REINFORCED CONCRETE BUILDINGS Published by the McGraw-Hill JBooIk. Company N//'/.>*/////>///////. ^s 1 IT TT FIGURE 20. " 1 Proposes to dispense with the use of ordinary joists and to make use of wrought-iron tie-rods extending from wall to wall. "The floor becomes one solid beam, having the tie-rods and hoop- iron in combination with the concrete to take the tensile strain, and the concrete to take the compressive action resulting from the weight of the floor." Lythgoe & Thornton, 1868, No. 640 (Figure 21). FIGURE 21. " The method of constructing floors with bars of J_-iron and concrete as shown." Johnson (Coignet) 1869, No. 884 (Figure 22). An invention relating to the facing of concrete blocks with cast-iron or steel protecting plates, to be used as street curbs, etc. Gedge (Monier) 1870, No. 1999 (Figure 23). BASIC PATENTS FOR INVENTIONS 23 " In short the iron is the skeleton and the cement its covering." Tall, 1871, No. 1001 (Figure 24). Iron hooping, wirework, or netting are interlocked between FIGURE 22. FIGURE 23. the lateral cross bars, and form a close lattice or basketwork. Portland Cement stucco is applied. Brannon, 1871, No. 2703 (Figure 25). FIGURE 24. " Wirework embedded in concrete, to give cohesive strength against transverse and tensile strains." Hyatt, 1871, No. 3124 (Figure 26). 24 REINFORCED CONCRETE BUILDINGS " The peculiar construction of floor which I designate an 'all-beam' floor, composed of a number of separate tubes laid side by side." Turner, 1872, No. 1396. On the iron beams "I strain my wire from the plates in the walls; these wires are intended to supersede the use of floor joists of wood, and will form beds for my concrete floors, and also answer on the underside instead of laths for the plastered ceil- ings, which work of plastering may be carried on at the same time as the laying on of the floors in concrete." Emmens, 1872, No. 2451 (Figure 27). FIGURE 25. FIGURE 27. " The employment of sheets of corrugated iron as founda- tion for roadways, paths, steps, and flooring." Lish, 1873, No. 1621 (Figures 28, 29). The drawing shows a sectional view of a floor and girder of concrete with tension .rods embedded therein, as indicated by the dotted lines. Hyatt, 1873, No. 3684. Asbestos combined with perforated, corrugated sheet metal or with crimped sheet metal or upon a hollow grate bar system. Coddington, 1873, No. 1004 (Figure 30). BASIC PATENTS FOR INVENTIONS 25 The figure shows a water pipe or tube, C being the cemented material, E the interwoven metal. FIGURE 28. FIGURE 29. Hyatt, 1873, No. 3381. " The system or mode of forming cellular or honeycomb structures by connecting together single cell blocks by means FIGURE 30. of tie-rods or crimped blades of metal, with or without addi- tional straight tie-rods." FIGURE 31. Hyatt, 1874, No. 2550 (Figure 31). "I form the tie in a way which gives it power to grip and hold the foreign material in a manner and by a method which 26 REINFORCED CONCRETE BUILDINGS brings the load and consequent strain upon the tie at the same instant it is felt by the concrete or foreign material, by which means the tensile and compressive forces act in harmony with each other." Hyatt, 1874, No. 1715 (Figure 32). FIGURE 32. " Making hollow metal beams of interlaced lattice or open- work, as the holder of a tie-rod, to connect the same with con- crete or equivalent material." Edwards, 1891, No. 2941 (Figure 33). 1892, No. 1415 (Figure 34). 1894, No. 15,466 (Figure 35). L FIGURE 33. i.--. FIGURE 34. Edwards' patents show a remarkable insight into the nature of reinforced concrete construction. It is proposed to cast BASIC PATENTS FOR INVENTIONS 27 the slabs separately and set them when hard, owing to the great cost of the centering; the bending up of the principal tension rods is described at great length, and stress is laid upon the benefit of many small rather than fewer but larger rods. The FIGURE 35. importance of preventing sliding of the reinforcement is shown, and it is described how the beams may be pierced by openings in much the same manner as done under the Visintini System. The benefits as well as troubles arising from the fixing of the ends of the beams into the walls are perfectly understood, and the entire argument advanced is illustrated by tests (by Kirkaldy). While in England the new construction made but scant headway, a considerable activity took place in Germany, where the Monier Patents were bought and exploited by G. A. Wayss, and where M. Koenen advanced the first rational method of calculation in 1886. The " straight line formula" was fully discussed by Koenen in " Centralblatt der Bauvervaltung," May 14, 1902, and is to this day the commonly accepted standard. In Holland, the first ribbed floors were erected in 1886 in connection with the Public Library in Amsterdam. In France, it seems that Monier's first patent was taken out in 1867, but it has been intimated that he had knowledge of the earlier patent granted to Lambot, who had made a reinforced concrete boat of small dimensions in 1855. This boat is said to be in existence today. Monier 's efforts toward the intro- duction of his inventions were not very successful, partly per- haps because he failed to realize the necessity of placing the reinforcement near the bottom; it is told that when Wayss showed him slabs so reinforced, Monier severely criticized this arrangement, and abruptly ended the argument by exclaiming, "Who is the inventor, you or I?" 1 As a matter of fact, little 1 Suenson: Jaernbeton, P. 5. 28 REINFORCED CONCRETE BUILDINGS was done in building construction until 1892, when Henne- bique and Coignet took the reinforced concrete construction up with great success, each introducing his own system. In the United States, the first indication of anything ap- proaching reinforced concrete may be found in a patent granted to P. Summer, 1844, No. 3566 (Figure 36), for a metal lathing, which was still further improved by J. B. Cornell in 1859, No. 22,939 (Figure 37). At this early date, a number of patents FIGURE 36. for cement pipes were granted, as to R. B. Stevenson, 1854, No. 11,814, for a combination of a pipe of sheet-metal and an exterior coating of hydraulic-cement mortar of " requisite thick- ness for strength." In the Wyckoff patent, No. 32,100, of 1861, the interior pipe is of wood wound with wire of iron or other metal; in the Knight Patent, No. 32,298, of the same year, a metal tube is disposed " intermediate between the inner and outer surfaces" of a cement pipe. In 1868, A. P. Stephens took out a patent, No. 78,336, on a similar pipe, in which the strengthening tube was made of corrugated iron; in 1872, Patent No. 127,438, the tube was changed to a spirally formed sheet metal tube, and in the same year J. A. Middleton, Patent No. 133,875, pro- posed to strengthen his cement pipes by a layer of wirecloth embedded in the cement, thus combining what we now consider the essential elements of a reinforced concrete pipe (Figure 38). The first reinforced concrete wall-patent appears to be one BASIC PATENTS FOR INVENTIONS 29 granted to S. T. Fowler, in 1860, No. 28,069, where the concrete wall is to be strengthened with vertical and horizontal timbers, to be buried in the concrete; a more rational construction is FIGURE 38. proposed in 1862, No. 37,134, by G. H. Johnson, for grain-bins: " a new construction formed of brick- work tied together by plates and rods of iron." In 1869, No. 87,569, G. H. Johnson FIGURE improved this construction, using " horizontal annular tension- bars . . . the ends of each bar being so united as that it shall form an endless, unbroken band ... in the combination . . . with . . . vertical connecting-rods so as to form a metallic 30 REINFORCED CONCRETE BUILDINGS frame within the walls of the structure." This invention (Fig- ure 39) was not the only important improvement of that day; in 1868, C. Williams, No. 75,098, invented the metal lattice- reinforcement for concrete walls. The lattice-work was built up by riveting the slats together (Figure 40). The first use of concrete in columns must be conceded to W. H. Wood who, in 1862, No. 36,747, patented an improvement in piers and bridges. The invention consists in the use of hol- low cast-iron columns filled with concrete or cement, and sup- ported on wooden spiles below the surface of the bed of the river. The first ceiling was proposed by J. Gilbert, 1867, No. 64,659. This patent shows corrugated iron plates filled with FIGURE 40. concrete, the concrete to extend an inch or so above the top of the corrugation (Figure 41). His solution of the problem " self-centering reinforcement" is not very inferior to those proposed by more recent inventors. Thus we see that around the year 1870 the combination of masonry of various kinds with a strengthening metal work was quite well known. The patent, No. 88,547, granted to F. Coignet, a Frenchman, in 1869, states the general principles very clearly: " In the body of artificial stones": "skeletons or metallic framework, linked or arranged so as to strengthen the same." This is the whole science of reinforced concrete construction in few words. As an example, he proposes to use a cylindrical web of small rod- iron or wire in combination with a cement envelope, for the pur- pose of resisting the interior pressure in pipes, as well as T- or BASIC PATENTS FOR INVENTIONS 31 L-irons for other purposes. The series of patents granted to Coignet in 1869 deserve more than usual attention, as they contain much good advice of value to engineers; they are No. 88,545, 88,546, 88,547, 88,548, and 88,549. FIGURE 41. The brick arch with abutment-shoe and tension bar between abutments was invented by C. Henderson, in 1871, No. 113,881 (Figure 42); the brick arch reinforced on the cantilever prin- ciple was invented by F. Alsip, No. 120,608, in the same year. It is not clear from the description whether Alsip really con- FIGURE 42. sidered his invention as a cantilever construction, but the fact remains that all the essential elements of a cantilever are pres- ent in this patent. A very interesting patent No. 122,498 is the one granted to W. H. Smith, in 1872, for a concrete pavement. On soft ground, the arched pavement is intended to be self- supporting. Tie-rods are then carried under the pavement from curb to curb, or "chords may be embedded in the composition to operate in lieu of abutments to the arch." In the drawing, the tension rod is shown provided with a large button on the end, evidently for the purpose of preventing slipping of the bar (Figure 43). 32 REINFORCED CONCRETE BUILDINGS The patent issued to Sisson and Wetmore, in 1872, No. 124,453 (Figure 44), shows "a combination of trussed and un- trussed frames of light bar-iron to form skeleton wall-posts, girders, etc., in combination with a filling of beton or other suitable concrete, to be poured in a state more or less liquid. Our object is to have the beton and iron frames furnish mutual support and protection to each other." Considered as a beam, FIGURE 44. the wall-post of this patent exhibits many of the essential fea- tures of present day practice: the top bar extending from one span to another, the trussed bar bent up over the support, the horizontal lacing of the verticals and the vertical lacing of the horizontals, etc. But generally speaking, the reinforced brick-arch continues to hold the interest of the inventors. In 1872, P. H. Jackson received a patent, No. 126,396 (Figure 45), for a peculiar con- struction of abutment-casting to be used in connection with reinforced arches, and in 1873, No. 137,345, N. Cheney proposes BASIC PATENTS FOR INVENTIONS 33 to make the tension reinforcement of light wires placed close together and interwoven with cross-wires, to serve the addi- tional purpose of a metallic lathing. The earthquake-proof house invented by D. L. Emerson, in 1873, No. 137,833, calls FIGURE 45. for vertical rods or plates in the walls, and anchors passing through them, the plates and anchors being connected with strap iron. In the same, year J. W. Basset, No. 138,118 (Fig- ure 46) shows a construction of individual plaster slabs with a metallic trellis work within, the ends of which extend beyond the block, for the purpose of locking the various blocks together. While not strictly within the scope of this paper, attention is called to the patent, No. 172,641, granted to O. C. Matthews, n FIGURE 46. in 1876, for a foundation, in which piles are driven and again withdrawn and the holes filled with concrete (Figure 47). In 1878, T. Hyatt, No. 206,112, ended the " period of dis- covery " and put the theory of reinforced concrete construction on a rational basis, and at the same time received a patent of remarkably broad scope, covering practically the entire field of reinforced concrete and masonry construction. The general purport of this invention is set forth in a volume entitled "An account of some experiments with Portland Cement concrete, combined with iron," of which a copy was deposited in the library of the Patent Office, but which was otherwise designed for private circulation. Hyatt appears to be the first to state specifically that the steel must be able to resist sufficient tensile 34 REINFORCED CONCRETE BUILDINGS stress to balance the compressive stresses on the concrete, that all metal may be dispensed with save the tension rod only, that both baked bricks and concrete possess in themselves cohesive power and strength sufficient to perform the functions ordina- rily performed by the metallic web. He realizes the value of deformed bars and says: " I prefer to use metal specially rolled for the purposes, with bosses or raised portions formed upon FIGURE 47. FIGURE 48. the flat faces of the metal. When I make use of common bar or hoop iron, I stud the slips with pins; or I make use of several blades threaded upon wires, as represented by Figure 1." In the book mentioned above, he laid down the results of his ex- periments which led him to bend some of the bars up, and also to use a rigidly attached separate " shear member." The analysis is very complete, both in his book and in his patent speci- fication. He reinforces his columns with longitudinals or hor- izontal hoops, as the case may require, or both. He says: "In BASIC PATENTS FOR INVENTIONS 35 constructing the columns or piers wholly of concrete, I make the structure solid, the concrete then bearing the load, and, giving way under compression, would naturally incline to yield in the first place, not from absolute crush of the materials, but from want of sufficient tensile resistance at the circumference of the column. But this tendency being resisted by the circular ties, such a concrete could give way only by the crush of its particles." In short, the whole theory of hooped columns. The only difference is that Considere prefers the use of spirally wound reinforcement, while Hyatt uses the individual bands (Figure 48). To what extent Hyatt was familiar with Pasley's tests, if at all, we do not know; in his book of 1877 he gives a brief ac- count of the history of fireproof construction, but gives no ref- erence whatever to the tests just mentioned. It appears that he had a test made in September, 1855, in New York, under the general supervision of Mr. R. G. Hatfield; the beam was about 9" square, and had a tie-rod passing through holes made for the purpose in the bottoms of the bricks. More important tests were made by Kirkaldy in London, from 1874 to 1877, on beams made by Hyatt. The Period of Improvement. Broadly speaking, the Hyatt Patent, No. 206,112, shows and describes everything necessary FIGURE 49. for the practical use of reinforced concrete, and the patents of the following period are therefore mainly for improvements, many of which are due to Hyatt. Most interesting is the one granted in 1883, No. 290,886, for a concrete floor, showing not only transverse arches between the ribs, but also the use of web reinforcement in a continuous sheet along the center of the beam. In 1881 a patent, No. 237,471, was granted to S. Bis- sell (Figure 49) for an arch-bridge, showing diagonal straight reinforcement within the masonry, the object being to construct 36 REINFORCED CONCRETE BUILDINGS "an arch of limited span without causing any horizontal thrust upon' the abutments." The Cubbins patent of 1883, No. 285,801, shows a circular cistern cover "of artificial stone, having a metallic band or tire" (Figure 50) or "consisting of FIGURE 50. a concavo-convex or arched disk . . . inclosed by a metallic band or tire." This appears to be the first slab with " cir- cular reinforcement." "Expanded metal" was patented, No. 297,382, in 1884, by J. F. Golding: "metallic screening formed of slashed and stretched metal." The particular use to which the invention was to be put is not specified, and at first it was used exclusively as a metal lath. Its use as reinforcement for structural concrete is of much later date (Figure 51). FIGURE 51. A number of interesting patents are granted at various dates to P. H. Jackson. The first, No. 302,338, in 1884, is not of interest in this connection; it shows principally the usual tie-rod construction in a brick arch. But the following year, 1885, he took out a patent, No. 314,677, showing, for the first time, the bent-up or "trussed" arrangement of reinforcement (Figure 52); the bars are carried to the support where they are anchored by means of nuts. The concrete and its reinforcement rest upon corrugated iron plates, and the bars may be secured BASIC PATENTS FOR INVENTIONS 37 or not at intervals to the bottom of the corrugated plates. Another patent, No. 320,066, of the same year, shows the rein- forcement continued into the adjacent bay and there hooked FIGURE 52. over the tops of the I-beams (Figure 53), which here have the function of the main girders. The patent, No. 339,296, of 1886, specifies an expansion joint in the construction of a reinforced concrete arch; evidently the troubles caused by expansion and shrinkage were well known at this early date. Two patents, Nos. 366,839 and 366,840, were taken out in 1887 for " series FIGURE 53. of arches composed of concrete, and a longitudinal tie on which the footings of the said arches are supported and to which they are fastened," and a construction of arches with longitudinal reinforcement near the bottom; these arches rest on one side on the front girder of the building, on the other side upon the area wall. Also from that year date the following patents: two, Nos. 367,343 and 370,625, showing the application of dove- tailed corrugated plates filled with concrete (Figure 54), and FIGURE 54. three, Nos. 371,843, 371,844, and ?71,845, showing the use of I-beam reinforcement in the bottom of the beam, as well as 38 REINFORCED CONCRETE BUILDINGS compression-reinforcement in the top (Figure 55). The reis- sued, RIO, 921, and the original patent, No. 375,999, issued in 1888, may be noted in passing. When we consider the state of the art as it appears from the patents mentioned above, the Monier patent of 1884, No. 302,664 (Figure 56), cannot be called much of an improvement. FIGURE 55. FIGURE 56. Nevertheless, the name Monier was for many years synonymous with " reinforced concrete," at least in Europe, where the Mon- ier patents were bought and greatly developed by German engineers. "My invention," he says, "relates to the use and sale of integral elements of construction of metal and concrete or mortar combined, the mortar forming the covering for a metal skeleton. This skeleton is composed of longitudinal bars or rods and transverse ribs, secured together by metal ligatures." The Monier patent, No. 486,535, of 1892, is practi- FlGURE 57. cally nothing but a series of special designs based upon this same principle, and contains little new material. Yet a great industry was based both here and in Europe upon the Monier patents. The Ransome patents have been described in an earlier chapter and are not referred to here. The "trussed" arrangement of the steel was, as stated above, invented by Jackson in 1885. The Gustavino patent, No. 336,048, of 1886 (Figure 57) shows the same feature, as BASIC PATENTS FOR INVENTIONS 39 well as the rod with a continuous curve between supports. In addition to this tie-rod which extends from wall to wall, " I may in practice use a straight tie-rod extending between wall and wall above the arch." The same year, 1886, saw the origin of another new type of construction which stands on the border between reinforced concrete and plaster work. The Rabitz construction, No. 339,211 (Figure 58), calls for a metallic skele- FIGURE 58. ton frame of vertical rods and a reticulated metallic netting, in combination with a suitable coating of cement mortar or sim- ilar material. In the patent issued to P. M. Bruner, No. 356,703, in 1887, something approaching the U-bar (Figure 59) is shown; FIGURE 59. although the construction would not be classed as reinforced concrete at this present time, the rods being disposed princi- pally on the compression side, from which rods transverse ties hang down in the beam. A telegraph pole was invented by D. Wilson, No. 374,103, in 1887; it was to be composed of a skel- eton frame having rods and horizontal hoops, and a coating or body of cement inclosing the frame. The same idea was pat- ented, No. 411,360, in 1889, by O. A. Stempel, who claims a post, rail-tie, or beam, composed of "a metal frame, the filling and inclosure of imperishable material that protects said frame 40 REINFORCED CONCRETE BUILDINGS from the inroads of moisture and rust, and said frame arranged to protect said structure from breakage." The drawing looks somewhat like what an engineer would prepare for a column at this time (Figure 60). The patents granted to M. F. McCarthy show again "the combination (with an I-beam supporting the slab) of the wire strands extending over and drooped between the same, and the concrete filling wherein said beams and strands are embedded." This quotation is from the patent issued in 1891, No. 455,687 (Figure 61); the four patents, Nos. 520,489, 520,490, 520,491, and 520,492, issued in 1894, show various combinations and variations of the same prin- ciple. The patent issued to P. Cottancin, No. 459,944, in 1891, is for a strengthening web " characterized by the union in a reticulated fabric of a warp and a weft, each composed of a wire, band, or bar bent on itself into a sinuous or like shape." This patent forms the base for a large industry especially in France. The J. Melan patent, No. 505,054, of 1893 (Figure 62), claims "a vault or arch consisting of abutments, beams, or girders, arched ribs rigidly con- nected with said abutments, beams, or girders, and a filling of concrete or the like between said ribs." A number of arch-bridges have been constructed under this patent. A. L. Johnson patented, No. 550,177 (1895), a construction of floors much used at one time in the West, comprising mainly I-beams with suspension straps fastened at the tops of the beams and drooping between the beams; the straps are flat and support the concrete rib of the beam (Figure 63), upon which in turn rests the concrete slab. Another important arch-patent, No. 583,464, was granted to F. von Emperger, in 1897, for an im- provement in the Melan patent described above; it FIGURE 60. consists mainly in using two ribs instead of one, each rib being placed near one surface of the concrete. Secondary members connect the top and bottom ribs (Fig- ure 64). Recent Patents. The idea of molding reinforced concrete BASIC PATENTS FOR INVENTIONS 41 members separately and afterwards erecting them in place appears to be almost as old as the art itself, and a number of FIGURE 61. the patents mentioned above refer to this possibility without going much into the details. In 1898, a patent, No. 606,696, FIGURE 62. was issued to G. B. Waite for a beam construction (Figure 65), the sole object of which is to provide members adapted to be FIGURE 63. molded in advance and erected in place after hardening. The individual sections are made of I-shape and reinforced in top FIGURE 64. and bottom, or in the bottom only; " shear" members of vari- ous forms are used in the beam-webs. The De Man twisted FIGURE 65. bar was patented, No. 606,988, in the same year; it consists 42 REINFORCED CONCRETE BUILDINGS of "a thin flat bar having twists formed therein at intervals" (Figure 66). The patents granted to F. Hennebique, in 1898, are three in number. The first, No. 611,907, shows the now almost univer- FIGURE 66. sally used combination of open, U-formed shear members with horizontal and trussed main reinforcement, with the main bars extending into the adjacent span (Figure 67). While the FIGURE 67. authorities seem to disagree in regard to the value of the pro- tection afforded by this patent, there is not the slightest reason to doubt that this construction has been of the greatest benefit to the art. The second Hennebique patent, No. 611,908, is for a system of separately molded members, claiming in sub- stance a combination of joists and " a plurality of slabs having projecting cores embedded in said joists " (Figure 68); the word FIGURE 68. core means here the reinforcing bar, and the slabs are placed with their ends resting upon the side-forms for the joists, so BASIC PATENTS FOR INVENTIONS 43 that, when concrete is poured in the joist-molds, the project- ing ends are embedded in the concrete. The third patent, No. 611,909, is for a pile of reinforced concrete having grooves in two faces, so that a tight cofferdam may be made by using the piles for sheet piling, and filling in the grooves with grout. The structures erected under the Hennebique patents are numbered by the hundreds in any one of the several civilized countries. The patent, No. 617,615, issued to E. Thacher, in 1899 (Fig- ure 69), for an arch construction, claims the combination of the FIGURE 69. concrete arch with its abutments, and reinforcing bars in pairs, one bar near the intrados, and one near the extrados, the two bars of each pair to be above one another, either both or only one of these bars to extend well into the abutment, and, in par- ticular, "each bar of a pair to be independent of the other." A comparison with the Melan and v. Emperger patents is of interest, as the bars in the v. Emperger patent extend into the abutments and are placed one above the other. In the same year, 1899, a patent (Figure 70), No. 634,986, was granted to FIGURE 70. A. Matrai for a system of wire reinforcement embodying many interesting features. One object of the construction is to unload as far as possible the middle of the supporting beam or girder, and these again are reinforced with a number of suspension cables or wires. This construction is in considerable favor in Europe. In 1900, a patent, No. 654,683, was issued to I. A. Shaler, for a construction embodying the use of longitudinal and transverse rods, the latter welded to the main bars at intervals, 44 REINFORCED CONCRETE BUILDINGS and in the same year, L. G. Hallberg had a patent, No. 659,967, issued for a foundation built on the principle of " circular rein- forcement " (Figure 71) in combination with radial bars. The Wayss patent, No. 673,310 - 72) of 1901, is of interest, on account of the rigidly attach ear members and other fea- tures, the purpose being to ^otain similar advantages as out- lined for the Hennebique patent without infringing the same; FIGURE 72. the construction is dissimilar to Hennebique in the particular arrangement of the parts. The well-known Thacher bar was patented in 1902, No. 691,416 (Figure 73), and in the same year BASIC PATENTS FOR INVENTIONS 45 a patent, No. 709,794 (Figure 74), was granted to W. C. Farm- ley for a concrete arch construction, in which the steel is so arranged as to make the same bar pass from the tension region FIGURE 73. near the intrados to the tension region near the extrados, etc. The Visintini patent, No. 735,920, of 1903, shows the peculiar type of construction known under that name; instead of the FIGURE 74. ordinary solid beam, a lattice-girder of reinforced concrete is used. The top and bottom flanges are reinforced with longi- tudinal bars, and the cross-bars are embedded in the concrete FIGURE 75. work of the lattices (Figure 75). The Visintini beam has been used but little in this country, but abroad a large number of structures have been erected under this patent. In 1903, FIGURE 76. the first Kahn patent, No. 736,602, was issued, to be followed by many more (Figure 76). The principal features are well 46 REINFORCED CONCRETE BUILDINGS known: The rigidly attached secondary members are manu- factured in one piece with the main tension rod, then sheared loose from the main body along the greater part of the length FIGURE 77. of the rod and bent up as desired. The Weber chimney-con- struction was patented in 1903, No. 748,242; the lower portion of the chimney is provided with a circumferential air-space open at its base to the outer air and leading at its upper end into the chimney flue at the base of the upper single flue (Figure 77). A. Considere took out a patent, No. 752,523, for his well- BASIC PATENTS FOR INVENTIONS 47 known column construction, claiming "a solid concrete core with independent helicoidal coils of metal surrounding said core, and arranged very close together," and also the combina- tion of these elements with separate longitudinal rods, in 1904 (Figure 78). With this patent we may consider the period of invention as coming to an end. A very large number of patents have been granted since, mostly for slight improvements, and an enumeration of all these details would be very tedious and without serious importance, although several patents of the greatest interest may be found in this great mass of dead material. PART II RATIONAL DESIGN OF REINFORCED CONCRETE BUILDINGS BY ALEXIS SAURBREY CHAPTER III INTRODUCTION 1. EXPERIENCE teaches that concrete beams may be greatly strengthened by introducing a comparatively small amount of steel within the concrete, according to certain principles of which the following is a discussion. This combination of concrete and steel is called Reinforced Concrete; the essential peculiarity of reinforced concrete structures is that both the concrete and the steel, if alone, would be grossly inadequate for the load which they will carry when combined; the load carrying capacity is not the sum of the individual capacities of the concrete and the steel. This general rule is not without exception, if structures like the ordinary reinforced concrete column are included; strictly speaking, only the hooped column is entitled to be clas- sified as reinforced concrete, because in that case a small amount of steel added to the concrete changes the structural properties of the column entirely. 2. The stresses in a reinforced concrete structure are neces- sarily complicated. Not only is the steel entirely dissimilar in nature to the concrete which it reinforces, but the concrete itself is not homogeneous in the strictest sense of the word. Yet two cubes of large size, cut from different parts of the beam, must be assumed to be theoretically alike; we make there- fore the necessary and justified assumption that the lack of homo- geneity of the concrete is of second order as compared with that of the structure as a totality: necessary, because otherwise we cannot advance any theory; justified, because the differences between the nature of steel and concrete are sufficiently large to overshadow completely the small differences which un- doubtedly exist within the concrete itself. 3. Generally, the properties of reinforced concrete are known when the properties of the two materials are known; there is 51 52 REINFORCED CONCRETE BUILDINGS no reason for believing that the properties of either material are changed in any way by the presence of the other. It is, however, necessary to expand the limits of our research when dealing with a combination of two materials, because the prop- erties of the combination depend primarily upon the ability of the two materials to co-operate, and only in second line upon their individual properties such as strength, elasticity, etc. This co-operative ability is of a somewhat obscure nature; without making any attempt of explaining it, we must admit its existence. In the following it is referred to as the "bond" or the "adhesion." When this bond is broken the structure fails. 4. The purpose of design is to produce not only a structure of adequate strength, but one of equal strength in its several parts. With consistent formulas for the various elements, the allowable stresses should therefore be the same for all elements of the structure considered. Experience shows, however, that the difficulties to be overcome in the erection are different for different parts; we can readily see that a local deposit of bad concrete as large as a hand will affect a 10" column and a 6" floor slab in dissimilar ways. This is the reason for variable allowable stresses in any case the purpose of fixing certain maximum stresses is to insure an ample factor of safety. For- tunately the investigation of stresses in a given beam is very much simpler under moderate loads than near ultimate failure; the coefficient of elasticity for steel E a is a constant, and that for concrete E c varies but slightly. For practical purposes the ratio E S /E C r is assumed to be a constant up to the limit of the allowable stresses. Within this same limit we assume sections plane before the load was put on to remain plane under load, and we assume proportionality between stress and deform- ation. The tensile strength of the concrete is entirely disre- garded. None of these assumptions can be called absolutely correct; they are, however, no more inaccurate than any other set of assumptions which we would be able to suggest in our present state of knowledge; moreover, they are the simplest possible. 5. As the tensile strength of concrete is much less than its compressive strength, the principle is to utilize the available compressive resistance and use steel bars to carry the tension. INTRODUCTION 53 Sometimes steel is also used in compression, although with less success, the object being to limit the size of the columns and to fortify them against excentric loads. We shall see later that it is possible to construct a column in which the steel is stressed in tension (Article 12). 6. In any kind of concrete structure the embedded steel has a tendency to displacement in its own longitudinal direc- tion under load. The value of the steel as reinforcement de- pends upon its ability to withstand any forces tending to either push or pull it out; reinforced concrete is an impossibility without adhesion between steel and concrete, and destruction of the bond or adhesion means failure. The law governing adhesion is therefore the foundation of all theoretical study of reinforced concrete. CHAPTER IV ADHESION 7. THE adhesion is measured in Ibs. per square inch of em- bedded surface of the rod; its value is different for pulling and pushing tests. As the latter is somewhat higher it is sufficient to investigate the laws governing the pulling resistance and apply these laws to the pushing resistance also. The mathematical analysis of the bond stresses is impossible with the material on hand; even the test-data are meager and often contradictory. We know, however, that the following statements are approxi- mately true, so that an embedded rod pulls out of the concrete block : (a) when the stress in the steel reaches the elastic limit of the steel. (6) when the tensile resistance of the concrete, in a lateral direction, is reached, because the block splits. (c) when, instead of splitting, the concrete around the rod expands sufficiently to let the irregularities of the rod pass through. (d) when the adhesion is destroyed. Obviously, then, the designer must keep the steel stress well below the elastic limit, allowing for this and other reasons an ample factor of safety, while, at the same time, the concrete must be strong enough to meet the demands made upon it. Hence the diameter of the concrete block, or the thickness of the piece, is a very important factor, but unfortunately nothing is known in regard to the minimum allowable diameter, except that it is greater for deformed bars than for plain round or square rods. We can readily see that both the tensile strength and the coeffi- cient of elasticity of the concrete has great influence upon the minimum allowable diameter; with a well-proportioned mortar and a mixture of say 1 :2:3J, we may perhaps suggest a diameter of concrete equal to ten diameters of the embedded steel as ' 54 ADHESION 55 reasonably safe. In floor construction the bars usually find their ultimate anchorage in much larger bodies, the slab bars passing through the beams, the beam bars through the girders, and finally the girder bars through the columns. In all these cases the concrete is reinforced in a direction transverse to the direction of the pull, and the expansion in a lateral direction is thus partially or entirely prevented. 8. In the beam theory to be outlined below, great importance is attached to the length of embedment beyond the supports of the beam, in fact, this length- represents the ultimate reserve of strength of the beam. It is usually considered good practice to imbed the bent bars from twenty-four to thirty-six diameters i 6 Diameters Y J, Diam.. V 1 FIGURE -f 79. beyond the support, using the lower figure for deformed bars stressed to 16,000 Ibs./sq. inch, and the higher figure for deformed bars stressed to 20,000 Ibs./ sq. inch. For plain bars, an addi- tional hook is made on the end of the bar, equal in length to six diameters. In many cases the length of embedment here recom- mended cannot be obtained for the reason that there is no adjacent concrete into which the bars may be extended, as, for instance, in the case of a beam finding its bearing in an outside brick wall. The bars are then hooked, and the length of embedment calcu- lated from the center of the seat to the end of the bar, including the curved end of the hook. Square hooks must be avoided, a gentle curve of, say, six times the radius of the bar is much more effective, and the more so the greater the radius of the bent (Figure 79). 9. The diameter referred to in the preceding paragraph is not the diameter of each individual rod or bar, unless the rods be spaced so far apart that each will pull out individually, leaving the concrete intact between. The diameter is that of a circle or other curved line in which all the rods may be enclosed if laid closely together. It follows that it is good practice to spread the rods out as much as possible; in a beam this is easily obtained by 56 REINFORCED CONCRETE BUILDINGS bending some of the bars up over the support, as is also done for other important reasons. It is a common but inexcusable mistake to use a number of small diameter rods bunched together; it is almost impossible to concrete such beams properly, and the fallacy of the argument leading to such construction should be evident. CHAPTER V COMPRESSION AND LATERAL EXPANSION 10. WITH few exceptions, materials submitted to deformation in one direction undergo deformations in all other directions. If the principal deformation is a shortening, the lateral deforma- tion is a swelling, which must be taken as evidence of certain interior stresses in the body in a direction normal to that of the principal stress. These transverse stresses are of the greatest importance for materials like reinforced concrete, because, if not restrained, they bring about the premature failure of the con- crete, while, if restrained, they may be used to increase the strength of the structure. Thus, as pointed out above, the transverse swelling affects the bond of an embedded rod; if restrained (by surrounding the bar with a coil of large diameter) , the value of the bond may be increased as much as fifty per cent, or more. Even a loose stirrup circling the tension rod at the bottom of a beam increases the sliding resistance of the rod, so that a rod, covered at the most with two inches of concrete, may have the same sliding resistance as one embedded in a large body of concrete. Similarly, the Ransome Coil Coupling may be used with good results when splicing rods, although the rods should always be made in one continuous piece whenever it is practically possible. The coupling is made simply of a coiled piece of very heavy wire or a light bar surrounding the splice for its entire length, which should be equal to at least fifty diameters of the rods to be spliced (Figure 14). 11. In figure 80 a short block is shown loaded and compressed in one direction, thereby shortening the length of the vertical side from aa to bb. We notice now that the block expands in a horizontal direction, the diameter increasing from cc to dd. It requires very careful observation to discover this swelling in a concrete block, which usually fails along a diagonal such as ae, but, in any case, experiments with greased surfaces have shown 57 58 REINFORCED CONCRETE BUILDINGS that when the friction is eliminated, the block fails along vertical planes such as ff. 1 It is therefore clear that longitudinal rein- forcement in the direction of the compressive force is not very efficient, because the longitudinal rods simply add their own 1 ,b strength to that of the concrete. The rods act as slender columns and have a tendency to buckle, so that if no other provisions are made, the strength of the rods is practically nil. To prevent buckling, horizontal ties or " hoops " are introduced, but it is evident that unless closely spaced the hoops are of little value. If therefore the column or block is to have vertical reinforce- ment, it must have closely spaced horizontal hoops, and these in turn prevent the concrete from breaking apart along the vertical planes ff described. In this way the hoops become a very efficient means of reinforcing. 12. In order to understand this fully, let us consider a cylinder filled with water, one end being equipped with a water-tight but frictionless piston. This piston will carry an immense weight on its upper surface ; in fact, the entire system cannot fail before the water pressure within the cylinder exceeds the capacity of the cylinder walls, so that the cylinder bursts. The pressure within the cylinder is the same in all directions per unit of area; more particularly there is a horizontal (lateral) pressure on each and every square inch equal to the vertical pressure produced by the load on the piston. If now the cylinder is filled with sand in- stead of water, the conditions are only changed to this extent that the lateral pressure against the walls is less than before, so that it takes a greater load on the piston to burst the walls. Finally, if the cylinder is filled with liquid concrete, and the con- crete is allowed to set hard, the pressure on the walls will be even 1 According to tests by Foeppel and Mesnager. See, for instance, Con- sidere, "Reinforced Concrete," page 120. COMPRESSION AND LATERAL EXPANSION 59 less than before, but the concrete will stand much higher pressure when enclosed in the cylinder than when free. This, then, is the principle of the " hooped column," that the horizontal metal jacket prevents the concrete from spreading and thereby in- creases its carrying capacity. 1 For practical reasons it has been found impossible to use a continuous sheet of iron around the concrete; the horizontal reinforcement is always in the shape of hoops encircling the body of the concrete. Under pressure the concrete is sometimes seen to ooze out between the hoops, indicating the failure of the column, but usually the column fails by the bursting of the hoops or the complete disintegration of the concrete. In practical construc- tion this need not concern us, as the stresses naturally always are low; more 'important is the relatively great shortening of the hooped column under working loads. This objection is overcome by the rational use of vertical rods, so that the true " hooped column " contains both hoops and verticals (Figure 81) FIGURE 81. 13. The computation of a hooped column naturally centers around the calculation of the lateral pressure against the hoops. With a given concrete area F and a given load X the unit stress on the concrete becomes ^ IK / K 5 wnere -X" is m Ibs. F lbs ' /Sq - mch (and F in square inches. If we were dealing with water, the horizontal unit pressure would be the same. For concrete this is not the case; according to 1 Attention is called to some very interesting tests by Prof. Ira H. Wool- son, Eng. News, 1905, Nov. 2, Steel tubes, 4" in diameter and 12" long, | " thick walls, were filled with concrete. When seventeen days old, the tubes were tested in compression under loads as high as 120,000 to 150,000 Ibs. The tubes bent out of shape, and shortened 3", while the diameter increased from the original 4" to 5". When the tubes were removed, the concrete was found unbroken, solid, and perfect. See also Trautwine, 1909, p. 1160. 60 REINFORCED CONCRETE BUILDINGS experiment, the ratio between intensity of vertical stress and transverse stress is as 1 to 1/4.8. In other words, if the load produces a direct compressive stress of 4801bs./sq. inch, the lateral pressure would at the same time be 100 Ibs./sq. inch. It is now a simple matter to write up an expression giving the resistance due to the hoops, in a granular material having this same coeffi- cient 4.8. Let us denote by u the ratio between this resistance and the volume of metal in the hoops, and let us denote by U a similar ratio obtained between the resistance due to vertical reinforcement, and the volume of the material in the verticals. The expressions u and U will then give the effect produced by a unit of material, used as hoops and as verticals. We find, assuming the same stress in hoops and verticals: ^-M_24 U 2 which' shows that pound for pound, the steel employed in the hoops is 2.4 times as effective as steel employed for longitudinal reinforcement. 14. The question is now to find the effect of the verticals. Assuming that they are well tied so as to prevent buckling of the individual rods, the unit stress on the verticals must be r times the stress on the concrete, if the sections are to remain plane as assumed in Article 4. It is easy to see that this assumption is on the safe side, because, if the sections curved, the stress in the steel might be very much more than r times that on the con- crete, which latter forms the starting-point for our investigation. The value of r can only be indicated in a general way, as the properties of concrete vary greatly with the circumstances; let us assume r = 20. Then, if the unit stress on the concrete is 500 Ibs./sq. inch, the stress on the steel becomes 10,000 Ibs./sq. inch. Let F be the area of the concrete inside the hoops, and the allowable stress on this concrete C Ibs./sq. inch. Let p denote the percentage of the verticals with reference to the volume of concrete, then the effective concrete area is F- - 1UU 7? and the area of the longitudinals F T^TT ; hence the load carried by the concrete is C - F fnf) ^s. COMPRESSION AND LATERAL EXPANSION 61 Let the stress on the longitudinals be S Ibs./sq. inch, then their 7) share of the load becomes S F - Ibs. Disregarding for the moment the influence of the hoops, the total carrying capacity of the reinforced column is while if allowance be made for the hoops, the percentage of which is q with reference to the concrete section, we have an additional strength due to the hoops equal to 2.4 S F -^ and the total carrying capacity of the column becomes . (2) 15. The formula (2) above is the true formula for a reinforced concrete column and should always be used except in localities where the building code prevents its use, in which case formula (1) may be used. In any case, hoops must be used, otherwise the column steel is of no value as reinforcement. For the hooped column, Considere, the inventor and first experimenter, recom- mends p = q = 2, which, with C = 600 Ibs./sq. inch, and r = 20, gives X 2 = 1400 F. The hoops are spaced as closely as possible, leaving 1" to 2" clear space between the hoops to facilitate concreting. The spacing should under no circumstances exceed 1/6 of the diam- eter of the core. Finally the core is protected with a suffi- cient thickness of concrete to prevent rust and fire danger, about 1 " to 2 " of protection being required according to location and exposure. The plain column has a vertical reinforcement varying from one to ten per cent, of the concrete area, although reinforcement in excess of say five per cent, should be avoided on account of the uncertainty of the strength of columns reinforced with large amounts of steel. It is evident that hoops are indispensable also in these columns; it is quite common to see the hoops spaced one or even two feet apart; such hoops are of no use. The steel cannot be depended upon to carry its load unless securely 62 REINFORCED CONCRETE BUILDINGS tied, say, 1/3 to 1/2 column diameters apart. With p = 4, 8 = 12,000, C = 600, we have, for r = 20: X l = 1060 F. 16. Owing to difficulties in filling columns of small diameter, the diameter should not be much less than 10 " in any case, although there are many 8" columns on record. On account of the danger of " column failure " the length should not exceed 15 diameters. It is possible to advance a theory for " long " columns, but experience shows that columns exceeding 15 diam- eters in length are rare indeed except in roof stories where the calcu- lations often give very light sections. Moreover, all such theories depend alone upon theoretical considerations and have never been conclusively tested in the laboratory, so that in the rare cases where " long " columns are required it is better to make the col- umn a little larger and avoid the uncertainties of the theory. 17. In tall buildings, or in warehouses, the column bars become quite heavy, and it is necessary to join the bars of the column above with those of the column below in a substantial manner. The most satisfactory way is to square the ends of the bars carefully and join them in rather closely fitting sleeves, taking care that each bar has a full bearing on the bar below. Absolute certainty is had by cutting threads on both bars and sleeves, and drawing the bars together tight with the sleeve, but this must be done with great care and under strict supervision in order to be at all effective; unless carefully made this joint is worse than useless. When light bars are used they may be spliced by lapping by the required number of diameters, say about thirty, but this method is hardly to be recommended. Each bar of the story above should find bearing on a bar below; the number of bars therefore increases downward in the building. The number of bars in each story should be such that the bars can be symmetrically arranged in the column, unless there is some extraordinary reason for arranging them otherwise (excentric loads). The proper arrangement of the column bars may sometimes cause the designer to spend a good deal of time in working out the correct solution, but he may feel assured that this time is well spent. The hoops may be made from round or flat stock; the round stock may be obtained in long lengths and lends itself more readily to the requirements of the hooped column, especially where the COMPRESSION AND LATERAL EXPANSION 63 reinforcement is manufactured in the shop, with permanent devices for coiling and fastening the hoops to the longitudinals. The hooped reinforcement may also be bought ready-made; quite frequently the manufacturer overlooks the importance of having the spiral hoops in one continuous piece from top to FIGURE 82. COLUMN REINFORCEMENT Loomis Building, Cleveland, Ohio. Alexis Saurbrey, Consulting Engineer bottom, or, where the wire is joined, he makes a flimsy joint. It must be remembered that the hoops are tension-reinforcement and subject to all the rules governing the design of such bars. The best joint is made by simply bending the ends of the wire to the center of the column, making the loose end long enough to secure the requisite grip. The hoops may also be made in individual pieces, slipped over the previously erected verticals and wired in place. If the hoops are neatly made an excellent job may be had in this way (Figure 82). 64 REINFORCED CONCRETE BUILDINGS 18. It follows from what is said above that a hooped column should preferably be made of a circular cross-section, because in that case the hoops are subject to direct tension only. In many cases the expense incidental to the use of circular forms is prohibitive; the concrete may then be made square or octagon in section while the circular form is retained for the hoops. In either case only the concrete within the hoops can be taken into account in the calculations. Sometimes the hoops are made square or rectangular, in which case they are less effective, but we do not know how much. 19. The top and bottom of each column deserves special attention as the tests made so far seem to indicate that these are the weakest parts of the column, although there are many ex- ceptions to this rule. Suitable caps and bases are inexpensive, improve the appearance and increase the strength. Special investigation is always necessary at points where the concrete column finds a bearing on another material; the weight carried by the reinforcing rods must be distributed over such an area that the concrete in the column is not over-stressed. This is particularly true where the column rests on the footing; a steel base plate must be used to distribute the load on the rods, and the concrete must be enlarged so as to bring the average pressure within the allowable. This will be considered in detail under "footings." 20. Before leaving the subject of hooped columns, attention is called to the possibility of strengthening existing concrete columns with hoops wound around the outside of the column. In many cases it would be impossible to obtain satisfactory results in this manner, but when the concrete is of good quality, and the existing reinforcement is such as to give a sufficient amount of longitudinal reinforcement in the finished column, there should be no theoretical objections to this procedure. In practice it would of course be difficult to wrap the core tightly, but this is not absolutely necessary, as grout rich in cement may be forced between the hoops and the old concrete. Great care would be required in this operation, but it is not at all impossible, as has been shown by actual experiments on a small scale. 1 1 Considere: "Reinforced Concrete," page 175. The prism tested in this manner was allowed to set for three months, then wrapped with hoops and covered with cement, and tested after ten more days. The crushing COMPRESSION AND LATERAL EXPANSION - 65 21. In many cases columns are subject to excentric loads, so that, in addition to the direct compressive force, a bending moment exists and must be taken care of. This will be con- sidered in detail in Article 81. strength was 10,500 Ibs. per square inch. There were no longitudinals in this prism. CL AFTER VI BENDING 22. THE theory of bending used for reinforced concrete beams is different from the ordinary " theory of flexure " as used for homogeneous beams in a few particulars only, and this difference is more apparent than real. We consider here only the point of maximum bending moment; this is also the point of maximum depth, and we may assume both the compressive and tensile resultant to be normal to a vertical section through this particular point, under the particular loading described below. The notations used are as follows (Figures 83, 84) : d or D = depth from top of concrete to center of steel, inches. xd = depth from top of concrete to neutral axis, inches. xd x = ~j~ = ratio between the two preceding items. di = distance center of compression to center of ten- sion, inches. E c and E s = coefficients of elasticity for concrete and steel. Tjl r = -pr-= ratio between these coefficients. tic t or T = thickness of a flange, inches. 5 = width of flange considered, inches. B = ^iS = width of flange considered, feet. \2i n = thickness of stem of beam, inches. r c and r s = deformations of concrete and steel, at extreme fiber. C = unit stress on concrete in outside fiber, compres- sion, Ibs. per square inch. S = unit stress in steel, tension, Ibs. per square inch. a = area of steel, square inches. St. = total pull in steel in tons. Clj c 2 , c 3 = coefficients relating to balanced design of the section. a, = coefficients relating to T-beams with greater than minimum depth. 66 BENDING 67 w = dead plus live load on. slab, Ibs. per square foot. / = span in feet. q = factor of continuity. M = bending moment in tons-inches. m = 2000 M = bending moment in Ibs.-inches. 23. In regard to the load, we wi^ let all loads act in the same vertical plane along the center line of the beam as is usually the case in practical construction. This excludes at once all loads which would cause the beam to rotate around its longitudinal axis and all loads which would cause the beam to slide in its own direction. 24. In regard to the deformations, we will consider these as very small in comparison with the dimensions of the beam, so that the stresses are considered as acting upon the original cross- sections, not upon the deformed cross-sections or upon the deflected beam. 25. This does not mean that the change of shape of the section is of no importance. In figure 83 a vertical section is shown \ Shortening ;= < of Top Fibre, * Concrete Elongatio | of~Steel FIGURE 83. with the deformations produced by the bending of the beam; we assume sections plane before bending to remain plane after bending. Inspection of the diagram shows that the upper fibers are shortened, the lower fibers extended under the load; the neutral axis forms the division line between shortened and ex- tended fibers. The assumption of plane sections is evidently equivalent to assuming that the deformation of any fiber is in 68 REINFORCED CONCRETE BUILDINGS direct proportion to its distance from the neutral axis, and thus we get the equation: r -< = ** = _^_ , 3) r a (l-x)d l-x 26. We further assume that the stress on any small unit is directly proportional with the deformation; this gives the equa- tions : Q for concrete C = r c E c or r c = ^r o for steel S = r 8 E s or r s = 7 ___ T S S' EC (4) 27. We shall later 1 have occasion to use the moment of inertia of the section. It is therefore necessary to note that the assump- tions made in the preceding paragraphs are identically the same as those used in the " common theory of flexure " which leads to the well-known expression a _ ^ . e where v = stress per unit (5) I M = bending moment / = moment of inertia e = distance from neutral axis to fiber considered. The new feature in a reinforced concrete beam is now that in writing up the moment of inertia we have to disregard the con- crete below the neutral axis entirely, and instead consider the steel area. To this we shall return later. 28. Combining now equations 3 and 4 we find S E c hence x which expression determines the location of the neutral axis. 29. If now a vertical section is laid across the beam and stresses added on and in the section to represent the removed portion of the beam, the beam will remain in equilibrium. Let us project all forces and stresses on a horizontal line: then the 1 Article 79. BENDING 69 loads, being vertical, give no projections, and similarly the stresses acting in the vertical section itself disappear. There remain only the normal stresses acting against the section; as equilibrium presupposes that the sum of all the projected forces and stresses is zero, we have horizontal component of stresses on tension side horizontal component of stresses on compression side. Referring now to Figure 84, the area stressed in compression is xd inches high, b inches wide, and the average stress J C Ibs. per square inch. Hence total compression = \ C xd b Ibs. Denoting by s t the total pull in the steel in tons, we have, neglect- ing the tension in the concrete, total tension = s t 2000 Ibs. Hence s t 2000 = i Cxdb , . , . Cxdb c 2 , , which gives st = TTTT or s t = Jo" where c 2 = - (6) 30. Two more conditions must be fulfilled in order to create equilibrium: (1) the sum of all stresses and forces must be zero when projected upon a vertical line (when the loads are vertical, Article 23); this condition we will consider later under " U-bars." (2) The sum of all moments around any arbitrary point must be zero. Select for this point the point of application of the com- pressive stresses; the moment of the loads is then the " bending moment " m inch-lbs. The moment of the stresses is 2000 s t d\ inch-lbs. We must then have = m - 2000 s t . di but according to the diagram (Figure 84) d, = (1 - J x) d hence = m - 2000 s t ' (1 - J x) d. Eliminating st we find m = -I Cxb (1 - J x) d* hence d = - V inches where Ci = V I Cx (1 -| x) (7) c\ o Finally the steel area: a = s t square inches. 70 REINFORCED CONCRETE BUILDINGS 31. The formulas apply to all rectangular beams and therefore also to slabs. As we disregard the tensile resistance of the con- crete, the concrete below the neutral axis does not in any way enter into the calculations at this point, and the formulas are therefore also correct for T-beams where the bottom of the flange* coincides with the neutral axis. In this case the thickness of flange simply becomes t = xd inches. 32. We have now everything required to proceed with the design: S t Tons FIGURE 84. The depth in inches: The pull in the steel, tons: *~^T a = The thickness of flange, inches: t = xd 2000 Ihe steel area, square inches: a = o (8) (9) (10) (11) Simple as these formulas are they can only be used when the values of the coefficients x, ci and c 2 are known, and these values in turn depend upon the allowable stresses and the factor r. The Tables I, II, and III give full information in regard to the values of the coefficients; it will be noticed that the same tables may be used for any value of r, by simply shifting the position of the S-column in relation to the values of the coefficients. On the left the ordinarily used ^-column is indicated, corresponding to r = 15; while on the right, the ^-columns corresponding to r = 12 and r = 20 are show r n. Usually existing building codes and engineers' specifications call forr = 15 in bending-problems, but BENDING 71 this selection is arbitrary, and other values of r may very well be used. It is impossible to predict the coefficient of elasticity of concrete beforehand, and even if determined by careful ex- periment there is no reason to believe that it would remain the same on the building to be erected as in the laboratory, while it is quite certain that it changes materially from day to day as temperature and moisture affect the mixture used for the concrete. In Table IV values of the coefficient c 3 = 1 i x are indicated; the use of this table will be clear from the analysis above. In Table V the percentage of steel in a rectangular beam is indicated corresponding to r = 15; when the allowable stresses are decided upon, the percentage of steel in the section is a fixed quantity. 33. In the formulas above all dimensions are in inches, the moment in inch-lbs., the pull s t in tons. In practical design it is usually convenient to have the bending moment in inch-tons, M, and the width in feet, B. The formulas then become : The depth in inches : d = y -^- (8a) The pull in the steel, tons: s t = c%Bd ; (9a) The thickness of flange, inches: t = xd (10a) 2000 The steel area, square inches: a = ~ Si (11) These formulas are different from those given above in this respect only, that the figures handled are much smaller and there- fore it becomes easier to avoid mistakes, as figures of two or three places may be multiplied and divided, etc., approximately, without the use of paper and pencil, so that all calculations are easily verified. 34. The formulas given above apply, as stated, to slabs, to rectangular beams, and to T-beams in which the neutral axis coincides with the bottom line of the flange. Usually these two lines do not coincide, so that it becomes necessary to make further investigation in order to derive a general formula. The formulas given above have this peculiarity, that, for a given width of beam, the dimensions derived are minimum dimensions which cannot be decreased without adding to the stress on the material, thus exceeding the allowable stresses on which the design was based. Briefly stated, the problem before us consists in finding 72 REINFORCED CONCRETE BUILDINGS TABLE I. DEPTH OF NEUTRAL Axis = xd 1 r = 15 TABLE I x r = 12 r =20 S = 24,000 .158 .200 .238 .272 .304 .333 .360 .385 S = 19,200 22,000 .170 .214 .254 .290 .322 .352 .380 .405 17,600 20,000 .184 .231 .272 .310 .344 .375 .404 .429 16,000 18,000 .200 .250 .294 .333 .369 .400 .429 .454 14,400 16,000 .219 .272 .318 .360 .397 .429 .458 .483 12,800 S = 24,000 14,000 .244 .300 .349 .392 .429 .463 .491 .519 11,200 21,300 12,000 .272 .333 .385 .429 .468 .500 .529 .556 9,600 18,600 10,000 .310 .376 .429 .474 .513 .546 .574 .602 16,000 C = 300 400 500 600 700 800 900 1000 r = 15 r = 12 r= 20 TABLE II. EFFECTIVE DEPTH 12.9 /"M 1 /m d = 4/ or d = ->. - ci y B Cl y 6 r = 15 TABLE II ci r = 12 r = 20 S = 24,000 4.7 6.1 7.4 8.6 9.8 11.0 12.0 13.0 -S = 19,200 _ 22,000 4.9 6.3 7.6 8.8 10.0 11.2 12.2 13.2 17,600 20,000 5.1 6.5 7.9 9.1 10.3 11.5 12.5 13.5 16,000 18,000 5.3 6.8 8.1 9.4 10.6 11.8 12.8 13.9 14,400 16,000 5.5 7.0 8.4 9.7 11.0 12.1 13.2 14.2 12,800 S = 24,000 14,000 5.8 7.3 8.8 10.1 11.3 12.5 13.6 14.7 11,200 21,300 12,000 6.1 7.7 9.2 10.5 11.7 12.9 14.0 15.1 9,600 18,600 10,000 6.5 8.2 9.6 11.0 12.2 13.4 14.5 15.5 16,000 C = 300 400 500 600 700 800 900 1000 r = 15 r = 12 r =20 TABLE III. TOTAL PULL IN STEEL s. = 02 Bd or s f = 02 6d * 12 r = 15 TABLE III c 2 r = 12 r = 20 S = 24 : 000 .14 .24 .36 .49 .64 .80 .97 1.16 S = 19,200 22,000 .15 .26 .38 .52 .68 .85 1.03 1.22 17,600 20,000 .17 .28 .41 .56 .72 .90 1.10 1.29 16,000 18,000 .IS .30 .44 .60 .78 .96 1.16 1.36 14,400 16,000 .20 .33 .48 .65 .83 1.03 1.24 1.45 12,800 S = 24,000 14,000 .22 .36 .52 .71 .90 1.12 1.33 1.56 11,200 21,300 12000 .25 .40 .58 .77 .99 1.20 1.43 1.67 9,600 18.600 10,000 .28 .45 .64 .86 1.08 1.31 1.55 1.80 16,000 C = 300 400 500 600 700 800 900 1000 r = 15 r = 12 r = 20 BENDING 73 TABLE IV. ARM OF "COUPLE OF STRESSES." r = 15 TABLE IV c 3 = 1 - \x r = 12 r = 20 S = 24,000 .95 .93 .92 .91 .90 .89 .88 .87 5 = 19,200 22,000 .94 .93 .92 .90 .89 .88 .87 .87 17,600 20,000 .94 .92 .91 .90 .89 .88 .87 .86 16,000 18,000 .93 .92 .90 .89 .88 .87 .86 .85 14,400 16,000 .93 .91 .89 .88 .87 .86 .85 .84 12,800 S = 24,000 14,000 .92 .90 .88 .87 .86 .85 .84 .83 11,200 21,300 12,000 .91 .89 .87 .86 .84 .83 .82 .82 9,600 18,600 10,000 .90 .88 .86 .84 .83 .82 .81 .80 16,000 c = 300 400 500 600 700 800 900 1000 r = 15 r = 12 r = 20 Amount of Steel In Section P i*T * 100 -b =12 TABLE V r = 15 p S = 24,000 .098 .167 .247 .339 .442 .553 .672 .801 22,000 .115 .196 .288 .393 .510 .636 .771 .920 20,000 .138 .231 .339 .465 .602 .750 .907 1.07 18,000 .166 .278 .406 .553 .714 .884 1.07 1.26 16,000 .204 .339 .493 .672 .865 1.07 1.27 1.50 14,000 .261 .428 .621 .839 1.07 1.33 1.58 1.85 12,000 .339 .554 .799 1.07 1.36 1.66 1.98 2.31 10,000 .463 .750 1.07 1.42 1.79 2.17 2.57 2.99 C = 300 400 500 600 700 800 900 1000 the effect on the T-beam of an increase in depth, which must, in order to balance the design, be accompanied by a corresponding decrease in thickness of flange and amount of steel. 74 REINFORCED CONCRETE BUILDINGS 35. Let, then, Figure 85a represent a section of T-shape of minimum dimensions, having a depth d, a thickness of flange t a , and a total pull in the steel of s a tons. Let, further, Figure 856 represent a new section with a new, larger depth D = ad. b=12B ta = xd /Neutral Axis \ <: n > c 9 * r ad= /Neutral Axis | FIGURE 85a. FIGURE 856. The given M and B remain the same; we wish to determine the new values s b and T pertaining to Figure 856. We observe, then, that the proportionate depth x of the neutral axis is the same in the two beams, because the allowable stresses are the same, so that the depth of neutral axis is calculated as t a = xd in the first beam and as t b = xD in the second. The " effective depth" in the first beam is di = (1 | x) d and in the second approximately Di = (1 - %X) D = (1 - J X) ad The approximation consists in disregarding the tendency of the center of compression to rise on account of the removal of the concrete near the neutral axis; the discrepancy is negligible in most cases and on the safe side. Since now M , -- and M we get the equation s a = t*s b where s a = and, by reference to Figure 86 T-6 + 4000 C T 4000 Introducing these values in s a an equation (at - T) - n. 4000 sb we get after some reduction BENDING 75 ST\ Solving for ( ) and denoting by \t a J p b ~H i t a 2 1 r ] ' ta r \ a ta/ ft the value of this ratio we find n (L ^ } b' 1 3FC |- K"~ *>*^ s Neutral Axis -Jn <&- 1 ^k C =r at a-T j T C ~^T- | FIGURE 86. Showing the stresses in the beam of Figure 856. The thickness of flange in our new beam is now T = pt a = Pxd = --x-D a. and the new total pull in the steel is a. a a* corresponding to the new depth D = ad. 36. We have now the following general formulas for any re- inforced concrete T-section: 12.9 I'M The depth, in inches D = a. IM VB The total pull in the steel, in tons s< = -| The thickness of flange, in inches The steel area a = 2000 (13) (14) (15) s ( where M is the bending moment in tons-inches, a an arbitrary coefficient larger than unit, while the width B feet may be given or selected. The coefficient ft is derived from by formula 12 but to facilitate calculations, Table VI has been prepared giving the values of fi for various combinations of a and n/b. This latter ratio has little influence on the result within the ordinary limits, and Table IX may also be used in cases where n/b is 76 REINFORCED CONCRETE BUILDINGS different from 1/4, if the variation is not too large, although prepared especially for n/b = 1/4. 37. The theory of T-beams is of great importance as all the floor systems in common use involve this principle. Lately, beamless floors have come into use, and to these we shall return later; the beam and slab floors may be divided into two groups, the first including solid concrete floors, the second what is known as " tile-concrete " floors. The first of these two is by far the oldest, but the " tile-concrete " is gaining in favor with every day, and justly so, as its cost is less for light buildings owing primarily to the simplicity of the form work. The long flat ceilings are well adapted to modern store building and office- structures, especially where the loads are light and distributed. These floors have a flat portion supported on main girders A (Figure 87), the flat portion consisting of ribs B built between I c I cTc I I I Section A-A FIGURE 87. rows of hollow tiles C and a top covering of two or more inches of concrete, thus forming series of comparatively light T-beams side by side. The main girders are also of T-shape, the flanges being formed by leaving out the requisite number of tiles next to the stem of the girder. Sometimes lighter tiles D are used near the stem, in which case the flange becomes thinner than when the tiles are omitted entirely, Figure 88. The commercial sizes FIGURE 88. of tiles are usually 12" x 12" in plan, the depth ranging from 4" to 12" or even 16". When designing, it becomes necessary to proportion the depth of floor so as to allow for these commercial BENDING 77 sizes; the function of the tiles is simply to create a void in the concrete, and they do not enter into the calculated strength of the floor. The calculations require considerable time if exact, and tables VII and VIII have therefore been prepared for C = 700 and 5 = 20,000 and 16,000 respectively. These tables show at a glance the depth of tile and thickness of concrete required for any given bending moment, together with the corresponding pull in the steel. Note, however, that the bending moment must be calculated for a width b of slab equal to the distance between centers of ribs. If other allowable stresses are assumed than those for which the tables have been prepared, we may easily prepare new tables. We have in all the preceding formulas, that the bending moment is directly proportional to the square of the coefficient ci, while the total pull in the steel is directly proportional to the coefficient c 2 . But we have - = a coefficient times c 3 A glance at Table IV shows that c 3 itself is practically a constant within fairly wide limits, so that, for the allowable stresses in ordinary use, we may make = constant. c 2 It follows that the new tables are prepared from the tables here given by multiplying both the bending moment and the pull in the steel of the old table with a factor; this factor is the same for both items and is the new value of c 2 the old value of c 2 A completed floor of this kind is shown in Figure 89. 38. Flat Slabs. If, in formulas Sa and 9a, we make B = l, we have the slab formulas d = ^?i IM and s t But the load on the slab is usually given and in Ibs./sq. foot; denoting by w the total dead and live load in Ibs./sq. foot, the bending moment per foot width becomes M = - - I 2 - 12 tons-inches 78 REINFORCED CONCRETE BUILDINGS TABLE VI. INCREASING THE DEPTH FROM d TO D = ad See Table IX for special case .- = * o a = 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.1 .64 .62 .59 .56 .50 .44 .38 .25 .08 1.2 .54 .50 .46 .41 .34 .26 .15 1.3 .47 .42 .37 .31 .23 .12 1.4 .42 .37 .30 .23 .14 .01 1.5 .38 .32 .25 .16 1.6 .35 .28 .21 .11 1.7 .32 .25 .16 .06 ' 1.8 .30 .22 .13 .01 1.9 .28 .20 .09 2.0 .27 .18 .06 n/b = 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 TABLES OF = S t TONS TABLE VII S = 20,000 15 700 T " =7.0 393/20.6 6.5 372/20.0 6.0 350/19.3 5.5 276/17.1 329/18.6 5.0 204/14.9 255/16.5 308/17.8 4.5 189/14.3 236/15.7 284/16.9 4.0 62/8.3 93/10.1 132/12.1 174/13.6 216/14.8 259/15.9 3.5 55/7.7 85/ 9.6 120/11.4 158/12.7 196/13.8 235/14.8 3.0 26/5.3 48/7.2 76/ 9.1 108/10.6 141/11.7 175/12.2 210/13.6 2.5 21/4.8 42/6.8 67/ 8.4 94/ 9.7 123/10.6 153/11.5 184/12.3 2.0 17/4.4 37/6.2 57 / 7.6 80/ 8.6 104/ 9.4 130/10.2 158/10.9 1.5 14/3.9 29/5.4 46/ 6.5 65/ 7.3 85/ 8.0 107/ 8.7 132/ 9.3 1.0 10/3.2 21/4.4 35/ 5.0 50/ 5.7 66/ 6.4 84/ 6.9 105/ 7.5 0.5 6/2.3 13/3.0 23/ 3.5 33/ 4.0 46/ 4.5 60/ 5.0 76/ 5.5 0.0 1/0.6 4/1.1 9/ 1.6 15/ 2.0 23/ 2.5 33/ 3.0 45/ 3.5 H " = 4 6 8 10 12 14 16 BENDING 79 q 2000 TABLE VIII S = 16,000 C = 700 T " = 7.0 _ 443/23.9 6.5 420/23.0 6.0 396/22.2 5.5 312/19.7 372/21.4 5.0 231/17.1 288/19.0 348/20.5 4.5 214/16.4 266/18.0 320/19.4 4.0 70/9.5 105/11.6 149/13.9 197/15.6 244/17.0 292/18.3 3.5 62/8.9 98/11.0 136/13.1 178/14.6 221/15.9 266/17.0 3.0 29/6.1 55/8.3 86/10.5 122/12.2 159/13.5 198/14.6 238/15.7 2.5 24/5.5 47/7.8 75/ 9.7 106/11.1 138/12.2 173/13.2 208/14.2 2.0 19/5.1 41/7.1 64/ 8.7 90/ 9.9 118/10.8 147/11.7 179/12.5 1.5 16/4.5 33/6.2 52/ 7.5 73/ 8.4 98/ 9.2 121/10.0 149/10.7 1.0 11/3.7 24/5.1 40/ 5.8 56/ 6.6 74/ 7.4 94/ 7.9 119/ 8.6 0.5 7/2.6 15/3.5 26/ 4.0 38/ 4.6 52/ 5.2 68/ 5.8 86/ 6.3 0.0 1/0.7 5/1.3 10/ 1.8 17/ 2.3 26/ 2.9 38/ 3.5 51/ 4.0 H" = 4 6 8 10 12 14 16 which gives the very convenient formulas (16) d =-.J inches Ci V q and s t = c%d tons. (17) Here the span I is expressed in feet, and the factor q is equal to 8 for non-continuous construction, and from 10 to 16 for con- tinuous construction. 39. If reinforced in both directions, and supported on all four sides, the slab is calculated by the formulas above, dividing the total load w into two portions w\ and w 2 where wi -f- w 2 = w. If we denote by LI and L 2 the span in each direction, we have the arbitrary formulas for the division of the load: and = W w<> = w Li 4 + I* 4 The heaviest of these is now assigned to the shortest span, and determines the depth and the reinforcement running the short way, the cross reinforcement is designed in a similar manner using the other load. In case of square panels the two loads become equal, each one-half of the total. The formula is entirely irrational and the only reason it is 80 REINFORCED CONCRETE BUILDINGS TABLE IX - =| o /S = a - \/M 2 - 1) a 2 a jS /3/a 1.00 .000 1.000 1.00 1.01 .005 0.890 0.89 1.02 .099 0.847 0.84 .03 .015 0.815 0.80 .04 .020 0.789 0.77 .05 .025 0.767 0.75 .06 .030 0.747 0.73 .08 1.039 0.712 0.69 1.10 1.049 0.684 0.65 1.12 1.058 0.658 0.62 1.14 1.068 0.636 0.60 1.16 1.077 0.616 '0.57 1.18 1.086 0.596 0.55 1.20 1.096 0.579 0.53 1.25 1.118 0.541 0.49 .30 1.140 0.508 0.44 .35 1.162 0.479 0.41 .40 1.183 0.452 0.38 .45 1.204 0.430 0.36 .50 1.225 0.409 0.33 .60 1.265 0.371 0.29 .70 1.304 0.338 0.26 .80 1.342 0.308 0.23 .90 1.378 0.283 0.21 2.00 1.414 0.258 0.18 2.25 1.500 0.209 0.14 2.50 1.581 0.167 0.11 2.75 1.658 0.130 0.078 3.00 1.732 0.099 0.058 3.50 1.870 0.045 0.024 4.00 2.000 0.000 0.000 a 2 a /3/a given here is that it is on the safe side and better than other existing formulas. The supporting girders are designed with reference to the load brought upon them by the particular direction which they support, and the bending moment is often increased above the BENDING 81 calculated because the load seems rather more concentrated towards the center. FIGURE 89. TILE CONCRETE CONSTRUCTION, READY FOR PLASTERER Wise Building, Cleveland, Ohio. Alexis Saurbrey, Consulting Engineer Discussion of Tables I to IX. 1 40. TABLE I. x = -3 rC The value of x determines the location of the neutral axis, xd being the distance from compression face to neutral axis. 82 REINFORCED CONCRETE BUILDINGS We have seen that the position of the neutral axis within the section of a T-shaped beam leads to the division of T-beams into two groups, according to whether the neutral axis falls, above or below the bottom line of the flange. In this latter case we introduce the coefficients a and /?, and the problems contain an arbitrary element which is absent in beams of the first type, where the dimensions depend mutually upon one another as in formula (8). The table shows that the neutral axis only in exceptional cases approaches the middle of the beam where it is located in all symmetrical beams following Hooke's Law (steel for example). It is obvious that a greater amount of steel is required for low steel stresses than for high; we therefore see that the neutral axis is lowered by increasing the amount of steel. 41. TABLE II. Cl = \J\Cx (\ - x , 12.9 I'M We have d = - The smallest possible value of d is obtained when a large value of Ci is used, or, in other words, when high concrete stresses are combined with low steel stresses. The influence of the concrete stresses is much more pronounced than that of the steel stresses; it is, therefore, not economy to increase the amount of steel in order to save on the concrete. It is not impossible to analyze this problem mathematically, but owing to variation in unit prices it seems hardly worth while. The possibility of decreas- ing the depth of construction by using high concrete stresses and low steel stresses may, however, be of importance in special cases where the head room is limited. Cx 42. TABLE III. C2 = 333 The total pull in the steel is st = 02. ^ d. The total amount of steel is a = s t 12 2000 S 2000 bd fa\fS so that -- ' C2 ' 12 r C2 = (bd) VT66 The coefficient c 2 , then, is a measure of the amount of steel used BENDING 83 for a given cross-section, bd being the area of the cross-section in square inches. We note that ^-, times 100 bd is the percentage of steel required for the beam; if we denote the percentage by p we have 16600 This expression has been used in calculating Table V. 43. TABLE IV. c 3 = 1 - I x In general the total pull in the steel is obtained by dividing the bending moment by a certain lever arm di, equal in length to the distance between the centers of compression and tension. Reference to Figure 84 gives at once di = | xd + (1 - x) d = (1 - \x) d = c,d When d is known we get d\ as c$d; Table IV gives the values of c 3 for various combinations of stresses. 44. TABLE V. p = ^ X 100 = c 2 The percentage of steel has but little interest for the prac- tical designer as the problems usually present themselves. The table is added for the convenience of those who are in the habit of selecting the percentage of steel rather than determine the allowable stresses. The table is correct only for such beams where a = 1 and r = 15. 45. TABLE VI will be found useful when designing T-beams of larger than minimum depth. When we have selected , as explained in connection with formula 12, etc., the correspond- ing ft is found by Table VI for any value of n/b. The method of design will be clearly evident from the example in Article 47. Table IX is a more extensive table for the special case where n/b = J, which is a common value in practice. The variation of n/b does not affect the values very much, so that for small values of a Table IX may be used for other values of n/b than J. 46. TABLES VII and VIII. Tile-concrete floors. The use of these tables is best explained by an example. The span of the flat portion is 20 feet; the total dead and live 84 REINFORCED CONCRETE BUILDINGS load is assumed to be 250 Ibs. per square foot. With 4" ribs we get the width of beam b = 4 + 12 = 16" and the corresponding bending moment in inch-tons 250 M = * x ^ x 2 2 X 16 = 100 inch-tons. If the allowable stresses are S = 20,000 and C = 700, we must use Table VII, and we see at once that we can use either a 10 " tile with about 2f" concrete, or a 12" tile with 2" concrete. As we do not wish to have less than 2" of concrete over the tiles, we cannot use the larger tiles economically. If we select 12" tiles and 2" concrete, the corresponding pull in the steel is 9.4 tons according to the table, requiring .94 square inches of steel, for instance, one f " square and one f " square bar. It should not cause surprise that the moments tabulated in Table VIII are larger than the corresponding values of Table VII, although the allowable stress on the steel is smallest in Table VIII. The explanation is given in the remarks under Table II, Article 41, and, in accordance with the statements made there, it will be seen that the larger moments of Table VIII are obtained only by increasing the steel areas. TABLE IX. See Table VI, Article 45. 47. EXAMPLE 1. T-Sections. Continuing the example given above under the discussion of Tables VII and VIII, we proceed as follows to design the girder: The load on the floor is 250 Ibs. per square foot, the span of the flat portion on each side of the girder is 20' 0", and the girder therefore carries a load of 250 X 20 = 5,000 Ibs. per lineal foot, to which should be added the weight of the girder itself. Assuming this item to be included in the 5,000 Ibs., and assuming a span of 24' 0" for the girder, the bending moment on the girder becomes M = J X x 24 2 X 12 = 2160 inch-tons. We decide to use high tension steel, for which S = 20,000, and we allow C = 700 Ibs. per square inch on the concrete. We get then from Table II : Ci = 10.3; from Table III : c 2 = .72; and from Table I: x = .344, and we may now proceed with the design, using formulas (8a), (9a), (lOa), and (11). The width of flange, BENDING 85 B, may be selected arbitrarily. Let us make B = 4' 0". Then, by (9a) ......... s = c 2 BD = .72 X 4 X 29.1 = 84 tons. (11) ......... a = ? - s t = X 84 = 8.4 square inches. (10a) t = xd = .344 X 29.1 = 10" We have to make the stem of the beam wide enough to accom- modate 8.4 square inches of steel, say n = 12", and the girder is then designed as far as concerns the bending moment. Ques- tions pertaining to shear, etc., will be considered later. We can, if we desire, reduce the thickness t of the flange by increasing the depth d. While this operation is not always neces- sary, or even desirable, we will nevertheless continue the ex- ample to show the method of procedure. If, then, we increase the depth from 29.1" to, say, 35", we get a = J^_ = a6a>2 the coefficient a indicating the proportionate increase in the depth. The value of ft is next obtained by Table IX, remem- bering that i _ OK 6 " *~ the stem being 12" wide and the flange 48". For - = .25 and a = 1.2, Table IX gives ft = .43; then, 6 by the theory outlined for this case, Article 36, we have The new depth D = ad = 1.2 X 29.1 = 35". s 84 The new pull in steel s& = = y-x = 70 tons. The new thickness of flange T = ftt = .43 X 10 = 4.3". It is of course unnecessary to calculate the dimensions of the " minimum " beam first, as done here, unless we expressly desire to have these dimensions. Let us, for instance, again consider a given bending moment of 2,160 inch-tons; let us 86 REINFORCED CONCRETE BUILDINGS further select arbitrarily the width B = 4' 0", and let us finally choose the coefficient a = 1.2, then, by Table IX, we 77 get ft = .43 for an estimated value of = .25; and we also o have a2 = 1.44. We may now find the dimensions directly, by 12.9 /M 12.9 2l60 '^'V- = 12.9 1, above the bottom of the flange. If we now wish to determine the stresses in a given beam, we begin by selecting r, next we determine the value of the cri- terion, so that, if equal to unit, we use formula (21), while if larger than unit, we use formula (19), and if smaller, the orig- inal formula. Then the location of the neutral axis is calcu- lated from 1 and the coefficient c 3 = 1 f x is determined from the value of x just found. The effective depth is then where d is the depth from ultimate compression fiber to the cen- ter of the steel. We have then s =- and, by (11) 2000 90 REINFORCED CONCRETE BUILDINGS We know now S and S/C, and it is a simple matter to determine C. 50. EXAMPLE 3. T-SECTION. Given the beam shown in Figure 92, find the stresses, when the bending moment is 2,160 inch-tons. We have T 2 4.5 X 4.5 = 20.25 2 VH ~ 2 X 2.19 X 30.5 " 113.5 V 7 sq. inches FIGURE 92. which is -evidently < 1. Using therefore formula (18) we find 325.4 _ 15X C ~ 5 X 168.4 Now x = 1 ^ = .342 and c 3 = 1 - J a; = 0.886. Hence di = c 3 d = 0.886 X 35 = 31" Bending moment 2,160 inch-tons, then 2160 31 = 69.6 tons. 69.6 = 20,100 20100 , c = W = 695 ' or about 20,000 and 700 Ibs. /square inch for steel and concrete, respectively. 51. EXAMPLE 4. T-SECTION. Special case. Given the beam shown in Figure 93. We have V = = 15 X ~ = 2.62. 10 X 10 = 100 2 VH ~ 2 X 2.62 X 19.1 == 100 = 1. BENDING 91 Use formula (21) which gives S 10 X48 C = = 28.6 2 X8.4 The balance of the calculations may now be continued ex- actly as in the preceding example. j-8.4 sq. inches FIGURE 93. 52. EXAMPLE 5. RECTANGULAR BEAM. Slabs. Given the rectangular beam shown in Figure 94; we use formula (19) which gives C 2 + 2\ V_0.85 sq. inches FIGURE 94. 225 + ^ ' 0.85 CHAPTER VII TRANSVERSE STRESSES. U-BARS 53. IN addition to the longitudinal stresses examined in the preceding articles, transverse stresses exist in reinforced concrete beams as well as in beams of other materials. But the transverse stresses are different in trusses and in solid beams : in the truss, each individual member is stressed in its longi- tudinal direction only, and there is no shear. In the solid beam, longitudinal stresses exist in the top and bottom chords or fibers, and the web is then subject to shear stresses both longi- tudinally and transversely. In special cases these shear stresses may vanish, as for instance in the I-shaped steel beam of vari- able depth, when the ratio bending moment at any point depth at the same point is a constant. This is the case in a parabolic girder loaded over its entire length with a uniformly distributed load. 54. In view of this difference between trussed beams and solid beams, it becomes necessary to decide whether to treat the reinforced concrete beam as the one or the other. To the eye a reinforced concrete beam certainly appears solid enough, and such is indeed the case when the beam is first made and the load is being put on. But when the load reaches a certain in- tensity the " solidity" of the beam is destroyed. Slight cracks soon become evident, at least when arrangement has been made to observe them, and that under loads corresponding to a steel stress of from 4,000 to 6,000 Ibs./square inch, or a concrete stress of 350 Ibs./square inch. It follows that under the ordi- nary working load our reinforced concrete beam is perforated with cracks extending from the bottom fiber up toward the neutral axis, without quite reaching the neutral axis, so that, under any circumstances, the beam is certainly not a "solid" beam. These hair cracks have been noted by all who have 92 TRANSVERSE STRESSES. U-BARS 93 taken the trouble to look for them with but one exception (Considere) ; they are not an occasional occurrence, but a uni- versally recognized phenomenon of the greatest importance for our understanding of the stresses within a reinforced concrete beam. The presence of these cracks is accounted for by the simple fact that concrete is unable to stretch as much as steel before cracking, so that, under a certain load, the concrete refuses to follow the steel in its elongation and goes to pieces. The cracks of this class appear throughout the length of the beam, fairly uniformly spaced, and increase in size with increasing load. 55. The crack of course is an open space existing between surfaces which at some earlier time were in close contact and united. We must now understand as a fundamental principle that stresses cannot be transmitted through open cracks. Com- pression may be transmitted through a contact only, and fric- tion may exist on surfaces pressed together, but no kind of stress will jump across an open space. It follows that shear in the ordinary sense of the word cannot exist in a reinforced concrete beam loaded above a certain limit, because the nature of shear requires equal intensity on a horizontal and a vertical plane, and this is of course impossible when the beam has ver- tical cracks. Or, we may simply say that the vertical shear cannot exist in the crack itself. Where a crack occurs there is therefore nothing but the compression flange and the tension steel to carry the shear, a distribution of the shear which is, to say the least, not easily reconciled with current ideas of shear in solid beams. 56. Entirely different from these hair cracks are the much larger, pronounced failure cracks which predict the approach- .1 t FIGURE 95. ing collapse of the test beam. If located at or near the point of maximum bending moment, they are undoubtedly due to excessive elongation of the steel disclosing a failure by tension in the steel; if near the end, the crack usually takes the shape shown in Figure 95, either with or without the horizontal crack 94 REINFORCED CONCRETE BUILDINGS D. The vertical crack E is wide open, especially at the bottom, decreasing in width as it approaches the top of the beam. As the steel stress at this point certainly cannot exceed the steel stress at the point of maximum bending moment, this crack is not due to excessive tensile stresses in the steel. It must be due to sliding of the reinforcement: the steel is pulling out of the end of the beam at the same time bursting its concrete envelope, and causing the horizontal crack. Let it be under- stood that no amount of shear will cause a gaping crack, but once sliding sets in and causes the vertical crack, it is clear that the one end of the beam will be compelled to revolve around the other end, causing in the first place the double-curved line of cleavage, and, secondly, great friction on the surfaces of contact. 57. The above remarks lead to the conclusion that a concrete beam is a solid beam up to a certain load at which point the tensile resistance of the concrete is exhausted, and a readjust- FIGURE 96. ment of stresses takes place within the beam. This readjustment is different for different types of beams.' In a rectangular beam (Figure 96) we may well assume that the com- pression follows lines as AC and BC when the load is placed FIGURE 97. at C; if the load moves to D, the lines change to AD and DB. Under these circumstances there is no shear, at least not in the ordinary sense of the word. We may compare a system of this TRANSVERSE STRESSES. U-BARS 95 kind to a triangular frame with hinged corners (Figure 97). The chords AC and BC will be in compression, and the chord A B in tension, hence at A and B the hinges are subject to severe stresses. The same is the case at A and B in the reinforced concrete beam, so that the " length of embedment " AE and BF in Figure 96 must be made long enough to prevent sliding of the rod. The shear existing in a system of this kind is that negligible quantity caused by the stiffness of the system as a whole, a kind of friction caused by the lack of flexibility at the supposed hinges. The system A BC is an equilibrium curve for the load C and the reactions A and B', this same argument would of course hold true for any number of forces, or even for uniformly dis- tributed loads, in which latter case the compression curve would be a continuously arched curve from A to B (when the load covers the entire span). But if the beam under consideration is a T-beam instead of a rectangular beam it becomes impos- sible to make the compression line curve down to the support- ing points, except for a width equal to that of the stem. A T-beam (Figure 98a) may be considered as consisting of two FIGURE 98a. FIGURE 986. FIGURE 98c. beams side by side; a T-beam proper (Figure 986) and a rect- angular beam (Figure 98c). In this rectangular portion it is quite possible for the compression lines to dip down at the sup- ports, but not so for the T-beam portion, there being no con- crete left to carry the stresses down to the steel. This leads to the idea of bending the steel up over the support to meet the compression flange, reversing the conditions shown in Figure 96. 58. Let us consider a portion of a reinforced concrete beam between two points a and b (Figure 99). The bending moment at a is denoted by M a , the distance between center of com- 96 REINFORCED CONCRETE BUILDINGS pression and center of tension by d a , then the total pull in the steel at a is M a Sa = -T- and at b s b = Mi d b FIGURE 99. The difference between these two is _M a _ M b ' Sa Sb - T a ~d b The prefix A simply denotes the difference in the item con- sidered so that As means the variation of s between the points in question. It is now evident that the portion abed is subject to two pulling forces acting near its lower end cd: one force s a pulling toward the left, another pulling toward the right, s b . If s a is larger than s b , the end cd must have a tendency to move toward the left in precisely the same manner as if pulled that way by a force equal to the difference of the two pulling forces; we may therefore consider the force As = s a s b as acting alone. This condition is represented in Figure 100, and this diagram 5 *f f F f 1 f ) h As^ 1 I f B ^j ^ Al FIGURE 100. shows at once that cdef is a cantilever fixed at its base ef, and loaded near its end with a load As. The depth ce we do not know at the present time; let us indicate this unknown quan- TRANSVERSE STRESSES. U-BARS 97 tity by h. The bending moment on the cantilever is then h As; the arm of " the couple of stresses " in the cantilever is c 3 AZ; hence if a vertical reinforcing rod is disposed near bd the pull on this rod becomes But this pull k can exist only when counterbalanced by a cor- responding compression, so that the beam becomes a trussed beam as shown in Figure 101. The vertical reinforcement J /] xx K |> h K XI Fl X X 1 ^ 1 I N N * 1 x 1^ \t \| \j A t R a B C!R C FIGURE 101. designed in this manner is usually made of a bar bent to U- shape and circling the main tension rod (Figure 102a, 6); they are therefore called U-bars or stirrups. The U-bar is unneces- FIGURE 102a. FIGURE 1026. sary when k = 0, which is always the case when As = 0: i.e., when either the tension chord, or the compression chord, or both together, follow the equilibrium curve. As shown above, this is always the case in a rectangular beam with well-anchored reinforcement, and it is also the case for such parts of a T-beam in which the reinforcement is bent up to follow the equilibrium curve. In all other cases k has a definite value. For straight reinforcement and straight top chord, we have (Figure 103) : d a = db = Cj,d M a - M b hence As = 98 REINFORCED CONCRETE BUILDINGS where and hence or M a = Ra - M b = R (a - AQ - 2P (p - AZ) M a - M b = AZ (R - A S = - (R - and k = ^ (R - It is now easy to understand that the length h cannot exceed 1r. r sr a-Al FIGURE 103. the distance from the center of the steel to the neutral axis. This gives h (1 - x)d 1 - x The value of this expression cannot exceed unit; for ordinary cases its value is about three-fourths. Hence the maximum possible value of k, for the conditions named, is: kmax == K Zr. (23) N A , mi: FIGURE 104. 59. It is now interesting to note that this same expression may be obtained directly in the simplest manner. Let, in Figure 104, the section A B remove the right end of the beam TRANSVERSE STRESSES. U-BARS 99 leaving the main tension bar and the U-bar projecting. The stress-resultants acting upon A B are then, when the chords are parallel: the horizontal compression X, the horizontal pull Y'j and the vertical force k in the U-bar. The loads are PI, P 2 , etc., and the reaction R. If we now project on a vertical line MN, the horizontal stresses vanish and we have R- (P! + P 2 + P 3 + ...... ) =k or k = R - 2P. 60. The beam with straight top- and bottom-chords is an exception. Usually the stem of a T-beam may be considered as in equilibrium, and in addition some of the bars are bent up in the T-beam portion to approximate the equilibrium curve, so that a material reduction in the value of k takes place in all practical beams. With the notations of Figure 98 we have, on account of the stem, a reduction equal to b n , j b n ro vrn r hence k = r [R 2P] If out of the total number x of bars in the T-beam, a certain number y follow the equilibrium curve, we have a further re- duction equal to fc = ?LH . *r_ . [R _ 2P] (24) x x b Thus, if b = 48" and n = 12", we have b - n _ 48 - 12 _ 3 ~b~ 48 ~ 4 so that, out of a total number of say eight bars, the six belong to the T-beam. If out of these six, two are bent as required, we have x = 6 and y = 2, hence k = i X i X [R - 2P] = } - [R - SP] 61. Thus, in order to calculate the stress on the U-bars, it becomes necessary to know the properties of the curve of equi- librium for the system. When the loads are stationary, the curve is drawn as a force polygon to the actual loads and reac- tions. For a uniform load, covering the entire span, this curve is a parabola; it is not practical to bend the bars to this shape, but it may be closely approximated by a system of bars with straight portions between the several bents. A uniformly 100 REINFORCED CONCRETE BUILDINGS distributed, moving load has no definite curve of equilibrium, so that in that case the most dangerous position of the load must be found and the U-bars proportioned according to For- mula 23 above, while the bent bars are arranged to meet the requirements of some particular type of loading, for instance, the total load. Similarly, concentrated loads may be either stationary or moving. In buildings the concentrated loads are usually stationary. The given load is a uniform load, so that the beams are loaded as explained above; these beams in turn frame into the girders, one, two, or three beams to each span, and these concentrated beam loads are stationary. It is a simple matter to bend the main tension bars to conform to this type of loading; examples are given in Figures 105 and 106. I I FIGURE 105. FIGURE 106. The moving concentrated load is usually found only in structures like highway bridges, subject to steam-roller traffic, in crane- track girders, etc. In such cases, the live load is large in pro- portion to the dead weight of structure and covering, so that the T-beams are usually not economical structures for this class of girders. They may be constructed by using the adequate number of U-bars; or rectangular beams may be used of the required cross-section. 62. The problem of designing a T-beam under a uniform load confronts the reinforced concrete designer every day. It is customary to consider the load as covering the entire span, except in cases where it is expressly stipulated that the most dangerous position of the load shall form the basis for the cal- culation of the U-bars. Arguments may be advanced pro et con., usually the load specified is a maximum load which seldom, if ever, covers the entire beam, and the designer will have to use his best judgment as to what constitutes proper practice in each individual case. It is hardly necessary to say that in other lines of engineering the most dangerous condition is always considered in making the calculations as a matter of course, and there is no reason why other professional ethics should prevail when dealing with reinforced concrete. TRANSVERSE STRESSES. U-BARS 101 In Figure 107 the moment-curve is shown corresponding to a uniformly distributed load covering the entire span. The maximum moment is taken as unit, and the several ordinates of the curve are given under the assumption that there is no continuity. The reinforcement must be made to conform to n 12 .75 .89 1X0 .97 FIGURE 107. this curve as closely as possible, hence we see that at points 3 and 9, only f of the total number of bars is required, at 2 and 10, slightly more than J, and less than J is required at points 1 and 11. The quota of bars not required may and should be bent up at the points specified, provided that no other kinds of loading can occur. In Figure 108 the corresponding curve is FIGURE 108. shown when the beam is considered as continuous, with q = 10. 63. The entire theory outlined for the calculation of the U-bars is based upon the assumption that sliding of the steel cannot take place. In such cases where the anchorage beyond or at the supports is insufficient to prevent sliding of the main tension bars the factor of reduction must be decreased, so that a correspondingly larger amount of vertical reinforcement is used for the U-bars. In the present state of our knowledge this must be taken care of by judgment alone, there being no way of cal- culating a beam with inefficient anchorage. It must here be sufficient to point to the fact that the U-bars retard the sliding of the reinforcement, and that, for that reason, light U-bars should always be used even in cases where the theoretical considerations show that they may be dispensed with. This applies particularly to rectangular beams. 102 REINFORCED CONCRETE BUILDINGS 64. Spacing of the U-Bars. It will be noted that the entire line of argument advanced in the preceding paragraph is based principally upon the inability of the concrete to resist tensile stresses, and that the entire problem finally resolves itself into one of tension carried entirely on the steel, and compression carried entirely on the concrete. The word " shear " is referred to incidentally only, and this is a natural consequence of the fundamental principle of disregarding the tensile stresses in the concrete. As this development leads to rather important results, it may be well to consider these matters a little more in detail. 65. Figure 109 shows the simplest conceivable system of material units, i.e., three particles, A, B, and C. Whatever the nature of the force uniting these par- tides, if the particle C is moved to the position D through the influence of some \ external force, the displacement CD rep- . \ resents in all cases the result of the influ- __ _\; ence of that force and is called the "shear A B deformation " if parallel with the line FIGURE 109. . T , . ,., , ,, AD. It is readily seen, however, that the more direct and more readily understood deformations are (1) the lengthening of AC to AD, and (2) the shortening of BC to BD. Hence this shear deformation CD is nothing but the resultant of the deformations along the original lines AC and BC, and we perceive that even in the most complicated system of particles any deformation may be reduced to a sys- tem of lengthenings and shortenings, that is, tension and com- pression, if we speak of stresses instead of deformations. The word " shear," therefore, has no real or material meaning, except as a pure figure of speech to express in one short word a rather intricate condition of tensile and compressive rela- tions, in precisely the same manner as the word " bending moment " is used to indicate a mathematical conception of the mutual condition of a number of forces acting upon a beam. Needless to say that nobody has ever seen, or will ever see, a bending moment in the realm of things as they are, and that whoever undertakes to explain the so-called " shear stresses " in a solid body will ultimately have to account for pure tensional and compressional stresses. TRANSVERSE STRESSES. U-BARS 103 66. If two material bodies are in contact, the stresses act- ing in the contact surface are termed frictional stresses which, as far as the materials themselves are concerned, are compres- sive stresses with no possibility of accompanying tensile stresses in the direction perpendicular to the contact surface. 67. Of the nature and extent of frictional stresses we know next to nothing. A force acting parallel with the contact sur- face will cause sliding of one body in relation to the other; if the force is inclined, the sliding becomes increasingly difficult as the angle of the force increases, and the sliding becomes impossible when the angle at which the force acts exceeds the " angle of friction," which has a definite value for each material, depending in part upon the character of the surface. For con- crete upon concrete, this angle appears to be near 41. 68. In certain types of reinforced concrete construction the floor beams are not made in one continuous operation with the floor slab resting upon the beams, and U-bars or similar mechani- cal devices are then resorted to in order to tie the slab and stem together, and to so unite them that they may be considered as acting as one piece. In this case, the slab would form the upper flange of a T-beam, and in order to insure this action, sliding between flange and stem must be prevented. Figure 110 represents a portion of a beam, the lines AC, CD, and CB c P 'Mf ;|i *" |, \M FIGURE 110. indicate the directions of the principal stresses. If now the line of diagonal compression BC is inclined so that the angle BCD is less than the angle of friction, the flange would slide in relation to the stem, on account of the joint along the line MM; the U-bars AC and BD would resist this tendency by virtue of their " shear " resistance (and this resistance we know is very small, and cannot exceed the compressive edge resist- ance of the concrete; see Figures 111, 112, where the black areas indicate the crushed concrete). If, on the other hand, the angle BCD is larger than the angle of friction, then there can be no 104 REINFORCED CONCRETE BUILDINGS sliding, and therefore no shear stresses on the U-bars, which will act directly in tension as described above. The rule derived from this argument may be briefly expressed thus: The spacing of the U-bars must not exceed the depth of the beam, in which case the angle of forces would be about 45. FIGURE 111. FIGURE 112. 69. If we now turn to the T-beam manufactured in one continuous operation, where no separation exists between stem and slab, we note that, theoretically at least, this beam is in the same condition as the one just considered, owing to the orig- inal assumption whereby the tensile stresses in the concrete are considered as non-existing. Each and every horizontal stratum must be considered as isolated and influenced by its neighbor through the medium of frictional resistance only, and the direction of the diagonal compression must be such that no sliding can take place. The same rule must therefore be imposed in this case. But this rule gives the maximum spacing possible: owing to the usual considerations of a margin of safety, the spacing must be made smaller, and we would therefore recommend that the spacing of the U-bars must in no case exceed one-half the effective depth of the beam. 70. Tensile Stresses in Concrete Disregarded. The ready- made reinforced concrete beam formulas now in common use are derived under the apparent assumption that the steel rein- forcement takes all the tensile stresses, and this is also the case in this book. In reality, we cannot wholly disregard these ten- sile stresses in the concrete, or, at least, we cannot deny their existence, because if we did, we would also rob the concrete of its cohesion, and we would have a granular mass such as sand or crushed stone, wholly unsuitable for our purpose. The true statement is that we disregard the tensile stresses in certain directions and for certain purposes. In this book, we have considered the concrete as fractured vertically along the planes TRANSVERSE STRESSES. U-BARS 105 of the U-bars (1) because the cracks in probability will appear in the weakest plane, there being less concrete to resist the ten- sion where the concrete is displaced by the steel of the U-bar; (2) because the U-bar encircling the main tension rod in a meas- ure acts as a washer on the rod, causing the somewhat resilient concrete to crack immediately behind this point of gripping; and (3) because such tests as throw any light upon the location of the cracks indicate that they occur very largely at just these points. 71. NOTE: For the gripping action of a loose U-bar encir- cling the tension rod, see Morsch, page 47. For the location of the cracks, see the same book, page 155. * 72. We have also considered the stem of our T-beam ' as composed of horizontal layers acting upon one another by con- tact only, and thereby determined the spacing of the U-bars. But between these vertical and horizontal lines of weakness, we have assumed the concrete to be solid. Hence, we have assigned to the concrete a certain amount of tensile resistance in certain locations and directions. 73. It follows that with increasing loads the compressive stresses in the beam do not increase as rapidly as the load, especially not in beams where the slab and the stem are sep- arately manufactured. In such beams, the compression at rupture must in many cases be uniformly distributed over the entire compressive zone, and we find here the explanation of the fact sometimes observed that the compressive strength of concrete is much higher in a beam test than in a cube test. An analysis of these conditions would be interesting and of great value practically. 74. Details of Reinforcement. The various arguments advanced above will lead to rational design of the steel if con- sistently applied, and there is but little new to add. The great principle in all beam construction is that there is a compres- 1 Morsc^i: Concrete Steel Construction, 1909. While the cracks do not all occur at the U-bars, the tendency is fairly pronounced, especially in the beams with U-bars in one half only, see Figure 149, Beam V; Figure 153, Beam VIII; Fig. 154, Beam IX; Fig. 157, Beam X; and compare the cracks in the U-bar end with those of the other end of the same beam. The draw- ings of all these beams show them just before final collapse, while our calcu- lations have reference to a much earlier stage, viz., under the working load, or at the most a load not more than twice the working load. 106 REINFORCED CONCRETE BUILDINGS sion and a tension, separate from one another, but with hori- zontal projections of equal intensity or magnitude, provided the loads are vertical. Whatever the arrangement, the com- pression and tension must ultimately meet one another and annihilate one another, whether this takes place gradually by increments, as in the plate girder of constant depth; or in one operation, as in the King truss, where the tension chord meets the compression chord at the ends of the beam; or in a number of places, all well defined, as in the Howe truss. We have seen that the rectangular beam is somewhat similar to the King truss, and that the T-beam is very similar to the Howe truss; we have also pointed out that the theory of stress-transmission by gradual increments is not tenable under high loads owing to the slight tensile resistance of the concrete. We must assume that the sooner the compression and the tension are brought to annihilate one another the better will our beams withstand the loads, hence the necessity of bending the rods up as soon as possible, and the desirability of closely spaced U-bars. A simple and effective way of bending the bars is shown in Figure 113. The point of bending should be deter- FIGURE 113. mined by the bending moment, so that there is steel enough to meet the requirements at all points. In this beam and the fol- lowing we must suppose that there are some straight bars, but these are not shown in the figures. Hence the principal stresses in Figure 113, disregarding the straight bars, are: a constant compression along the slab, a constant tension in the rod, and certain vertical resultants. The rod has a curve under the load A, against which the concrete is pressing. The resultant of all these pressures should go through the point of application of A, hence the rod should be bent to a circle with center in the point of application. The same applies to the reaction, B, and in addition the rod should be extended beyond the support to develop the full adhesive resistance. A somewhat more complicated method is shown in Figure 114 TRANSVERSE STRESSES. U-BARS 107 where there are two systems of bent rods (aside from some straight ones). The " first " rod, AC, is curved under the load P for the reasons explained above; in addition, the resultants C and D must be made to meet one another in the same point and with the same direction and same force. Hence the num- ber of rods in each chord should be the same. The length of rod in compression flange ab should be sufficient to develop the full strength of the bond, in the same way as for the " second " \ \ FIGURE 114. chord over the point of support. The slope or angle of the bent bars would seem to be of no importance; but many authorities are of another opinion and recommend an angle of about 45 degrees. (In practice the bars are seldom bent to such large radii as shown in Figure 114, this diagram being purposely exaggerated.) 75. The shape of the U-bars should be as shown in Figure 102a, 6, with curved top and bottom, and hooked over. The downward projection of the end makes it easy to support the U-bar on the form work, and the entire U-bar is firmly anchored against sliding, both top and bottom : the top on account of the curves, the bottom because it passes around the reinforcement. The direction of the U-bar should be vertical. The sloping or slanting U-bar is said to strip the concrete away from the ten- sion rod, as we might expect if our theory is correct, and it does not give as efficient reinforcement in the small cantilevers as the vertical U-bar. Round U-bars appear to be better than flat bars; but there is a great amount of information along this and similar lines which will have to be furnished before rein- forced concrete design can be perfected. But our lack of in- formation in this and similar cases is not different from that existing in other lines of engineering. 76. When we now finally combine all these elements to one beam, Figure 115, we have a structure of a very complicated nature, and we must ask ourselves if all these stresses can travel through and between one another as here assumed without 108 REINFORCED CONCRETE BUILDINGS upsetting our calculations and assumptions entirely. To this we must answer that we do not know, but if we compare our problem with those met in other lines of engineering we must admit that there is no fundamental difference between the diffi- culties. Thus a combination of two simple Pratt trusses is treated as if the two trusses were really present individually instead of combined into one structure, and many other instances FIGURE 115. could be cited to show that we often have to dissolve a struc- ture into its apparent elements in order to solve its problems. Assuming the reinforced concrete beam to be similar to a Howe truss, as here proposed, seems to be no more of a mistake than to assume the connections in a riveted truss to be frictionless, movable joints. But the approximations made in steel con- struction are so old that they seem almost part and parcel of the art, while the comparatively new assumptions made for reinforced concrete have hardly had time to solidify, and they are therefore supposed to be of a more questionable nature than the older ones, which have indeed had the profit of the test of time. Yet there is a number of reinforced concrete buildings about thirty years old which stand up as well as any- body could wish, and the modern steel sky-scraper is of no older date. CHAPTER VIII APPLICATIONS OF THE BENDING THEORY 77. Continuity of Reinforced Concrete Beams. The dif- ference between the beam with simple supports and the continuous beam is that the continuous beam is subject to a " reverse" bending moment over the support, while in the simple beam there is no such reverse moment. The cantilever beam is an example of the beam in which only reverse moments exist, and as we have found it feasible to construct reinforced con- crete cantilevers we cannot deny that continuity may exist in reinforced concrete beams. In fact, unless special precautions are taken to eliminate reverse moments over the supports, we know that continuity must exist and should be taken into ac- count. The question is then: to what extent are the ends of a reinforced concrete beam restrained? When this question is answered we must make the beam strong enough to resist the bending moment at the column, and then it is a matter for further investigation to decide in how far the beam is actually benefited by the restraint to such extent, that the moment at the middle of the beam may be reduced. In Figure 116 a beam is shown in which the ends are per- p Ibs. lin. foot FIGURE 116. fectly restrained, and where the uniform load covers the entire span. The bending moments are over the supports: MA = MB = & pi 2 at the center: Me = aV pi 2 - Hence M A + M c = M B + M c = ( T V + A) P? = i P?; 109 110 REINFORCED CONCRETE BUILDINGS or: the total amount of bending moment to be taken care of in the beam with " built in " ends is the same as in a simply sup- ported beam. The bending moment carried by a reinforced concrete beam is m a constant X bd 2 (Formula 8); hence for constant depth the allowable bending moment is directly proportional to the width 6. At C, Figure 116, the width is = 6, but at the support where the reverse moment must be taken care of, the width of beam is only that of the stem = n. Hence if we assign a moment M c to the middle of the beam, the end will only carry a moment M A = so that M A + M c = \M C + M c = \ pi 2 ', u o b 4 we have MC = % i pi 2 = iV pi 2 and M A = MB = I- T V pi 2 = T, n 1 For 6 = 6 we have M c = f 1 pi 2 = - pi 2 and M A = MB = J - -& pi 2 = A pi 2 - The moment at the center of the span, in the case of a T-beam, will therefore be about and at the end ^ pi 2 . i 4U If greater depth is provided near the support the reverse moment may be increased and the moment at the center of the span may be decreased a corresponding amount. 1 In Europe it is quite Attention is called to the obvious fact that no degree of "restraint" can be allowed at wall ends; this is especially true for beams resting in brick work. APPLICATIONS OF THE BENDING THEORY 111 common to make the beams and girders deeper at the columns; in America the beams and girders are usually of the same depth throughout. The American practice is to be preferred, because the continuous effect depends entirely upon the stiffness of the supports: the slightest yielding of the footings, or even the com- pressibility of the columns may destroy the continuity entirely, and too much dependence upon the continuous effect may lead to serious trouble. In a slab, the depth and the " width of beam "is the same at the middle of the span and at the supports. If the supports are unyielding there may be some excuse for allowing a higher degree of continuity for slabs than for beams; the more so because tests on reinforced concrete buildings point distinctly to such effects. Let us assume a degree of continuity leading to the following bending moment: M e = - p I 2 . In Figure 117 the equivalent system of construction is shown FlGURE 117. in which the center portion is considered as a simple beam resting upon cantilevers of span R. We have then (L - 2 R)* = - L\ hence R = % L 1 - (25) It is hardly necessary to say that we have no absolute certainty that the slab will adjust itself to conform to this arbitrary divi- sion of the bending moment. Yet if the cantilever is made strong enough to carry its load, and the central portion strong enough to carry its share, it is difficult to see why such a system should not be perfectly safe. Other assumptions may be made and carried through in the same manner; this analysis will be used later for the calculation of the " mushroom " system as invented by Mr. Turner. 112 REINFORCED CONCRETE BUILDINGS The formulas usually given for continuous beams depend upon the factor EL The value of / for a reinforced concrete beam is not a constant; in Article 79 we shall consider this in detail. We will find that the moment of inertia depends upon the maximum unit stresses in the point considered, and we cannot expect these stresses to be uniform throughout the length of the beam. The usual application of the formulas for continuous beams presupposes that the moment of inertia is constant throughout the length of beam, and we cannot there- fore apply the formulas used for homogeneous beams to the reinforced concrete beams with any degree of certainty. 78. While, then, the exact degree of continuity cannot be determined, continuity does nevertheless exist in many cases if not in all, and the stresses thus created must be taken care of. These are, primarily, tensile stresses over the supports, requiring reinforcement in the top of the girders over the col- umns, in the beams over the girders, in the slabs over the beams. The top bars may be loose bars, but it is rather difficult to main- tain such bars in their proper position; the bent-up bars may be utilized as top reinforcement with good results, especially as they extend a distance into the next bay in any case. It is evident from the remarks made above that the top reinforce- ment over the support should not be less than 25 per cent, of the bottom reinforcement; usually more bars are bent up, but they need not all extend as far beyond the support as the bars designed to resist the reverse moment. For a uniform load covering the entire span, the point of inflexion is evidently deter- mined by the Formula 25: R = JLfl v/S so that 25 per cent, of steel mentioned above should be carried at least that distance out from the center of the support. The bars must be embedded in a sufficient amount of concrete to develop the bond, not less than four diameters from the face of the concrete, or, if closer to the face of the concrete, they should be provided with inverted U-bars. The stress on these U-bars cannot be calculated, it is their presence rather than their strength which benefits the beam. 79. Moment of Inertia. The moment of inertia in a rein- APPLICATIONS OF THE BENDING THEORY 113 forced concrete beam is of interest only because certain prob- lems connected with continuity of the beam, deflection, etc., cannot be solved except through a knowledge of its value. The expression given below is of indirect value only, showing that the ordinary formulas for continuity do not apply to reinforced concrete beams, because the moment of inertia is not a con- stant for the length of the beam, as is usually assumed in the solution of such problems. The moment of inertia with reference to the neutral axis may be found as the sum of two moments: /i, referring to the concrete above the neutral axis, and J 2 , referring to the steel below the neutral axis, the concrete below this line being dis- regarded as usual. We have then (Figure 118) "a sq. inches FIGURE 118. and Jl=l 7 2 = rad 2 (1 - x) 2 the steel being considered as equal to ra square inches of con- crete. But according to the formulas given in Articles 25 ff. we have i Cxdb = aS a _xdbC_ Mb . 2 S 2(l-x) r' hence (after some reduction) : 7 = /i + 7 2 = J fed 8 (1 - J a) x 2 (26) By means of Formula 5 in Article 27 the expressions derived above for d and s t , etc., may now be verified. The real import- ance of Expression 26 is, however, that it shows that the moment 114 REINFORCED CONCRETE BUILDINGS of inertia depends upon the location of the neutral axis which again changes with the stresses in the various points of the beam. 80. Beams with Reinforcement in the Compression Side. Sometimes it is found impossible to make the compression flange of the beam wide enough to bring the concrete stress down to the allowable maximum. In that case some engineers use compression reinforcement, but as a matter of fact, our knowl- edge of the properties of such beams is very slight, and there is grave doubt as to the advisability of using this method of construction in important cases. The calculations are simple: to the bending moment sustained by the beam with its ordi- nary amount of reinforcement is added another bending moment due to extra reinforcement in top and bottom, this latter cal- culated as for an ordinary steel beam, but with quite low stresses (not to exceed 10,000 Ibs./square inch). The compression bars must be laced carefully to r the tension bars, but under any cir- cumstances it seems hardly possible to provide properly for the excessive shear stresses set up in this kind of beams. A steel I-beam is cheaper and better in places where this kind of con- struction is actually necessary. 81. Combined Bending and Compression. The section is best designed by trial. In the case of an arch ring, the sec- tion is rectangular, and the symmetrical reinforcement is of small area compared with the concrete sec- tion. The bars on the compression side must therefore be disregarded, as it would require too many hoops to make this rein- forcement effective in compression. We must select the depth and the reinforcement by judgment; the stresses due to the bending moment alone are then easily found by For- mula 19 in Article 49. Let Figure 119 rep- resent the section; let C m and S m denote the stresses just found due to the moment alone. If now in Figure 119 gh is made equal to S m /r and ef is made equal to C m , then the line gf will represent the distribution of stresses on the section due to the bending moment alone. The stress due to the pressure P is now P/bd Ibs./square inch; this is represented by the line ik parallel with FIGURE 119. APPLICATIONS OF THE BENDING THEORY 115 fg. The total pull in the steel is then equal to the area of the triangle khl times the width 6 of the section. For slabs or arches the width is usually taken as 12 ". The final concrete stress ei must not exceed the allowable stress; we can therefore arrive at a preliminary estimate of the dimensions required by Formula 16, assuming a materially lower " allowable stress " for the concrete, and a higher stress for the steel, when making the first trial. If the section is one in a column the calculations are essen- tially different. The eccentrically loaded column is of fre- quent occurrence; in fact, few columns are always loaded centrally. In practical cases it is almost always impossible to calculate the eccentricity of the load, and elaborate formulas are therefore of little or no use. Tension should never occur in the column; if there is tension with the selected arrangement it is better to change the lay-out. The percentage of steel will always be much greater than in the case considered above, and, as there is no tension, we may perhaps calculate our col- umn as a homogeneous section, using, however, for the moment of inertia the expression / = I e + r I s (27) where ] T c ~ i. l s = I c = mom. of inertia of concrete alone, mom. of inertia of steel alone. The cases where the condition of loading can be ascertained with any degree of certainty are very few indeed, and when they do occur the bending moment is likely to be very small. If such is the case it is simpler and probably as correct to cal- culate the column as a pure column, using a correspondingly higher factor of safety, and then, if necessary, finally investi- gate the problem assuming the neutral axis to be disposed at the center of the section, and take the moment of inertia with reference to the center line. 82. Chimneys. As an example of approximate methods of calculating a piece subject to bending and compression, let us consider a single shell chimney of uniform thickness. The diameter d (in feet) of the flue is given, and so also the height H (in feet). Let the outside diameter be D (in feet); the area presented to the wind pressure (w Ibs. per sq. ft.) is then DH 116 REINFORCED CONCRETE BUILDINGS square feet, and the total pressure DHw Ibs. Hence the bend- ing moment at the base (the overturning moment) becomes DHw times J H = \ wDH 2 Ibs. X feet. If now the total allowable compressive stress on the concrete is C Ibs./sq. in. and the compressive stress due to the weight of a column of concrete 1" square and H feet high is (approxi- mately) H Ibs./sq. in., then the compressive stress due to the overturning moment must not exceed C -- H Ibs./sq. in. Assuming the neutral axis to go through the center of the sec- tion, which indeed is not true, and disregarding further the bene- fit derived from the steel in the compressive side (which is on the safe side), the moment of inertia of the ring is hence (C - H) 144 = 64 (D 4 - d 4 ) D 2 which, when solved, gives the outside diameter D J V - | 187T (C - H) ' V V187T C - H, and the tension per inch of circumference becomes \ (C - 2 H) (D - d) Ibs. 83. Footings. In Figure 120, 2R (inches) denotes the side of the footing, 2r (inches) the side of the column. The bending FIGURE 120. moment on side ab (considering the footing as a cantilever- slab) corresponds to the loaded area dabc. We have, for a load p Ibs./sq. in. : Load dcef = (R r)2Rp; arm of bending moment around ef - i (R - r). APPLICATIONS OF THE BENDING THEORY 117 Hence bending moment A = pR (R - r) 2 ; load aed plus bcf = (R r) 2 p ; arm of bending moment around ef = 402-0. Hence bending moment B = I p (R - r) 3 The total bending moment due to the area abed is then the difference between A and B; and the depth of footing becomes, according to Formula 8, for a width of beam 2r = b ' *--'-v*/ +1I -R-rJp ci VG to which corresponds a pull s, in the steel, for the distance ab s t = 77: 2 r 1- It is, however, quite necessary to provide reinforcement for the portions ae and bf ; for this reason the amount found above may be multiplied by a factor estimated at about 2, which gives : s = C ^rd (29) for each layer of steel (Figure 121). The radius of the column i 314 FIGURE 121. should be made as large as possible, because a material saving in depth of footing is obtained thereby; usually the column must have an enlarged base for other reasons as well. In Ar- ticles 14 and 15 we found the cross-sectional area of column: X = 1400 F for a hooped column X = 1060 F for a plain reinforced column, 118 REINFORCED CONCRETE BUILDINGS so that the average pressure, under the conditions assumed, is 1400 and 1060 Ibs./sq. in., respectively. With higher per- centages of reinforcement these pressures may become mate- rially higher; the column base is. therefore enlarged so that the pressure on top of the footing does not exceed the allowable unit pressure, and a steel plate is put under the bars in order to distribute the pressure over the requisite area. According to tests by Bach this allowable pressure may be somewhat increased, owing to the reinforcing effect of the surrounding concrete of the footing, but it does not seem wise to exceed say 1000 Ibs. per square inch. The thickness of the plate may be approximately determined by means of a formula by Grashof: t = ^Vp (30) where t = thickness of plate, in inches, r = radius of reinforcement (= | of diameter of column, less 2"), in inches, p = pressure on plate, in Ibs. per square inch. The dimension Q in Figure 129 may be found by Formula FIGURE 122. 28 above, using for p the allowable pressure on the concrete. 84. Circular Reinforcement in Plates. The circular plate in Figure 122 is supported on a central column. The load is APPLICATIONS OF THE BENDING THEORY 119 uniformly distributed over its surface, or symmetrically and continuously disposed along the circular circumference. A segment, Oab, will then be subject to a certain bending moment, which moment determines the depth D at the circumference of the column. It is now easy to show that when the load is uni- formly distributed over the entire surface, the same formula applies in regard to depth as was derived above for a square footing; the calculations are practically the same and need not be repeated here. When the load is distributed along the edge, the Expression 32 in the following article may be used. In any case, we will assume that the depth is known in the thickest part of the plate (at the edge of the column). If now the distance dc is one inch long, we have by Formula 9 s t = T V c 2 D, which expression leads to the amount of steel required along the radii, the bars being I" apart on the circumference of the column. Imagine now that all these radial bars be cut asunder over the top of the column disregarding the tensile strength of the concrete, each bar will then have a tendency to move outward, so that if a steel ring surrounded the entire plate, each bar would exert a pressure s t against the inner face of the ring. If now dc equals one inch, then db equals one inch times R/r , hence the pressure on the ring, measured in pounds per lineal inch of circumference, equals s t X r /R. The tension on the ring is then SR = s t ' r ^.R = ^czr D. (31) It is now evident that the ring with radius R and designed to resist the tension S R is, mathematically, sufficient reinforce- ment, so that the radial bars may be dispensed with. In actual practice this is somewhat modified owing to the fact that con- crete shrinks when setting, so that it would pull away from the ring; the ring would therefore exert no pressure against the concrete until a substantial, and perhaps dangerous, deforma- tion had taken place. But when the ring is used in combina- tion with a radial reinforcement and when at the same time the depth D is not too small compared with the radius, say D larger than ^ R, then the ring would seem to be a very efficient rein- forcement. Direct proof of this statement is indeed missing, 120 REINFORCED CONCRETE BUILDINGS but the " Mushroom " floors furnish at least some indirect information in this respect, as they probably owe their strength in a great measure to the intelligent use of circular reinforce- ment. That this type of reinforcement is successful in other types of structures may be seen from the remarks made under " columns," where hoops are extensively used to take care of stresses somewhat similar to those existing in a plate, although the plate at the same time acts as a beam. Exact analysis is of course difficult in these structures which border upon the class where reinforcement may sometimes be omitted entirely : it is well known that tapering footings are often constructed with- out steel, and the same may be true of columns in special cases. 85. Theory of plates. The " Mushroom " System. A reinforced concrete floor without beams or girders is first indicated and patented by Mr. C. A. P. Turner of Minneapolis. As far as known there is no perfectly satisfactory way of finding the stresses in constructions of this kind, although buildings actually constructed on this principle have given good satis- faction, according to the published records. The stresses must necessarily be of a very complicated nature, especially under concentrated or unsymmetrical loads; the following analysis does not pretend to solve the problem in anything approaching a general way, and the formulas apply only in case the entire building is loaded with a uniformly distributed load. The formulas are not inconsistent with the assumptions made for reinforced concrete construction, and they are therefore pre- sumably a step in the right direction. It is well known that most of the proposed formulas are based upon the theoretical strength of the plates with equal tensile and compressive resist- ance, and reinforced concrete does not possess any such qualities. Figure 123 shows the general scheme for a floor of this kind: the floor slab is simply a flat plate resting upon columns, the tops of which are enlarged. Let the uniformly distributed load be w Ibs./sq. foot and the span I feet. The slab is divided into six strips: two diagonal strips AD and BC, and four strips along the sides AB, BD, DC, and AC. If we now suppose the panel to be square, the load on each of the crossing diagonals may be taken as \w, while the span AD = BC = I V2. Then, by Formula 30 for AB: d = lJ and for BC: d= ^ . t/SL 1 - 1/? Ci V q a V q a V q APPLICATIONS OF THE BENDING THEORY 121 so that the depth is uniform, and our problem centers around the design of a side strip like AB. The notations are shown in Figure 124, where If FIGURE 123. FIGURE 124. L inches is the span between column centers. p the load in Ibs./sq. inch. R the radius in inches of a certain circular plate. r the radius in inches of the support under the plate, here referred to as the " cap." p the radius of the column in inches. d the depth of the slab in inches. D the depth of the cap in inches. We will now proceed as follows : We consider the floor slab as supported on the edge of the circular plate with radius R; this plate will then have a uniform load on its surface and a concen- trated load along its circumference. Finally the " cap " with radius r will be designed for a load concentrated on its circum- ference, disregarding the uniform load on its surface. The total area of the floor panel between the four column centers is L 2 square inches; the total weight corresponding to this area is pL 2 Ibs. The area of the circular plate with radius R is irR 2 , the total weight on same ptrR 2 . Hence the weight of the portion outside the circular plate becomes p (L 2 irR 2 ) Ibs. 122 REINFORCED CONCRETE BUILDINGS In the following computations, L 2 is always large compared with TrR 2 so that this quantity may be neglected, which is also on the safe side. The load is therefore pL 2 , and as the circumference of the circular plate is 2 TrR, the load per lineal inch of circum- ference becomes producing a bending moment equal to _ 27TR (R r ). If measured per lineal inch of the circumference of the cap with radius r it becomes, by multiplication with R/r c (Figure 125), p L 2 -7T R 2 27TK FIGURE 125. FIGURE 126. and the corresponding depth, for 6=1" L R - r ) (32) for the circular plate. For the slab portion we have, by Formula 16: 7 LT, T. ' l~ V < 33 ) q c,\ q 'The value of r must now be such that the two depths become alike, which gives q r n (34) at the same time, the value of R is determined by the selected value of q by formula APPLICATIONS OF THE BENDING THEORY 123 see Article 67, Formula 25. The depth of the cap. is found by the formula above, as the load again is pL 2 , substituting only r for R and p for r ot hence D--J^-(r.-P). (35) Ci V 2-n-p According to Article 84 the reinforcement may be disposed in a ring with radius r ; the tension in this ring becomes: S T = ^ c,r D (36) LA The arrangement is shown in Figure 126, where the thickness t should be about 4" so as to cover the ring thoroughly. The cap should be cast in one piece with the column, but there is no reason why a joint may not be made between the top of the cap and the bottom of the flat portion along line a a in Figure 126. The reinforcement for the flat portion is designed as for any other slab. We have the depth - (33) q and the corresponding pull in the steel $/ tons, for a band one foot wide, is therefore, according to (17) S f = c z d (37) This reinforcement should be disposed near the bottom of the slab at the center of the span, and near the top over the columns. It will be seen that this leaves a considerable space around the column without reinforcement near the bottom, which should be avoided. We may therefore follow the prevailing practice and bend every alternate bar up, leaving the balance of the steel straight near the bottom. The reinforcement over the column is then inadequate, and we will have to introduce addi- tional steel at that point; if we decide to use rings we may use one ring with radius R, the strength of which is determined according to Article 84 by the formula (38) We have now (Figure 127) Thickness of slab, in inches d = y - (33) Pull in slab steel per foot wic Radius of upper ring, inches Tension in upper ring, tons Radius of cap-ring, inches Tension in cap-ring Depth of cap, inches q is the factor of continuity, L- r - - ith, tons Sf = c . , ft ( Beam Type B No. ent Straight 1\ fl o gs U-bars Type of Loading "S-S u 03 111 I P>3 3 Diam. in No. Diam. in m/m m/m IS -2 H2 IV T 3 15 and 1 18 14 none Uniform 42.0 VI T 3 15 and 1 18 14 full supply load 37.8 V S -2 15 and 2 16 14 one end only covering 31.0 entire span Two VII T 3 16 and 1 16 14 full supply concentrated 34.0 VIII S 2 16 and 2 16 10 one end only loads 23.4 IX S 2 16 and 2 16 14 one end only at third 25.6 points X T 3 16 and 1 16 14 one end only One 27.0 XI S 2 16 and 2 16 14 none concentrated 26.0 XII T 3 16 and 1 16 14 none load 26.0 at center The tests naturally divide themselves into three groups, according to the manner of loading: (1) Uniform load, beams IV, VI, and V. It is at once apparent, by comparing beam IV (without U-bars) with beam VI (with U-bars), that the influence of the U-bars is very slight, if any, the difference in ultimate load being accounted for by the fact that the ends of the straight bar in beam IV were hooked, while those in beam VI had no hooks. Both of these beams were of the trajectory type, and if we compare them with beam V of the suspension type, the superiority of the trajectory type seems clearly established. But we must not lose sight of the fact that in the two first beams three of the four rods were bent up, while in the latter, only two of the four rods were bent up. This beam failed in the end without U-bars, and while therefore this group does not prove the author's theories, as outlined in a preceding chapter, it does not disprove them, and still leaves the question open whether or not the bending of one additional rod, or the proper use of U-bars, would not have changed the results materially. It will be remembered that in a T-beam, the straight reinforcement is effective only as reinforcement of the stem, while the bent bars correspond to the flange; if the rods are not so arranged, stirrups must be introduced to again bal- 212 REINFORCED CONCRETE BUILDINGS ance the design, the size of the U-bars being in direct ratio to the violation of the principle outlined. (2) Two concentrated loads, beams VII, VIII, and IX. Here again, beam VII of the trajectory type, with a full supply of U-bars, is compared with two beams of the suspension type, the two latter being without U-bars. Again the traject- ory type seems superior to the suspension type, and again we find the reason to be that in the trajectory beam, the proper amount of rods have been bent up, while in the suspension type, only two of the four rods have been bent, and no U-bars have been introduced to overcome the deficiency. It is rather interesting to note that the difference in width of stem between beam VIII (10 cm.) and beam IX (14 cm.) affects the ultimate strength but slightly, both beams being of the suspension type. In his discussion of these tests, Prof. Morsch has taken occasion to criticize the suspension, or Hennebique, type. It is to be regretted that the tests were not carried out so as to have the same number of bent-up bars in both of the types considered, in which case the suspension type would probably have stood up as well as the trajectory beams. It is only fair to note that the U-bars or stirrups have always been considered as an essential part of the Hennebique system, and that such tests as these, however valuable otherwise, give no indication whatever as to the merits of this system. (3) One concentrated load at center, beams X, XI, XII. In this group, the two systems give the same carrying capacity, owing undoubtedly to the fact that in no one of these beams the reinforcement is arranged according to the equilib- rium curve, while in no case U-bars have been introduced to compensate for the deviation. Bent Bars in T-Sections Author's Tests. The beam tests just referred to were published by Prof. Morsch in "Deutsche Bauzeitung," April 13, 1907. It occurred to the author of the theory of this present volume that the description of the action of the suspension rods was subject to doubt, for the reasons outlined above, and that additional information might possibly be gained by tests on beams with trussed rods only. The author designed a series of nine test beams which were tested in the winter 1907-1908 at Case School in Cleveland, in co-operation THEORY OF BEAMS AS ILLUSTRATED BY TESTS 213 Type A ams 1,4,7 n TypeB- ams 2,8. \ >> c "I r U omag jo JlH 9U I 5" ^ ^H^ |< 214 REINFORCED CONCRETE BUILDINGS with Prof. F. H. Neff. It will be seen from Figure 151 that these beams had no straight reinforcement, and that the sloping stem terminated at the supports, so as to make the system one of equilibrium under two concentrated loads. The results were first published in the Engineering Record, August 22, 1908, from which the following is an extract: " Three different molds were made, types A, B, and C, respectively, each one of which was used three times with a different percentage of steel for reinforcement of the beam. In this way, three beams of type A were made, one of which was reinforced with 0.5 per cent., one with 0.75 per cent., and one with 1.0 per cent. In the same way three beams B and three beams C were made, reinforced as described, so that of the total number of nine beams no two were alike in all respects, but any one beam would have a corresponding one which was different in one detail only. In this way, it would be possible to compare the beams and find the exact effect of a certain change, which is a safer way than to try to obtain absolute results from so few tests. "All the beam swere provided with U-bars in one end only, the object being to show that the stirrups were of no conse- quence at all. The stirrups made no difference in the results obtained, four of the nine beams failing in the end equipped with U-bars. "Two short cross bars were placed in the slab at the points where the loads were applied, and three similar bars were placed in the slab near the support. These bars were i-inch square twisted bars. The main tension bars were 1-inch square twisted Ransome bars. It was found that the elastic limit of these bars averaged about 56,000 Ibs. per square inch, and their ulti- mate breaking strength was 73,600 Ibs. per square inch.- " The concrete was made quite wet and very carefully placed. The mixture used was 1:2:3^, Lake Erie sand and Euclid bluestone being used for the aggregates. The strength of the cubes was low, as might be .expected with the aggregates used, and the average of the 6-inch cubes in pounds per square inch was as follows: Age, days 7 14 28 60 Strength, pounds 660 1,065 1,440 1,787 "The beams were all tested when sixty days old. In the table here below the results are given, and this table, together THE THEORY OF BEAMS AS ILLUSTRATED BY TESTS 215 with the diagrams of the beams, should give all the information needed. Attention is called to the ways of supporting beams 5 and 6. While all the other beams are supported at the point where the sloping stem begins, these two beams are supported further out from the stem, making the overhang shorter for them than for the similar beams of same type. "As to the column headings used in the table, the percen- tage of reinforcement is calculated with reference to the ' enclos- ing rectangle' proposed by Professor Talbot. Under 'lever' the distance from point of support to point of application of the load is given, while 'overhang' means the length of the pro- jecting end beyond the support. "The bending moment given in this table is found by mul- tiplying the 'lever' by one-half of the ultimate load, disre- garding entirely the weight of the beam itself. The lever arm of the internal stresses is assumed to be 0.85 times the distance from the top fiber to the center of the steel, which distance is approximately 9 in., giving a lever arm of 7.65 in. This, of course, is not quite correct, as the position of the neutral axis varies with the percentage of steel and the coefficient of elas- ticity, which latter again depends upon the stress on the concrete. It is, however, sufficiently accurate considering the un- avoidable variations in the position of the steel bars and in the elastic properties of the concrete, and the 'total stress in the steel' may therefore be found by dividing the bending moment by 7.65, giving the values shown in the table as well as the stress in the steel per square inch of its cross-section. RESULTS OF TESTS AT CASE SCHOOL. Beam. Type. Per cent. Lever. Overhang. Ultimate load. 0.5 0.5 0.5 0.75 0.75 0.75 1.00 1.00 1.00 30 in 26 in 22 in 30 in 30 in 30 in 30 in 26 in 22 in 10 in. 14 in. 18 in. 10 in. 10 in. 10 in. 10 in. 14 in. 18 in. 12,500 16,000 27,800 11,900 16,200 16,000 13,950 22,000 28,900 Bending - Stress in steel moment. Total. Per sq. in. 187,500 24,500 49,000 208,000 27,200 54,400 305,800 39,900 79,800 178,500 23,300 31,000 243,000 31,800 42,400 240,000 31,400 41,800 209,250 27,300 27,300 286,000 37,400 37,400 317,900 41,600 41,600 "Eef erring now to the several photographs of the beams after failure, it will be noticed that the failures are of uniform nature. Comparing the figures given in the table above, it will be seen that the ultimate load varies greatly as well as the total stress and the stress per square inch. If the failure has 216 REINFORCED CONCRETE BUILDINGS Beam 1 Beam 2 Beam 4 FIGURE 152. FIGURE 153. FIGURE 154. Beam 5 FIGURE 155. THE CASE SCHOOL BEAMS AFTER TESTING. TtiE THEORY OF BEAMS AS ILLUSTRATED BY TESTS 217 a common cause in all these beams it cannot be due to either tension or compression in the usual sense of the word. It may also be assumed that shear had little to do with the failure. Beam 6 FIGURE 156. Beam 8 FIGURE 157. Beam 9 FIGURE 158. THE CASE SCHOOL BEAMS AFTER TESTING. On account of the trussed form of the beams, the steel follows the curve of equilibrium of the external forces acting upon the beam, and the only stresses possible are tension in the steel and compression in the concrete. 218 REINFORCED CONCRETE BUILDINGS "This is also evident from the behavior of the beams under load. The cracks started on the tension side and opened slowly with increasing load, at the same time becoming longer, until finally the compressive area left above the top of the crack became too small to carry the stress on it and crushed. A shear crack cannot grow in this manner. It is well known that the maxi- mum shear stress does not occur at any fiber near the extreme top or bottom of a beam. Therefore, when the crack extends up into the stem and reaches the neutral axis, the shear resist- ance of the beam is practically exhausted. " The beams also made it evident in other ways that no ver- tical shear was active. In some cases the beams had received a vertical crack in handling, the crack being located about 3 in. inside the support, and extending clear through the concrete. At first, it was believed that these beams would not give a fair test, and it was taken under consideration to leave these beams out. It proved, however, that the crack closed up as soon as the load was put on, and after the load was increased to a cer- tain amount, the cracks were hardly visible, while the final failure took place some distance from the injured section. If there had been any vertical shear acting on the beam, the ulti- mate load would have reached a comparatively small value only, and in all probability the injured section would have sheared off at once. "The tension in the steel must be constant from end to end of the beam between the supports. The steel would have a tendency to pull out of the overhanging ends with a force equal to the total pull in the steel, which is the same near the supports as at the center of the beam. The overhanging ends furnish the necessary anchorage for the bars on account of the grip of the concrete around the bars, which increases with the com- pression in the concrete, and, therefore, also with the load, the horizontal cross-bars giving the required horizontal restraint of the concrete to produce the desired effect. The numerical value of the length of the anchorage may therefore be expressed in figures by simply dividing the length of the overhang into the total pull on the steel, the quotient giving the value of the bond in pounds per lineal inch of embedment, regardless of the amount of steel. This figure is given in the accompanying table, the length of the anchorage being the length of the over- THE THEORY OF BEAMS AS ILLUSTRATED BY TESTS 219 hang, and disregarding the extra length of the hook at the ends of the bars. Beams 5 and 6 are not included in the table, as these beams had an overhang of only 10 in., leaving a hori- zontal space inside the support, and this, of course, makes it impossible to compare these two beams directly with the rest. VALUES OF BOND OBTAINED Total pull Bond Beam Type Overhang in steel per lin. in. 1 A 10 in. 24,500 2,450 2 B 14 in. 27,200 1,940 3 C 18 in. 39,900 2,230 4 A 10 in. 23,300 2,330 7 A 10 in. 27,300 2,730 8 B 14 in. 37,400 2,670 9 C 18 in. 41,600 2,310 "This table, it is believed, is remarkable when the uniform- ity of the results is considered. The beams tested here had reinforcement varying from ^ of 1 per cent, to 1 per cent., spans varying from 74 to 80 in., and tension stresses in steel varying from 27,300 to 79,800 Ibs. per square inch. It seems safe to say that these beams all failed by sliding of the steel. " So far, no attention has been paid to beams 5 and 6. The overhang for these beams was 10 in. in each case, the slab con- tinuing for a distance inside the supports. The bond stress developed in the overhang, if figured as for beams above, be- comes 3,180 and 3,140 Ibs. per linear inch, or quite high when compared with the results of the table above. Remembering, however, that the straight portion of the bar is continued inside the supports for a distance of 4 and 8 in., respectively, the bond, if distributed over the total distance of 14 in. for No. 5 and 18 in. for No. 6, becomes: Total stress Bond Beam Type Overhang in steel per lin. in. 5 B 10" + 4" = 14" 31,800 2,270 6 C 10" + 8" = 18" 31,400 1,745 " If any importance can be given these two isolated results, they would show that the bond inside the support is quite as effective as that outside the support, but for a short distance only, and that its value decreases rapidly with the distance inside the support." 220 REINFORCED CONCRETE BUILDINGS The lessons to be drawn from these tests are: (1) That, with the arrangement used, the presence or ab- sence of U-bars does not influence the strength of the beam. (2) That "shear," properly understood, does not exist in beams of this kind. (3) That, with proper arrangement of the end supports and of the anchorage, such beams will not fail until the com- pressive strength of the concrete, or the tensile strength of the steel, is exhausted. (4) That such beams are rational structures capable of prac- tical and economical use. (5) That the sliding resistance of the steel does not depend upon the number or size of the individual rods, but only upon the anchorage of the group of rods, the length of embedment being much more important than the diameter of either each rod or of the group of rods. Effect of Joint between Slab and Stem, Tests by Professor Johnson. In connection with the introduction of the Ransome Unit System (p. 162 ff.) in Boston, a series of very interesting tests were made on T-beams of both the monolithic and unit types, reinforced with straight bars only, and with both straight and bent bars. All the beams had U-bars. In the "Unit" beams, the slab was cast from four to nine days later than the stem. A total of twenty-eight beams were prepared, of which eleven have so far been tested, the balance being held for a longer-time test. 1 The beams were all reinforced with Ran- some steel, that is, cold-twisted squares. The U-bars were round except in Type C, where square twisted U-bars had been used. Type A, Beams 1, 2, 4, 5, 9, 10, and 12. See Figure 159. 1 The authors are indebted to Prof. L. J. Johnson, M. Am. Soc. C. E., for the following data, and for permission to publish the same. The beams were designed by Prof. Johnson, by Mr. J. R. Worcester, M. Am. Soc. C. E., Consult- ing Engineer, by Mr. J. R. Nichols, Jun., Am. Soc. C. E., by the Concrete Engineering Co. of Boston, and the Ransome Engineering Co. of New York, each having designed one series of beams or contributed to the design by suggestions. Professor Johnson, who made the tests on the testing machine in the Harvard University laboratory, expects to publish in due season a complete report of both this series and of the long-time tests. The authors of this present volume, eye-witnesses of these tests, are solely responsible for conclusions reached herein. THE THEORY OF BEAMS AS ILLUSTRATED BY TESTS 221 In the Unit beams, the top of the stem was either left fairly smooth, as it would be in usual every-day practice, or corru- Toggle for lifting beam 8-8 Rods "13-"^ Bars y \ t 1 4^ ^ 6 * Stirrups 5^ Straight' FIGURE 159. gated as shown in Figure 160. The ends of the stem rested in previously prepared seats (Figure 161), and the joints were Y FIGURE 160. sealed with grout to ensure a similar action as obtained in actual construction, where the Unit beam rests in a pocket in the gir- der. Nine days after the casting of the stem, the slab was put on, while in the monolithic beam, the entire amount of concrete was, of course, deposited in the forms in one operation. Beam No. 5 was a Unit beam, with the top of the stem corrugated. The age of the stem was forty- five days, that of the slab thirty-five days. At a total load of 12,000 Ibs. the first tension crack ap- peared near the middle of the span. Inclined cracks became evident near the ends under a load of 23,000 Ibs.; the ultimate load was 41,000 Ibs., when failure occurred, through compression of the slab between the loads, and slipping of the straight tension bars (see beam No. 12 below). Beam No. 4 was a Unit beam, the top of the stem being fairly smooth; that is, no attempt had been made toward getting a particularly .rough sur- face. The age of the stem was forty-five days, of the slab t t i 1 I i 1 FIGURE 161. 222 REINFORCED CONCRETE BUILDINGS thirty-six days; the first crack was observed under a load of 20,000 Ibs.; ultimate failure took place under 48,000 Ibs. in precisely the same manner as in No. 5. Beam No. 1 was exactly similar, except that slab and stem were both one day older than in No. 4; the ultimate load was 49,400 Ibs., and the beam failed in the same manner as the fore- going. Beam No. 2 was of the same age and detail as No. 1 ; the first crack was observed at 10,000 Ibs. loading; the ultimate failure occurred in the same manner as above under 54,300 Ibs. total. The higher load on this beam is perhaps due in some measure to the fact that the rocker-supports for the beam came to a bearing, making possible some horizontal thrust on the beam. Beam No. 12 was monolithic, forty-one days old; of same design as the foregoing Unit beams, except that the fillet between stem and slab was slightly reduced (see Figure 162). The first FIGURE 162. crack occurred at 4,000 Ibs., the ultimate load was 45,800 Ibs., and failure occurred through a slip of the straight reinforce- ment, causing the sudden collapse of the left end. Beam No. 9 was of the same general design, cast in one piece, and forty-one days old. The first crack occurred at 3,000 Ibs.; the beam failed suddenly at 50,000 Ibs. by slipping of the rods at the right end. Beam No. 10 was also a monolith forty-one days old, show- ing a tension crack at 6,000 Ibs., with ultimate failure at 52,400 Ibs. from a combination of initial sliding of the tension rods with compression at the center. It will be seen from these data that the Unit beams stood up as well under the load as the monolithic beam, so that the joint between slab and stem was perfectly adequate, whether cor- rugated or plain. The general behavior of all these beams up to the point of failure was so much the same that no one, from THE THEORY OF BEAMS AS ILLUSTRATED BY TESTS 223 observation of the beams in the machine, could have pointed out which beams were unit and which monolithic. In fact, they all failed in the customary manner, exhibiting the usual inclined and vertical cracks, and no sliding was noticeable between slab and stem, although carefully looked for. Type B, Beams 25 and 27. See Figure 163. 2- 3 /- Bars'T t I Ho Stirrups 2-K" Straight Ho Stirrups 5K FIGURE 163. The beams of Type B were built exactly as the beams of Type A, except that the tension rods had been reversed, being 2|" bars bent and 2^" bars straight. This reinforcement would, under the theories- advanced in this book, be more effi- cient, and the U-bars were therefore reduced from T V' round stock in Type A to i 3 s " round stock in Type B, thus having about one-third of the area of the former. Beam No. 25 of Unit construction had a stem twenty-nine days old and a slab twenty-five days old; the first crack was observed under a load of 9,000 Ibs., and ultimate failure took place under simultaneous compression of the slab and of the side of the stem, at the point where the tension rod was bent, under a load of 42,500 Ibs. (See Figure 164.) 2 - Straight FIGURE 164. Beam 27 Beam 25 Beam No. 27 was monolithic, of same design, and twenty- seven days old. The first crack was seen at 10,000 Ibs. ; while the ultimate load was 47,500 Ibs. Also in this case was com- pression in both slab and stem evident as shown in Figure 164. 224 REINFORCED CONCRETE BUILDINGS It is interesting that these two beams carried practically as much load as the older beams of Type A, in spite of the great reduction in the weight of the U-bars. The explanation is to be found in the theory set forth in Chapter VII of this book, where the relation between the bent bars and the U-bars has been considered at length; in fact, the design of beams 25 and 27 was made to prove, or disprove, these theories as far as possible. Type C, Beams 13 and 20. See Figure 165. i /i 1 ,,0-10 ,,-J 'oggle f orjif ting beam k-8 >r< 8 >i 9_ S/' 1 -f *^ /8-X Rods _^. r T"~T~7~r~7~ 1 i ! i ! , ! ! ! i V f7 J.-L_L_L_L _o_L 1 X i J J J J ._ 1 -L/ 3 CO i V/ ///\ V 2-1" w/ y///k H " Stirrups *'Y *Y///A FIGURE 165. In designing these beams, Professor Johnson had endeavored to secure a high strength in compression and tension. The special feature was the absence of bent bars so that the stresses in the stem and in the joint between slab and stem were especially severe under the common theory of shear. The beams rested in concrete supports shown in Figure 166. Beam No. 20 was of the Unit type with corrugated top (Figure 167). The stem was forty-two days old, the slab thirty- 4 1 i "^ i 1 1 %" _/ \ / \ / ' \ / j t i Wj [ ! ! FIGURE 166. FIGURE 167. two days old. Tension cracks developed in the usual manner, beginning under a load of 10,000 Ibs. ; at 22,000 Ibs. small diag- onal cracks appeared. At 50,000 Ibs. it was noticed that the visible end of the curved tension rods began to slide, and the THE THEORY OF BEAMS AS ILLUSTRATED BY TESTS 225 ultimate failure occurred under a load of 55,600 Ibs., when the compression area was crushed. Beam No. 13 was again of the Unit type, with smooth top of stem, which was forty-three days old; the slab was thirty-four days old. The first tension craok occurred at 9,000 Ibs., and the cracks then developed in the usual manner. Failure took place at 50,000 Ibs., when the adhesion between concrete and steel was broken; the rods began to pull through, and the slab was crushed at the center. The analysis of the stresses follows: Types A and B By reference to formula (18), page 38, we have 2 = 1; T = 4"; H = 9"; D = 13"; V = ~ X 1.62 = 2 hence S 82.0 and ^ = 7^M = ^ = - 436;1 -? = - 8 ^ Now, the bending moment is f L.42 = 21 L, and the arm of internal stresses approximately .855 X 13 = 11.1 inches, hence the pull in the steel s = jj-L = 1.89Llbs. The beams had 1.62 square inches of tension steel, hence the unit tension on steel: S L89 T 1 17 T = ' L = 1 ' 17 ' L and the unit compression on the concrete Type C | = i ; T = 5"; H = 6"; D = 11"; V = 15 - ^ = 2.5; 226 REINFORCED CONCRETE BUILDINGS hence ~ = 15.9 and x = 1 + 15.9 15 - .485; 1 - l -x = .838 Again, the bending moment is \ -L-30 = 15 L, and the arm of internal stresses approximately 11 X .838 = 9.2", hence the pull in the steel is The beams had 2.0 square inches of steel, hence the unit tension on steel and the unit compression on the concrete It is evident that these calculations do not give the true stresses existing at rupture, because r is not equal to 15 at that time, and the assumption of plane sections probably does not hold good. For the sake of comparison, however, they may be useful. The results are indicated in the table. The testing ^ Corresponding 6 Age (days) calculated stresses "SI fc Ultimate 165 sq. in. a" S How made Load > 2 Ibs. 1 pq Stem Slab C s A 1 Unit, Smooth Top 46 37 49,400 3060 57,800 A 2 Unit, Smooth Top 46 37 54,300 3370 63,500 A 4 Unit, Smooth Top 45 36 48,000 2980 56,200 A 5 Unit, Corrugated Top 44 35 41,000 2540 48,000 A 9 Monolith 41 41 50,000 3100 58,500 A 10 Monolith 41 41 52,400 3240 61,200 A 12 Monolith 41 41 45,800 2840 53,600 B B 25 27 Unit, Smooth Top Monolith 29 27 25 27 42,500 47,500 2640 2950 49,700 55,600 C 13 Unit, Smooth Top 43 34 50,000 2560 40,700 C 20 Unit, Corrugated Top 42 32 55,600 2840 45,300 Harvard Test Beams. Summary of Results obtained THE THEORY OF BEAMS AS ILLUSTRATED BY TESTS 227 machine was equipped with means for registering the deflec- tions automatically; the diagrams are shown in Figures 168, FIGURE 168. 169, and 170. Generally speaking, there is little difference in the deflection of the Unit and monolithic beams. A number of interesting observations were made during these tests. First, the feasibility of the Unit beam was estab- lished beyond doubt, contrary to what many engineers would Deflection ^L 40000 ^25 ^<^ V 30000 > \ 20000 \ \ 1.0* 0.9" 0.8 " 0.7" 0.6" 0.5" 0.4" 0.3" 0.2" 0.1" \ FIGURE 169. probably have expected. In fact, many building regulations throughout the country specify positively that the beam and its superimposed slab must be concreted in one continuous opera- 228 REINFORCED CONCRETE BUILDINGS tion. Where improperly designed, or otherwise inadequate, U-bars are used, this rule is undoubtedly highly beneficial, but where proper U-bars are used, the rule is wholly unnecessary. The progress report of the special committee of the American Society of Civil Engineers recommends that the slab be con- sidered effective in compression when " proper bond" is pro- vided between slab and stem; it will be appreciated that this is a much more consistent requirement, although somewhat indefinite. The beams tested so far have shown that the bond provided was adequate, whether the more elaborate method of Deflection 0.9" 0.8" 50000 30000 0.6' 0.5' 13 FIGURE 170. corrugating the top of the stem was used, or whether the top of the stem was simply left as it was upon completion. It would be very interesting to learn what would happen when the bond was " inadequate," and just where the limit may be found, and in this particular the present tests furnish no information, as the bond remained intact in all cases. See Figures 171 and 172, showing the Unit Beam No. 25. In the second place, these tests confirm in a remarkable degree the theories set forth by the author in Chapter VII in regard to the action of U-bars. Compression failures of the stem were observed in beams 25 and 27; these are shown in Figure 164 and in Figures 171-174. It was observed that the compression failure of the stem was on the same side as the corresponding bent bar, the two bent bars being each near the opposite face of the beam ; in beam 27 THE THEORY OF BEAMS AS ILLUSTRATED BY TESTS 229 FIGURE 171. HARVARD BEAM No. 25. The black lines are ink marks indicating the principal cracks Photo by Mr. J. R. Nichols, Jr., Am. Soc. C. E. FIGURE 172. A CLOSER VIEW OF BEAM No. 25, SHOWING CRUSHING OF THE CONCRETE AT THE ROD. This beam was a Unit beam, it will be noticed that there was no indication of slipping between stem and slab. Photo by Mr. J. R. Nichols, Jr., Am. Soc. C. E. 230 REINFORCED CONCRETE BUILDINGS FIGURE 173. BEAM No. 27, SHOWING PRINCIPAL CRACKS AT LEFT END, AND THE CRUSHING OF THE CONCRETE AT THE ROD. Photo by Mr. J. R. Nichols, Jr., Am. Soc. C. E. FIGURE 174. CLOSER VIEW OF BEAM No. 27. Photo by Mr. J. R. Nichols, Jr., Am. Soc. C. E. THE THEORY OF BEAMS AS ILLUSTRATED BY TESTS 231 crushing took place at both bent bars, one spot on each side, but in different locations, corresponding to the position of the curves in the bars. It is self-evident that this upward pres- sure of the rod must be resisted by an equal downward pres- sure (from the load) thus dissolving the beam into a number of well-defined compressive zones in a manner very different from what takes place in a " solid" homogeneous beam. The same observation was made by Prof. Morsch in regard to his test beam No. VI. Third, a deep, gaping crack was observed in the top of beam No. 20 (Figure 175), near the support. The explanation of this FIGURE 175. crack may be found in the distribution of internal stresses indicated in the drawing, the horizontal arrow at the steel indi- cating the pulling of the steel, the inclined arrow indicating the sum of the compressive forces in the concrete. It will be noted that if these two do not intersect on the vertical line of the reaction, a " re verse" bending moment is created at the end which would cause just such a crack. Here again we have a fact showing that a reinforced concrete beam cannot be consid- ered as a " solid" beam, in which such stresses' are impossible. Considering the beam as a truss, we see at once that the crack comes outside the "end panel," and so would have no influence on the load-carrying capacity. In addition, it must be admitted that "shear," so called, would have caused the instantaneous collapse of a beam with such a crack. As an actual matter of fact, this beam, with the gaping crack in the top, carried a total load of 55,600 Ibs., or more than any other beam of the entire series. The stem was perforated with inclined and vertical cracks so that the only portions of the beam which could actually carry some shear were the main tension rods. This proposition has been consid- ered above and cannot be maintained. The truth is that there 232 REINFORCED CONCRETE BUILDINGS was no active shear in this beam, the system consisting approx- imately of members as shown in Figure 176. 1 Fourth, it was established that the quarter turn given the straight tension rods at the ends was not sufficient to develop the desired amount of sliding resistance. Thus, beams 9 and 12 failed suddenly by the entire separation of the lower rods from the concrete, while the beams of Type C (13 and 20) showed a sliding of from i" to f " (in these beams, the ends of the rods I I FIGURE 176. could easily be observed by breaking away a thin shell of con- crete). The behavior of the balance of the beams, and especially inspection of the deflection diagrams, makes it, however, appar- ent that only a very small additional margin of sliding resistance was required in order to prevent the sudden collapse. Without doubt, the large turn of the upper bar might profitably have terminated at its lowest point, as the last fourth of the circle materially weakened the concrete along the lines of cleavage 1 The authors are aware of the fact that other observations were made during the testing of the Harvard series which strongly support the theory advanced in Chapter VII of this volume. We are, however, requested to withhold this matter from publication at the present time, and we must refer to the later report to be published by Prof. Johnson. INDEX Accidents, 191 Acid, carbonic, for hardening con- crete, 12 hydrochloric, for joining con- crete, 10 joint, 10 joint, specifications for, 198 Adhesion, 54 Allowable stresses, 52, 128 Alum, 190 Arches, allowable stresses in, 131 Assumptions, homogeneity, 51 in general, 52 tensile resistance of concrete disregarded, 104 Basic inventions, 18 Beam formulas, 70, 71, 75, 88 Belt course, Ransome patent, 13 reinforcement of, 153 Bending, 66 combined with compression, 114 Board marks, 176 Brushing, 179 Cement, 137 Chimneys, approximate formula for, 115 Chloride of calcium, 2, 190 of sodium, 2, 190 Clay, effect on concrete, 12 Coil joint, concrete to concrete, 9 for joining rods, 15 Columns, allowable stresses, 130 hooped, 58 least diameter, 62 repairs of defective, 189 strengthening existing, 64 Concrete, dry or wet mixture, 16 mixing and placing, 149 shrinkage and swelling, 126 specifications, 196 Conflagrations, 185 Continuity of beams, 109 Core boxes, permanent, 15 Corners chamfered, 176 Cracks, in cement finish, 181 in slabs and beams, 127 in tile-concrete floors, 182 repairs of, 188 structural significance, 92 Delayed placing of concrete, 7 Design, general remarks, 153 Earthquake, San Francisco, 6 Embedment, required length of, 55 Expansion joint, 2, 128 Facing of concrete, 177 see also Veneer Factor of safety, 129, 132 Falsework, general design, 156 instructions, 196 standardization, 15 Finish, cement, general, 181 instructions, 198 Fireproofing, 183 effect of salt, 16 Floor coverings, cement finish, 181 wood, 154 Footings, formulas, 116 foundations, 171 Forms. See Falsework Frost protection, 151 effect of salt, 10 233 234 INDEX Hooked ends of reinforcement, 55 Hooped columns, 58 Inertia, moment of, 112 Injurious agencies, 17 Illuminating panels, 8 Joining new concrete to old, 9, 198 Laitance, 149, 181, 189 Lateral expansion, 57 Lime, slacked lime in concrete, 12 Mixing and placing of concrete, 149 Molds. See Falsework Monolithic construction, 156 Notations used in bending theory, 66 Overmixing of concrete, 7, 8 Patentees, Alsip, 31 Aspdin, 19 Basset, 33 Bissell, 35 Brannon, 23 Bruner, 39 Bunnet, 21 Cheney, 32 Coddington, 24 Coignet, 22, 30, 31 Considere, 46 Cornell, 28 Cottancin, 40 Cubbins, 36 De Man, 41 Dennet, 21 Edwards, 26 Emerson, 33 Emmens, 24 Fowler, 29 Gedge, 22 Gilbert, 30 Golding, 36 Gustavino, 38 Hallberg, 44 Henderson, 31 Hennebique, 42 Patentees, Hyatt, 23, 24, 25, 26, 33, 34, 35 Jackson, 32, 36, 37, 38 Johnson, 22, 29, 40 Kahn, 45 Knight, 28 Lambot, 27 Lish, 24 Lythgoe, 22 McCarthy, 40 Matrai, 43 Matthews, 33 Melan, 40 Middleton, 28 Monier, 22, 27, 38 Parker, 19 Parkes, 22 Parmley, 45 Rabitz, 39 Ranger, 21 Ransome, E. L., 2, 3, 5, 8, 9, 10, 13, 15, 16 Ransome, Fk., 22 Shaler, 43 Sisson, 32 Smith, 31 Stempel, 39 Stephens, 28 Stevenson, 28 Summer, 28 Tall, 23 Thacher, 43, 44 Thornton, 22 Turner, 24 Visintini, 45 von Emperger, 40 Waite, 41 Wayss, 44 Weber, 46 Wetmore, 32 Wilkinson, 21 Williams, 30 Wilson, 39 Wood, 30 Wyckoff, 28 Piling, 171 cost of, 174 Plastering, 176 INDEX 235 Plates, concrete, 118, 120 steel base, 118, 171 Reinforced concrete, defined, 51 elements of invention, 18 Reinforcement, details of, 105 double, 114 circular, 118 kinds of, 145, 147 requirements, 147 placing, 159 Repairs to buildings, 188 Rolling of floors, 8 instructions, 196 Rubbing of surfaces, 179 Salt, effect on concrete, 10, 184 instructions for using, 196 Sand, 143 Sidewalk lights, 8 Silicate of lime, 2 of potash, 190 of soda, 2, 190 Slab formulas, 79 Slag, aggregate, 145 for making joints, 9 Sliding, of concrete upon concrete, 103 of reinforcement, 101 see also Adhesion, 54 effect of U-bars, 57 Specifications, superintendents', 195 Stand-pipes, 132 Stone, 144 Steel, 145 specifications, 198 see also Reinforcement Stresses, longitudinal, 66 transverse, 93 tensile (in concrete), 104 initial, 126 temperature, 126 allowable, 128 Stirrups. See U-bars Tables, T-beams of minimum depth, 72, 73 T-beams of increased depth, 78, 80 tile concrete floors, 78 discussion of, 81 Tanks, 132 Tensile stresses disregarded, 104 Tile-concrete construction, 76 tables, 78 description of, 148 Tooling, 178 Twisted bars, invention of, 3 effect of twisting, 147 U-bars, 97 spacing of, 102 Unit construction, Ransome's, 16, 162 types of, 161, 162 Veneered buildings, 187 Water, consistency of concrete, 149 consistency of finish, 181 dry versus wet concrete, 16 hot water used, 21 importance of sprinkling, 127 14 DAY USE RETURN TO DESK FROM WHICH BORROWED LOAN DEPT. ZV&&SZ %S ' s=E-u b *s Tel. No. WP INT MAY JVJN JUL Si 5 i^ - > -SENT ON ILL- MAY 2 LD 21-. JUN ?. o 8 FFR 9 1995 * General Library ^jniyej-sity O f California Berkeley ;ULATION 403 V GENERAL LIBRARY - U.C. BERKELEY BDDDfl4D?Dt,