Mobile Edition
Print
Get the Mag
Weekly UpdatesThe final test on the small specimens was an evaluation of the density variation across the channel. The 16-in by 16-in corrugated panels were produced from a flat mat that must elongate to assume the shape of the corrugated dies. Hence, density variation across the corrugations may occur. Four corrugated panels were cut into 17 strips parallel to the corrugations to determine the lateral density profiles. All strips were approximately 1 inch wide, except for the first and the last strips which were 0.5 inch wide. Each strip was weighed, and the volume was determined from the measured dimensions.
Small panel test results and evaluation
The mechanical testing results are summarized in Table 1. The flexural response of the corrugated panel is affected by plate action, also present in flat panels. In addition, there is significant shear deformation in the side walls of the corrugations. The bending stresses in the horizontal elements of the section are not constant across the width of the element. They are highest near the junction with the side walls and lowest midway between side walls, due to a phenomenon known as shear lag. It is convenient to express flexural behavior in terms of beam theory. The conventional expression for the deflection of a simple span beam loaded at the mid point is
P/D = [([L.sup.3] / 48EI + L / 4G[A.sub.s]).sup.-1] [1]
where P applied load at midspan (lbs), D = midspan deflection (in), L = span (in), E = modulus of elasticity (MOE) (psi), I = moment of inertia (in4), G = shear modulus (psi), [A.sub.s] = shear area ([in.sup.2]). The shear area is approximately the cross-sectional area divided by 2.2 for the Type A and B sections and by 1.75 for the Type C section (Pang, 2005).
To correct for plate action, shear deformation and shear lag, a parameter study was done using a detailed finite element model of the test configuration. The results of the analyses were fitted with a multiple linear regression to produce a modified expression
P/D = ([L.sup.3] / 45.3EI + L / 0.79G[A.sub.s]).sup.-1] [2]
For the panel geometry of Figure 1, this equation gives P/D values accurate to within 5 percent over a wide range of E/G ratios and spans from 14 to 48 inches. The average value of E/G for the Type A specimens was found to be 6.5. Details of the methodology leading to Equation [2] and the E/G ratio can be found in Pang (2005). The change in the flexural coefficient is only about 6 percent. However, the change in the shear coefficient indicates that shear deformation effects are about five times larger in this particular panel geometry than predicted by Equation [1]. The modulus of rupture (MOR) values are nominal stresses, calculated as the ratio of maximum moment to the section modulus of the panel.
There is very little difference in the primary stiffness and strength for the Type A (random mat) and Type B (OSB aligned mat). The average nominal El of the Type B specimen is only 4.4 percent higher than the El of Type A. The average nominal bending strength, S[F.sub.b], for Type B is only 5.5 percent higher than that of Type A. This is a consequence of the corrugated shape. In a flat panel, alignment concentrates stiffness and strength in the extreme fibers and the strain gradient through the depth is relatively steep. In a corrugated panel, the horizontal elements of the section are completely in tension or compression, largely negating the benefits of a three-layered alignment. So there is no significant advantage in using an OSB aligned mat. For primary stiffness and strength, unidirectional alignment might be beneficial. However, experience has shown that such a mat tends to tear badly during moulding. Even if the panel could be successfully molded, its weakened secondary properties would make it more fragile and potentially unsafe during construction activities, prior to underlayment installation.
The secondary direction bending properties are shown in Table 2. Bending stiffness is based on an equivalent straight span derived from a frame model for the specimen shown in Figure 3. Shear deformation effects were ignored. The values in Table 2 are below the listed baseline secondary stiffness of 3600 pounds-[in.sup.2]/ft and the baseline secondary moment capacity of 930 pounds-in/ft for 24/0 sheathing (APA 2004). However, the corrugated panel is intended to be used only in the primary direction. Stiffness and strength in the secondary direction are primarily factors during handling of the panel prior to installation.
The shear, bearing, and edge loading strengths of the Type A panels are shown in Table 3. The failures of all the shear specimens appeared to be flexural. Even though results must be considered a lower bound on shear strength, the fifth percentile value corresponds to an ultimate uniform loading of 600 psf on a 48-in span. Thus shear strength of the section poses no practical limit on capacity. Bearing strength is difficult to interpret. Much like the case for lumber, there is no well-defined failure. Assuming a deformation limit of 0.02 inch from short-term testing as the basis for a bearing limit leads to a bearing value of 225 pound/ft. This is adequate to transmit floor loads at any practical joist spacing and to support non-load-bearing walls, but indicates a need for blocking if the panel is layered between load-bearing walls. The edge loading results show the expected advantage of placing the free edge down. Interpreting these results is also difficult. First, it would be very difficult (and hazardous) for a person to stand on the free edge in the lower position, because it is so narrow. So the 520 pound capacity may have little practical meaning. It is possible to stand on the free edge in the upper position. The 385 pound mean strength in that orientation is conservative in that the panel would almost always be longer than the 16-in test specimen. However, trying to stand on such an edge must be avoided, since if the panel is not yet fastened to the joists the panel will tip.
The final small-specimen test was for density variation across the channels. Prior studies (Sandberg et a1.1989, Haataja et a1.1991) have shown that mat movements during the moulding process can induce significant density differences in the various parts of the cross section. This can be controlled by special mat forming techniques and/or varying the final thickness of the parts of the section. However, the goal in this study was to produce practical corrugated panels without special forming or die adjustments. The transverse density distributions for four Type A panels are shown in Figure 5. The plotted densities were calculated at EMC, not oven-dry conditions. Some of the variations are due to the natural variations in density that occur during mat formation, especially in the relatively shallow mat for a 0.375 in thick panel. The lack of left to right symmetry may also be a consequence of hand forming a small mat. Most important is what is not evident. There is no systematic pattern of density variation across the panel. This is an indication that little, if any, mat tearing occurred during moulding.
Full-sized panel production
To produce full-sized, 4-ft by 8-ft, panels, a set of 58-in by 108-in dies were fabricated from 1.5-inch aluminum plate. The dies were designed to produce the same basic geometry as in the 16-in panels. These were mounted on the electrically heated platens of the press at Michigan Technological University's School of Forest Resources and Environmental Science.
Due to the large amount of material required, aspen strands were purchased from nearby GFI Strandwood Corporation. These were very similar to those used in the small specimen study and averaged 3.5 in long, 0.75 in wide, and 0.03 in thick. The flakes were received at 3 percent MC and adjusted to 5 percent by addition of atomized water prior to adhesive application. As in the small panels, a 5 percent pMDI adhesive content was used. A rotating drum blender with airless spray heads was used to apply adhesive. No wax or other water repellent was included in the furnish.
The mat was prepared using a belt and picker roll former that dropped the strands onto a constant speed carriage that moved back and forth under the forming head. Mat weight was controlled by weighing a number of trial passes through the former. A canvas belt loader rested on the carriage and a forming box was positioned over the belt loader to contain the edges of the mat. When the mat was complete, the forming box was removed. Then the belt loader was rolled off the carriage and onto the dies, guided by small flanged wheels on the edges of the loader. Once the mat was in position, the back end off the upper side of the belt loop was locked against the press frame and the belt loader frame was retracted back onto the carriage. The mat drop of the end of the belt was approximately 1.5 in to the top surface of the lower die. The mat deposition onto the die was without any indication of disturbance. Only a few strands dropped into the recesses of the die so that the top surface of the mat remained planar. Observation of the press stroke from the ends indicated no mat tearing during closing. The press conditions were the same as those used for the small specimens.
A total of 53 panels were produced. After pressing, the panels were trimmed to 4 ft by 8 ft, after being allowed to cool to room temperature. Each trimmed panel was weighed and measured for thickness across the ends, at the mid points of the horizontal and sloped elements of the cross section. Four specimens were rejected at this point due to low weight caused by mat forming problems. All full-sized panels were stored and tested at ambient room conditions.
Mechanical properties of full panels
Primary direction flexural testing.--The panels were first flex tested in the 8 ft direction by means of weights placed on a narrow strip at the mid point of a 7-ft span. The distribution of nominal bending stiffness vs. panel weight is shown in Figure 6. The five panels with densities below 38 pcf were considered to be rejects because of underweight mats. The trend line shown has an [r.sup.2] = 0.51.
FEED