CONTINUOUS TIMBER DIAPHRAGMS

Transcription

1 CONTINUOUS TIMBER DIAPHRAGMS By Thomas S. Tarpy, Jr., 1 M. ASCE, David J. Thomas, 2 and Lawrence A. Soltis, 3 M. ASCE ABSTRACT: Current design assumptions for diaphragms assume support conditions which are either simple span or fully continuous. The building codes require a design based on the highest values for moment and shear obtained under either of these two support conditions. More practical criteria for assessing continuity conditions at supports for wood diaphragms are needed. This investigation is to determine experimentally the effects of continuity conditions on timber floor diaphragms with plywood sheathing subject to inplane loads. Previous testing programs have evaluated simply-supported diaphragms subject to uniform loading; this study evaluates the effects of other support conditions and non-uniform loads. Static loading conditions were used to evaluate the response of the diaphragm for both deflection and ultimate strength. Six 8 ft 16 ft (2.44 m 4.88 m) floor diaphragms typical of certain residential construction techniques with three different sets of boundary and loading conditions were tested in accordance with ASTM E72. The tests demonstrated that: (1) Continuity over a rigid support apparently does not increase the unit shear resistance values of the diaphragm; (2) concentrated loads on the diaphragm produce lower load factors than moment-equivalent uniform loads at a given load level; (3) there is not an apparent direct relationship between relative panel displacement and overall diaphragm deflection for the size diaphragms tested; and (4) local panel buckling has a minimal effect on overall diaphragm failure patterns. INTRODUCTION Lateral forces on buildings due to wind or earthquake are often resisted by combinations of end walls and interior walls at one or more locations. The floor diaphragm behaves like a beam, with the end and interior walls acting as supports. Building codes require design based on support conditions which are either simple span (thus, the diaphragm is flexible) or fully continuous, as over an interior support. The highest moment and shear values obtained under either of these conditions is required for design. The actual diaphragm behaves in a manner somewhat between these extremes. This research is a pilot study to assess the effects of continuity conditions for a typical residential construction size horizontal wood diaphragm under different loading and boundary conditions. Cantilevered horizontal diaphragms, continuous spans of substantially differing widths, or end spans of continuous diaphragms are often designed for two different sets of forces corresponding to either simple or continuous assumed support conditions. The first assumes that the load from half the tributary span is applied to the next interior span, 1Research Assoc. Prof., Dept. of Civ. Engrg., Vanderbilt Univ., Nashville, Tenn.; Struct. Engr., Stanley D. Lindsey and Associates, Ltd., 1906 West End Ave., Nashville, Tenn Former Grad. Student, Vanderbilt Univ., Nashville, Tenn. 3Project Leader, Forest Products Lab., Madison, Wisc. Note.-Discussion open until October 1, To extend the closing date one month, a written request must be filed with the ASCE Manager of Journals. The manuscript for this paper was submitted for review and possible publication on June 13, This paper is part of the Journal of Structural Engineering, Vol. 111, No. 5, May, ASCE, ISSN /85/ /$ Paper No

2 while the second assumes that more than half of the load is applied to the next interior span. In many cases, this approach results in nail spacings so close that splintering occurs and boundary members need replacing. This conservative approach often increases construction costs. Thus, more complete criteria for assessing continuity conditions for horizontal wood diaphragms in current design practice are needed at this time. A comprehensive bibliography (3) and state-of-the-art publication (2) are available for plywood and timber diaphragms. Design values for plywood diaphragms were developed through tests on large diaphagms (4,5,7,8,10,11) which considered the diaphagm as simply supported, flexible elements. This research used static loading conditions to be compatible with experimental test results of previous research. These tests describe the inplane loading response for deflection and ultimate shear strength with three sets of boundary support and loading conditions: 1. Simple span horizontal diaphragms with the loads and reactions located symmetrically-thiscondition serves as a control specimen condition for comparison with earlier research. 2. Cantilever horizontal diaphragms which are resisted by two rigid supports with the diaphragm extending over one of the supports-the loads and reactions are located symmetrically on opposite sides of the diaphragm. 3. Cantilever horizontal diaphragms as previously mentioned, but with loads and reactions located asymmetrically on opposite sides of the diaphragm; this loading results in inplane torsional behavior. Two specimens were tested for each of the three conditions. These two specimens differed only in that the long dimension of the plywood sheathing was perpendicular to the applied loads in one specimen and parallel to the applied loads in the other specimen. The results of these tests and their implications for design practice are considered in this study. A more detailed presentation of the results is presented in Ref. 9. TEST METHODS AND PROCEDURES Six 8 ft 16 ft (2.44 m 4.88 m) horizontal plywood diaphragms were constructed as shown in Figs The individual plywood panel sheet numbers are shown in circles, along with the reaction points, R, and load points, F, for each test in Figs Tests 1 and 2 (Fig. 4) were simple span specimens loaded at the third points with reaction points at the ends of the diaphragm. Tests 3 and 4 (Fig. 5) were cantilever diaphragms, loaded 1-1/2 in. (38 mm) from the end of the cantilever, and at a point on the same side 6 ft (1.83 m) from the exterior support condition. Tests 3 and 4 reaction points were located at the exterior support condition and at a point 4 ft, 2 in. (1.27 m) from the cantilevered end of the opposite side. Tests 5 and 6 (Fig. 6) were cantilevered diaphragms loaded 1-1/2 in. (38 mm) from the end of the cantilever on one end of the long side of the diaphragm and at a point 6 ft (1.83 m) from the exterior support condition on the opposite side of the diaphragm. Re 993

3 FIG. 1.-FramingPian and Plywood Panel Orientation for Tests 1, 3 and 5 FIG. 2.-FramingPlan and Plywood Panel Orientation for Tests 2, 4 and 6 action points for Tests 5 and 6 were located at the exterior support condition and at a point 4 ft, 2 in. (1.27 m) from the cantilevered end on the opposite side. Test Specimens 1 and 2, 3 and 4, and 5 and 6, respectively, taken as pairs, differ in orientation of the plywood sheathing. While Specimens 1 and 2 have the same framing pattern, the plywood sheathing in Specimen 1 is oriented so that the continuous panel joint is perpendicular to the applied loads. In Specimen 2, the plywood sheathing is oriented so that the continuous panel joints are parallel to the applied loads-contrary to code recommendations, but necessary to make comparisons to Specimen 1. The same holds for Specimens 3 and 4, and 5 and 6, respectively. Thus, each set of two specimens corresponds to only one loading condition, and each of the two members of the set differs from the other only in the orientation of the plywood sheathing. The plywood sheathing used was Structural I, 4 ft 8 ft (1.22 m 2.44 m) nominal sheets, 3/8-in. (9.5-mm) thick, C-C, Exterior Grade, 24/ 0. The sheathing was fastened to the sub-framing members with 8d common nails spaced at 4 in. (102 mm) around the perimeter of the diaphragm, at 6 in. (152 mm) along panel edges and at 12 in. (305 mm) over interior joists inside the panel boundaries. This conforms to a design load level of 320 lb/ft (4.67 kn/m) per the Uniform Building Code (6). The minimum edge distance for the 8d common nails in the sheathing was 1/2 in. (12.7 mm). Nails were driven on a slant where necessary to avoid splitting of the joists and blocking. The gap distance between panels was 1/8 in. (3.2 mm). 994

4 FIG. 3.-FramingDetails: (a) Typical Corner Framing-Detail A ; (6) Framing Details, All Diaphragms; (c) Blocking Details, All Diaphragms The sub-framing material was nominal 2 in. 8 in., (actual mm) No. 1 Dense Surface Dried (19% or less moisture content) Southern Pine, pressure treated. The No. 1 Dense was selected to minimize the effects of knots or other defects normally found in timber testing. Only treated material was readily available in No. 1 Dense and, thus, was used. The perimeter chords of the diaphragms were two nominal 2 in. 8 in. (actual mm) members laid horizontally and nailed together with 12d common nails at 4 in. (102 mm) spacing along the perimeter except for the middle 8 ft (1.22 m) of the long sides where 8 in. (203 mm) spacing was used. Vertical framing consisted of five 16 ft (2.44 m) lengths of nominal 2 in. 8 in. (actual mm) joists, placed in the long direction of the diaphragm, perpendicular to the intended direction of the loads and reactions. Blocking consisted of nominal 2 in. 8 in. (actual mm) sections toe-nailed to the joists. Intervals for placement of the blocking were: (1) Standard blocking at ends and at 4 ft on center, and 995

5 FIG. 4.-Panelidentification and Orientation for Simply Supported Diaphragms: (a) lest l; (6) Test 2 FIG. 5. -Panelidentification and Orientation for Cantilever Diaphragm with Loads and Reactions Symmetrically Located on Opposite Sides of Diaphragm: (a) Test 3; and (6) Test 4 (2) "load support" blocking immediately in line with the applied point loads. The load support blocking location vaned with the loading conditions for the individual tests. All wood framing was assembled and tested at between 10 and 14.5% moisture content. The framing plans for each test showing joist direction, blocking location, and load and reac- FIG. 6.-PanelIdentification and Orientation for Cantilever Diaphragm with Loads and Reactions Asymmetrically Located on Opposite Sides of Diaphragm: (a) Test 5; and (6) Test 6 996

6 FIG. 7.-An Example of Sub-Framing and Panel Layout: (a) Sub-Framing; and (6) Plywood Panel Layout tion points are shown in Figs. 1 and 3. Fig. 7 shows an example of diaphragm sub-framing and plywood panel layout located in the test frame. Test procedures were based on ASTM E72-74a, Standard Methods of Conducting Strength Tests of Panels for Building Construction (1) and on the methods detailed in the 1966 APA Report (10). Since one of the purposes of the test was to determine the load-deformation curves for the different loading conditions and panel orientations, considerable care was taken to insure a large number of readings of both loads and deflections. At each load level, the load was maintained for five (5) min prior to recording diaphragm deflections and moving to the next higher load level. Load increments were such that a load rating of less than 300 plf/min (4.38 kn/m/min) was maintained. Load increments were continued until the diaphragm was unable to carry additional load. The last load level for which deflections were recorded is defined as the ultimate load. TEST RESULTS AND ANALYSIS The results of Tests 1-6 are summarized in Table 1. Two salient results are apparent. The load factors (ultimate load/allowable load) for the control specimens are greater than 2, while the highest unit shear resistance load level is for Test 5-atest in which continuity over a single support 997

7 TABLE 1.-Summaryof Test Results was present and in which the loads were applied from opposite sides of the diaphragm. It is notable that the failure modes for Tests 2, 4 and 6 were virtually identical, thus, indicating the importance of the orientation of the continuous plywood panel joint location relative to the direction of application of the loads. Particular attention is made to the ultimate strength of Test 5 since an edge joist failed rather than failure in the plywood. The principal significance of this failure mode is that the diaphragm could possibly have sustained even higher unit shears in this loading and panel configuration if the sub-frame joist had not split. The failure was basically a result of difficulties experienced with the travel distance of the apparatus for applying the loads to the diaphragm and not with the actual diaphragm itself. The deflected shapes of the individual diaphragms at ultimate load (solid line) and recovery (dashed line) are shown in Figs A review of previous test results with larger diaphragm sizes (5,10) indicates that the load factor ranges were on the order of three-four for all of the aspect ratios considered, and that deflections at ultimate load appeared to be more a function of aspect ratios than of design shear strength. The influence of fastener spacing (nail or staple schedule) was evident in all the tests undertaken. There was not an apparent way to assess the influence of quality and size or depth of framing members, 998

8 FIG. 8.-DeflectedShape (-)and Recovery (---)at Ultimate Load for: (a) Test 1; and (b) Test 2 FIG. 9.-DeflectedShape (-)and Re covery (---) at Ultimate Load for: (a) Test 3; and (b) Test 4 since neither depth of framing nor grade of framing were the same throughout these previous tests. Given the difference in the ratio of deflection at ultimate load to deflection at design load, it is difficult to definitely conclude that deflection was mainly a function of aspect ratio, with all other factors being equal. However, the range of values of the FIG. 10.-DeflectedShape (-) and Recovery (---)at Ultimate Load for: (a) Test 5; and (b) Test 6 999

9 deflection ratio seems to be from 8-11 for blocked and nailed specimens with an aspect ratio of 4:1. A comparison of the load factors obtained in these tests with the results of the 1966 APA tests (10) which used 3/8 in. (9.5 mm) plywood indicates that the load factors for the tests which used uniform loads were considerably higher. The APA load factors were in the range of , while that for the concentrated loading case reported herein was 2.34 for the similar configuration (Test 1). The difference between APA results and those in this report are due to the method of loading, i.e., uniform versus concentrated as the following calculations will show. In Test 1, for example, the failure load was 6,000 lb (26.7 kn) per reaction, and the maximum pure bending moment is kip-ft (43.5 knm). Using a static moment relationship for the uniform load case gives plf (14.6 kn/m) or a total equivalent load of 16,000 lb. (71.2 kn) on the diaphragm-greaterthan the 12,000 lb (53.4 kn) able to be sustained by the diaphragm. The relative reduction in load capacity is, thus, 25%. The results of Test 5 show the importance of panel orientation and the relationship of the applied loads to the line of the continuous panel joint. This condition, in accord with the requirements of UBC Table 25- J (6), has been a part of the design process for some time. Its importance for the tests reported here is that the shear strength resisting inplane torsion was seen to be a direct function of panel orientation. Extensive panel buckling took place in Test 5, but this buckling did not result in early failure of the specimen. The orientation of the panels increased the shear resistance to inplane torsion of the diaphragm. This resistance was notably different from the resistance in Test 6, where the continuous panel joint was parallel to the direction of the applied loads. The load factor for Test 5 was 3.05, while the load factor for Test 6 was Thus, design for inplane torsional effects, which is a consideration in such locations as offsets and L shapes of buildings, must acknowledge this difference in response. CONCLUSIONS The purpose of this investigation was to determine the effects of continuity conditions on timber diaphragms with plywood sheathing subject to inplane loads. This is an important concern in the design of floor and roof diaphragms for low-rise timber and composite structures subject to lateral forces. Current design assumptions, as manifested in the Uniform Building Code and other model codes, assume support conditions for diaphragms which are either simple span, considering the diaphragm to be fully flexible, or fully continuous, as over an interior support. These codes require a design based on the highest values for moment and shear obtained under either of these two conditions. The research reported herein used static loading conditions for determining the deflection and ultimate strength of six 8 ft 16 ft (2.44 m 4.88 m) horizontal diaphragms under three sets of boundary and loading conditions. 1000

10 The results of these tests demonstrate that: 1. Continuity over a rigid support apparently does not increase the unit shear values which can be resisted by the diaphragm, thus, justifying current design assumptions. 2. Concentrated loads on the diaphragm produce lower load factors than moment-equivalent uniform loads for a given load level. 3. There is not an apparent direct relationship between relative panel displacement and overall diaphragm deflection for the size diaphragms tested. 4. Local panel buckling has a minimal effect on overall diaphragm failure pat terns. Since only one size diaphragm (and, thus, one aspect ratio) was tested, it is the writers opinion that future testing should be directed at developing a table of continuity conditions versus aspect ratio for a given panel thickness, plywood orientation, and framing patterns which can be used directly by designers. The effects of openings in horizontal plywood diaphragms on shear transfer and continuity should also be addressed. APPENDIX I.-REFERENCES 1001

11 APPENDIX II.-NOTATION The following symbols are used in this paper: F = applied forces; P = panel number; and R = reaction. 1002