Steel deck floors supported by protected steel beams and girders have been used in multistory buildings in the past because, with their relatively light weight compared to some other types of fire-resistive floors, the structural load on the girders, columns, and foundation of a building could be reduced, with consequent increase in allowable live loads. Further advantages were the ease of framing the steel decks into irregularly-shaped panels, and the rigidity which the floors, when suitably attached to structural members, could impart to the building as a whole.

The tests described in this paper were designed to determine the fire endurance performance of such floors under design load. Trials were made with the fire either above or below the floor assembly. The work was performed between 1931 and 1933 with the cooperation of the American Institute of Steel Construction. The apparent promising future for steel deck floor systems at that time failed to materialize because of the development of other lightweight steel products such as open-web joists and light-gage decking.

This report is being issued with the purpose of making generally available the large body of still pertinent data secured in the tests. In some of the tests, the floors were subjected to fire exposure by means of controlled burnout of combustible materials from above the floor. It appears that these are the only laboratory tests so conducted. In the tests with the fire conventionally below the floor, the performance of the ceiling structures and floor toppings are of

interest in the design of floor assemblies now in

use.

For the purpose of the tests, the steel deck floors were installed in a specially built furnace, and subjected to 16 separate trials of fire endurance, 10 with fire exposure below the floor, and 6 burnout tests of combustible furnishings typical of office occupancies placed on the floor. The 10 tests with the fire below the floor were conducted under requirements for fire tests substantially the same as those of the currently applicable standard [1]'. In two of the set of 10 tests, however, a hose-stream application was made on the floor assemblies subsequent to the fire exposure. This is not listed as a requirement or an option in the standard now in use. As steel in itself has little resistance to the effects of exposure to high temperatures, the floor systems were necessarily protected with various insulating coverings applied as a ceiling beneath the beams, and also in most cases on top of the steel floor deck.

The purpose of the six burnout tests where the fire was above the floor was to determine if heat from fires in combustibles in several concentrations representing the furnishings, supplies, and records of office-type occupancies would be transmitted through the floor to an extent sufficient to weaken the plates, beams, and other structural members to the point of failure under the applied load. In four of these tests, insulating floor-covering materials were applied to the steel deck surfaces to retard the passage of heat to the structural members.

1 Figures in brackets indicate the literature references at the end of this paper.

The floors were constructed of materials of commercial grade supplied by the manufacturers or purchased in the open market. The workmanship was representative of that normally obtained in construction. Welding and plastering operations were carried out by skilled craftsmen in the employ of local contractors. Certain special materials such as linoleum, mastic floor finishes, and ceiling tiles were installed by representatives of the manufacturers. Casting and finishing of ordinary concrete on the floor surfaces were done by regular employees of the National Bureau of Standards.

The floors are described, together with sketches of the various constructions, in table 1 (Tests 1-6, fires above floors), and table 2 (Tests 7-16, fires below floors). All of the floors were approximately 1311⁄2 ft wide by 18 ft long and were built in place in the floor furnace.

2.1. Floor Structure

In every case, the floor structures were 1/4 in steel plates 18 ft long attached to the upper flanges of small section I- or H-beams by either continuous or intermittent welds along their longitudinal edges. From the tables it will be noted that the steel beams varied in size and spacing. In Tests 1-14, seven beams, either in 4-in, 7.7-lb, or 5-in, 10-lb size, were used on 24-16 in centers. The floor assembly for Test 11 was also used in Tests 13 and 14. In Tests 15 and 16, four beams were spaced 48-5/16 in on centers. Two of these were 5-in, 18.9-lb I-beams placed near the centerline of the panel, with side support furnished by two 5-in, 10-lb I-beams, each about 8 in distant from the restraining frame of the furnace. The same structure sufficed for the two tests.

Steel girders were used in six of the tests. They were 12-in, 31.8-lb I-beams spanning the width of the floor 5 ft from one end of the furnace restraining frame. Thus the girder divided a test floor into two sections, one having a supported clear span of 13 ft, the other consisting of a 5-ft span cantilevered from the girder. In one test, No. 10, the floor beams were carried over the top flange of the girder. In the other tests, shelf angles welded to the girders supported the beams, placing the tops of the beams and the girder in the same plane.

In floor assemblies designed with restrained or fixed end supports, the beams were welded along both sides of the lower flanges where they rested on supporting angles securely bolted to the furnace frame. Additional welds were made under the flanges at the edge of the supports. Other restraining angles, bolted to the furnace, were welded to the top of the plate above the

beam supports. The beams of freely supported floors were not welded to the shelf angles, although temporary tack welds were used to secure proper spacing of the parts during assembly of the floors.

2.2. Floor Finish (Surface Insulation)

Three of the tests (1, 2, 15) were conducted without a covering of any kind on the steel plates. The floor in Test 3 was covered with 3/16in thick battleship linoleum while that in Test 4 had a 12-in coating of an asphalt emulsion concrete. The floors for the remaining tests had from 1 to 3 in of concrete, or concrete base and topping. Concretes had gravel or cinder aggregate, or were gas-expanded. Toppings, where used, were 12 in thick, and principally of cement and sand mix, although limestone concrete and mastic were also used. Wherever a concrete flooring was to be applied, except in Test 7, an expanded metal reinforcement binder was tack welded to the steel plates at 2-ft intervals, and raised from the surface between welds. Details are given in tables 1 and 2.

2.3. Ceiling Protection

Details of lathing, plastering, and application. of ceiling tile on the underside of the floors are also given in tables 1 and 2.

Metal lath, with or without reinforcing rods. was used in all tests except Nos. 6, 10, 12, and 15, which had 2-in precast gypsum ceiling tile. This was usually supported on 1-in hot rolled channels with wire ties or clips, but in Test 12 was installed directly on the lower flanges of the beams.

Plaster was applied to plain diamond mesh metal lath (3.4 lb/yd2) or to lath reinforced with stamped ribs or welded wires (3.4 or 4.2 lb/yd2). The ceiling protection varied in materials and thickness for the several tests. Twocoat sanded gypsum plaster 7% in thick was used in Tests 1, 3, 4, 5, 7, while an average thickness of 13% in (range 1 to 15/gin) was applied for Test 2. For the ceiling in Test 13, three different materials were used, portland cement, hair-fibered gypsum, and wood-fibered gypsum. Other ceilings had three-coat plaster built up to thicknesses of 1 to 111⁄2 in.

Heavier insulation was provided in the four tests in which 2-in gypsum ceiling tile was used. For these, an application of 1/2 in of unfibered sanded gypsum plaster was applied to the tile, producing a total thickness of 211⁄2 in. Where the girders were protected by 2-in (Tests 6, 10) or 3-in (Test 11, 13, 14) hollow gypsum tile installed along the web, approximately 1/2 in of sanded gypsum plaster was

TABLE 1. Description of structures for fire exposure tests of steel plate floors-fire above floor.

1/ Where protection is indicated, a 12-in., 31.8-lb. girder was used in the assembly.

4-in. 7.7 Continuous lb.

4-in. 7.7 Continuous 1b.

3/16-in. battleship linoleum

Reinforced metal lath clipped to beam flanges; 7/8 in. 1:2, 1:3 fibered gypsumsand plaster

1/2-in. asphalt emulsion concrete 1 1/2:2:3:2 cement, sand, limestone, asphalt emulsion

Metal lath clipped to beams; 7/8-in. 1:2, 1:3 gypsum-sand plas

ter

Exp. metal lath; 7.8-in. 1:2, 1:3 fibered gypsumsand plaster

Expanded metal floor binder; 1 1/2-in. gasexp. concrete; 1/2-in. concrete topping

Exp. metal lath fasted to beams with clips; 7/8-in. 1:2, 1:3 f1bered gypsum-sand plaster

25

25

15

15

80

105

50

50

20

115

165