. If the jet is inclined it describes a parabolic.curve, Fig. 6, the height or maximum ordinate of which will be: h vu? sin? a + 29 In which: a the angle of projection, the other sym- bols being, of course, the same as in the formula for the vertical jet. When this is applied to a jet of flame, that is to say, a light fluid in motion within a heavier fluid, the parabola must be inverted. The force which acts upon the jet is not its weight, acting in a downward direc- tion; but the difference in weight between a unit vol- ume of the fluid in motion and a corresponding volume of the medium through which it moves. That is the resultant force upon the jet acts in an upward direction, and it describes a parabola, as shown in Fig. 7. LLILMLLLI LLL. Fig. 7—The Inverted Parabola of a Flame in Making Bottom Professor Yesmann has given the following expres- sion for an infinitely thin jet projected from the port of an open hearth furnace:* Let: v initial velocity of the fluid in motion; Dm specific weight of the fluid in motion ; pi = specific weight of the fluid at rest. The weight of each unit volume of the fluid in motion being less than pi, the ascensional force which acts upon each unit of volume is equal to pi-——-pm. The acceleration caused by this force is: pi pam pm The middle ordinate or depression of the jet for this zon- dition is: vw sin’ a pm h=- ~~ 29 Pi—Ppm When developed and simplified this formula becomes: v? sin’ a 273 + ts h _— — —_——___ 29 tm — ti And for a vertical jet, or when a = 90 deg.: v 273 + ts ‘aso >X —.... 29 tm — ti *Fours a Flamme, by Prof. Groume-Grjimaiio, p. 56 and following. ' ; ' : ' 120 THE IRON AGE In which: tm = the temperature in degrees centigrade of the gas in motion; ti = the temperature of the gas at rest. In the open hearth furnace the ports are just below the roof, and the hearth forms a pocket below the level of the port and door sills and above the level of the tapping hole. This pocket will become filled with the coolest gases in the heating chamber. In order to sweep these gases out of the pocket and permit the cintering of the bottom of the furnace in place it is necessary to utilize the jet of flame issuing from the ports. This jet must describe a parabola having a middle ordinate equal to the vertical distance from the tapping hole to the port sills. In the earliest types of open hearth furnace it was endeavored to attain this result by depressing the roof of the furnace. The serious objec- tions to this poor construction arose from its tending to strangle the furnace during the heat and the ex- tremely short life of the roof. Later experience showed that the flame might be jetted into this pocket by the direction of the ports and the velocity of the gas and air. In European furnaces these velocities range be- tween 12 and 18 meters (40 to 60 ft.) per second, while in American furnaces the velocity is as high as 50 meters per second. If the jet theory is true, all open hearth furnaces in operation will conform to it. The air and gas enter- ing the heating chamber of the furnace are preheated to a temperature of between 1000 and 1200 deg. The theoretical flame temperature under these conditions is between 2100 and 2200 deg., but the Wanner pyrom- eter indicates a temperature of 1800 to 1850 deg. for the jet of flame. Knowing the quantity of fuel which is transformed into gas in the producer, it is not diffi- cult to approximate the volume of gas supplied and the amount of air necessary for its combustion. From these data the velocity of the gas and air issuing from the ports may be computed. For example, a 30-ton furnace making four heats per day, consumes 0 kg. 347 of coal per second. At zero temperature and 760 mm. atmosphere pressure this corresponds to 1 m.* 81 of gas and 2 m.* 52 of air per second. At a temperature of 1000 deg. the volumes of gas and air entering the heating chamber will be respectively 8 m.* 44 and 11 m.* 77. The combination of the gas and air gives a temperature of 1850 deg. to the jet of flame. Professor Yesmann’s formula shows that the value of h will become smaller and smaller as the gases at rest within the heating chamber are colder, and will become infinite when tm = ti. In this case tm—ti— 0. Those who are familiar with the starting up of a furnace are familiar with the fact that when a furnace Table V—Applying Yesmann Formula tv Designs of 30-Ton Furnaces e i - @ © os b a a = 4 s n o> a = as s e 3 = 0" M a* <= og &E~ = < = 2% €22 “=... | h6;lCU Ee 22 25% 2S < > < > mS a ; Sq.M. M./Sec. Sa.M. M./Sec. Deg. Deg. Deg. M./Sec. Mm. Deg. 20 0.285 29.6 0.54 21.8 15 38 27 24.0 1150 74 21 0.358 23.5 0.57 20.6 13 40 29 21.2 980 58 23 0.440 18.9 0.72 16.3 33.033 33 17.3 980 22 24 0.260 32.4 0.56 21.0 10 38 25 24.5 1020 70 25 0.320 26.3 0.64 18.4 15 41 29 21.0 1370 158 is cool the flame bathes the roof of the furnace, and that it drops lower and lower as the temperature in the furnace increases, the drop of the flame being a good indicator of the progress of the heating up of the furnace. A basic bottom cannot be thoroughly burnt in unless the temperature is 1600 to 1700 deg., and it January 8, 1920 ; cannot be done unless the flame licks the hearth of the furnace. Professor Pavlow, in his “Album de Dessins Con- cernant la fabrication de l’acier Martin” gives the drawings of a number of open hearth furnaces. Apply- ing Professor Yesmann’s formula to the 30-ton fur- naces enumerated in Table V, the velocities given in columns 8 and 5 are obtained. Columns 8 and 9 give the resultant angles and flame velocities. The last column gives the minimum temperature of the gases in the heating chamber at which the jet of flame will lick the bottom. Except in the case of furnace No. 23 the jets of air and gas have different velocities and slopes; the mass of the air and the gas is also different. When the velocities and inclinations of the two jets forming the flame are different their resultant. must be obtained by the parallelogram of velocities, in which the velocities of the air and the gas are multiplied by coefficients pro- portional to their weight and temperature. This resultant is then used in Yesmann’s formula to deter- mine h or the minimum temperature at which the flame will reach the bottom of the furnace. In these 30-ton furnaces it is evident that the heads Table VI—T wo 60-Ton American Furnaces Analyzed Area of gas port, sq. m 0.7300 0.3935 Veloc'ty of gasin port, m. per sec. 23.70 42.90 Area of ar port, sq. m. 2.70 3.75 Veloc ty of a'r in port, m. per sec. 6.20 6.28 Inclination or slope of gas port, deg. 6 12 Inclination or slope of air port, deg. 17 26 Drop of flame=difference in elevation between port sill and tapping notch 0.920 1.000 Resultant velocity of flame, m. per sec. 11.80 20.50 Resultant slope or flame angle, deg. 9.5 14.5 Temperature at which flame commences to lick bottom of fur- nace, deg. C. 1531 646 Ee) give a velocity and slope to the flame which bring it down to the bottom very quickly, and white other portions of the furnace will still be at a low tempera- ture. This is very useful when making bottom or patching, but has disadvantages when melting and refining. : Furnaces 35 and 36 in Professor Pavlow’s album are two American 60-ton furnaces. The hearths are long. The heads as shown have a very slight slope, and the velocity of the air is low and that of the gas high. At a temperature of 1000 deg. the volume of the air and gas supplies is respectively 23 m.* 54 and 16 m.’ 88. Table VI gives the data regarding the ports and the temperatures at which the jet of flame com- mences to lick the bottom of the furnace. This calculation shows that the cintering of the bottom jin furnace No. 35 would be extremely difficult. It is very possible that the published drawing of this furnace is incorrect. Furnace No. 36, however, should not give any diffi- culty in the making of the bottom and the flame im- pinging upon the surface of the bath will not be driven or reflected back against the roof of the furnace. When this furnace is in operation erosion will occur in the gas port and it is desirable to learn to what extent this erosion may proceed or how much the velocity of the gas may diminish before it will com- mence to interfere with the patching of the bottom. Assuming that during the time the gas is off the gases in the heating chamber may drop to 1400 deg. without interfering with the cintering. The flame velocity may drop to 9.2 meters per second. With the original port section the flame velocity was 20.5 m. per sec. By proportion it may be seen that the area of the gas port may increase to 2.2 times its original section before it will be necessary to repair it. (To be continued) 3 4 & Pt Kg *® Measurement of Rolling-Mill Reactions Apparatus for Accurate Determination of Friction Losses on Roll Necks and Spreading Forces on Rolls and Roll Housings — Loads BY W. B. HE great need of accurate knowledge of the forces ; acting on rolls and the friction losses on roll necks in different types of rolling mills has prompted the design and construction of a mechanism herein de- scribed, and which is soon to be tried out on the experi- mental rolling mill of the bureau of rolling mill re- search at the Carnegie Institute of Technology, Pitts- burgh. By means of this equipment it is expected that ex- perimental data will be secured that will give positive, accurate knowledge, not only of the relative values of dD Inner face of... Fol! Hous: 1g fe foller D Vo eee Fig. 1.—Diagrammatic Arrangement of the Two Hydraulic Testing Cylinders and Special Roll Carrier with Roller Bear ing on Inner Face of Roll Housing the various lubricants and bearing metals used on roll necks, but also of the spreading forces acting on the rolls, from which the stresses in rolls and housings may be calculated to a far greater degree of accuracy than is at present possible. The device consists of two hydraulic supporting cyl- inders marked No. 1 and No. 2 in Figs. 1 and 2, lo- cated in the bottom of the roll-housing window and so arranged that they can easily be removed and replaced by a filler. The carrier for the bottom roll rests on the plungers of these cylinders in such a manner that the forces acting on the roll neck are recorded by the re- sulting hydrostatic pressure in these cylinders. Fig. 1 shows a diagrammatic arrangement of these cylinders and their relation to the rolls. It shows a *Director, bureau of rolling mill research, Carnegie Insti- tute of Technology, Pittsburgh. 121 SKINELE* —.— bloom between the upper and lower rolls with the lower roll carrier and cylinders. The full lines show the posi- tion of this carrier under static conditions (condition of rest). The conditions existing at the lower roll neck being greatly exaggerated for the purpose of illustra- tion. The total spreading force Ft, ¢ue to the bloom pinched between the rolls, will cause a heavy force P, to be applied about as indicated in Figs. 1 and 2. Under static conditions this force would be transmitted from the roll bedy to the neck and would be applied to the — — 5 { 1. Inner face of froll Housing - Roller D \ Pont of Applicaton nf force P Between ' Noch @ Bearing -* Hydraulic . 5 rf ~ — om YUppori +. A eeaan iS Pe aw 7 Connection for ® ; fecording Jha LY J Pressure Gage Pa Pog eae gh te If. t LS oS ES oh ‘ (viinder \« D ofa D >] ?/ Fig. 2 Analysis of Force “P"” Applied at Point of Sliding Friction, “X,"" Between Roll Neck and Housing. Hydraulic cylinders shown in section bearing as indicated by the arrow “A” in Fig. 1. Such an application of forces would cause one-half the load P, on each bearing, to be carried by cylinder No. 1 and the other half by cylinder No. 2. The proportion of the total force F, which is dis- tributed to each neck of the rolls will depend on the location of the pass in the roll body and can readily be figured by the well-known principles of beam reactions. Where rotation starts, and the bloom is being rolled, the point of contact between roll neck and bearing changes to some other point, the force P is then ap- tit is not the purpose of this article to enter into a dis- cussion of the complicated system of forces acting at the contact surfaces between rolls and bloom. Inasmuch as the rolls are able to move up or down and are restrained from horizontal movements in any direction by the bearings and roll housing, the resultant of all forces acting must be in a vertical direction. — Oe Oy eee AT Ae en entanaatin 122 THE IRON & : © /6- Equally Spaced Bo/fts A about 22" Centers my ‘i [f Dia Bolt Circle 7 2 - & Da 6 > ~< 4°D8 E oe , 1% ann, aaa Cylinder % 2.000 Grind to Fer pe) vy ‘ ZF 2 ‘2 Ball Races 83 #000 Die. Polish to Perfect Fit Porger 4% * 000 Die a ~ 4'Dia 27 Outside Dia ° Fig. with a pressure gage plied to the bearing as indicated by the dotted arrow “B,” and the lower roll carrier has a tendency to as- sume a position indicated by the dotted lines, using the roller “C” as a fulcrum point; the roller “D” lifting off its seat. The position indicated by these dotted lines in Fig. 1 is of course greatly exaggerated for the purpose of illustration. It is, however, vitally important that the bearing be perfectly free to rotate as indicated, and that no forces other than the cylinders Nos. 1 and 2 be interposed to prevent this rotating tendency. (In actual practice it is expected that not more than 0.003 in. or 0.004 in. of movement will be obtained on these cylinders between no load and their maximum carry- ing capacity of 250,000 lb. each.) When the contact point has moved to a place where it slides back as rapidly as it tends to climb, the angle of sliding friction has been reached and a condition of equilibrium is again established. These conditions are shown and analyzed in Fig. 2. The pressures on the supporting cylinders will no longer be equally distributed, as in the case of static conditions, but will be much heavier on cylinder No. 1 than on cylinder No. 2. In the analysis of these conditions the following symbols will be used: P = Total resultant force on the roll neck, in pounds. R Radius of the roll neck in inches. The angle whose tangent is the coefficient of friction. D Distance from center-line of roll housing to center line of supporting cylinder d Distance from center line of roll-neck bearing to the point of contact between the neck and bearing. N Some fraction of P greater than % M Some fraction of P less than \. If equilibrium in a “free body” is maintained, the following conditions must be met: 1. The sum of all horizontal or vertical forces acting on the body must be equal to zero. 2. The sum of the moments of all forces acting on the body taken about any point must equal zero. Fulfilling the first condition, as applied to the ver- tical forces on the lower roll-neck bearing, we have -P+NP+MP=0 P, and cancelling the P’s, we have N+M™M=1 then N P MP This shows that for each increase over % of the total load carried by cylinder No. 1 there is a corre- sponding decrease in the load carried by cylinder No. 2 and that the sum of the pressures on cylinders Nos. 1 and 2 must be equal to the force P. This same fact can be shown by taking moments about the point X where the roll neck applies its load to the bearing. In this case the force P has no lever arm. Its moment Special Stretched Stee! Daphragin C -Hydraulic Testing Cylinder Shown in Section The the thin steel diaphragm C upon the fluid film beneath, which is connected directly AGE January 8, 1920 would therefore be zero, and it may be neglected. Taking moments about X we have +N P (D—d) —M P(D + ’ *) 56-3° Balls to d) = 0, _+ One Circle 112 Balls Regd which expands into NP D—N P d—MPD— MPd=0...... (1) Taking moments about Y, a the center of the roll neck, Kk we have ; —Pd+NPD—MPD= wy Dakvibe bax ae (2) ‘ ae Subtracting equation (2) WN from equation (1) we have PD—N Pd—MPd- O,;or NPd + M9P4d& = P * d; and, cancelling the P d’s, ie we have N+ M =1. } Z SLMA LD LDS DADA L ILE uN Y The force P applied at X Tay A may be resolved into its + ce component forces, one nor- Tr mal to the tangent of the ro contacting surfaces and the other parallel to this tan- gent. As shown in Fig. 2 the force normal to the tangent will be equal to P cos #, and the tangential force will equal P sin z. These two component forces may then replace the original force P and are indicated by dotted arrows applied to the point X. If moments are now taken about the point Y it will be noticed that the force “P cos »” has no lever arm, therefore a zero moment, and the equation becomes plunger D acts against —PsinnR+NPD—MPD=—O, or by transposing PsinnR =NPD—MPOD and by rearranging we have Psing R = (NP—MP)D.......... (3) In the above equation N P is equal to the total pres- sure on cylinder No. 1 and M P is equal to the total pressure on cylinder No. 2; also D is equal to the lever arm of one cylinder. It will therefore readily be seen that the difference between the total pressures on the cylinders, when mul- tiplied by the lever arm of one cylinder, is the exact measure of the resisting forces due to friction on the roll neck and it is only necessary to multiply this quan- tity by the surface speed of the roll neck in feet per minute and divide this result by the radius of the roll neck, in inches, to secure the number of foot-pounds of work lost in friction at this point. It will also be noticed that it is only necessary to divide this same quantity, (VP—M P) D, by the sum of the recorded pressures and also by the radius of the roll neck in inches, to obtain the sine of the angle whose IK Rh 4 YM tipsy 4 tf / LI, 3 Ys Fig. 4—The Special Bolt, Marked “3’' in Fig. 5, Has Ball Joints at the Inner and Outer Ends to Avoid Any Retarding Action Which Would _ Result From the Thrust Plate, Marked “1” in Fig. 5, if Stationary eco ae use ae ¥ RO tot ea ect. A es, “aac ae cals: Dees Roa January 8, 1920 tangent is equal to the coefficient of friction between roll neck and bearing. This can be expressed mathe- matically as follows: Re-arranging equation 3 we have (NP—MP)D Sin « —__—_—_—_——_——- .... ; ; (4) (NP + MP)R The sum of these quantities, N P+ MP, also gives the load applied to the housing which carries that par- ticular roll neck, and by adding these quantities ob- tained from the corresponding cylinders in the housing ° on the other end of the roll, the total load on the roll is obtained. The foregoing is true only if the hydraulic support cylinders are frictionless. If these hydraulic cylinders were of the ordinary hemp-packed plunger type, the results obtained would be almost useless, as the fric- tion of the packing on the plunger would be very large and the exact amount of this friction would be very difficult if not impossible to obtain. It will therefore be necessary to produce a friction- less hydraulic support cylinder. This, the author be- lieves, has been obtained in the cylinder shown in Fig. 3. While it is recognized that no machine can be abso- lutely frictionless, this cylinder so nearly approaches this ideal state, that for all practical purposes it may be considered as frictionless. The cylinder consists of an upper or guide cylinder “A” securely bolted to, its base “B” by 16-in. bolts, “G.” The base, which is really the cylinder proper, has a groove, 9% in. inside diameter, cut into it. The center portion of the base is then faced off to allow just a film of fluid not over 1/16 in. in thickness to act on a special stretched steel diaphragm, “C,” which is securely held in place by the friction between the parts A and B of the cylinder and is made tight on its lower side by a soft lead gasket. By this means the fluid is completely confined in the cylinder, the only escape be- ing through the orifice leading to the recording pres- sure gage. The plunger “D” never comes in contact with the fluid in the cylinder but rests on the upper side of the steel diaphragm “C.” These contact- ing surfaces must be very flat and true. dite tn In order to avoid friction between the L plunger “D” and the side walls of the oy guide cylinder “A” the plunger is turned at 4 in. smaller in diameter than the guide x cylinder on its upper parts, and 0.005 in. smaller in diameter where it rests on the diaphragm. In order to prevent lost motion, or the plunger from tilting and shearing the diaphragm, and further to prevent the plunger from getting “off center” and having a rubbing contact with the cylinder walls, two rings, each of 56 steel balls of % in. diameter, marked “F,” are provided. The inside diameter of the guide cy!- inder “A” is ground to as nearly a per- fect circle as possible and the grooves for the balls are ground and polished to a perfect fit. When the plunger moves downward, due to the application of a load, these mo steel balls tend to lift off their seats and tH establish a rolling contact between the plunger and cylinder walls. ( The only friction on this cylinder is the friction of these ball bearings, the Baca internal friction of the steel in the diaphragm and the friction of the fluid on the pipes leading to their recording gages. These quantities are so small when compared with the maximum ioad of 250,000 Ib., which each cylinder is de- signed to carry, that they may be neglected. The displacement of fluid by the plunger when under load is equally THE IRON AGE 123 small. Marks’ handbook, page 251, gives the following information: “A pressure of 1 lb. per sq. in. compresses liquids in volume as follows: Water 1 part in 300,000; mercury 1 part in about 4,700,000. For water in an iron pipe this corresponds to a compression of 2 in. per mile of length for a pressure of 10 lb. per sq. in.” It will easily be seen that such a compression of the fluid will not disturb the roll setting. The only other factors entering into the plunger displacement will be the quantity of fluid necessary to move the needles of the recording gages and the expansion of the short pipes leading to these gages. Both of these factors are so small that they may be neglected. The idea of this friction‘ess diaphragm cylinder is not new, but has had several stages in its development. It appears to have been originally ‘invented by Mr. Emery, who first applied it to the well-known testing machine bearing his name. The general application of this invention was delayed by the fact that Mr. Emery measured the pressure in his hydraulic supports by means of a complicated balancing and weighing mechanism. The hydraulic support was made accessible to general use by the research at the German Govern- ment materials testing station of Dr. A. Martens, who measured the pressure in the hydraulic support by means of Bourdon pressure gages and who proved con- clusively that the combination of hydraulic support and pressure gage forms the most accurate and desir- able means for measuring, without friction or lost mo- tion, forces of any magnitude and of any suddenness of application. Professor W. Trinks, head of the department of mechanical engineering at the Carnegie Institute of Technology, first conceived the idea of applying cylin- ders of this type to roll necks, sometime early in 1916 and made drawings of a special roll housing for this purpose. When the author was placed in charge of the organ- ization of the Bureau of Rolling Mill Research and of o i S © uJ Fig. 5.—Detailed Design of Hydraulic Testing A tus as Applied to a Three-High 18-In. Bar Mill ie es ee as 124 THE IRON AGE the designs of the experimental! mill, he carried the de- velopment one step further by designing the present cylinders in such a manner that they would fit in the window of an ordinary roll or pinion housing, and could easily be replaced by a filler casting in case experi- ments not requiring their use were to be made. The first cylinder of this type is now in the course of construction and will be carefully tested and cali- brated under a 600,000-lb. compression testing machine, before the final design is produced for the experimental mill. Diaphragms varying in thickness from 1/64 in. to 3/64 in. are on hand and their relative merits together with the relative merits of mercury or water for the fluids will be carefully investigated before the final selections are made. There is one more important item which should be considered. That is the friction of the end of the roll- neck brass on the roll body. The “thrust plate” marked “1” in Fig. 5 is held firmly against the inside of the roll housing by the bolts “2” and “3” and, unless a flexible connection between the lower roll carrier and this thrust plate was pro- vided, would offer a very substantial resisting force acting against the rotating tendency of the lower ro!l carrier. To overcome this the bolts “3” are made with ball joints at their inner and outer ends. In this man- mer any retarding action due to the stationary thrust plate is avoided. The importance of careful experiments on roll-neck friction will be realized when, according to the best in- Main Hoists for Open January 8, 1920 formation at present available, between 40 per cent and 50 per cent of the total work of a blooming or bar-mill engine is absorbed in friction which is developed on the roll necks. These figures represent the best working conditions in rolling-mill practice. For mills, such as plate, skelp or sheet mills rolling wide, thin sections, which are often cold, the work lost in friction on the roll necks is frequently between 70 per cent and 80 per cent of the total work of the engine or motor. The designs of the experimental mills are being worked out with extreme care, which of necessity in- volves a slow rate of progress. They offer a great many problems similar to the one just described. It is expected that these designs will be completed sometime during the next three or four months and will be given to the public very shortly thereafter. The Carnegie Institute of Technology is throwing open to members of the bureau, nearly $400,000 worth of building and equipment, free of all charges for ren- tals, ete. More than half the money necessary to add this ex- perimental rolling mill has already been definitely sub- scribed and the remainder is in sight, approved by of- ficials of many different companies interested, but awaiting ratification of their various boards of di- rectors. The completion of this experimental mill will make the Bureau of Rolling Mill Research what is probably the largest and most complete steel research laboratory in the world. Hearth Ladle Cranes Tests of Two Motors in Series—Reserving One Motor as Spare—Series SERIES of tests and observations made in the main hoists of two overhead electric traveling open-hearth ladle cranes was recently presented by W. W. Garrett, Jr., of the Tennessee Coal, Iron & Railroad Co., Birming- ham, before the Philadelphia monthly meeting of the Association of Iron and Steel Electrical Engineers. Tests were made on the main hoists of two cranes, both of which were designed for 150-ton loads with a hoisting speed of 6 ft. per min. The cranes had a span of 65 ft. and were built to handle their rated loads at the center of their girders with a factor of safety of five. Hoisting was accomplished by two 100-hp. (crane rating) size G.W. 450 r.p.m. 220 volts, Crocker Wheeler motors wired in series on a 440-volt circuit and geared so that each motor can operate independently all of the four cable drums, thus giving to each motor one half of the total hoisted load. Braking was accomplished by the use of the standard Reliance friction brake, which will at all times permit the load to be raised, but sets at the instant the motors are stopped and the hoist begins to lower. After one of the cranes had been in service for sev- eral weeks, a full load test employing a ladle of molten steel was made. To start the hoist up required 105 electric hp. as against 136 e.hp. when the crane was new. To raise the load at normal hook speed required 99 e.hp. as against 147 e.hp. when the crane was new. Hoisting With One Motor Some of the operating men, Mr. Garrett said, had advocated handling the hoist with one motor. “After the crane has been broken in and as the most severe service will necessitate the rated capacity of only one motor, would it not be better to use one of these motors as a spare only?” they asked. In order to try out this plan the pinion was removed from the shaft of one of the series hoisting motors and the wiring was so rearranged as to make it possible by changing switches to transfer the hoisted load to either motor, reserving the idle unit as a spare. The plan, however, had not been in operation for sufficient time, he said, for data to be available. vs. Parallel Operation In answering the points advanced in favor of using the one motor, Mr. Garrett said he granted that one motor has sufficient power to operate the hoist, but that this motor must run at its full load capacity at all times, consequently the chances for failure are in- creased. “If one motor is used,” he said, “the hoisting speed will be increased about 18 per cent while the no load hook speed will be approximately doubled. This means higher operating speeds under full load condi- tions and much higher speeds under light load ser- vice; hence a marked increase in the general wear and tear of the cranes and a shorter life for the motors. By increasing the hook speed we will save time in han- dling the steel, but our possibility for delays and acci- dents will be greatly increased. Since we must have a duplicate motor as a spare to insure positive handling of the steel the intial outlay in dollars will be the same, so we have saved nothing on the investment by the use of one motor. In the scheme of two motors in series, each unit will at all times operate below its rated capac- ity and at slower speeds, regardless of the nature of the service. Under these conditions we can assume that we have at least doubled the life of the motors.” The speaker pointed out that it might be suggested that one larger unit be used thus to operate the motor below its rating. “Our experience in steel production,” he said, “has taught us that by preventing a short delay in which the loss might be only one ladle of steel, we have saved in dollars the price of a spare, hence two motors are carried regardless of the wiring plan.” Relative to the point that the change to the spare unit can be made in a few minutes’ time, Mr. Garrett pointed out several objections. These included delay in assembling the men from the various parts of the mill after the accident has happened and their getting ready to make the change; difficulty in working over a ladle of molten steel; and liability of the motor failing to run after being idle for months. Other advantages emphasized for the use of the two motors in series were: That it is easier to reverse the motors; they have less tendency to race; braking from two separate units is more positive; commutation of ocala rs bois. tea 7 January 8, 1920 current peaks is more easily accomplished; and the motors in series act as a double booster to the line when the hoist is running down. It was Mr. Garrett’s opinion that the two-motor series scheme is by far the safest, cheapest and most dependable plan. “It may not appear,” he said, “to be the most efficient at first glance, but after taking both sides of the question into consideration, I am convinced that it is just as efficient as the less safe plan might be.” Discussion In a discussion of the paper, George W. Richardson stated that he was in favor of two motors in series. “I have one case,” he said, “of two motors working in series. When we bought the crane we had one motor and had a considerable amount of trouble at that time, and we replaced it with two smaller motors, and we have very good results. In fact, we very seldom have any trouble with the hoist motors on that crane, but we also have two speeds on the crane in case we have one motor get out of order. We can throw it over and run one motor and we will still accomplish our work. That is not on our heaviest work, but on the lighter loads. This particular crane I speak of was used for changing loads. We can throw one motor over and run the other motor at reduced speed. That is done by a clutch arrangement. I have seen a number of installa- tions of two motors operating especially on travelling motions, and we find we have far better results from the travelling cranes when we use two motors for travelling motions than we did with one motor.” A. J. Standing stated that “We have operating on 175-ton ladle cranes two 150-hp. motors in parallel on the hoist with switches, provided in control, so that we can cut out either motor in case of trouble with the other one, and the remaining motor will take care of the load. On bridge travel on 10-ton soaking pit cranes we have gone to two motors in series, mounted near end trucks on one side of the bridge, doing away with the motor in the center of the crane span in two cases, and we have better results in that way.” Motors in Series or Parallel Parallel versus series operation was discussed by R. H. McLain. “Suppose you require 6 ft. per min. hoist speed for your crane, and suppose the actual horsepower is 100 hp. to hoist. It makes no difference as to power input whether you put two motors in par- allel or in series. There is a slight difference in effi- ciency, but we can neglect that. If you take the two 100-hp. motors and connect them in series, you are sending full load current through each motor. Sup- pose in case of a breakdown on one of these motors it is necessary to cut it out and run with the other motor. Your speed will come up to something like 10 ft. or 11 ft. per min. That would depend on the torque speed curve. If it was 9 ft. per min. you would draw 150 hp. during the emergency and operate at higher speed. Now let us consider the scheme of con- necting the two motors in parallel. If you spend the same amount of money for two 100-hp. motors and design your crane to hoist 6 ft. per min., it will take 100 hp. to hoist* it just the same. Since you have two 100-hp. motors and only 50 hp. will go through each one of these motors, from a current standpoint and from the standpoint of the torque on the motor, you will be operating at half load on the motors when you have two motors in parallel; whereas, when you have them in series, from a torque standpoint and from the standpoint of the current, you will be operating at full loa