EXPERIMENTAL INVESTIGATION OF WOOD NAILER TOP CHORD ATTACHMENTS FOR OPEN WEB STEEL JOISTS

Transcription

1 EXPERIMENTAL INVESTIGATION OF WOOD NAILER TOP CHORD ATTACHMENTS FOR OPEN WEB STEEL JOISTS Michael J. Martignetti 1, Shawn P. Gross 2, Sameer S. Fares 3, David W. Dinehart 4, Joseph R. Yost 5 ABSTRACT: Presented herein is an experimental program to determine the structural effect and potential steel savings achieved when attaching a wood nailer to the top chord of open web steel joists. Partially composite behavior between the wood nailer and the steel top chords changes the properties of the top chord and the performance of the joist. The investigation consisted of a series of ten tests, which simulated a full-scale model floor system by utilizing two joists and either wood or steel decking depending on whether the joists had a wood nailer top chord attachment or not. Load, displacement, and strain readings were recorded as a uniform load was applied monotonically over the two joists in each test. Material properties of the steel chords and wood nailer were also verified experimentally. The results demonstrated that the wood nailer top chord attachment exhibits partially composite behavior with the steel top chord angles of the joists, allowing a wood nailer joist with smaller chord angle sizes to perform comparably with a traditional steel joist. Utilizing the partially composite behavior in the design of wood nailer joists can optimize the economy of the joists by reducing the amount of steel required in the top chord. KEYWORDS:, Open Web Steel Joists, Partially Composite System 1 INTRODUCTION Open web steel joists are a very efficient means of spanning long distances to create floor and roof systems through both reduced material and field labor costs [1]. A traditional open web steel joist, which is typically used in conjunction with steel deck to create a floor or roof system, consists of steel angles for the top and bottom chords and has either angles or rods that make up the webs of the joist. This study investigates the benefit of attaching a wood nailer to the top chord of the joist to create what is known as a wood nailer joist. The presence of the wood nailer allows wood decking to be utilized in conjunction with the joists as the wood 1 Michael J. Martignetti, E.I.T., Project Engineer, CMC Joist & Deck (Formerly Graduate Student, Villanova University), 1651 Cedar Crest Blvd., Allentown, PA michael.martignetti@cmc.com 2 Shawn P. Gross, Ph.D, Associate Professor of Civil Engineering, Villanova University, 800 Lancaster Avenue, Villanova, PA 19085, shawn.gross@villanova.edu 3 Sameer S. Fares, P.E., S.E., Research and Development, CMC Joist & Deck, 619 North 1 st Street, Jacksonville, AR sam.fares@cmc.com 4 David W. Dinehart, Ph.D, Professor of Civil Engineering, Villanova University, 800 Lancaster Avenue, Villanova, PA david.dinehart@villanova.edu 5 Joseph R. Yost, Ph.D, P.E., Associate Professor of Civil Engineering, Villanova University, 800 Lancaster Avenue, Villanova, PA joseph.yost@villanova.edu decking can be nailed directly to the joist. This system is commonly used in the Pacific Northwest. The primary purpose of this study is the examination of the structural benefit of the wood nailer top chord attachment through the investigation of the partial composite behavior of the wood and steel top chord components. The wood nailers are screwed to the steel top chord of the joist, resulting in a partial composite behavior. The addition of the wood nailer raises the centroid of the top chord; thereby, increasing the moment arm between the top and bottom chord of the joist. Theoretically, lower forces are then developed in the chords of the joist; resulting in a reduction in the area of steel needed in the top chords and offering a more efficient joist design. Figure 1 illustrates the cross sections of a wood nailer joist and a standard joist. Figure 1: Joist Cross Sections

2 2 EXPERIMENTAL Experimental testing was conducted in three phases to determine the partial composite behavior, and the basis of calculating the effectiveness of the wood nailer in increasing the moment arm between the top and bottom joist chords. Joists were tested both with and without the wood nailer top chord attachment and with different top chord steel angle sizes to produce the data needed to develop equations for the partial composite ratio [2] [3]. Testing of the material properties of the joist components such as the elastic modulus of the wood nailer and the yield strength of the steel was also necessary. 2.1 JOIST TEST SPECIMENS A total of ten full-scale tests, summarized in Table 1, were conducted to evaluate the performance of the wood nailer joists. All of the joists were 40 long, 28 deep, and contained 2 x 2 top chords, a combination of crimped angle web members and double angle web members, and 1 diameter exterior rod web members. The nominal dimensions and thickness of the bottom chord angles varied to examine the effect of the wood nailer. Table 1: Test Joists PHASE III-A Tests 1 & 2 (J1) Yes Top Chord 2.0" x 2.0" x 0.176" (A s = in 2 ) Bottom Chord 1.75" x 1.75" x 0.170" (A s = in 2 ) lbs PHASE III-B Tests 3 & 4 (J2) No Bottom Chord 2.0" x 2.0" x 0.163" (A s = in 2 ) lbs PHASE III-C Tests 5 & 6 (J3) No Bottom Chord 1.75" x 1.75" x 0.170" (A s = in 2 ) lbs Tests 7 & 8 (J2) No Bottom Chord 2.0" x 2.0" x 0.163" (A s = in 2 ) lbs Tests 9 & 10 (J1) Yes Bottom Chord 1.75" x 1.75" x 0.170" (A s = in 2 ) lbs Four tests utilized wood nailer joists and six tests investigated the performance of traditional steel joists. Four of the tests on traditional joists included joists with a larger bottom chord than the wood nailer joists and two tests used joists with the same chord sizes. All wood nailer joists had a 2 ½ tall x 5 wide wood nailer attached to the top of the steel top chord angles with ¼ x 2 self tapping screws spaced at approximately 12 on center, on alternating sides. 2.2 TEST SETUP AND DATA ACQUISITION General Test Setup The joists for this study were tested by constructing a 40 by 8 floor system comprised of two joists and the appropriate decking material. The two joists were bolted to angle supports on steel columns, which were anchored to a strong-floor. Diagonal bracing was installed at midspan to resist lateral deformation of the joists. The decking material was then spanned across the joists and fastened to the top chord to complete the floor system and provide further lateral stability Joist Test Setup The wood nailer joists were tested with a deck comprised of 2 by 4 wood purlins and ½ thick plywood sheathing. Steel joist hangers were attached to the sides of the wood nailer top chords to provide a means of fastening the wood purlins to the joists. The purlins were spaced 2 on center to support the plywood. The sheathing was nailed to the purlins and top chords with 10d common nails spaced at 2.5 at the perimeter of the floor, 4 at the plywood seams, and 12 at intermediate purlins. Steel angles were bolted to the bottom joist chord at mid-span and nailed to the wood purlin above to provide additional lateral stability. Figure 2 shows the wood nailer joist specimen. Figure 2: Joists and Plywood Deck Standard Joist Test Setup The floor system for the traditional joists was constructed using 20 gage 1½ standard wide rib painted steel deck. The strength and geometry of the steel deck eliminated the need for any purlins spanning between the joists. The deck was attached to the top chords of the joists using puddle welds at every other low flute, and the side laps of the deck were attached with #10 screws. Additional lateral stability was achieved by using diagonal cross bracing bolted between the top and bottom chords of the opposing joists.

3 action; consequently, testing was conducted to verify the modulus of elasticity of the Douglas Fir wood nailers. After the joist tests were complete, the wood nailers top chord attachments were removed and cut into 5 lengths for use in a four point bending test per ASTM D4761 [4]. Each piece was set on two roller supports spaced 4 apart in a self reacting test frame and loaded at the third points. The load and displacement were measured with the same testing equipment used in the joist tests. The span and loading conditions were known and the actual load and deflection results were recorded; therefore, it was possible to determine the modulus of elasticity of the wood. Figure 3: Traditional Steel Joists and Steel Deck Loading A uniform distributed load was accomplished through the use of load pyramids, each of which used a stiffened beam and a series of steel plates to distribute the load from hydraulic pistons over an 8 length of the joist. With four pistons on each side, a 32 uniformly distributed load was created on each joist. Load was applied directly to the top chord of the joists. Figure 4 illustrates the distributed loading pattern. Each of the hydraulic pistons had a capacity of 50 kips and a stroke of 14.25, allowing for very significant loading and displacement. A static load was applied by a 10,000 psi manual hydraulic pump. For each test, increasing load was applied to the floor system until joist load capacity was attained. Figure 4: Load Distribution Setup Data Acquisition Load, displacement, and strain data were recorded in order to characterize the performance of the joists. Over each joist, a 50 kip load cell was placed between one of the hydraulic pistons and the load pyramid to determine the load applied per cylinder. Displacement was measured through the use of a series of linear variable displacement transducers (LVDTs) placed at the midspan of each joist. Multiple strain gages were applied at the mid-span of each joist chord to provide an indication of the strains and stresses in each chord. A total of 12 strain gages were applied on each steel joist, and 18 gages were applied to the wood nailer joists. The signals from the load cells, LVDTs, and strain gages were all recorded by three Vishay Micro-Measurements scanners and the associated computer software. 2.3 WOOD NAILER MATERIAL TESTING The properties of wood are known to vary considerably. The modular ratio is critical when examining composite 2.4 DETERMINATION OF STEEL PROPERTIES Since the primary structural material in this study was steel, testing was performed to verify the actual strength of the steel in the chords of the joists. Steel coupons were made from each of the chords per ASTM E8-04 for tensile testing [5]. Each coupon was then tested in tension using an 810 Material Test System machine, which recorded load and displacement data until the coupon fractured. The elastic modulus, yield stress, and ultimate stress of the steel were determined from this data. 3 RESULTS 3.1 JOIST TESTING RESULTS Phase III-A and III-B Limit States Phase III-A consisted of Tests 1 and 2, which examined the first two sets of wood nailer (J1) joists [2]. As the test setup for Test 1 was loaded, one of the partially crimped interior web members began to buckle in compression at the transition between the crimped and uncrimped portion of the angle. The buckling led to the eventual fracture of the weld attaching the web member to the bottom chord. With the web member essentially ineffective, the top chord then buckled, as shown in Figure 5, to reach the ultimate failure point at a total uniform load (applied load plus dead weight) of k/ft. At the point of failure the bottom chord of the joist had not begun to yield. Test 2 also experienced a web member failure; the weld connecting a tension web member to the top chord fractured leading to buckling of the top chord at a total load of k/ft. Prior to the top chord buckling, yielding occurred in the bottom chords. Figure 5: Test 1 Web Member and Top Chord Failure

4 Phase III-B included Tests 3 and 4, the first two tests that looked at J2 joists, or standard steel joists designed to have a comparable strength to the wood nailer joists in Tests 1 and 2 [3]. Test 3 displayed similar results to Test 1 as one of the partially crimped compression web members buckled at the location between the crimped and uncrimped portion of the web angle. This again led to fracture of the weld securing the web member to the bottom chord followed by buckling of the top chord under at a total load of k/ft. Test 4 exhibited the identical failure mode as Test 3 at a total load of k/ft. In both tests the top chord buckling occurred prior to yielding in the bottom chords Phase III-C Limit States In order to compare the performance of the different joists it was necessary to assure that the joists reached the point of bottom chord yielding as evidence of true chord failure. Three of the four tests in the first two phases did not reach bottom chord yielding; consequently, the joists for Phase III-C were designed with stronger web members and additional welds to guard against preliminary web member failure. The Phase III-C tests consisted of Tests 5-10 and included both wood nailer joist and standard joist tests. Tests 5 and 6 consisted of the J3 joists, or standard steel joists fabricated with the same steel member sizes as the wood nailer joists. The tests displayed similar behavior, and bottom chord yielding was achieved. The joists were loaded and once the bottom chords began to yield the joist deflection accelerated. The tests were stopped when the system was deemed potentially unsafe at a deflection of over 7 inches. The tests were stopped at a total load of k/ft and k/ft for Tests 5 and 6 respectively. Tests 7 and 8 utilized the J2 joists that were tested in Tests 3 and 4 except with webs designed to prevent a web member compressive buckling. The performance was similar for the two tests, and yielding was achieved in the bottom chords for both tests. The joists were loaded and once the yielding began to occur, the deflection rate increased. The large deflection eventually resulted in a sudden buckling of the top chord, shown in Figure 6. The buckling occurred at a total load of k/ft for Test 7 and k/ft for Test 8. Tests 9 and 10 utilized J1 wood nailer joists. The joists displayed similar behavior to Tests 5 and 6; the tests were stopped when the stroke of the pistons was maximized. The yielding in the bottom chord allowed for the excessive deflection. The total loads at the point of maximum deflection were k/ft for Test 9 and k/ft for Test Load and Deflection Results The load and deflection data provided a means of comparing the performance of the different joists as well as determining the stiffness of the joists. Table 2 summarizes the maximum load and deflection for each joist in all ten tests. The applied load is the force that was actually measured by the load cell as being produced by the hydraulic piston distributed over the 8 length of the load pyramid. The total load accounts for the self-weight of the joist and test setup in addition to the applied load. The self weight of the joists and test setup were similar for each joist type, and, therefore, 54 lb/ft was used for all tests. The deflection values are the maximum displacement readings recorded at the center of each joist. Note that after the first test that maxed out the pistons the height of the test setup was adjusted to allow for additional stroke in subsequent testing. Table 2: Load and Deflection Results Phase Test III-A III-B III-C Applied Load Total Load Deflection / (k/ft) (k/ft) (in) The load-deflection plots illustrate the behavior of the joists as load is applied. The slope of the plots in the elastic region, or the stiffness, showed a good degree of consistency throughout all ten tests. The point at which the bottom chords began to yield is also evident for those tests in which yielding occurred. The long inelastic range of some of the plots indicated strain hardening in many cases. The plots for the joists that failed prior to chord yielding do not show an inelastic range. Figure 7 contains the load vs. deflection plot for the joist with the lower capacity of from each of the ten tests. Figure 6: Test 8 Top Chord Buckling

5 the stiffness based on the load and deflection data at 10% and 60% and then the stiffness was used with the known moment of inertia to determine the modulus of elasticity using Equation K E = (2) 3I Table 4 contains a statistical analysis of the calculated modulus of elasticity values for the 28 wood specimens. Table 4: Wood Modulus of Elasticity Test Results Figure 7: Load (lb) vs. Deflection Plots Stiffness Results The stiffness values for the joists shown in Table 3 were obtained using the load and deflection data. The slope of the elastic portion of the load-deflection plot is used per Equation 1 to determine the stiffness values. The load and deflection are examined at 20% and 60% of the ultimate load to eliminate the effects of any seating issues during the initial loading and to assure that the values being used were still within the elastic range. Table 3: Stiffness Results Phase Test / Stiffness Avg. Avg. (k/ft/in) (k/ft/in) (k/ft/in) III-A III-B III-C K Load 60% 10% = (1) Deflection 60% Load Deflection 3.2 WOOD NAILER TESTING RESULTS The four point bending test performed on the wood nailer top chord attachment samples provided load and the corresponding deflection data for each test. The modulus of elasticity was calculated by first computing 10% E (ksi) Mean 1,604 Minimum 1,096 Maximum 2,147 Median 1,609 Std. Dev STEEL TESTING RESULTS The results of the tensile tests were determined by dividing the load at which the coupon yielded or fractured by the measured cross sectional area at the center of the coupon. All of the steel coupons tested produced yield and ultimate strengths exceeding those required for A572 (50 ksi yield) steel. Table 5 presents the statistical analysis of the 48 coupons that were tested. Table 5: Steel Coupon Test Results Yield (ksi) Ultimate (ksi) Mean Minimum Maximum Median Std. Dev DISCUSSION 4.1 FAILURES All of the tests provided data for use in comparing the performance of the different joists; however, Tests 1 4 did not reach the point of bottom chord yielding, meaning that the true capacity of the chords was not achieved. The joist stiffness data for these tests were valid; however, the capacities were considered to be lower bound values. The modifications made to the joist designs following the first two phases assured the desired primary limit state of bottom chord yielding in Phase III-C. 4.2 LOAD CAPACITY The ultimate applied load was recorded either at the point of collapse or the when the test was stopped due to potential instability in the system. Table 6 provides an analysis for the load results for the joist that held the lesser amount of load. The design load takes the output load from the joist design software for the 40 joist and converts the load to a 32 uniform load centered on the joist and normalizes it such that the bottom chord

6 stresses are The joist capacity load is the ultimate applied load plus the self-weight of the joist. The factor of safety is the failure load divided by the design load. Note that the minimum and target factor of safety for a joist is 1.67 [6]. For the tests in which web member failures occurred prior to bottom chord yielding, the safety factors are conservative. The maximum moment is calculated using the uniform failure load over the center 32 of the joist. Table 6: Loading Analysis Phase Test Design Load Failure Load Max Moment Stiffness FS (k/ft) (k/ft) (k-ft) (k/ft/in) III-A III-B 1 (J1) 3 (J2) (J1) 4 (J2) (J3) (J3) III-C 7 (J2) (J2) (J1) (J1) Note: For each test the joist with lower capacity was chosen + Because of web member failure this is a minimum value The factors of safety ranged from 1.70 to 2.43; all joists exceeded the required 1.67 safety factor. The J1 joists had the highest factors of safety out of the three types of joists, but the factor of safety values were not unreasonably high. The performance of the J2 and J3 joists were relatively similar in terms of factors of safety. The strengths exceeded the desired factor of safety values, but were also not excessively high. The maximum moments directly correspond to the failure loads; therefore, the J1 joists had the highest moment values. From the strength analysis it can be concluded that the joists were all adequately designed for strength because the factors of safety exceeded 1.67 by a reasonable amount. The strength of the wood nailer joists may be slightly more conservative, but the stiffness and chord force analyses must also be considered before that judgment is made. 4.3 STIFFNESS The stiffness results from Table 3 show that the wood nailer joist exhibits similar or greater stiffness than the J2 joists that are designed with more steel in the chords and the J3 joists that have the same amount of steel, but no wood nailer top chord attachment. The stiffness results of the J1 joists in Phase III-A are nearly identical to those of the J2 joists in Phase III-B. The J1 joists display a greater stiffness than the J2 joists in Phase III- C because the J1 joists had a slightly heavier top chord than the design called for due to a material substitution at the fabrication facility. Had the J1 joists in Phase III-C had the same chord sizes of those in Phase III-A comparable stiffness values would be expected. Therefore, it is evident that the wood nailer contributes to the overall stiffness of the joist so that the steel savings in the wood nailer joist will not adversely affect the stiffness compared to that of the standard joist. 5 CONCLUSIONS The experimental program has clearly demonstrated that the wood nailer joists exhibit an equal or greater load capacity as well as a comparable stiffness despite having a reduction in the amount of steel in the chord angles. This behavior proves that some composite behavior exists between the steel top chords and the wood nailer top chord attachment. The outcomes from this test program can be refined to develop the design methodology for wood nailer joists to take advantage of partial composite behavior in the top chords and realize the steel savings that can result. Additional details on the experimental testing, results, and conclusions are available in Martignetti, 2008 [7]. ACKNOWLEDGEMENT The authors are grateful to CMC Joist & Deck for providing the financial support of this research, and Joe Pote for his technical guidance. Jeffery Cook and Christopher Gudas, former graduate students at Villanova University, are acknowledged for their contributions to Phases III-A and III-B of this project. REFERENCES [1] CMC Joist & Deck [2] Cook, J. J Behavior of Open Web Steel Joists with Top Chord Attachments. Master of Science Thesis, Villanova University, Villanova, PA. [3] Gudas, C. J Full-Scale Tests of Open Web Steel Joists With and Without Top Chord Attachments. Master of Science Thesis, Villanova University, Villanova, PA. [4] American Society for Testing and Materials (ASTM) Standard Test Methods for Mechanical Properties of Lumber and Wood-Base Structural Material. ASTM D4761, ASTM International, West Conshohocken, PA. [5] American Society for Testing and Materials (ASTM) Standard Test Methods for Tension Testing of Metallic Materials. ASTM E8-04, ASTM International, West Conshohocken, PA. [6] Steel Joist Institute Standard Specifications, Load Tables, and Weight Tables for Steel Joists and Joist Girders. 42 nd Ed., Steel Joist Institute, Myrtle Beach, SC. [7] Martignetti, M. J Strength and Stiffness of Open-Web Steel Joists With and Without Wood Nailer Attachments, Master of Science Thesis, Villanova University, Villanova, PA.