Mobile Edition
Print
Get the Mag
Weekly UpdatesAbstract
To improve our understanding of localized density effects in wood-based panels on the holding capacities of fasteners commonly used in furniture, a comprehensive study was conducted using static and cyclic tests of withdrawal and head pull-through of screws and staples and lateral resistance of screws in oriented strandboard (OSB), medium density fiberboard (MDF), and particleboard. In this paper, results of cyclic tests are presented and comparisons are made with the static test results reported in Part 1. Similarly to static tests, cyclic test data indicated that density variation in OSB panels had a significant effect on screw withdrawal, head pull-through, and lateral resistances, but the effects were less evident with staple withdrawal and head pull-through. For particleboard, density variation had a significant effect on screw and staple face withdrawal and head pull-through resistances, but the effects were less pronounced for screw and staple edge withdrawal and screw lateral resistances. For MDF, no significant correlations were found, which was likely due to the low density variation in these panels. The data from this study will be useful to the panel industry and furniture manufacturers for optimizing use of panel products and fasteners in furniture frames.
**********
To reduce the cost of framing components, upholstered furniture manufacturers are always interested in alternative materials that are less expensive, but as strong and reliable as the traditionally used materials. With the development of computer numerical control technology, composite panels, such as oriented strandboard (OSB), medium density fiberboard (MDF), and particleboard, potentially can be used as replacements for solid wood in upholstered furniture frames. But, the suitability and performance of traditional fasteners used in panel products for such applications has not been well studied. One of the concerns is density variation and non-uniformity across the thickness and in the plane of the panel, and how it could affect the holding capacity of fasteners. Available reference values are based on static tests; only limited information is found about the density effects on fastener holding capacities in wood-based panels under cyclic loading conditions.
In-service upholstered furniture frames are subjected to a wide range of loads, which act as repetitive events of loading and unloading. Typically, in the furniture industry, long-term fatigue loading is studied using a large number of non-reversed loading cycles (25,000 cycles on each load level depending on the performance acceptance level) with an average rate of 20 cycles per minute (GSA 1998). This performance test regime is based on a zero-to-maximum (one-sided or non-reversed) cyclic stepped fatigue load method rather than a static or constant amplitude cycling load method (Eckelman 1988). For example, Zhang et al. (2006) studied bending fatigue life of metal-plate-connected (MPC) joints in furniture-grade pine plywood by subjecting the joints to one-sided stepped cyclic bending loads.
Several studies were conducted on the performance of wood or wood-based material assemblies under cyclic or fatigue tests using different test protocols. For example, reversed and non-reversed cyclic loading (Hayashi et al. 1980) were used to evaluate the fatigue properties of wood butt joints with metal-plate connectors in timber. De Melo Mouria et al. (1995) used a non-reversed cyclic load schedule (varying tension) followed by a sinusoidal function to examine the influence of wood density on the mechanical behavior of MPC joints.
This study is part of a broader research program to examine the localized density effects on fastener holding capacities in wood-based panels. This paper presents cyclic test data and complements the static test results reported by Wang et al. (2007). The key objective of this study was to evaluate the holding capacity of screws and staples in commercial wood-based panels under cyclic loading in comparison with static loading. This paper also discusses the correlation between fastener holding capacity and density distribution in panels. Technical information generated in this study will be used to provide recommendations to the panel industry on the use of fasteners in their products and how to optimize the construction of furniture frames.
Materials and methods
The specimens for cyclic tests were prepared using the same materials and mapping and cutting techniques as described in the previous paper (Wang et al. 2007). The following panels were used in the study:
1. MDF: 16 mm thick, grade 150,
2. Particleboard: 16 mm thick, grades M2 and MS (two of each),
3. OSB: 11 mm (7/16 in.) thick,
4. OSB: 15 mm (19/32 in.) thick, and
5. OSB: 18 mm (23/32 in.) thick.
The OSB panels were made of aspen and were grade O2. In each of these five categories, four full-sized (1.22 by 2.44 m) panels, marked A, B, C, and D, were chosen for specific tests, so that one half of the specimens were cut for static tests and the other half were cut from the same panel for the matching cyclic tests. Table 1 provides information on the types and number of cyclic tests performed. Each specimen was used for two tests. Fasteners were driven into the specimens following the same techniques as described in Part 1 (Wang et al. 2007).
The tests were conducted in accordance with ASTM D1037 (ASTM 2005a) and ASTM D 1761 (ASTM 2005b) standards with the following modifications. The test set-up used for the screw lateral resistance was modified by adding a screw supporting device on each side, as shown in Figure 1, to allow for cyclic loading without slack. Load rate was adjusted to produce 15 cycles per minute under load-controlled conditions at a uniform rate of loading at every step. Ultimate loads ([P.sub.ult]) determined from static monotonic tests were used to calculate the reference load levels ([P.sub.ref]) needed for cyclic stepped loading (Tables 2 through 5) Cyclic loading was applied in three steps with 30 cycles at each load level: 15 percent, 35 percent, and 70 percent [P.sub.ult], after which the specimens were loaded to failure (Fig. 2). Note that for the staple edge withdrawal tests, load levels in the first step were 30 percent [P.sub.ult] and 25 percent [P.sub.ult] for parallel and perpendicular orientations, respectively (Table 3). A preload of 40 N was applied to eliminate slack in the system during cycling.
The cyclic load regime used in this study is referred to as short-term cyclic to distinguish it from a typical fatigue loading which is usually conducted with a large number of cycles until failure occurs. Due to the amount of time needed to perform a single test following the General Service Administration (GSA 1998) procedure, it was decided to conduct the short-term cyclic test that lasts approximately 6 minutes. This allowed for testing a sufficient number of specimens from various panel types and thicknesses.
An analysis of variance (ANOVA) general linear model procedure was performed on individual fastener holding capacities for all types of panels to examine the correlation between localized density and ultimate holding capacity. In order to classify the holding capacities of the fasteners in the panels, Duncan's multiple tests were performed on the average values.
Results and discussion
An example of typical load-displacement curves of the cyclic and corresponding static tests of screw head pull-through are shown in Figure 3. In this study, ultimate fastener holding capacity has been used as an indicator of connection resistance. Other parameters associated with the load--displacement relationship, however, could be established (e.g., initial stiffness and the slip of the fastener in the panel at a certain load level or at failure). Summarized test results including classified average resistance values for each type of panel are presented in Tables 6 and 7 and Figures 4 through 7. Comparisons between static and cyclic test data are shown in Table 8.
[FIGURE 1 OMITTED]
Withdrawal resistance of screws
Figure 4 shows the average withdrawal resistance of screws for each tested panel. With the exception of the 11-mm and 18-mm OSB panels, face withdrawal resistance was significantly higher than that from the edge. There were no significant differences between the edge withdrawal resistances parallel and perpendicular to the long axis of the panel for all of the panels tested. The average face withdrawal resistance was highest with the 15-mm OSB panel and lowest for the 11-mm OSB panel. MDF panels showed resistances similar to the 15-mm OSB panel. Particleboard panels performed similarly to the 18-mm OSB panels. In edge withdrawal, the 11-mm OSB and 16-mm particleboard panels showed the lowest resistances, which can be explained by the low core density of particleboard and the small thickness of OSB. Also, some specimens were split during the insertion of screws prior to testing. Edge withdrawal resistance in the 15-mm and 18-mm OSB panels were similar to that of the 16-mm MDF panel.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
ANOVA was performed to verify if a relationship exists between screw withdrawal resistance and localized density of the panel. Results indicated that for cyclic face withdrawal in all of the OSB and particleboard panels the relationship was significant at the 95 percent confidence level. The r-values of the linear regression model ranged from 0.47 to 0.83 (Table 6). A poor relationship, however, was observed for MDF panels. This could be attributed to the uniformity in the MDF localized density in comparison to OSB or particleboard panels. For edge withdrawal of screws under cyclic load, the relationship was found to be significant for all of the OSB panel specimens perpendicular to the long axis of the panel, while in the parallel direction, the relationship was only significant in the 11-mm and 18-mm OSB panels (Table 6). Relationships were poor for both MDF and particleboard panels.
FEED