INCREASING SERVICEABILITY OF WOOD PALLETS USDA FOREST SERVICE RESEARCH PAPER FPL 215 1973

U.S. Department of Agriculture Forest Service Forest Products Laboratory Madison, Wis.

ABS TRACT

More than 10 percent of the annual U.S. cut of timber is used for constructing pallets. To reduce this percentage,which is a drain on timber resources, evaluations were made of the following: Effectiveness of minor alterations of pallet and of the powered, hand-type forklift truck, nailed joints versus mastic joints, and recessing the leading-edgedeckboards of pallets from the ends of notched stringers.This work determined that (2) service life of reusable pallets could be extended greatly when handling loads up to 750 pounds, mostly by using an "impact panel" on the fork­lift truck; (b) nailed joints are much more resistant to impact than are comparable mastic joints; and (c) recessingleading-edge deckboards probably does not improve impactresistance of pallets. The information should benefit all users of wood pallets because pallet cost per trip will be reduced and pallet grade lumber will be used more efficiently.It is likely that the severity of handling by the forklift truck can be reduced further by employing cushioning devices and techniques. Based on this study, the U.S. Forest Service is applying for a Government patent on the impact panel.

Supplementary information for USDA Forest Service Research Paper FPL 215, "Increasing Serviceability of Wood Pallets"

The sketch below illustrates details of the impact panel described on page 6, but not fully shown by figure 3, As indicated, the three steel bars were bolted rigidly to the bottom of the channel. The 1/2-inch relief between the bottoms of the projecting bars and the forks allowed the fork assembly to slide easily over surface irregularities during the tests.

R. K. Stern

- - - -

INCREASING SERVICEABILITY OF WOOD PALLETS

By R. K. STERN, Technologist 1

Forest Products Laboratory,2 Forest Service U.S. Department of Agriculture

INTRODUCTION

Tests for both laboratory and field evaluations of the serviceability of reusable wood pallets include the following: Free fall drop, edgewise drop, cornerwise drop, revolvingdrum, conventional incline impact, static bending, and others. However, under service conditions there is of ten appreciable uncertainty about correlation between the resultingdata and pallet performance. This is complicated by the difficulty of obtaining trulyrepresentative information on the nature and the levels of physical stress associated with use. Conventional handling of a pallet with both laboratory-loading and data-monitoring apparatus attached is shown in figure 1; the forklift truck is a powered handtype.

Reeves and others of the Eastern Forest Products Laboratory, Department of Environment,Ottawa, Canada, studied the problem of accelerated laboratory simulation of rough handlingthe pallets used by the Grocery Products Manufacturers Association of Canada. After a 2-year study they concluded that about 75 percent of all initial damage to reusable wood pallets (two-way entry type) required repair because of splitting of or loss of the top leading-edge deckboards or of both.3 They also observed that these failure patterns were caused by the nature of the loading action of the powered, hand-type forklift trucks that handle the pallets. 3

The test apparatus developed by Reeves is shown in figure 2 wherein the fork assemblyis mounted with a 4° forward inclination from the backstop of an incline impact-testing system. The test pallet is loaded and allowed to travel down the incline from a predetermined height. At 8 inches from the apex of the angle formed by projecting the horizontal and the vertical faces of the forks, the pallet contacts the upper face of the horizontal forks. Additional upward stress is applied as the pallet continues to move toward the vertical portion of the forks, Ultimately, the top leading-edge deckboard is impacted directly by the vertical fork assembly, and a rapid transfer of the kinetic energy occurs (as happens in actual forklift truck handling operation).

This study was conducted to evaluate the effectiveness of the following changes in pallet construction and handling technique: 1The author gratefully acknowledges the experimental research on this project by

J. D. Wiese, D. P. Schroeder, and G. Wilson, Engineering Technicians at the Forest Products Laboratory.

2Maintained at Madison, Wis., in cooperation with the University of Wisconsin. 3J. R. Reeves. The Development of Improved Grocery Product Pallets: A Preliminary

Evaluation. Forest Prod. Lab. , Ottawa, Ont. Inf. Rep. OP-X-31, Canadian Forest? Serv., Dep. Fish. and Forest. March 1970.

Figure 1.--Handling of a loaded pallet by an ordinary forklift truck without an impact panel.

M 141 428

FPL 215 -2-

Figure 2.--Diagram of test equipment used by J. R. Reeves to produce typical failure of the top leading-edge deckboard as happens frequently in service. (Drawing by courtesy of Eastern Forest Products Laboratory, Ottawa, Canada, file No. 2-2129.)

M 141 424

-3-

Change

Joining pallet parts with mastic instead of nails.

Adding an impact panel to the forks of a forklift truck.

Making top leading-edge deckboards recessive from ends of stringers and attaching an impact panel to the fork assembly.

Principle

Shock transmitted to load reduced because of increased cushioning action.

Reduction of unit stresses applied to all pallet joints by applying stresses to the stringer ends during handling.

Reduction of unit stresses applied to pallet joints of the top leading-edgedeckboards by applying stresses to the stringer ends during handling.

Materials and Assembly

Materials

The following materials were used for the pallets:

Lumber.--Green, red oak (Quercus rubra L.) logs, purchased from K&M Tie Company,Sauk City, Wis., and transported to FPL by truck, were stored underwater in a holdingtank until they were sawed into lumber.

Nails.--2-1/2-by 0.120-inch helically threaded pallet nails.

Adhesive.--Synthetic elastomeric sealant applied with a calking gun to a thickness of about 1/32 inch.

While still green (a moisture content well above the fiber saturation point), the red oak logs were ripped, trimmed, and planed into nominal 1-inch deckboards or 2- by 4-inch stringers. The stringers were notched. The pallet parts were then selected by randomiza­tion. The boards and stringers were then wrapped in polyethylene sheets, stored in an atmosphere of 36° F. and 82 percent relative humidity, and nailed or glued together as notched-stringer, nonreversible, reusable 48- by 40-inch pallets. The completed pallets were stored for at least 30 days, and reached a moisture content of about 14 percent before testing. Equal numbers of pallets were made (1) with the top leading-edge deckboards flush with the ends of the stringers and (2) recessed 3/8 inch. All top deckboards were uniformlyspaced.

RESEARCH

Modification of Testing Equipment

The equipment with modification based on this research is shown in figure 3. A powered, hand-type forklift truck was used to handle (impact) the loaded pallets at 1.97 miles per hour, the maximum speed attainable.

Simulated load.--A load of 750 pounds, consisting of a 1/2-inch steel, open-top retain­ing box with steel and wood plates, was assembled as shown in figure 3. The individual plates were arranged in the box to prevent secondary impacts between the individual plates.The rear vertical panel extended downward 3-1/2 inches from the top surface of the pallet.Therefore, inertial loading (in response to the handling stress applied by the forklift truck) simulating cargo was always applied through the stringers.

FPL 215 -4-

Figure 3.--Modified fork tine assembly developed by the Forest Products Laboratory (USA) to lower unit stresses caused by handling. Inset: Impact panel construction (A, impact panel; B, steel bars; and C, steel channel).

M 141 427

-5-

Impact panel.--Construction of a suitable impact panel to achieve the basic energytransfer from the forklift truck to the pallet stringer involved the following principalrequirements:

a. Sufficient stiffness to withstand the rigors of testing. The compression of the panel when loaded anywhere between support points could not allow for proportional limit to be exceeded. Should this occur, the panel would sustain permanent deformation that varies with overstress, and would lead to severe misalinement.

b. Compact, and adjustable in service.

A 2-inch-thick red oak panel with portions extending below and beyond the forks was tried initially. This panel had insufficient strength in shear for the span between the forks, and failed early in the program.

A 46-inch length of American Standard channel steel 8 by 2-1/4 inches and 13.75 pounds per foot, was mounted with the open side adjacent to the vertical portions of the forks. The fillet at the apex of the right angle formed by the horizontal and the vertical portionsof the forks was removed to allow stable support of the impacting surface of the impactpanel. Exploratory calculations and testing confirmed that the impact panel (thus attached)had sufficient strength to resist the handling forces within the elastic region, but some of the pallets failed after only a few impacts. The mechanism for this failure was shear parallel to the grain at the top of the stringer ends. Restated, therefore, the wood (with an allowable slope of grain of 1:10) failed in shear parallel to the grain for the small area of wood loaded under this arrangement. This led to the addition of 2-inch-deep and 2-3/8-inch-wide steel bars bolted to the bottom of the channel so that the load would be distributed over a greater area at the stringer end. This addition is shown in figure 3. Exploratory research with the panel and its attached downward projections indicated that this combination successfully overcame the deficiencies of earlier prototypes.

Monitoring Test Results

Research data were recorded by three techniques: (1) An accelerometer-oscilloscope land process camera system; (2) high-speed motion picture photography with suitable reference lines; and (3) a count of the cumulative number of handling impacts required to cause failure or repair. Visual observations of effects were made before, during, and after impacts.

The initial element of the electronic recording system used was a strain-gage acceler­ometer mounted at the center of the rear panel of the simulated load, The acceleration registered by the accelerometer was coupled through a bridge balance into the y-amplifierof an oscilloscope. With the proper gain settings, acceleration-time records were recorded for each set of impact conditions. Acceleration varies with the force applied to the pallet .

A high-speed motion picture camera, set to operate up to 4,000 frames per second, was used with appropriate lighting, triggering, and measurement reference systems to photo­graph significant parts of the exploratory portion of the work, especially during impact.

The last and simplest, but most extensively used data-recording system, involved counting the elapsed number of handling-type impacts with a full-size forklift truck.

Major Test Procedure

With the exploratory work concluded, the experimental work was planned so that the principal variables and their interactions could be compared on a statistically valid

FPL 215 -6-

basis. Therefore, a 23 full-factorial experimental design that involved a total of 32 pallets was used. Each pallet was "handled" (impacted) by a forklift truck on one major entry point (end) until (1) a substantial part of the pallet became separated or (2) the operator decided the pallet would have been withdrawn from services for repair.

For the first group of eight pallets--each with a different set of majorcharacteristics--acceleration-time (a-t) records were made for the first, fifth, 50th,and 300th impacts (if the test series lasted that long). To facilitate data interpreta­tion, photographs of pallet condition were also taken before and after testing wherein a-t records were made. If a pallet remained usable after the 300th impact, further testing was discontinued.

The principal experimental design included nailed or glued pallet components, palletswith leading-edge deckboards either flush with the ends of the stringers or recessed 3/8 inch, and pallets handled by a forklift truck with either conventional forks (withthe usual curvature at the inside right angle of the forks) or with the modified fork with the impact panel attached (table 1). The advantage of proceeding with testing on this basis was that the resulting data could probably be interpreted on a statistical basis.

RESULTS

The performance of the test pallets and forklift truck with the test variables are shown graphically in figure 4. Nailed pallets when handled by the forklift truck with an impact panel performed best. All eight of these pallets (fig. 5, clear bar) were still serviceable after 300 impacts; further testing was discontinued. A photograph typicalof this condition is shown in figure 5.

In sharp contrast to pallets with nails handled by a forklift truck with the modified panel, pallets with mastic handled by a conventional forklift truck averaged only four impacts. The maximum value obtained by any of eight pallets in the group was only eight(fig. 6). Figure 7 illustrates the typical condition of a nailed pallet with recessed stringers after 37 impacts in testing to failure.

Representative a-t recordings for the electronically monitored position on the simu­lated load of a pallet with deckboards flush with its stringers and handled by a modified forklift truck are shown in figure 8. The pulses representing the first impact rose sharply from 0.0 to about 20.5 G (ratio of absolute acceleration to that of free-fallingbody at sea level); then the pulses oscillated through a period of about 20 milliseconds. Typically, the joints loosened progressively with additional testing so that the duration of the a-t pulses for the 300th impact of this pallet had extended beyond 30 milliseconds. However, the peak acceleration for the 300th impact was not greatly different from that of the first impact. Unavoidable (but distinguishable), shock-excited ringing in the 1/2-inch steel plate of the simulated load caused a primary plate resonance of about 900 hertz. This complicated interpretation of the performance of the pallet proper.The oscillation on the leading edge of the pulse without a corresponding increase of the mean pulse along the y-axis signified that (in this impact) only one fork had contacted the loaded pallet initially, About 8 milliseconds later the other fork also made contact,and the full energy transfer began. This caused a rapid increase in the acceleration of the pallet and its load to about 19 G. The trace then oscillated back to 0 about 21 milli­seconds later. Slight misalinement between the motion direction of the forklift truck and the surface contacted in pallet handling caused this behavior frequently during the test­ing, but this was considered typical of service-handling conditions. In summary, while the pallet joints loosened with successive impacts, the shock pulses lengthened but did not appreciably increase in magnitude.

-7-

Figure 4.--Performance of nailed or mastic-jointed pallets with deckboards flush or recessed with the stringers when handled by a forklift truck, either conventional or modified. (H, high; A, average; and L, low.)

M 141 426

FPL 215 -8-

Figure 5.--Serviceable condition of a typical loaded, nailed, flush pallet when testing was arbitrarily halted after 300 impacts by a modified forklift truck. M 140 876-9F

Figure 6. --Unserviceable condition of loaded, mastic-jointed, flush-type pallet after five impacts by a conventional forklift truck. M 140 873-9F

-9-

Figure 7.--Unserviceable condition of loaded, nailed, recessed-type pallet after 37 handling impacts by a conventional forklift truck. M 140 875 4

Figure 8.--Typical acceleration (force) time pulses from rough handling of a loaded pallet by a modified forklift truck: A, after first impact; B, after 300th.

M 141 429

FPL 215 -10-

Table 1.--Performance of p a l l e t s wi th three assembly var iab les

Table 2.--Analysis of variance f o r comparisons between the p r i n c i p a l va r iab les and in te rac t ions

-11-

Statistical Importance

A summary of the statistical analysis of the data is presented in table 2. As indi­cated, statistically significant results at the 0.99 level (at least 99:1 certainty that the same relationship would exist if the experiment were repeated) are indicated in descending order of certainty for the following conditions:

1. Use of an impact panel on the forklift truck provided a definite advantage over conventional forklift.

2. Nailed pallets were significantly more resistant to handling impacts than were similar mastic-bonded pallets.

3. Nailed pallets when handled by a forklift truck with an impact panel providedsignificantly better handling resistance than did any other interaction studied.

CONCLUSIONS

Based on the data of this work, the following can be concluded:

Performance of wood reusable pallets loaded with up to 750 pounds can be significantlyimproved during handling by powered, hand-type forklift trucks if handling stresses are applied simultaneously to all joints by contact with the stringers. This can be achieved by adding an impact panel to the forks of a forklift truck.

Nailed conventional pallets withstand handling by a forklift truck much better than do similar mastic-bonded pallets. However, additional work is needed to evaluate this conclusion at much higher loading--like 3,000 pounds. Also, it is likely that incorpora­ting any type of cushioning device, such as a cushioning material next to the impact panel or shock absorbers coupled to the panel, will greatly enhance the effectiveness of the impact panel in reducing handling shock to the pallet.

Apparently, recessing leading-edge deckboards from the ends of stringers does not markedly increase the resistance of pallets to shocks imparted during handling operations.

Mention of trade or proprietary name is for information purposes only and does not imply any endorsement by the Forest Service, U.S. Department of Agriculture.

FPL 215 -12- 4.0 - 13 - 11 - 73 U.S. Government Printing Office: 754-550/48