Design and Testing of Environmentally Friendly Wooden School Chairs for Developing Countries

Transcription

1 Design and Testing of Environmentally Friendly Wooden School Chairs for Developing Countries By E. Haviarova, C. Eckelman, and Y. Erdil Introduction In many, if not most of the underdeveloped and developing countries of the world, school furniture is poorly designed, of low quality, and often unfit for school use, yet it is costly and consumes a disproportionate share of limited educational budgets. This situation need not continue. Attractive, well-designed, durable, maintenance-free furniture can be produced from locally available woody residues including materials such as plantation thinnings. Durability is ensured by the total design. Furthermore, none of the processes involved in the construction of the frames are inherently costly. Thus, attractive, sturdy, easily-maintained furniture can be produced at modest cost in any developing country. Only the simplest machining and joinery processes are required to construct the furniture. Production techniques may vary from those best-suited to cottage industries to those more appropriate for small factories in more developed areas. Thus, school furniture and the accompanying production processes can be matched to the level of development of essentially any targeted country. The intent of this study was to evaluate two types of construction that appeared to satisfy the above criteria, namely, solid wood construction and cross lap construction. In the solid wood construction, the stretchers, rails, and posts were constructed of solid wood. In the cross lap construction, the cross stretchers and cross rails were constructed of solid wood, but the side frames consisting of the side rails, side stretchers, and front and back post were of laminated construction. The configuration of the chair that more or less served as a model for the construction of all of the chairs used in the study is shown in Figure 1. Individual chairs differed in detail, some chairs had straight backposts, for example, and the number of stretchers varied, but in general, all of the chairs were of more or less the same size and of similar configuration. Overall, sizes of the chairs were such that they would be suitable for many, but not all, children in the 8 to 9-year old age group.

2 In the case of solid wood chairs, Figure 1, the stretchers and back rails were first ripped from 7/8-inch thick stock. After they were cut to length, round tenons were machined on the ends of each part with a 3/4-inch outside diameter hole saw. These saws cut tenons with a nominal diameter of about 19/32 inches. Excess material is removed from around the tenons as needed. Front rails were constructed of 7/8-inch thick by 1-3/4-inch wide material. Round tenons were cut off-center on the ends of this member in order to allow the matching mortise in the front post to be cut slightly further down from the top of the post. The top rail iss also constructed of 7/8-inch thick by 1-3/4-inch wide material, but the tenons were centered on the ends of this member. Holes (mortises) to receive stretchers or rails were drilled completely through the posts. The drill used to bore the mortises in the construction of the furniture described in this paper is chosen to provide a shrink&swell fit between tenon and hole. The backposts in the chairs with sloping back legs were either creep bent or bandsawn to shape. Cross sectional dimensions of all the members ae given in Table 1. From this point, assembly techniques varied as a best method of assembling the chairs was sought, but, in general, the side frames were constructed first. Walls of the holes and tenons were first coated with adhesive. The tenons on the ends of the stretchers were then inserted in the holes and the assembly pulled together by means of bar clamps until the desired front to back side-frame dimensions were obtained. Because of the slant of the back legs, a close fit between the shoulders of the tenon and the wall of the post

3 adhesive was used in construction of all chairs. Frames were allowed to cure for several days before testing. Moisture content of the wood at the time the chairs were constructed varied from 12 to 15 percent. Before assembly, however, the stretchers were dried to 6 percent moisture content. The completed chairs were then conditioned to a moisture content of 7 percent. The first two chairs were constructed of yellow poplar. This species was chaosen because it is one of the weaker woods. Chair #3 was constructed entirely of shagbark hickory. Shagbark hickory was selected for this chair because it is one of the strongest hardwoods and test results would provide an estimation of the maximum strength to be expected from this type of chair configuration. In the case of chair #4, the posts were constructed of yellow-poplar, whereas the stretchers were constructed of shagbark hickory. Mixed woods were used in this chair to determine what percentage of strength of Figure 2. Chair #5. Cooperator chair. the hickory chair could be obtained in a chair in which a weaker wood was used in the posts in combination with hickory in the stretchers. The woods used in the remaining chairs are listed in Table 1. Chair #5, Figure 2, was unique in that it was constructed with hand tools alone by a cooperator in the study. It differs somewhat in configuration from the other chairs because this was the cooperator s design. Construction of chairs #6a through #6d were of slightly simpler design with straight backposts and were constructed for testing purposes only. In general, a close fit was obtained between the shoulders of the tenons and the sides of the posts, in contrast to chairs 1 through 4. Tenon-Mortise Fit All of the stretchers were attached to the posts or frames with round mortise and tenon joints. This construction provides a simple yet durable means of connecting members and side frames together. Round tenons may be produced on a conventional wood lathe or they may be cut with a simple hole saw.

4 regains moisture, either from the adhesive on the walls of the mortise or the atmosphere, it swells and a tight shrink&swell fit is obtained. Bending to Shape of Green Backposts The backposts used in chair #1 were bent to shape while they were in the green condition. In this process, the back post, while green, is supported at each end and subjected to a load perpendicular to its longitudinal axis at midspan as shown in Figure 4. In practice, a backpost is bent as much as possible without fracturing and allowed to remain in this deflected condition for some length of time. The nut is periodically tightened and the backpost Figure 4. Green bending of back posts. Note that posts are bent in pairs. allowed to creep until the desired degree of curvature in the back post is obtained - note that an allowance should be included for springback. The post is then allowed to dry. In some cases, the wood may be bent to the desired shape in one stage; in other cases, it may be necessary to allow the wood to creep so that the bending process must be carried out in stages. It should be noted that in this process, as the wood creeps, most of the creep occurs in areas of high stress. Thus, a round load head of small diameter will produce relatively sharp bends. Conversely, a large shaped load head with a gentle curvature may be used to produce sweeping curves. Seats The seats used with these chair frames were constructed of 1/2-inch thick plywood. This was done largely as a matter of convenience since the primary focus of this study was chair frames. Nonetheless, it is likely that such seats would be specified by many schools. In keeping with the objective of maximum utilization of locally available resources, however, it is anticipated that chair seats and backs in many school districts will be constructed of slats. The general shape of the slats can be obtained by green bending, and the slats can then be attached by means of mechanical fasteners to the front and back seat rails. Alternatively, it is anticipated that the ends of the slats can be bent around the front and back seat rails through the use of a simple ammonia treatment process so that the slats can be essentially locked in place as the chair is assembled. Initial tests indicate that the process is entirely feasible, and the process will be developed in the next stage of the study. Cross Lap Joint Chair Construction Cross lap laminated joint construction provides one of the simplest and most

5 to post joints. In form and function, these joints are equivalent to corresponding mortise and tenon or multiple mortise and tenon joints. Construction of chair #7, a 3- lamination cross lap joint chair constructed of yellow poplar is shown in Figure 5. Dimensions of the members are given in Table 1. To obtain a sloping bacl leg, backposts were sawn to shape after the side frames had been assembled. In constructing the side frames, the joint centers of the frame were first located on a flat Figure 5. Cross-lap chair construction. plywood panel. Holes were then drilled through these points, and bolts were inserted through the holes and fastened securely in place with nuts. The lap areas of the laminations were then coated with adhesive, and the pre-drilled laminations slipped over the ends of the bolts. Large washers were then slipped over the ends of the bolts, and the nuts were then threaded onto the ends of the bolts and tightened to apply pressure to the lap joint areas. Shorter lengths of material were glued in place in the slots between laminations in the front and back posts in order to produce solid legs. Thus, the front and back post consisted of three laminations each, whereas the side rail and side stretcher consisted of only one lamination each. Once the adhesive dried, the frames were removed from the forms. The holes at the joint centers were then re-drilled to the same diameter used in the solid wood frames, and the side frames joined together in essentially the same manner as with the solid wood chairs. Construction of Chair #8 was

6 left open. Overall, the chair was somewhat larger than the others and was constructed of mixed hardwoods. Table 1. Cross sections of members and species of wood used in construction of chairs. Chair No. Front Rail Front Stretcher Side Stretcher Top Side Stretche r - Middle Side Stretcher- Bottom Back Rail Back Stretcher Top Rail Top Stretcher Front Post Back Post 1 Fig. 1 2 Fig. 1 3 Fig. 1 4 Fig. 1 5 Fig. 2 6a Fig. 3 6b Fig. 3 6c Fig. 3 6d Fig. 3 7 Fig. 5 8 Fig. 5 9 Fig D R -Oak.675D R-Oak 7/8x7/S- Hick.675D R-Oak 3/8x1-1/2 3/8x1-1/2 Gmelina na.675d R-Oak na na 3/8x1-1/2 3/8x1-1/2 Gmelina S-Hich.675D R-Oak.675D R-Oak 7/7x7/8 7/7x7/8 1/4x3-1/2 W.-Ash 7/8x2-1/2 Bandsaw n 7/8x2-1/2 Bandsaw n na 7/8x1-1/2 Bandsaw n 7/8x1-1/2 Bandsaw n 1-3/8x1-1/4 W-Ash 1-1/4x1-1/2 1-1/4x1-1/2 Gmelina 1-3/8x1-1/4 Creep Bent 1-3/8x1-1/4 Bandsawn 1-3/8x1-1/4 Bandsawn 1-3/8x1-1/4 Bandsawn 1-3/8x1-1/4 W-Ash Straight Straight Straight Straight Straight 1-1/4x2-1/2 1-1/4x2-1/2 Gmelina Performance Tests - Front to Back Load Test and Side Load Test on Seats

7 The test itself consists of pushing from front to back on the seat of a chair (or on the front rail). This action produces internal resisting forces in the side frame of the chair similar to those caused by the action of someone tilting backward. A strap passes over the seat and is pulled backward by an air cylinder (or by a similar loading mechanism) located behind the chair which tends to tip the chair over backward. As the chair begins to tilt slightly, however, its motion is resisted by that portion of the strap that hangs vertically from the front edge of the seat and is anchored below -- in effect, the vertical portion of the strap always provides the exact force needed to keep the chair from overturning. The chair is mounted for testing as shown in Figure 7. Reaction brackets are placed behind each of the back legs to prevent the chair from sliding backwards. A strap is then passed over the seat from front to back and attached to a small clevis connected to the rod end of an air cylinder that is used to apply loads to the chair. The other end of the belt is dropped over the front edge of the seat, allowed to hang vertically, and attached to a crossbar located directly below the front edge of the seat. Horizontal loads are applied to the chair seat in a front to back direction at a rate of 20 cycles per minute. For the purposes of this test, i.e., based on a consideration of the Figure 7. Front to back load test on seats. anticipated size of the users, testing of the chairs was begun at the 50 pound load level. Loads were increased in increments of 25 pounds after 25,000 cycles were completed at each preceding load level. Tests were conducted until a chair suffered disabling damage. The hoorizontal side load test on seats is identical to the front to back load test exzcept that the load is applied to the seat in a sideways direction, and reaction brackets are placed on one side of the chair rather than behind the chair. Results and Discussion Results of the tests are given in Table 2. Both the total number of cycles completed and the cycles completed at the load level at which failure occurred are given in Table 4. Front to Back Tests Chairs #1 and #2 failed at 200 lbs. and 250 lbs., respectively. An examination of

8 stretchers but low strength posts may be expected to perform as well as chairs constructed with all high strength material. Chair #5 failed after 21,000 cycles had been completed at the 250 pound load level. Cause of failure was fracture of the side stretcher to post joints. This chair had only two side stretchers so that its performance was regarded as quite good. Diameter of the tenons on the ends of the stretchers was inches compared to 0.59 inches for Chairs #1 through #4. Since the bending strength of the tenons is related to the cube of their diameters, even this small difference in diameter could account for a 38 percent increase in tenon strength in bending. Chairs #6a-6b failed at 275 and 350 pounds respectively. The reason for their better performance as compared to Chairs #1-2 is likely the close fit between the shoulders of the tenon and the walls of the posts. In all four chairs, the walls of the posts were indented by the shoulders of the tenons during the source of testing. This result indicates that the strength of the joints was not dictated by the bending strength of the tenon alone. Chairs #7 failed at 425 pounds. The adhesive in the joints on the left side of the chair failed after a few cycles had been completed at the 400 pound load level. Importantly, the left side frame continued to carry load although it deflected more than the right. The adhesive in the right side frame also failed, but at the 425 pound load level. Finally, the tenons in both side frames fractured and the test was stopped after 15,780 cycles had been completed at the 425 pound load level. These results tend to indicate that even if the adhesive in the joints fail, the joints will continue to carry load by mechanical action alone up until the point that the tenons fracture. Chair #8 failed after 24,500 cycles had been completed at the 450 pound load level. Cause of failure was fracture of the side stretchers at the stretcher to front and back post joints and fracture of the side rails at the rail to back post joint. This result tends to indicate that chairs constructed of gmelina, considered to be a relatively weak wood, could be expected to be as strong as chairs constructed of yellowpoplar. It should be noted that the side rail to front post joints, i.e., L-shaped joints as opposed to T-joints, did not fail. In the case of both of the above two chairs, the holes for the tenons of the front and back stretchers were drilled through the centers of the cross lap joints in the side frames. Clearly, these holes lessened the mechanical strength of the side frame tenons. Specifically, once the adhesive in a joint failed, the corresponding side frame tenon tended to rotate about the round stretcher tenon that framed into the side of the joint. As this action occurred, there was also a tendency for the side frame tenon to withdraw from the mortise. This action

9 Acceptance Levels To provide a basis for an intuitive feeling for the strength of the chairs evaluated in this study, it is useful to compare the strength values obtained with the acceptance levels specified in the American Library Association test method for chairs used in library reading rooms, a severe use environment. In the case of front to back load tests, values of 250, 350, and 450 pounds are specified. Essentially, these values correspond to what are regarded as light, medium, and heavy duty. Experience indicates that if chairs do not have at least 250 pound front to back strength, a significant number will fail during the first two years in an adult library environment. Few chairs with a strength of at least 350 pounds ever fail, however. Comparable values for the side load test are 200, 250, and 300 pounds for low, Table 2. Results of front to back and side load seat tests. Cycles Initial Load Ultimate Complete Total d Chai Type of Load Increase Cycles r e No. Test (lbs.) (lbs.) (lbs.) Load Complete d 1 Front to Back do do do , ,000 3 do do do , ,000 Side Load 50 50? 450 3, ,956 4 Front to Back do do 400 7, ,000 5 do do do , ,000 6a do , ,000 6b do ,150 6c do 275 6d do do , ,780 Side Load ,100 8 Front to Back , ,500 6 do do do , ,000 medium, and high categories, respectively. The higher values tend to reflect the strength of chairs produced rather than strength needed to survive in service.

10 To put these comparisons into perspective, it must be realized that adult library reading rooms (especially those located at universities) represent a severe use environment and acceptance levels, accordingly, are set high relative to other use environments. Chairs that meet only the low ALA acceptance level, for example, have given good service in fast food restaurants. Furthermore, chairs intended for home use often fall considerably short of satisfying the low ALA acceptance level. Thus, a very high level of strength has been achieved in the test chairs that presumably can be attributed both to their design and method of construction. Structural Analysis - Solid Wood Chairs Following testing, structural analyses were conducted in order to determine the distribution of forces in the side frames under the action of the ultimate loads reached in the tests. Once these bending forces were determined, the corresponding bending stresses were then determined and compared to the modulii of rupture (MOR s) of the materials used. In evaluating results, the cyclic, as opposed to static, loading of the chairs must be taken into account. Of particular interest, the peak load in each load cycle was held for less than a second. According to the Wood Handbook (1974), the load required to cause failure in a wood member in one second is approximately 25 percent higher than that obtained in ASTM standard strength tests. Thus, calculated stresses in some cases may be considerably higher than published modulus of rupture values. Fatigue, on the other hand, may have also influenced the results. The number of cycles completed in these test may have not been high enough for fatigue to be a factor. On the other hand, the fatigue characteristics of the wood when subjected to a stepped load schedule, particularly at high percentages of ultimate strength, is unknown. Analyses of the chairs was carried out with a commercially available finite element software, namely, MICROSAFE, a product of the Microstress Corporation. Bending forces acting on the ends of the stretchers at their points of entry into the posts at the ultimate load levels achieved by the frames are summarized in Table 3. The stresses developed in the tenons may be found by means of the flexure formula 32? F4 s4? 3? D where s 4 = flexural bending stress, psi; F 4 = the bending force, lb-in; D = the diameter of the tenon, inch; and?? Substituting the appropriate value into the above expression, i.e., D = 0.59 inches and solving for a unit bending force gives 32? 1 32? 1 s4??? psi.? 3 3 D 3.14? (0.59) The ultimate bending stresses developed in the tenons corresponding to ultimate bending

11 The lowest calculated breaking stress for these two chairs was 9,324 psi; the highest was 12, 461 psi. These two values appear reasonable; however, there was a question concerning the quality of this material, as previously discussed. In the case of Chairs #3 and #4, the calculated modulus of rupture of shagbark hickory at 7 percent moisture content is 23,230 psi with a corresponding cyclic MOR of 29,038 psi). Referring to Table 3, it is seen that the calculated ultimate stresses developed in the tenons ranged from a low of 18,152 psi to a high of 21,183 psi. These results again indicate that the chair is developing a relatively high percentage of the strength of the materials of construction. In the case of the cooperator chair, Chair #5, the modulus of rupture of red oak at 7 percent moisture content is 16, 445 psi with a cyclic MOR of 20,556 psi. Calculated ultimate stress values varied from 18,184 psi to 18,992 psi. This chair is of particular interest because the tenons had no shoulders. Hence, the strength of the joint would be expected to be governed by the bending strength of the tenon alone. Results tend to indicate that this is a very efficient type of construction. In the case of Chairs #6a through #6d, the calculated stresses developed in the tenons ranged from a low of 12,479 psi to a high of 17,445 psi. In general, some of the values obtained with these chairs are higher than would be expected, even when cyclic loading is taken into account. The back legs on these chairs were straight rather than curved, however, so that a close fit was obtained between the shoulder of the tenon and the wall of the post. As a result, the shoulder likely contributed to the strength of the connection so that the strength of the joint was stronger than would be predicted from a consideration of the bending strength of the tenon acting alone. Structural Analysis - Cross Lap Chairs Structural analyses also were carried out on the cross lap chairs of the type shown in Figure 5. Bending forces acting on the ends of the stretchers at their points of entry into the posts at the ultimate load levels achieved by the frames are summarized in Table 4. The bending stress produced by these moments may be calculated by means of the expression s? 4 td where s 4 refers to the bending stress, f 4 refers to the bending force acting on the end of a member, t refers to the thickness of the rail, inch, and d refers to the depth of the rail, inch. Substituting the appropriate values into this expression, i.e. t = in. and d = 1.5 in, gives 6 f 4 2 6? f4 s 4?? 7. 1? f ? (1.5) 4

12 In the case of Chair #8, the tenons failed at calculated stress levels ranging from a low of 3982 to a high of 4338 psi. When forces are redistributed as described above, calculated bending stresses range from a low of 3293 to a high of 6748 psi. The mode of failure of the joints during testing is of interest because of concern over possible weakening of the laminated joints in service owing to shrinking and swelling that may result from climatic changes in regions with well-defined wet and dry seasons. If, in fact, the adhesive in the four cross lap joints in a side frame should fail owing to climate related factors, the L-shaped front post to side rail joint would become a pinned joint, i.e., it would have no resistance to bending forces, but it would be held in place by the round tenon of the front rail that mates with it and passes through its center. The remaining T-shaped joints would behave as mechanical mortise and tenon joints rather Table 3. Front to back load bending forces and stresses acting on the ends of the stretchers in solid wood frames at ultimate side frame strength levels. Ultimate Tenon Front End Rear End Chair Stretcher Load Diam. Force Stress Force Stress No. Position (lbs) Species (inch) lb-in (psi) (lb-in) (psi) top 200 y pop middle bottom top 250 y pop , ,655 middle , ,903 bottom top 425 s hick ,838 middle , ,235 bottom , ,183 4 top 400 s hick , ,648 middle , ,045 bottom , ,937 5 top , ,272 bottom , ,632 6a top 275 y pop , ,820 middle , ,093 bottom , ,706 6b top 350 y pop , ,317 middle , ,664 bottom , ,445 6c top 275 y pop , ,820

13 of the side rails and stretchers. This mode of behavior was infact observed during the tests although adhesive failure was caused by external loading of the structure rather than shrinking and swelling of the joints. At first glance, the rectangular side rails and stretchers used in this chair did not develop as high an apparent efficiency as the solid wood chairs constructed with round tenons. In point of fact, however, only half as much material was required for the rectangular members in the lap joint chairs as for the 7/8-inch square stretchers in the Table 4. Bending forces and stresses acting on the ends of the stretchers (on the tenons) in cross lap frames at ultimate side frame strength levels. Ult. Tenon Front End Rear End Chair Stretcher Load Sect. Force Stress Force Stress No. Position (lbs) Species (inch) lb-in (psi) (lb-in) (psi) 7 top 425 y pop.375x bottom top 425 y pop.375x bottom top 450 gmelina.375x bottom top 450 gmelina.375x bottom and 8 1. Analysis with the L-shaped joint pinned. solid wood chairs. The result simply indicates that a rectangular section is much more efficient than a round or square section in resisting bending forces. In addition, as discussed above, the cross hole drilled through the center of the tenon to accommodate the cross member, significantly altered the mechanical characteristics of the section, and presumably reduced them. Hence calculated stress values based on an undisturbed tenon cross section would be expected to be lower than the actual stresses developed in the modified tenons. Conclusions Results of the study indicate that strong durable school chair frames can be constructed using the simple techniques employed in the construction of the chairs evaluated in this study. The high strength of the chairs results both from the favorable distribution of internal resisting forces among the stretchers and rails and also from the strength of the round mortise and tenon joints and the cross lap joints. Deep hole saws provide a convenient method of cutting round tenons of uniform diameter without the

14 Eckelman, C. A Library Chairs: An Overview of the Library Technology Reports Test Method with Test Reports on 30 Chairs. Library Technology Reports 31(2): American Library Association, Chicago. Sennett, Robert E Matrix Analysis of Structures. Prentice Hall. Englewood Cliffs, N.J. 228 pp. Gere, James M. and Stephen P. Timoshenko Mechanics of Materials. 3rd Ed.PWS Publishing Co. Boston. 807 pp. USDA Forest Products Laboratory Wood Handbook: Wood as an Engineering Material. Agriculture Handbook No. 72. USDA Forest Serv., Forest Prod. Lab., Madison, Wis.