Joints are connected by bolts, then grouted. The grout mix is not specified. The following grout mix was used in the test structure: 1 part Type I cement 2 parts of masonry sand 3 oz per 1 lb of cement of a polyvinyl acetate emulsion. The floor channel details are shown in figure 5.16. These elements are standard, commercially available precast concrete roof tile. Concrete used in channel units consists of 3/8-in maximum size lightweight aggregate ("block-mix"). Air-dry unit weight of concrete is 103 lb/ft and 28-day strength ranges from 4000 psi to 5,500 psi (71⁄2 sack mix). Reinforcement consists of a No. 4 deformed intermediate grade (ASTM-A15) steel bar in each leg of the channel and a 34-1412 wire mesh (ASTM-A185) [7] in the back of the channel with the 14-gage wire in the longitudinal direction. The top of the channel is very rough to develop resistance to horizontal shear between the supporting channel and the topping slab. 5.1.3.1. Fire Walls Figure 5.18 shows a typical cross section of the fire walls. The walls are installed in every second bay in the "short" direction of the building (N-S in the test structure). These walls are continuous in all spaces between columns and have no openings in these spaces. The full width of a building will therefore contain two such uninterrupted firewalls in every second bay. (See figure 5.17.) The fire walls are standard dry-wall construction. Metal channels are attached to the concrete members with power-actuated fasteners at a 6- to 8-in spacing. Metal studs 211⁄2 in X 25 gage) are spaced 16-in on center. Wallboards on either side of the metal studs consist of one 12-in gypsum backing board (ASTMC442) [8] and one %-in gypsum wallboard (ASTM-C36 [9]. The wallboards are fastened to the studs by screws spaced 8 to 12 in o.c., which is a closer spacing than that used in standard practice. The details of the actual fire wall installation in the test structure are illustrated in figure 5.19. 5.1.2. Cast-in-Place Topping Slabs The topping slabs have a specified nominal thickness of 2 in. Concrete is made of 344-in maximum size lightweight aggregate, with a weight of 110 lb/ft and a nominal 28-day strength of 3,000 psi. Reinforcement is 66-1010 wire mesh (ASTM-A185) set one inch from the top of the slab. Additional reinforcement is provided at the main beams by the shear connectors and by two No. 4 bars on the first floor and two No. 3 bars on all other floors, as shown in figure 5.8. This reinforcement is ASTM-A61 steel. 5.1.3.2. Exterior Walls Figure 5.20 shows a typical section of the exterior walls. "Exterior walls" as defined here are the outer walls in the long direction of the proposed building (E-W direction in the test structure). Each building will thus have two exterior walls. (See fig. 5.17.) Exterior walls are located in the outer rows of columns and fill the 10 ft space between columns. The walls are not continuous and each panel may contain a door or a window. Exterior walls are standard dry-wall construction. Channels and studs are as in the fire walls. Facing consists of 5-in gypsum wallboards (ASTM-C36 on either side of the stud; screw spacing is as in the fire walls. The wall surface exposed to the atmosphere will be protected by optional siding. 5.1.3.3. Interior Walls "Interior walls" as defined here extend along the two interior rows of columns in the long direction of the building (E-W direction in the test structure). Each building thus has two interior walls in the long directions. These walls fill the 10-ft panels between columns (see fig. 5.17). A 3-ft door may be expected in every second panel. Several types of interior walls are used in the Mitchell System; of these the standard 212in "structicore" partition wall construction was deemed to have the least resistance to lateral load and was thus chosen for the laboratory structure. Figure 5.21 shows typical sections of an interior wall, and figure 5.22 shows a typical interior wall partially dismantled. 5.1.4. Foundations, Grade Beams, and Slabs on Grade Foundation plans are shown in figure 5.23. All foundations are specified as ready mix concrete with a 28-day strength of 3,000 psi. Slabs on grade are ready mix concrete with a specified 28-day strength of 2,000 psi. Reinforcement is ASTM-A15 intermediate grade steel and ASTM-A185 welded wire mesh. Lower-story columns are encased in the foundations. (See fig. 5.7.) 5.2. The Test Structure 5.2.1. Structural Simulation The test structure before and after installation of the walls is illustrated in figures 5.24 and 5.25, respectively. It comprises a part of the complete structure, made up of full-scale components and erected in the laboratory. The test structure as part of the complete structure is illustrated in figure 5.2. The performance of the complete structure is simulated in the test structure by: (1) applying to the test structure all live loads which under field conditions would act directly on the test structure; (2) simulating all forces caused by dead, live, and wind loads which would be exerted on the test structure by the rest of the structure under field conditions. The test structure is thus treated in the laboratory as a "free body." The test structure was so chosen that all aspects of structural performance in the field could be simulated under laboratory test conditions. The test structure corresponds to a part of the total structure which is cut off below the slab on grade. 5.2.2. Description of the Test Structure The test structure was constructed under the provisions of the plans and specifications of the 5.2.2.2. Cast-in-Place Topping Slab The specified thickness of the cast-in-place topping slab is 2 in. The average slab thickness as built in the test structure was 21⁄2 in in measured from the top of the main beam. The top of the floor channels was irregular and tended to be somewhat higher than the top of the main beam, producing a somewhat lesser average thickness than the measured 21⁄2 in. But even after allowing for a thickness reduction due to floor channel irregularity, the as-built thickness was still in excess of the specified 2-in thickness. 5.2.2.3. Walls The test structure had the following walls: (a) East and west walls were "fire walls" as described in section 5.1.3.a, except that 3gin gypsum backing boards and 1-in. wallboards were used instead of the thicker sizes called for in the plans. (b) The south wall was an exterior wall as described in section 5.1.3.b except that the exterior siding was omitted and 12-in thick wallboards were used instead of the 8-in thickness shown in the plans. (c) The north wall was an "interior wall" as described in section 5.1.3 (3). All channels for the wall system were attached by power-actuated fasteners to the floor slab and the structural frame. Vapor seals and insulation between walls were omitted since these materials do not add to the strength of the structure. The omission of exterior siding on the south wall may have slightly decreased the stiffness of that wall, which would cause the test results to be lower than they might otherwise have been. Each panel in the south wall contained a 5 ft 7 ft aluminum doorframe on its west side. This represents the least stiff condition that may be encountered in the field. The western panel of the north wall contained a 3 ft x 7 ft wooden doorframe on its east side. This simulates field conditions. 5.2.2.4. Floor Slab The cast-in-place floor slab was poured on top of a vinyl sheet which was spread on the laboratory floor. The floor slab slab was subsequently post-tensioned against the laboratory floor by four 1/2-in-diameter bolts in order to prevent sliding due to lateral test forces applied to the structure. Tests indicated that the floor-slab concrete had a 17-day compressive strength of 5,600 psi. Slab reinforcement consisted of a 66-1010 mesh (ASTM-A185). The slab was poured around the columns which were lined by 1/2-in asphalt-impregnated fiberboard, thus forming full-depth pockets at the column seats to permit column rotation at the base. A g-in-thick neoprene sheet was inserted between the column base and the laboratory floor. 5.2.2.5. Materials Standard compression tests (ASTM C33-66) were carried out on cylinders of concrete from the "cast-in-place" slabs and the precast members with the exception of the floor channels. In all cases concrete strength exceeded the strength specified in the plans. Reinforcing bars were ASTM-A61 (60 ksi) steel wherever the plans permit the option of using 50 or 60 ksi steel. 5.3. Fidelity of Simulation of Field Conditions by the Test Structure Complete full-scale structures can be and have been tested in the field. While such field. tests provide a means for the observation of the performance of a complete structure, it should also be noted that when compared with laboratory tests, field tests have many disadvantages. Some of the more obvious disadvantages are: cost; the time required to erect and test a full-scale structure in the field; changing conditions of temperature and wind; and the difficulty of precise application of loads and measurement of deformations. The major advantages of field testing are the ability to test an entire structure and a better simulation of foundation conditions. For the case reported here, the entire test was performed inside the laboratory facilities. of the National Bureau of Standards. Since it was impractical to erect a complete structure in the laboratory, it was decided to construct a portion of the structure and to test it in a manner that simulated the performance of the complete structural system. A lower-story section was selected, since lower-story components are subjected to the most critical loading conditions. The load program to which the test structure was subjected is discussed in section 6. The fidelity of the simulation is discussed in the following sections. 5.3.1. Interaction Between the Test Structure and the Complete Structure Figure 5.2 illustrates the test structure as part of a complete structure. The test structure with the testing equipment installed is shown in figures 5.26 and 5.27. In an actual building, the test structure would be connected to the remainder by: (a) Columns, (b) Abutting tie beams and main beams, At all of these connections forces are exerted on the test structure, either by direct transmission of loads carried by the connected members or by restraining effects on motion of connected members. It is neither feasible. nor necessary to simulate all these effects. Simulation of the most adverse conditions will generally lead to simplified approximations which are on the conservative side. Simulation of structural interaction at these four points of continuity is discussed in the following: 5.3.1.1. Columns Upper-story columns will transmit to the beam-column connection most of the dead loads generated by the stories above and the live loads acting on these stories. For the laboratory model it was assumed that the upper-story columns would transmit the following loads to the joint at their base: (1) Dead loads of the upper stories. (2) Vertical live loads on the upper stories. In reality the columns between the fire walls. will also transmit a certain amount of horizontal wind-induced shear load. However, as will be noted later, in the presence of the partition walls, only a negligible amount of the total wind shear was carried by the columns. The wind shear from the upper stories was assumed to be carried by the walls to the top slab, which in turn will transmit the shear to the partition walls below. It will also be noted later that some of the vertical loads are carried by the wall system. directly into the foundations. The assumption that the entire vertical load is carried by the columns is a conservative assumption with respect to columns. The fact that the walls could potentially be more highly stressed in the complete structure than in laboratory simulation does not appear to be of significance, since a wall failure by vertical loads would not occur without a simultaneous column failure. Column loads were applied vertically by rams at the center line of the lower-story columns as illustrated in figure 5.28. Rollers were inserted to roll in the direction of racking and to minimize frictional forces which might resist racking while vertical loads were applied. It is recognized that upper-story columns would transmit moments as well as vertical loads, while the rams applied only axial vertical loads. It is demonstrated in Appendix B that this application of column axial loads is conservative. 5.3.1.2. Abutting Tie Beams and Main Beams Main beams are discontinuous at both of their ends in the real structure, and this was correctly reflected in the test structure. Tie beams may be either continuous or discontinuous depending on their position in the structure. If tie beams were continuous on either or both sides of the test structure, this would result in increased load-carrying capacity and decreased deflections. Thus, it may be stated that with respect to structural continuity the test structure represents a conservative approximation. 5.3.1.3. Continuity of Topping Slab In the complete structure, topping slabs may be continuous on three sides of the test section -west, north, and east- or on two adjacent sides of the test section. The severing of this continuity in the test structure represents a conservative approximation with respect to both load-carrying capacity and deflection. 5.3.1.4. Walls The wind load is imparted to the wall by (1) shear along its upper connection to the beam above it, and (2) bearing of the windward column against the wall. Since the floor system is very rigid in relation to beam column joints and walls, the horizontal forces acting above any floor are transmitted into this floor by the walls and in turn essentially equally distributed among the walls below this floor by a uniform displacement of the entire floor. In the test structure, simulated wind loads equal to one-half the wind loads generated by the entire contributory portion of the threestory building were imparted at the end of each main or tie beam by a ram load, as illustrated by figure 5.26. In the case of the north direction, a wind load was also applied at the main beam on top of the column between the two fire walls. Due to the stiffness of the floor system, these wind loads have a net effect equal to the effect that may be expected on a structure in the field. The reason for applying only one-half of the wind force to each wall is the above-discussed assumption of great floor stiffness, which would distribute the wind load to two wall panels in the north direction and to more than two wall panels in the east direction. Test results also indicate that the walls participate in the support of vertical loads. This was demonstrated by the fact that deflection of main beams connected to fire walls increased almost fivefold when these walls were removed. As will be noted later, the loading applied in Test No. 9 more than compensated for any adverse effect of vertical loads on the walls under service load conditions. Column loads were computed without regard to possible wall participation in load support. It is therefore concluded that the simulation of wall action adequately represented the most adverse conditions that may be expected in a complete structure. 5.3.2. Simulation of Foundation Conditions Column foundations in the proposed building extend to a 6-ft depth below grade for exterior columns, and 3-ft depth below the top of the floor slab for interior columns (see fig. 5.23). Exterior column footings are also tied into the perimeter wall for added fixity. This configuration provides some degree of fixity at the column base, the degree depending on prevailing soil conditions. In the test structure, the columns were "cut off" at the bottom of the floor slab. The lower ends of the columns were provided on all sides with a 1/2-in-thick asphalt-impregnated fiberboard expansion joint against which the floor slab was cast, thus providing a detail similar to that of the real structure where a 3g-in premolded filler is placed around the column. The base of the column was set on a 18-in-thick neoprene bearing pad which rested on the laboratory test floor. The resulting column connection permitted the column base to rotate, and therefore was a conservative simulation of the real structure, where the foundations provided partial column base fixity. 5.3.3. Simulation of Live Loads Vertical live loads on the top slab of the specimen were simulated by air-bags which were held down by a suitable reaction system (see fig. 5.26). This created a uniformly distributed load which was able to follow the deflections of the slab. Air bags were made of 20-mil polyethylene and were designed to withstand 300 psf (7 times live load). The live loads applied represented a valid simulation of live load conditions as used in structural design. Horizontal live loads were applied by horizontal 10-ton rams as illustrated in figures 5.26 and 5.29. Occupancy loads (floor)-40 psf Snow loads (roof)-30 psf 6.3. Loading Schedule Figure 6.1 shows schematically how the test loads were applied to the structure. Table 6.1* explains the symbols used to represent the test loads and the magnitude of these loads. Table 6.2 summarizes the magnitude of test loads which represent the performance criteria. Tests were conducted on the test structure with walls installed, and subsequently on the same structure after the walls were removed. All load tests were conducted between May 10, 1968, and May 22, 1968, and are listed hereafter. 6.3.1. Tests Conducted on the Structure With Walls Installed The validity of wind load simulation has been discussed in section 5.3.1.d. 6. Load Program Column loads of 0.9D South wind load to 25 psf (0.9D 1.25H) The load program in this test had three objectives: (1) Evaluation of the structural adequacy of the proposed system and determination of its ability to satisfy the performance criteria established in section 4. (2) The acquisition of additional information about the behavior of complex structural systems and the interaction of their components. (3) The development of suitable methods of performance testing for complex structural systems. Section 6.2 explains the assumptions which were made with regard to the magnitude of applied live and wind loads and section 6.3 explains the load schedule. Load computations and the detailed sequence of loading used in each test, are presented in Appendix C. 6.2. Applied Loads All applied loads were determined in accordance with "Minimum Design Loads in Buildings and Other Structures," Structures," USASI A58-1955 [11]. The following unit service loads were used: Windloads for average Midwestern conditions were selected. * All tables mentioned in the text appear in the section beginning on p. 26. Test 4: Test 5: |