EMPIRICAL THRESHOLD VALUES OF ROOF CONNECTIONS USING PULL- OUT TEST ON NAILS AND SCREWS

Transcription

1 EMPIRICAL THRESHOLD VALUES OF ROOF CONNECTIONS USING PULL- OUT TEST ON NAILS AND SCREWS Harvey O. BISA, Engr. Romeo Eliezer U. LONGALONG, Engr. Raniel M. SUIZA and Engr. Christian R. OROZCO ¹ University of the Philippines Diliman, Quezon City, Philippines Abstract: Philippines is a pathway of typhoons. Strong typhoons inflict damages in residential house especially on its building envelope. Due to negative pressures produced by the strong winds, the roofing materials are being pulled away and the only type of connection that holds the roofing is its fasteners. The study deals with the determination of threshold values of these connections in Philippine context. These threshold values are important on simulations involving wind loads. In addition, disaster risk management relies on experimental data to predict how large the damage would be on single or group structures. The determination of fastener pullout resistance (FPR) in static loading and cyclic loading tests were done. Combination of roofing, fasteners and purlins were made to determine the governing modes of failure for the roof connections. For single fastener test static loading, wood nail, metal screw and wood screw had FPR of 0.89 kn, 0.87 kn and 5.6 kn respectively. The decrease in FPR from the static loads to cyclic loads in the cycles test for wood nail, metal screw and wood screw was 2.8%, 9.4% and 23% respectively. When the type of connection is nail, the nail is pulled out without experiencing tearing of GI sheet. For metal screw, mostly tearing occurs rather than pullout. If the fastener in the connection is wood screw, tearing will always be the mode of failure for the connection. The maximum loads are 1.43 kn, 1.26, kn and 1.37 kn for wood nail, metal screw and wood screw respectively. Key words : Pullout, Static Loading, Cyclic Loading, Roof Connections 1 INTRODUCTION 1.1 Background Structural roof systems are an important part of a building. It protects the interior parts of the building from heat, rain, wind, humidity and others. One of the elements on the environment that is enough to damage the structural roof system is the wind loads. Wind loads during typhoons apply suction in the roof such that the negative pressure is enough to pull out one or more parts of the system. The Philippines is a typical pathway for typhoons. It experiences 20 or more typhoons annually with varying winds speeds. Residential buildings are the most commonly constructed structures here in the Philippines. It is frequently observed that low rise residential buildings suffer largely during tropical cyclones Li (2005). The most vulnerable part of a residential building is its envelope. It includes the roof, walls, doors and windows. Failure of one these components may lead to the decrease of the structural integrity of the building (Veron, 2012). The common failures that roofing experience are pullout of the connections on roof-to-wall, tearing of the roofing material, and pullout of the fasteners connecting the roof to the purlin. Researches were conducted for roof-to-wall connections (Shanmugama, et al. 2009), tearing of the roof (Henderson, 2011) and pullout of fasteners in a roofing field (Baskaran, et al. 2008). The results may be different for the structures in the Philippines since researches were made in other countries. Currently, the threshold values used for the uplift resistances were based on foreign researches and not based on the conditions here in the Philippines. Thus, it is important to determine these values in the Philippine context. 1.2 Statement of the Problem Roofing materials are always subjected to negative pressure during typhoons which cause uplift to the fasteners. Therefore, this research will focus on developing uplift threshold values in the Philippines that will be used in developing fragility curves of structures subjected to wind loading. The study specifically involved materials commonly used in roofs in the Philippines. The research also focused on the difference on static and cyclic loading and behavior of multiple fasteners during uplift.

2 1.3 Objectives The objective of the study is to conduct pullout test on the commonly used fasteners in the roof-to-purlin connections to observe the behavior of the components of the roof systems under pullout and to determine uplift threshold values for the resistance of the fasteners. Another objective is to conduct two types of tests: static and cyclic loading to determine the decrease in resistance of the fastener to pullout Significance of the Study Threshold values are important since they are used in preliminary design or simulations of structures. Simulations use threshold values in order to predict possible failures in the building envelope, which are then used to develop fragility curves. Uplift threshold values vary depending on the type of materials and their properties. The developed threshold values from this study are suitable based to the Philippine setting. Another importance of the research is to compare the decrease on fastener pullout resistance (FPR). Lastly, the design for the fastener spacing in the National Structural Code of the Philippines 2010 (NSCP) did not state what kind of fastener to be used. Different fasteners have different FPR therefore fastener spacing should vary depending on the type of fastener. After obtaining values of FPR, the research would be able to recommend design fastener spacing for roofing systems based on empirical data. 1.5 Scope and Limitations The study only focuses on the nail and screw resistances during pullout. Tearing of the roof cladding during pullout was not considered. Wood panel sheathing and metal cladding were not used during pullout. Other factors that were not considered include angle of penetration, depth of penetration, moisture in the system, and variation of material to be used. The factors not considered may be included in succeeding researches. The common types of fasteners such as wood nail, wood screw and metal screw were used. 1.6 Conceptual Framework Quantitative risk assessment methodology involves computations based on the magnitude and probability that loss will occur. One of the factors that is used in risk assessment is vulnerability which is quantified by fragility or vulnerability curves. Fragility curves, in view of severe wind loadings, show the relationship of the probability of occurrence of each damage state as a function of wind speed. Damage states vary with different building types (e.g. residential buildings, industrial buildings, etc.) and different hazards (e.g. wind speed for tropical cyclones, spectral acceleration for earthquakes) (Veron, 2012). For the development of fragility curves, simulations were done in computer applications such as ANSYS. Before running simulations, parameters should be specified such as the threshold values for uplift resistance of nails. The values of these parameters are critical since they will change the resulting fragility/vulnerability depending on the assumptions. Figure 1 Conceptual Framework

3 2. METHODOLOGY 2.1 Test Types For the study, there were 5 types of tests conducted. The first two were static and cyclic loading of single fasteners (Single Fastener Test). The purpose of these two types of tests is to determine whether there is significant decrease in fastener pullout resistance (FPR) of the fasteners if subjected to series of progressive cycles. The third type of test, Static loading for GI Roof (Roofing Test), was done to determine the strengths of GI roofing when upward pressure is applied on it. After the third test, it was compared to the results of the static loading of single fasteners in order to determine what kind of failure will govern depending on the maximum pull out resistances. The fourth test, which was a combination of first and third, was performed to determine whether increase in performance exists when the strengths of roofing and purlin were combined. Lastly, the fifth test, the multiple fastener test was performed to observe the behavior of group of fasteners during pullout. Of particular importance in this test is the sequence of pullout and the maximum resistance the group can hold. Table 1. Loading types and corresponding test types for pullout Loading Type Static Loading Cyclic Loading Test Types Fastener pullout resistance (FPR) Roofing Test Combination Test Multiple Fastener Test Fastener pullout resistance (FPR) 2.2 Test Specimen Fasteners Properties Three types of fastener were used in the experiment. The fasteners used were wood nail, metal screw and wood screw. Below are the dimensions of the fasteners that were used. Table 2. Specifications of fasteners used Inner Diameter Outer Diameter Head Diameter Threaded Total Length Length Wood Nail N/A N/A 20 mm N/A 75 mm Metal Screw 2.5 mm 4 mm 12 mm 35 mm 80 mm Wood Screw 3 mm 5 mm 13 mm 60 mm 85 mm Purlin Properties The purlin sample dimensions are 2 x 3 x 6 in per sample. Wood S4S and galvanized iron c channel were the types of material used. The thickness of the C Channel is 1 mm Roof Properties The roofing used in the experiment was gauge 24 galvanized iron with a size of 200 x 200 mm. The thickness of the roof is 0.76 mm. The fastener is placed on the crest of the wave of the roof 2.3 Design and Test Set Up Since the Universal Testing Machine's wedge grips are not capable of holding the specimen, attachments were designed such that they are capable and flexible enough to conduct different types of test. Figures shows the experimental setups for the different tests performed.

4 Figure 2 Design and setup of single fastener test (Static loading) Figure 3 Design and setup for single fastener test (Cyclic loading) Figure 4 Design and setup for roofing test Figure 5 Combination of roofing, fastener, purlin test set-up

5 2.4 Testing Process and Parameters Since there are five types of tests, each of the tests needs specifications for loading that will be set in the Bluehill software. Therefore the researcher made some protocols for testing based on ASTM D , ASTM D7332/D7332M 09 and modified CSA Loading Cycles from (Baskaran et al, 2012) Static Loading Parameters For static loading, the loading rate is 2 mm/min. The movement of the crossbeam is constant until the fastener reached total pull-out. Figure 6 Static loading rate Cyclic Loading Sequence For cyclic loading, the crossbeam is displaced up and down with increasing amplitude. The loading rate for the crossbeam is 75 mm/min. For certain amplitude, there are 200 cycles applied on the specimen. The sequence is similar to Baskaran but the amplitude increases by 0.5 mm after every 200 cycles to determine the behavior of the load displacement curve at a certain loading sequence.. Figure 7 Sample cyclic loading sequence for metal screw 3. RESULTS AND DISCUSSION 3.1 Static Loading for Fasteners Three types of fasteners were used and tested until pullout. During the test, the load displacement curve of the crossbeam was automatically generated by Bluehill 3. It was observed that the wood nail after pull out test did not tear the fibers the purlin. The fibers of the wood remained intact since the texture of the nail is smooth.

6 Figure 8 Force vs. displacement of wood nail sample (Static loading) The force vs. displacement graph of wood nail (Figure 8) has a spike where the maximum load is located. These spike, which is due to the slipping of the nail, are similar to the results discussed by (Rammer, 2010) for smooth surfaced fasteners. Table 3 Maximum Loads per Fastener during Static Loading Sample number FPR (Newtons) Nail Screw Metal Average SD Based on the table, the FPR of wood nail and metal screw are identical with values of 0.89 kn and 0.87 kn respectively. On the other hand, wood screw is 5 times stronger than wood nail and metal screw. It should be noted, however, that the presence of moisture and corrosion were not considered and may affect the FPR in actual conditions. 3.2 Cyclic Loading for Fasteners In cyclic loading tests, the fastener was displaced up and down due to the movement of the crossbeam. As the crossbeam moves, it makes a cycle which is repeated for many times. Cycle is the completed up and down movement of the crossbeam from two lowest points. Since the loading cycle used has increasing amplitude, every sequence produce cycles which are changing in sizes. Every sample takes about 2 hours since cycles were made. To save time, static loading was then applied when the fastener displacement is more than 10 mm. This is because most of the peak loads occur from 4-6 mm extension. Figure 9,10 and 11 shows the load displacement of the wood nail, metal screw and wood screw specimen under cyclic loading respectively.

7 Displacement: 3 mm Figure 9 Force vs. displacement of Wood nail specimen (cyclic loading) Figure 10 Force vs. displacement of metal screw (Cyclic loading) Figure 11 Wood screw specimen after cyclic loading test Table 4 summarizes all the maximum loads obtained in 17 samples. The average FPR of wood nail and metal screw are 0.86 kn and 0.76 kn, respectively. On the other hand, the FPR of the wood screw is about 4 times stronger than the two. Comparing it with the FPR from static loading, Table 4 shows that there is a decrease in fastener resistances: 2.9%, 9.4%, and 23% for wood nail, metal screw and wood screw, respectively.

8 Table 4 Maximum loads per fastener during cyclic loading Sample Number FPR (Newtons) Nail Metal Screw Average SD Table 5 Comparison of average FPR (Static and cyclic Loading) per fastener type Static (N) Cyclic (N) Decrease (%) Wood Nail Metal Screw Wood Screw Roofing Test The roofing test was performed to determine the maximum load a roofing field can hold before tearing occurs. From the load displacement curve (Figure 12), the graph has reached a maximum load of approximately 2.4 kn. Based on the graph, the formed curve has some bulges which are possibly due to the creasing of the GI sheet. Figure 12 Wood nail + GI sheet specimen after roofing test In summary, wood nail with GI sheet has the largest maximum load while metal screw and wood screw are identical. With a difference of 60 N in average maximum loads, the metal screw and wood screw can be considered close while wood nail has larger maximum load with an average 1.96 kn. Wood nail has a larger maximum load because the head has a larger area. This area allowed more capacity for the roof to resist force compared with the screw with small heads. Sample Number Table 6 Maximum loads per fastener during roofing test Fastener Types in Roofing Test (Newtons) Wood Nail Metal Screw Wood Screw Average

9 3.4 Combination Test After the roofing test, combination of roofing, fastener and purlin test was done. This is to check whether the there is increase in strength when the materials were combined. Table 7 Expected governing mode of failure for combination tests Fastener type used in combination test Expected governing mode of failure Wood Nail Pullout Metal Screw Pullout Wood Screw Tearing Table 8 Maximum loads per combination type Combination Types Sample Number Roofing + Wood Nail + Purlin Roofing + Wood Nail + Purlin Roofing + Wood Nail + Purlin kn (P) kn (P) kn (T) kn (P) kn (T) kn (T) kn (P) kn (T) kn (T) kn (T) Average (kn) SD (kn) P: Pull-out T: Tearing Figure 13 Comparison of average maximum loads on fastener, roofing and combination test From Fig. 13, the results in combination test showed that there is an increase in fastener resistance. This pattern will lead us to conclusion that there is an increase in fastener resistance when all materials are combined in one system. In contrary, one of the results differed from the hypothesis made. In the combination test, failure in wood nail (pullout) and wood screw (tearing) tallied with the hypothesis while metal screw experienced 1 pullout and 3 tearing failures. The possible explanation for this is that the tearing and the pullout have no substantial difference in values. If there is an increase in strength due to the combination of roofing strength and fastener resistance, then the maximum loads for the both pullout and tearing will be almost the same. If they are almost the same, then there is uncertainty in the mode of failure. Only 4 samples were made, therefore, it is recommended for future studies to conduct more samples. 4 CONCLUSIONS After the series of tests conducted in different set up of roof connections, threshold values were obtained. For single fastener test static loading, wood nail, metal screw and wood screw had FPR of 0.89 kn, 0.87 kn and 5.6 kn respectively. The percentages of the standard deviation for the static loading range from 10-20%. The decrease in FPR from the static loads to cyclic loads in the cycles test for wood nail, metal screw and wood screw was 2.8%, 9.4% and 23% respectively. The percentage of standard deviation of FPR in cyclic loading ranges from 4-13%. Therefore, based on the

10 changes in FPR, it could be concluded that cyclic loading gives significant decrease in FPR. This study points out that the decrease in FPR was based on testing under controlled condition and do not consider factors such as corrosion, water immersion, and realistic wind pressures. When the type of connection is wood nail, the nail is pulled out without experiencing tearing of GI sheet. The explanation for this is because the FPR is lower than the capacity of the roof before tearing. Before the load reached the maximum load for tearing, the nail will be the first to be pulled out. For metal screw, mostly tearing occurs rather than pull-out. The reason for this is due to the fact that the FPR for metal screw is close to the capacity of roofing against tearing. If the fastener in the connection is wood screw, tearing will always be the mode of failure for the connection because the FPR of the wood screw is stronger than the capacity of roofing against tearing. The maximum loads are 1.43 kn, 1.26, kn and 1.37 kn for wood nail, metal screw and wood screw respectively. ACKNOWLEDGMENT I would like to acknowledge and thank the following for making this research possible: Mr. Romeo Longalong, Dr. Jaime Hernandez Jr., Mr. Raniel Suiza, Mr. Christian Orozco, Dr. Benito Pacheco, Dr. Fernando Germar, Dr. Tingatinga, Ms. Liezl Tan and the rest of the structural Engineering Group, for their comments and suggestions. To my SEG batchmates, To my parents, Wilfredo and Eliaquim Bisa, for their support. To my friends, churchmates and CE batchmates for constant support and ideas. REFERENCES ASEP (1961), National Structural Code of the Philippines 2010, Volume I, Fifth Edition, ASEP, Panay Avenue, Quezon City, Philippines. ASTM International (2009), Standard Test Method for Measuring the Fastener Pull-Through Resistance of a Fiber-Reinforced Polymer Matrix Composite, D7332/D7332M 09, ASTM International, 100 Barr Harbour Drive, West Conshohocken, Pennsylvania , United States. ASTM International (2009), Standard Test Methods for Mechanical Fasteners in Wood, D , ASTM International, 100 Barr Harbour Drive, West Conshohocken, Pennsylvania , United States. Baskaran A, Ham H and Lei W, (2006), New Design Procedure for Wind Uplift Resistance of Architectural Metal Roofing Systems, Journal of Architectural Engineering, pp , Institute for Research in Construction, National Research Council of Canada, Ottawa, ON, Canada Baskaran A& Om, D 1997, Performance of Roof fasteners Under Simulated Loading Conditions, Journal of Wind Engineering and Industrial Aerodynamics Vol. 72, pp , viewed 13 July Baskaran B, Ko S and Molleti S, (2008), A novel approach to estimate the wind uplift resistance of roofing systems, Building and Environment Vol. 44, pp , Institute for Research in Construction, National Research Council of Canada, Ottawa, ON, Canada Henderson, D and Ginger, J (2011), Response of Pierce Fixed Corrugated Steel Roofing Systems Subjected to Wind Loads, Engineering Structures Vol. 33, pp Elsevier, viewed 13 July Henderson, D, Ginger, J, Sumant, B and Leitch C, (2011), Timber Connection Strength Characterisation, Report TS815, Cyclone Testing Station, School of Engineering, James Cook University, Townsville, Queensland. Hill K., Datin P., Prevatt D., Gurley K. and Kopp G., (2009) A Case for Standardized Dynamic Wind Uplift Pressure Test for Wood Roof Structural Systems, ACWE, San Juan, Puerto Rico. Instron (2009), 5980 Series Dual Column Floor Frames Manual, Revision B, Illinois Tool Works Inc., Norwood, USA Mahendran, M. and Mahaarachchi, M., (2008), A strain criterion to Pull Through Failures in Crest Fixed Steel Claddings, Engineering Structures Vol. 31. pp , School of Urban Development, Faculty of Built Environment and Engineering, Queensland University of Technology, Brisbane Australia. Veron, MC (2012), Development of Computational Fragility Curves for Residential Buildings Considering Severe Wind Loadings, Undergraduate Thesis, University of the Philippines, Diliman, Quezon City. Shanmugama B, Nielson B and Prevatt D (2009), Statistical and analytical models for roof components in existing lightframed wood structures, Engineering Structures, vol.31, pp