. . 4 y v v .1... :,_... IG ll lllllllllllllll " This is to certify that the thesis entitled LOSS IN COMPRESSION STRENGTH OF CORRUGATED CONTAINERS AS A FUNCTION OF STACKING PATTERN AND DYNAMIC COMPRESSION presented by EUNGJOO KIM has been accepted towards fulfillment of the requirements for Masters degree in Packaging ‘ jor professor Gary J. urgess, ph.D. Date m— 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE JUN 0 9 Kill .J 4! , uvvi-x MSU Is An Affirmative Action/Equal Opportunity institution c:\circ\datedm.pm3—p,1 LOSS IN COMPRESSION STRENGTH OF CORRUGATED CONTAINERS AS A FUNCTION OF STACKING PATTERN AND DYNAMIC COMPRESSION BY Eungjoo Kim A THESIS Submitted to MICHIGAN STATE UNIVERSITY in partial fulfilment of the requirements for the degree of MASTER OF SCIENCE SCHOOL OF PACKAGING 1993 ABSTRACT LOSS IN COMPRESSION STRENGTH OF CORRUGATED CONTAINERS AS A FUNCTION OF STACKING PATTERN AND DYNAMIC COMPRESSION BY Eungjoo Kim This study investigated the compression strength loss of bottom containers after vibration due to different pallet patterns using three kinds of pallet patterns, corrugated boards, and package weights. This study also measured dynamic forces on the bottom box in a pallet load at resonance on a vibration table using a specially designed instrumented shipping box. The measured dynamic forces were related to measured accelerations through Newton's Law to see if the instrumented shipping box could.be replaced by accelerometers. The results show that column type pallet pattern may be the strongest stacking type during long term storage but not in transit. The results also show that the dynamic forces can be accurately predicted by accelerations if there is no bouncing of individual boxes during vibration. Dynamic compression levels at resonance could not be predicted. Copyright by EUNGJOO KIM 1993 I would like to dedicate this thesis to my parents: CHUNGUN KIM and CHABUN CHUNG, for their love and prayer my family : BYUNAH KIM (Wife) and WOOSUK KIM (Son) for their support through this period. iv ACKNOWLEDGEMENT I wish to express my sincere gratitude to my major professor Dr. Gary J. Burgess for his expert guidance and useful suggestions. His assistance throughout my research findings is gratefully acknowledged. I would also like to extend my sincere thanks to Dr. S. Paul Singh for his valuable advice and to Dr. Galen Brown for his help and time. I appreciate the support of the School of Packaging, Dr. Harold A. Hughes, and Dr. Ruben J. Hernandez for offering me financial assistance as a Teaching Assistant, package consultant, and grader. I also thank Dr. Bernard Fehr and ALVCO CONTAINER CO. for their assisting in my data collection. Grateful acknowledgement is extended to Ms. Beverly Underwood for her kind help and time. Finally, I would like to thank all the faculty and staff of the School of Packaging, all my friends, and all those who helped me in one way or another during the course of my graduate studies. They made me feel like a member of a big family while I was away from home'country. TABLE OF CONTENTS Page LIST OF TABLES Vii LIST OF FIGURES ix 1.0 INTRODUCTION 1 2.0 LITERATURE REVIEW 3 3.0 EXPERIMENTAL DESIGN 8 4.0 DATA AND RESULTS 36 5.0 CONCLUSION 59 LIST OF REFERENCES 62 APPENDIX 64 vi LIST OF TABLES Table 10. ll. 12. 13. 14. 15. 16. 17. Measured Static Loads and External Accelerations External Acceleration Predicted from Measured Force for Metal Brick External Acceleration Predicted from Measured Force for EPE Corrugated Box Specifications Measured Burst Strength (lns/inz) Measured Forces vs Predicted Forces for a 157.7 lb Static Load Measured Forces vs Predicted Forces for a 105 lb Static Load Measured Forces vs Predicted Forces for a 213 lb Static Load at Stack Resonance Measured Forces vs Predicted Forces for a 105 lb Static Load at Stack Resonance Column Type Stacking Pattern : SW-A Interlock Type Stacking Pattern : SW-A Pinwheel Type Stacking Pattern : SW-A Column Type Stacking Pattern : SW-B Interlock Type Stacking Pattern : SW-B Pinwheel Type Stacking Pattern : SW-B Column Type Stacking Pattern : DW Interlock Type Stacking Pattern : DW Page 11 18 19 24 26 38 39 41 42 45 46 47 50 51 52 55 56 18. 19. Pinwheel Type Stacking Pattern : DW Static Load vs Compression Strength Loss viii 57 59 LIST OF FIGURES Figure 1. Instrumented Shipping Box (ISB) 2. Test Setup for Dynamic Force Relative to Accelerations 3. Newton's Second Law 4. True Peak vs Measured Peak 5. Test Setup and Results 6. Box Specifications Label 7. Column Type Pallet Pattern 8. Interlock Type Pallet Pattern 9. Pinwheel Type Pallet Pattern 10. Test Setup for Vibration 11. Test Setup for Comparing Forces and Accelerations 12. Mean Percent Compression Strength Remaining : SW-A 13. Mean Percent Compression Strength Remaining : SW-B 14. Mean Percent Compression Strength Remaining : DW ix Page 15 16 21 23 25 28 29 3O 34 35 48 53 58 1.0 INTRODUCTION The bottom package in a pallet load must be designed to withstand the maximum compression level that will occur during transportation and storage due to the load on top. ygrious researchers have studied the effects of static compression on packages during storage in warehouse conditions (Maltenfort, 1989). The effects of various environmental factors like creep during long term storage, temperature, humidity, and stacking pattern have been investigated./The dynamic forces that the bottom package experiences while being transported in a stack has not been previously studied. Previous studies reveal that the column type pallet pattern is the strongest stacking pattern during long term storage compared to other types (Maltenfort, 1988). It has not yet been proven that the column type stacking pattern is also the strongest pallet pattern during transit. Therefore, the L‘rmsfi’f thisif‘ifill’ffit‘?ii“SfEifiéEeuifla9139933 of different stacking patterns on the compression strength of the bottom:packageafter it has experienced vibration during transit. An additional goal of this study was to measure the dynamic compressive forces that occur during vibration. Three kinds of pallet patterns, corrugated boardsimand l 2 package weights were used. Boxes made from these boards were OM palletized and subjected to resonance tests for approximately 15 minutes. The compression strengths of the bottom boxes before and after vibration were measured. An instrumented shipping box (ISB) was also used to measure the dynamic compression levels and external accelerations. The primary objectives of this study were to : 1. Measure the compression strength loss of bottom containers after vibration due to different pallet patterns. 2. Measure the dynamic compression forces on the bottom packages in the pallet using an instrumented shipping box specifically designed for this purpose. 3. Correlate the measured dynamic forces to accelerations measured simultaneously to see if the instrumented shipping box can be replaced by accelerometers. 2.0 Literature Review A board for any corrugated box can be selected based on the anticipated stacking load which the box may have to support during shipment and warehousing. In making this decision, it is assumed that no abnormal influences will act upon the container, thereby claiming a share of the potential stacking strength. While it is impossible to anticipate the exact magnitude of these influences, an understanding of their effect is valuable in deciding on how strong the box should be. A considerable amount of work has been done to determine the structural behavior of a corrugated box (Maltenfort,1988). Top-to-bottom compression strength of a single regular slotted container (RSC) can now be accurately predicted based either on material strength properties of components or with the modified IPC formula using the combined board (Maltenfort, 1989). Kellicutt found that a B—flute RSC lost 23% of its compression strength when in a three—box-high aligned stack, and 51% when a box in this stack was misaligned by 1/2 inch (Maltenfort, 1989). For A-flute boxes, he observed an 18% strength loss when the box was in a vertically stacked 4 palletized arrangement and a 55% loss when the stacking was in an interlocking pattern. Hillenious found a 49% strength loss for A-flute containers in an interlocking stacking arrangement, but for vertical-aligned stacks, losses ranged from 13% for a single stack down to 5% for multiple stacks while palletized (Maltenfort, 1989). An overhanging stack of boxes suffers strength losses from two sources. First, there is a loss of approximately 10% from vertical stacking and then there is an overhanging strength loss (Fibre Box Association, 1989). The latter was found to be approximately the same as that for single boxes, when related to their single or stacked box counterparts. For example, a single box overhanging one inch on two adjacent edges will suffer a 40% loss in compression strength compared to a single box on a solid surface. A stack of three boxes, aligned with respect to each other but overhung in the same manner, will compression test approximately 40% less than a non-overhung stack of three boxes. Uldis Ievans also confirmed that there is approximately a 45% strength loss when the straight stacking pattern is changed to an interlocking one (Maltenfort, 1989). The bursting strength of combined board is a requirement of the various carrier regulations and federal specifications for shipping containers. The bursting strength of the combined board is only a general indication of the character of the 5 materials used in manufacturing a fiber board box and has some value in this respect. On the other hand, it gives no direct information regarding the ultimate performance of the finished container and correlates very poorly, if at all, with compression strength (TAPPI T 810, 1985). Because of this, it is necessary to perform tests and reproduce the actual environment as closely as possible. Shipping containers are exposed to complex dynamic forces when subjected to vibration present in all transportation vehicles. This is due to the random nature of the vibration (ASTM D 4728, 1987). Approximating the actual damage experienced in shipping requires subjecting the containers and contents to vibration input. Vibration tests should be based on representative field data (ASTM D 999, 1986). Exposure to vibration can affect the shipping container, its interior packaging, means of closure, and contents. Design modifications to one or more of these components may be utilized to achieve optimum performance in the shipping environment. Resonance responses during shipment can be severe and may lead to package or product failure. The determination of the effect of the resonant frequencies of the product and package can aid in designing the proper packaging system to provide adequate protection for the product. This also provides an understanding of the product as it relates to expected 6 transportation vibration inputs (ASTM D 3580, 1990). Marcondes (1988) found that acceleration levels around 1.5 g's for the products were very common in less than truck load shipments. However, levels as high as 10 g's were recorded indicating severe bouncing. The average magnification factor from the input at the wheels to the truck bed was 2 for the front and 3 for the back in less—than-truck load shipments. Low natural frequency packages (below 10 Hz) showed more bouncing and larger accelerations than high natural frequency packages. Antle (1989) found that lateral and longitudinal vibration levels in commercial truck shipments were much lower than vertical levels. At frequencies above 20 Hz, the lateral and longitudinal levels were similar to the vertical vibration levels. The heavier loaded trucks showed higher lateral and longitudinal levels compared to the lightly loaded ones. The data acquisition system (instrumented shipping box) developed for this study is capable of measuring dynamic compression levels that packages experience when being transported in a stack (Singh and Leinberger, 1992). The dynamic compression level measurement errors were less than i5% and the static load measurements errors were less than :5 lbs. The maximum dynamic compression level measured in a pilot study was found to be about five times the static load when 7 the composite truck spectrum was used to drive the vibration table. Random vibration testing methods of shipping containers are recommended to simulate actual conditions. These methods can be divided into three types: Closed Loop-Automatic Equalization, Closed Loop-Manual Equalization, and Open Loop —Magnetic Tape (ASTM D 4728, 1987). Automatic equalization systems allow the enter of Power Spectral Density (PSD) data as the input test levels via a control panel. Manual equalization systems allow predetermined PSD data to be controlled. Magnetic tape systems enable equalized PSD data to be incorporated into the tape's original preparation. Any of these can be used to drive random vibration controllers which operate electrohydraulic vibration tables. The problem with using random vibration for this study though is that random vibration spectra are all different in intensity levels. This would make the results for dynamic compression levels and box compression strength dependent on the particular spectrum used. It is for this reason that the standard sine wave dwell test at the resonant frequency of the stack of boxes was used (ASTM D - 999). 3.0 EXPERIMENTAL DESIGN In order to accomplish the objectives of the study, three kinds of corrugated board, package weights, and stacking patterns were selected. Standard ASTM tests for bursting strength, vibration, and compression strength were conducted. An instrumented shipping box (ISB) specially designed for measuring forces was also used. The respective packaging materials, static loads, stacking patterns, and test methods used are described in this chapter. 3.1 Instrumented Shinnina Box (ISB) 3.1.1 Explanation of Features : The instrumented shipping box used to measure instantaneous compression forces was constructed at MSU specifically for the purpose of measuring dynamic loads. It consists of electronic hardware housed inside a reinforced plywood outer container as shown in Figure 1. This instrument measures 16 x 13 x 10 inches and is designed to fit one ninth of a standard GMA pallet. The total system weighs 45 lbs and is capable of withstanding a 5000 lb compression load. Four internal shear beam load cells are used to measure the load at each top corner of the instrument and an internal accelerometer is used to measure the acceleration history of Figure 1. Instrumented Shipping Box (ISB) 10 the unit. The ISB has a microprocessor that is capable of accepting a maximum of 16 inputs, some of which are used for system checkouts, system status, load at each corner, internal acceleration, temperature, and humidity. The ISB also accepts a signal from an external accelerometer. The processor is capable of recording all the individual inputs at a maximum sampling rate of 200 Hz. The loads and internal and external accelerations are also measured simultaneously. All data is saved in an internal memory which can be later downloaded into a personal computer through a R8232 interface. The data can then be analyzed to determine the dynamic compression forces and external accelerations that the ISB measured. 3.1.2 Procedure for Checking the ISB Accuracy for Static Weight Measurement The accuracy of the ISB was evaluated first using static weights. An accurate laboratory balance graduated in ounces was used to measure the weights of three metal bricks. They weighed 55.8, 53.5, 55 lbs for a combined weight of 164.3 lbs. The balance had its own errors of only i 1/32 lbs because it was graduated in ounces. Three bricks were then placed on the ISB. A 5.5 lbs weight difference was found between the balance and the ISB (Table 1). The 32 readings shown in Table 1 are 32 successive "samples" recorded by the ISB with a sample rate of 200 Hz. Since the four internal shear beam load cells of the ISB measure the load at each top corner of the instrument by rounding up or down to the nearest pound, there is automatically a precision error of i 2 lbs (4 x 0.5 = 2 lbs) 11 Table 1. Measured Static Loads and External Accelerations No J K L M Static External (lbs) (lbs) (lbs) (lbs) Load (lbs) Acc(G) 1 35 49 42 33 160 0.01 2 35 49 41 35 160 0.01 3 35 47 41 35 159 0.01 4 35 47 41 32 156 0.01 5 35 47 41 33 157 0.01 6 35 46 41 35 157 0.01 7 35 46 41 32 155 0.00 8 35 46 41 35 157 0.00 9 36 47 42 35 161 0.01 10 35 46 41 33 156 0.00 11 35 47 42 33 159 0.01 12 35 47 42 35 160 0.01 13 35 47 44 33 160 0.00 14 35 46 42 35 159 0.02 15 35 47 42 33 159 0.01 16 36 47 41 33 159 0.00 17 35 47 42 33 159 0.01 18 35 47 41 33 157 0.00 19 35 47 42 35 160 0.00 20 35 47 44 32 159 0.01 21 35 47 42 33 159 0.01 22 35 47 42 35 160 0.01 23 35 49 42 35 161 0.00 24 35 49 42 33 160 0.01 25 36 47 41 33 159 0.01 26 35 47 41 33 157 0.01 27 35 47 42 35 160 0.00 28 36 47 41 33 159 0.01 29 35 47 42 32 157 0.01 30 35 49 41 35 160 0.00 31 35 49 42 33 160 0.01 32 35 47 42 36 161 0.01 AVG 35.1 47.2 41.7 33.7 158.8 0.007 * J,K,L,and M refer to the four load cells. 12 in the worst case. The accuracy of the ISB is taken to be i 5.5 lbs for static weight measurements. Table 1 also shows the 32 successive acceleration readings. The numbers represent sampling errors coming from the external accelerometer because there was no movement so the accelerations should be zero. Therefore, the external accelerometer readings have an inherent precision of i 0.02 G. 3.1.3 Procedure for Checking the ISB Accuracy for Dynamic Force Measurement The next check on the ISB accuracy was to estimate the error in measuring dynamic forces as opposed to static forces. This can be done in two ways. One way would be to use an external accelerometer attached to a mass which is secured to the top of the ISB. Through the use of Newton's law which relates forces to accelerations, an estimate of the dynamic measurement error of the ISB can be obtained, provided of course that the ISB is able to accurately record external acceleration. Since it is not known ahead of time that external acceleration is measured accurately, a different method was used to estimate dynamic measurement error. The application of Newton's law will then be used to estimate the error in the exact acceleration. The ability of the ISB to accurately measure dynamic forces can be evaluated by finding out how fast it responds. To do this, the natural frequency of the shear beam mechanism must be measured because the shear beam acts like a spring—mass system. If its own natural frequency is large, then it responds quickly and is very 13 accurate for measuring dynamic events. As long as these events do not occur over time periods on the order of 1/fn where fn is equal to the natural frequency of the shear beam, then dynamic events will be recorded accurately. If the event occurs over very short time periods, then the ISB will report M times the true value: ISB reading = M x (True force) where M is the magnification factor(Brandenburg and Lee, 1985) defined by M = 1 / t 1 - (rev... / fsh... beam 3-1 The natural frequency of the shear beam (Timoshenko, 1974) can be estimated using. fn = 3.515 ((EIG/DAL4) / 2n 3.2 where E modulus of elasticity = 30,000,000 lbs/in2 for steel. D = weight density = 0.284 lb/in3 for steel I = moment of inertia = bh3/12 = 0.203 in4 (b=h=1.25 in) A = cross-section area = bh = 1.56 in2 L = beam length = 3 in G = acceleration due to gravity = 386.4 in/sec2 Using these values in equation 3.2 gives fn = 4530.1 sec"1 = 4530.1 Hz Because the vibration table will be operated at frequencies characteristic of truck trailers (around 3 Hz), the magnification factor from equation 3.1 is 14 M = 1 / [1 - (3/453o.1)21 = 1.000000439 Therefore, the error due to dynamic effects is around 0.00004%, which is negligible. Based on this result, the ISB should have no problem measuring dynamic forces. The error in the force reading should be the same as that for static forces. The ISB was next tested by itself for its ability to measure external accelerations. An external accelerometer with a sensitivity of 1 mv/g and a Piezotron coupler (Kistler 5112) were used to check the ISB. A metal brick with the accelerometer attached to it was placed on top of the ISB as in Figure 2 and the table was set to vibrate at 4 Hz and 0.5 G. The ISB was set to sample at 20 Hz. The compression forces at each corner of the ISB and the acceleration of the brick were also measured simultaneously. According to Newton's second law (Figure 3), the relationship between the dynamic force F and the acceleration G should be approximately F=(1+G)W. The relationship F=W(1+G) is not exactly correct because the weight of the platform is not accounted for. In a static weight measurement, the true static force on the shear beams is: F = Wb + Wp where F is the sum of four corner shear beam forces up on platform, Wb is the weight of brick, and Wp is the weight of ISB platform. But, the ISB does not read this F value because the weight of the platform was zeroed out during calibration. Instead, it reads just Wb or F-Wp. In a 15 r______—— Accelerometer Fixed Metal Brick ISB (f/’- E] U? 4_] L___.l | Electrohydraulic Vibration Table l ;::}—- Seismic Base Coupler Figure 2. Test Setup for Dynamic Force Relative to Acceleration 16 Wb 2F = F-Wb-Wp = WbG I Wp ISB Reading = F - Wp = (1+G)Wb+WpG F Figure 3. Newton's Second Law 17 or F - (Wb+Wp) = (Wb+Wp)G F = (Wb+Wp)(1+G) This force (F) is the true force down on the shear beams, but this is not what the ISB reads because the weight of the platform was zeroed out. The ISB always reads an amount Wp less than the true force. In other words, ISB reading = F — Wp = (Wb+Wp)(1+G) - Wp = Wb(1+G) + WpG 3.3 where Wb(1+G) is the static weight of brick plus force to move brick and WpG is the force to move platform. The correct external acceleration can therefore be calculated in the following manner: G = (ISB force reading - Wb) / (Wb + Wp) 3.4 Based on equation 3.4, Table 2 shows measured accelerations, measured forces, and actual accelerations taken over 32 successive peaks from the raw data. The measured accelerations and the actual accelerations averaged 0.4 G and 0.38 G respectively. From a statistical point of view, these two accelerations can be regarded as nearly the same value because they have large standard deviations. The data in Table 3 was measured in the same way as the Table 2 but with a piece of expanded polyethylene cushioning between the metal brick and the ISB. This was done to measure the effect of cushioning material on measuring dynamic compression levels and external accelerations. Table 3 shows 18 Table 2. External Acceleration Predicted from Measured Force for Metal Brick Measured Measured Actual No Peak Acc. Force(1bf) Peak Acc. ( G ) (F-Wb) / (Wb+Wp) 1 0.6 87 0.46 2 0.3 73 0.24 3 0.4 85 0.43 4 0.5 86 0.45 5 0.1 75 0.27 6 0.1 73 0.24 7 0.2 75 0.27 8 0.3 73 0.24 9 0.6 87 0.46 10 0.3 87 0.46 11 0.3 62 0.07 12 0.4 86 0.45 13 0.1 74 0.26 14 0.1 75 0.27 15 0.4 85 0.43 16 0.1 73 0.24 17 0.6 86 0.45 18 0.3 86 0.45 19 0.4 85 0.43 20 0.7 88 0.48 21 0.6 87 0.46 22 0.6 88 0.48 23 0.4 86 0.45 24 0.5 86 0.45 25 0.6 87 0.46 26 0.5 83 0.40 27 0.4 71 0.21 28 0.5 85 0.43 29 0.5 85 0.43 30 0.5 85 0.43 31 0.6 86 0.45 32 0.3 83 0.40 AVGiSTD 0.430.175 81.7i6.65 0.38:0.11 * The table was set at 0.5 G and 3H2. The ISB was set to sample at 20Hz. 19 Table 3. External Acceleration Predicted from Measured Force for EPE Measured Measured Actual No Peak Acc. Force(1bf) Peak Acc. ( G ) (F-Wb) / (Wb+Wp) 1 0.5 92 0.54 2 0.5 92 0.54 3 0.5 90 0.51 4 0.6 94 0.57 5 0.6 91 0.52 6 0.6 89 0.49 7 0.5 94 0.57 8 0.7 90 0.51 9 0.5 92 0.54 10 0.5 91 0.52 11 0.5 91 0.52 12 0.5 92 0.54 13 0.4 94 0.57 14 0.6 82 0.38 15 0.6 93 0.57 16 0.4 88 0.48 17 0.4 81 0.38 18 0.7 91 0.52 19 0.5 97 0.62 20 0.3 92 0.54 21 0.7 92 0.54 22 0.6 89 0.49 23 0.4 94 0.57 24 0.4 93 0.56 25 0.4 92 0.54 26 0.6 92 0.54 27 0.5 91 0.52 28 0.4 92 0.54 29 0.5 93 0.56 30 0.5 91 0.52 31 0.8 92 0.54 32 0.8 92 0.54 AVGiSTD 0.53:0.12 91.2i3 0.53:0.05 * Same as Table 2 but with a expanded polyethylene foam between the metal brick and the ISB. 20 that the measured and actual accelerations are the same on average. On the basis of the results in these two tables, it can be concluded that the external accelerometer measures accelerations accurately. Since the relationship between force and acceleration must be true at every instant, it should not matter what the sampling rate of the ISB is. Sampling rate does matter however when the ISB is used to measure such things as peak force and peak acceleration. Therefore, an estimate of the error in measuring peak values will be made because it affects what follows. A low sampling rate can cause the ISB to miss the peak. The worst possible situation is when two consecutive samples are on either side of the true peak as shown in Figure 4. A perfect sine wave has been used to provide an estimate even though a true sine wave was never recorded. Because the ISB and the brick were not rigidly connected to the table, the motion was smooth enough to be considered close to that of a true sine wave. The force varies over time according to Force = A sin (2nt/T) 3.5 where A is the true amplitude, t is the time, and T is the period. The reported peak is just the sine wave evaluated at t = T/4 — At/2 where At is the time between samples: Reported Peak = A sin [2n(T/4 - At/2) / T] A sin (”/2 - ”At/T) II 3’ cos (”At/T) Acceleration (G) True Peak Measured Peak Sampling Rate Error Time Figure 4. True Peak vs Measured Peak 22 The maximum percent error is defined here to be (true - reported)x100 / reported, or % error = [A - A cos(nAt/T)]x100 / A cos(nAt/T) = 100[sec(nAt/T)-1] = 100[sec(nfv/fs)-1] 3.6 where fv is the frequency of vibration of the table, and fs is the ISB sampling rate. Using a typical table frequency of 4 Hz and the maximum ISB sampling rate of 200 Hz, the maximum error in measuring the peak is: Maximum error = 100[sec(n4/200)= 0.2% This means that the true peak may be as much as 0.2% higher than the reported peak, but only if the table vibrates at 4 HZ and the ISB samples at 200 Hz. Finally, the external accelerometer was used to check the vibration table. The table was set to vibrate at 4 Hz and 0.8 G peak and the ISB was set to sample at 200 Hz. The results of the external accelerometer are shown in Figure 5. The average measured acceleration was 0.82 G peak which differs from the input acceleration level by only 2.5%. 3.2 Compression Strength Loss after Vibration : 3.2.1 Materials Two kinds of single wall "C" flute corrugated board and a "B+C" flute double wall corrugated board were used to make the test boxes for this study. All of the containers were RSC style boxes with the same external dimensions of 19.1x13.4x9.1 $0.3 inches (length x width x height). According to the type Acceleratzioxn-(Ci) C) Accelerometer ‘ _ I Fixed to table 1 LI] Electrohydraulic Vibration Table I—._.__d— Seisnic Base Coupler I“ 7-1 la III 1 IIII i- '1 f QII ;d I .r 'i I II 7” ‘l I‘ “‘1 ii III F! 1 II II 1 II I=fliI ii. ,i’i '~. iii 2) u.“ . 7-“ ”1‘qu IE7! . I; I 55 ms H 10 2.0 30 40 50 60 70 80 90 100 110 Random Samples from Portion of Sine Wave Figure 5. Test Setup and Results 24 of corrugated board used for the box, the boxes were designated as SW-A, SW-B, and DW which stand for single wall- board A, single wall-board B, and double wall. The specifications shown in Table 4 were taken from the manufacturer's labels shown in Figure 6. According to Rule 41 (Fibre Box Handbook, 1989), the gross weight in Figure 4 represents the maximum box weights for which no visual damage should occur. Bursting strength tests were done on samples of the board to check the manufacturer's specifications. The bursting test results using TAPPI T 810 0m-85 for ten samples are shown in Table 5. Table 4. Corrugated Box Specifications (Fibre Box Handbook) SW-A SW-B DW Bursting Test 200 275 350 (lbs/inz) Min Comb Wt Facings 84 138 126 (lbs/1000ft2) Size Limit 75 90 100 (L + w + H) (inches) Gross Weight 65 90 120 (lbs) * SW-A stands for Single Wall Board A, and SW-B for Single Wall Board B, and DW for Double Wall Board. ON m£¥, ; .«..._.:msienr CtASfiLFIQATIQ if»; -. BURSTING; veg-$50 'igggsnsgg 1: 'ii’i seat... 426%?“ ' ‘25 25 unn’iloo’ififu ‘ :B‘souiaguz‘fifl rsor’m Asia regicmicussmcmgu ‘tBS PE: 0 INC III? ffélfifi84%‘m "I: Figure 6. Box Specifications Label 26 3.2.2 Static Loads Three different package weights were examined in this study. The first was 35 lbs and consisted of five house bricks tightly packed with two pieces of expanded polyethylene foam on opposite long sides with crumpled newspapers put in between all void spaces to prevent movement during vibration. The second was 52.5 lbs with eight house bricks and the third Table 5. Measured Burst Strength (lbs/inz) 327 347 346 323 317 323 291 293 317 l 2 3 4 5 6 7 8 9 194 265 357 10 187 315 300 AVG i STD 191.6:12.47 314.6:22.12 318.6i21.67 was 71 lbs with eleven house bricks. These package weights were chosen to reproduce the most common pallet load weight range of 560 lbs to 1420 lbs. These box weights exceed the Rule 41 gross weight recommendations in Table 4 only for the SW-A box with the 71 lb load. 27 3.2.3 Stacking Patterns Pallet loads of these three boxes were constructed in different stacking_arrangements. The various arrangements or pallet patterns are shown in Figure 7, Figure 8, and Figure 9. Each pallet pattern had four boxes per layer with four layers with the exception of the interlock pattern which had five boxes per layer. Two identical pallet loads of boxes were used so that a total of seven bottom boxes could be used for compression testing after vibration: the eighth box was unavailable because it was replaced by the ISB so that dynamic forces could be recorded. In the case of interlocking pattern where nine boxes were available, only seven were chosen so that comparisons between the other two pallet patterns could be made on a uniform basis. All boxes within a pallet load had the same weight and were made of the same corrugated board. For the SW-A board then, there were two identical pallet loads for each of the three types of pallet patterns and each of the three box weights: SW-A : 2 pallet loads of 35 lb boxes in the column type. N pallet loads of 35 lb boxes in the interlock type. N pallet loads of 35 lb boxes in the pinwheel type. N pallet loads of 52.5 lb boxes in the column type. N pallet loads of 52.5 lb boxes in the interlock type. N pallet loads of 52.5 lb boxes in the pinwheel type. N pallet loads of 71 lb boxes in the column type. Figure 7. 28 .mm‘ SIN \“\§mwnf3 \\\ . "ms \ >>o.\>\\\\‘ “‘ \' W). N v \ y i?” Column Type Pallet Pattern 29 sat-4N *th t\.\\‘\§\ __ I ...... , __.-\\ _==' x___. : {\‘{\\ \u . ‘- . ( u . \ - \.‘\ \ n . ...‘w‘ \\\‘\§\ ~ \\\N\\\\\ \\\ .\\\ Figure 8. Interlock Type Pallet Pattern 30 Side u \n\\u\u \ .\\\m\\. ... ~.\. w .\\ «fin - \ ‘ \\ \_ “RNA? “mm \ Figure 9. Pinwheel Type Pallet Pattern 31 2 pallet loads of 71 lb boxes in the interlock type. 2 pallet loads of 71 1b boxes in the pinwheel type. SW-B : same as SW-A. DW : same as SW-A. Because there were four layers of either 35, 52.5, or 71 lb boxes in any pallet configuration, there were always three boxes on top of each bottom box. This means that the static compression load on a bottom box was either 3x35=105 lbs, 3x52,5=157.5 lbs, or 3x71=213 lbs, depending on the box weight used to make the pallet load. Based on the maximum dynamic load factor of 5 found in an earlier study (Singh and Lienberger, 1992), the dynamic compression load on a bottom box could therefore be as high as 5x105=5751bs, 5x157.5=787.5, or 5x213=1065 lbs, depending on the box weight used to reproduce the pallet load. As the compression test results in the next chapter will show the compression strengths of new boxes made from SW-A, SW-B, and DW board are around 800 lbs, 1000 lbs, and 1300 lbs respectively. The dynamic loads on the bottom boxes can therefore be greater than the static compression strengths of the SW—A.and SW-B boxes. Even though the individual box weights were not intentionally chosen to make this happen, it is fortunate that it did turn out this way because it raises another important question: will a dynamic load which is greater than the static compression strength but which.is presently only for short time periods on a reported basis cause the box to fail. This will be addressed 32 in the next chapter. 23.2.4 Vibration Tests Vibration tests on whole pallet loads were performed on an MTS 840 electrohydraulic vibration table and a Lansmont Model 10000-10 Touchtest Vibration System. A sine sweep from 3 to 100 Hz at 0.5 G peak was used to look for resonance. The bottom boxes were packed on all four sides with expanded polyethylene foam to prevent the collapse of stacked boxes during vibration. These boxes were placed at the bottom layer on the vibration table and three layers boxes loaded with house bricks were put on the top of the bottom boxes. Pallets were not used and all boxes were conditioned at 73°F and 50 % RH for 24 hours prior to testing. Guide rods with wooden boards fixed to them were attached to the table as restraining devices to prevent the specimens from horizontal movement off the platform, and to prevent toppling and excessive rocking. The restraining devices were adjusted to permit free movement of the test specimens to within 10 mm (0.4 in.) in any direction. The test setup is shown in Figure 10. The vibration tests were performed for approximately 15 minutes at the resonant frequency of the pallet load using a table acceleration of 0.5 G's in accordance with ASTM D-999. 3.2.5 Compression Strength Tests The compression strength tests were performed before and after the vibration test according to ASTM D 642-90: 33 Determining Compressive Resistance of Shipping Containers, Components, and.Unit Loads. For these tests, A Lansmont.Model 1210-AF Compression Tester was used. The boxes were maintained at 73°F and 50% RH following the vibration test, and compression tested under the same conditions. 3.2.6 A Related Test An accelerometer was attached to the middle box in the column of three boxes resting on top of the ISB. The accelerometer was taped.to the side of box. The middle box was chosen because it represents the mass center of the column of boxes on the ISB. The output of the accelerometer was fed into the ISB so that it could measure acceleration and force simultaneously to see if there was in fact the expected relationship between the two. The same box weights of 35 lbs, 52,5 lbs, and 71 lbs used in the pallet load vibration tests were used in this test. The test setup is shown in Figure 11. 34 Stacked Packages Restraining Device Blectrohydraulic Vibration Table Seismic Base Figure 10. Test Setup for Vibration 35 Restraining Device Stacked Packages Accelerometer raped Blectrohydraulic Vibration Table Seismic Base Figure 11. Test Setup for Comparing Forces and Accelerations 4.0 DATA AND RESULTS There were two different kinds of tests performed in this study: 1)compression tests on bottom boxes in a pallet load after vibration t resonance for 15 minutes, and 2) simultaneous force and acceleration measurements on.aicolumn of boxes. The results of the force and accelerations will be presented first. 4.1 Dynamic Compression Level vs Acceleration : 4.1.1 Package Weights The data in Table 6 shows the relationship between the actual force measured by the ISB and the predicted force using the measured acceleration and Newton's law. A box weight of 52.5 lbs was used to produce a static load of 157.7 lbs on the bottom box in a stack of three (shown in Figure 11). For this test, the vibration table was set at 3 Hz and 0.5 G and the ISB was set to sample at 20 Hz. No bouncing was found during'the test" Measured forces and predicted forces averaged 230 lbs and 204 lbs respectively. From the statistical point of view, these two forces can be also regarded as a same value because they have big standard deviations. The data in Table 7 was measured in the same way as Table 6 but with a different static load of 105 lbs. The measured forces and predicted 36 37 forces averaged 158 and 142 lbs respectively. This shows that measured forces weighed around 12 lbs more than that of predicted one. The data in Table 6 and 7 show that dynamic forces can similarly be predicted by measuring acceleration and applying Newton's law. An obvious consequence of this is that dynamic force measurements can be made without the need for an expensive instrument like the ISB. But the tests so far have been done in such.a way that no bouncing of boxes on top of each other took place. At resonance, where bouncing will occur, there may not be such a relationship between force and acceleration. 38 Table 6. Measured Forces vs Predicted Forces for a 157.5 lb Static Load Successive Measured Predicted No Peak Acc. Force(1bf) Force(1bf) (1 + c) Wb(1+G)+WpG 1 1.18 209 187 2 1.29 237 205 3 1.22 244 194 4 1.19 212 187 5 1.34 241 213 6 1.26 246 200 7 1.29 209 205 8 1.38 240 220 9 1.26 244 200 10 1.28 206 203 11 1.35 238 215 12 1.26 244 200 13 1.28 207 203 14 1.33 238 211 15 1.31 246 208 16 1.33 208 211 17 1.41 240 225 18 1.18 225 187 19 1.28 207 203 20 1.41 236 225 21 1.32 248 210 22 1.37 236 218 23 1.33 242 211 24 1.30 204 207 25 1.43 238 225 26 1.35 244 215 27 1.34 207 213 28 1.47 240 235 29 1.33 241 211 30 1.34 203 213 31 1.43 235 228 32 1.37 242 218 AVGiSTD 1.32 230i16.1 204i35.7 39 Table 7. Measured Forces vs Predicted Forces for a 105 lb Static Load Successive Measured Predicted No Peak Acc. Force(1bf) Force(1bf) (1 + c) Wb(1+G) +WpG 1 1.47 163 157 2 1.35 159 144 3 1.40 154 149 4 1.42 165 152 5 1.39 159 148 6 1.42 150 152 7 1.45 163 155 8 1.38 159 147 9 1.37 153 146 10 1.41 166 150 11 1.32 153 140 12 1.39 165 148 13 1.33 162 142 14 1.34 149 143 15 1.38 163 147 16 1.35 160 144 17 1.34 154 143 18 1.38 163 147 19 1.33 161 142 20 1.37 148 146 21 1.39 162 148 22 1.37 160 146 23 1.31 149 139 24 1.39 166 148 25 1.32 160 140 26 1.33 150 142 27 1.39 161 148 28 1.32 159 140 29 1.34 149 143 30 1.40 163 149 31 1.39 161 148 32 1.33 146 142 AVGiSTD 1.37 158:5.9 146:4.3 40 As the table frequency approaches the natural frequency of the stack of boxes during the sweep, the value of the vibration magnification factor M increases; M = 1 / [1 - (Ft/Fn)2] where Ft stands for the table frequency and Fn for natural frequency. At the resonance where Ft/Fn = 1, the value of M becomes infinite. This means that a very small input 'vibration will cause an extremely large response: in other words, the stack of boxes will bounce and separation will occur. To test the relationship between force and acceleration at resonance, a static load of 213 lbs on the bottom box was used. The vibration table frequency was adjusted so that the stack resonated. The measured forces and predicted forces averaged 402 lbs and 560 lbs respectively (Table 8). The data in Table 9 was measured in the same way as Table 8 but with a different static load of 105 lbs and a different vibration table frequency of 7.5 Hz. The measured forces and predicted forces averaged 184 lbs and 269 lbs respectively. These results show that the dynamic forces can not be predicted by measuring accelerations when there is separation between individual boxes. The reason is that Newton's law requires the acceleration to be measured at the mass center of the stack of boxes. When there is separation, the mass center is no longer located at the middle box. 41 Table 8. Measured Forces vs Predicted Forces for a 213 lb Static Load at Stack Resonance. Successive Measured Predicted No Peak Acc. Force(1bf) Force(1bf) (1 + G) Wb(1+G)+WpG l 2.30 405 497 2 2.99 404 648 3 2.76 404 598 4 2.04 401 441 5 2.75 402 596 6 2.23 401 482 7 1.75 399 377 8 1.92 405 414 9 1.82 401 392 10 2.56 406 554 11 4.88 400 1062 12 3.20 401 694 13 3.15 406 683 14 2.80 405 607 15 3.53 400 766 16 1.85 406 399 17 1.57 396 338 18 1.92 400 414 19 2.61 396 565 20 1.85 401 399 21 2.57 401 556 22 3.15 402 683 23 2.94 404 637 24 4.12 402 896 25 3.01 404 653 26 2.31 398 500 27 2.93 402 635 28 2.77 396 600 29 1.83 402 395 30 2.70 402 585 31 1.63 401 351 32 2.38 400 515 AVGiSTD 2.59 402:2.8 560i160.3 42 Table 9. Measured Forces vs Predicted Forces for a 105 lb Static Load at Stack Resonance Successive Measured Predicted No Peak Acc. Force(1bf) Force(1bf) (1 + G) Wb(1+G)+WpG 1 2.80 233 304 2 2.69 217 292 3 2.46 213 267 4 2.62 195 284 5 2.45 188 266 6 2.46 169 267 7 2.41 176 261 8 2.79 185 303 9 2.87 186 312 10 2.55 189 277 11 1.71 190 184 12 2.48 181 269 13 2.64 159 287 14 2.31 127 250 15 2.49 226 270 16 2.44 203 264 17 2.99 192 325 18 2.40 167 260 19 2.52 172 273 20 2.52 188 273 21 2.24 185 242 22 2.53 184 274 23 2.26 189 245 24 2.54 180 276 25 2.34 157 253 26 2.55 235 277 27 2.02 83 218 28 2.61 212 283 29 2.50 190 271 30 2.11 167 228 31 2.61 174 283 32 2.36 184 256 AVGiSTD 2.48 184i29.0 269i26.8 43 4.2 Compression Strength Loss : The seven bottom boxes in the two pallet loads of identical boxes were compression tested after 15 minutes of vibration at resonance. The results are shown in Table 10-18. Each table shows the compression strength results for a particular pallet pattern and board type. 4.2.1 Corrugated Box (SW-A) Table 10 shows the percent compression strength loss of the SW—A corrugated box in the column type stacking pattern. Thgflggmpressionstrength of the box before vibration averaged “EESLEMlbs and the compression strengths after vibration at resonance for 15 minutes with a stack load of 105, 157.5 and 213 lbs on top averaged 666.7, 650.1, and 391.1 lbs respectively. The percent loss in compression strength using the 806.9 lb value as a reference is therefore 17.3%, 19.4%, and 51.5% respectively on average. Table 11 shows the percent compression strength loss of the SW-A corrugated box in the interlock type stacking pattern. The loss in compression strength of the SW-A averaged 17.7%, 18.3%, 19.9% for static loads of 105 lbs, 157.5 lbs, and 213 lbs respectively. Table 12 shows the percent compression strength loss of the SW-A corrugated box in the pinwheel type stacking pattern. The loss in compression strength of the SW-A averaged 0.0%, 20.7%, 22.5% for static loads of 115 lbs, 157.5 lbs, and 213 lbs respectively. The results in Table 10, 11,and 12 for the average 44 compression strengths are graphed in Figure 12. It would appear that the greater the static load, the greater the loss in compression strength. The specific loss did not seem to be related to the type of pallet pattern until the static load exceeded 157.5 lbs. Beyond 157.5 lbs, the column type stacking pattern had the highest percent loss of compression strength. 45 Table 10. Column Type Stacking Pattern : SW-A ' Total Static Load (lbs) No 0 105 157.5 213 1 666 *959 548 478 2 762 751 618 330 3 904 548 659 398 4 656 591 656 375 5 815 707 709 437 6 856 721 672 378 7 789 684 692 342 Average Compression 806.9 667.7 650.1 391.1 Strength lb STD 150.4 79.8 53.1 52.1 Resonance No vib. 7.5Hz 7.2Hz 6.5Hz * not included in the average. 46 Table 11. Interlock Type Stacking Pattern : SW-A Total Static Load (lbs) No 0 105 157.5 213 1 666 *1021 666 *1039 2 762 669 692 642 3 904 516 747 501 4 656 742 800 659 5 815 738 609 698 6 856 627 465 641 7 789 692 634 736 Average Compression 806.9 664.0 659.0 646.2 Strength lb STD 150.4 84.8 107.6 80.0 Resonance No vib. 7.3Hz 7.0Hz 6.3Hz * not included in the average. 47 Table 12. Pinwheel Type Stacking Pattern : SW-A Total Static Load (lbs) No 0 105 157.5 213 1 666 703 *1057 *1029 2 762 654 604 686 3 904 760 578 695 4 656 789 636 428 5 815 695 713 712 6 856 586 639 636 7 789 *1057 671 594 Average Compression 806.9 711.2 640.2 625.2 Strength lb STD 150.4 81.0 47.9 105.9 Resonance No vib. 7.5Hz 7.2Hz 6.5Hz * not included in the average. 48 110 8 90 8O 7O 60 50 Percent Compression Strength Remaining 40 ‘ ‘ ‘ ‘ O 50 100 150 200 250 Static Load (lbs) ... Column Type ... Interlock Type .... Pinwheel Type Figure 12. Mean Percent Compression Strength Remaining : SW-A 49 4.2.2 Corrugated Box (SW-B) Table 13 shows the percent compression strength loss of the SW-B corrugated box in the column type stacking pattern. The compression strength of the box before vibration averaged 1068.6 lbs and the compression strengths after vibration at resonance for 15 minutes with a stack load of 105, 157.5 and 213 lbs on top averaged 1011.3, 987.9, and 886 lbs respectively. The percent loss in compression strength using the 1068.6 lb value as a reference is therefore 0.0%, 0.0%, and 17.1% respectively on average. Table 14 shows the percent compression strength loss of the SW-B corrugated box in the interlock type stacking pattern. The loss in compression strength of the SW-B averaged 7.9%, 11.8%, 13.8% for static loads of 105 lbs, 157.5 lbs, and 213 lbs respectively. Table 15 shows the percent compression strength loss of the SW-B corrugated box in the pinwheel type stacking pattern. The loss in compression strength of the SW-B averaged 0.0%, 12.1%, 16.2% for static loads of 115 lbs, 157.5 lbs, and 213 lbs respectively. The results in Table 13, 14,and 15 for the average compression strengths are graphed in Figure 13. It would appear that the greater the static load, the greater the loss in compression strength. The specific loss did not seem to be related to the type of pallet pattern. But with a static load of 157.5 lbs, the column type stacking pattern.had the highest percent loss of compression strength. 50 Table 13. Column Type Pallet Pattern : SW-B Total Static Load (lbs) No 0 105 157.5 213 1 1051 1145 929 783 2 1083 1059 921 847 3 1057 1038 1027 1088 4 1113 1056 762 1086 5 1086 959 1130 826 6 1106 918 1057 792 7 999 904 1089 780 Average Compression 1068.6 1011.3 987.9 886.0 Strength lb STD 38.4 87.4 126.3 131.3 Resonance No vib. 7.5Hz 7.2Hz 6.5Hz 51 Table 14. Interlock Type Pallet Pattern : SW-B Total Static Load (lbs) No 0 105 157.5 213 1 1051 886 851 904 2 1083 898 1033 991 3 1057 906 932 951 4 1113 1045 995 862 5 1086 994 891 932 6 1106 1110 950 918 7 999 1045 943 891 Average Compression 1068.6 984.4 942.1 921.3 Strength lb STD 38.4 88.0 60.7 42.1 Resonance No vib. 7.3Hz 7.0Hz 6.3Hz 52 Table 15. Pinwheel Type Pallet Pattern : SW-B Total Static Load (lbs) No 0 105 157.5 213 1 1051 1117 874 959 2 1083 1109 1041 959 3 1057 998 810 845 4 1113 1026 1039 854 5 1086 979 966 885 6 1106 1012 901 864 7 999 914 945 900 Average Compression 1068.6 1022.1 939.4 895.1 Strength lb STD 38.4 71.6 85.1 47.4 Resonance No vib. 7.5Hz 7.2Hz 6.5Hz 53 110 8 \O O 0) q 00 O O O I I I Percent Compression Strength Remaining Vt CD I 1 J 1 1 A o 0 so 100 150 200 250 Static Load (lbs) ... Column Type ... Interlock Type + Pinwheel Type Figure 13. Mean Percent Compression Strength Remaining : SW-B 54 4.2.3 Corrugated Box (DW) Table 16 shows the percent compression strength loss of the DW corrugated.box:in the column type stacking pattern. The compression strength of the box before vibration averaged 1358.1 lbs and the compression strengths after vibration at resonance for 15 minutes with a stack load of 105, 157.5 and 213 lbs on top averaged 1432.6, 1300.1, and 1407.3 lbs respectively. The percent loss in compression strength using the 1068.6 lb value as a reference is therefore 0% for all three stack loads on average. Table 17 shows the percent compression strength loss of the DW corrugated box in the interlock type stacking pattern. The loss in compression strength of the DW averaged also 0% for three static loads of 105 lbs,157.5 lbs, and 213 lbs. Table 18 shows the percent compression strength loss of the DW corrugated box in the pinwheel type stacking pattern. The loss in compression strength of the DW averaged 0% again for three static loads of 115 lbs, 157.5 lbs, and 213 lbs. The results in Table 13, 14,and 15 for the average compression strengths are graphed in Figure 14. It would appear that the compression strength of DW corrugated board box is not affected by the stacking pattern and vibration test used. Table 16. Column Type Pallet Pattern 55 : DW Total Static Load (lbs) No 0 105 157.5 213 1 1259 1400 1471 1364 2 1288 1529 1271 1377 3 1368 1340 1332 1505 4 1257 1424 1409 1418 5 1210 1383 1034 1336 6 1589 1448 1210 1395 7 1536 1504 1374 1456 Average Compression 1358.1 1432.6 1300.1 1407.3 Strength lb STD 148.3 66.8 145.7 57.8 Resonance No vib. 7.5Hz 7.2Hz 6.5Hz 56 Table 17. Interlock Type Pallet Pattern : DW Total Static Load (lbs) No 0 105 157.5 213 1 1259 1217 1378 1300 2 1288 1373 1235 1649 3 1368 1333 1030 1306 4 1257 1362 1385 1327 5 1210 1285 1362 1326 6 1589 1471 1241 1381 7 1536 1526 1471 1443 Average Compression 1358.1 1366.7 1300.3 1376.0 Strength lb STD 148.3 105.3 145.5 123.3 Resonance No vib. 7.3Hz 7.0Hz 6.3Hz 57 Table 18. Pinwheel Type Pallet Pattern : DW Total Static Load (lbs) No 0 105 157.5 213 1 1259 1377 998 1412 2 1288 1215 1514 1545 3 1368 1341 1388 1214 4 1257 1250 1301 1464 5 1210 1291 1500 1567 6 1589 1270 1409 1420 7 1536 1285 1361 1467 Average Compression 1358.1 1289.9 1364.4 1441.3 Strength 1b STD 148.3 54.5 186.9 116.0 Resonance No vib. 7.5Hz 7.2Hz 6.5Hz 58 110 60 .e a g 100M 2 4: 9o .— so a g 80 . e S .2 70 _ i. 2‘ 60 r 8 a g 50 — a2 40 l L 1 1 0 so 100 150 200 Static Load (lbs) ... Column Type ... Interlock Type ... Pinwheel Type 250 Figure 14. Mean Percent Compression Strength Remaining :DW 5.0 CONCLUSIONS Based on the test results, the following conclusions were reached : 1. The ISB has a static and dynamic measurement error of r 5.5 lbs and an external acceleration measurement error of i 0.02G. 2. The percent compression strength losses of SW-A, SW- B, and DW boxes as a function of the static load expressed as a percent of the box compression strength are described in Table 19 below. Table 19. Static load vs Compression Strength Loss Static loads Percent loss in compression strength as a percent due to different stacking patterns of compression strength Column Interlock Pinwheel 13.1% 17.3% 17.7% 0.0% 19.5% SW-A 19.4% 18.3% 20.7% 26.4% 51.5% 19.9% 22.5% 9.8% 0.0% 7.9% 0.0% 14.7% SW-B 0.0% 11.3% 12.1% 19.9% 17.1% 13.8% 16.2% 7.7% 11.6% DW No Loss 15.7% 59 60 3. According to the compression test results in Table 19 for SW-A and SW-B, the interlock type stacking pattern.has the maximum compression strength loss at the lightest package weight; the pinwheel type stacking pattern has the maximum compression strength loss at the intermediate package weight; and the column type stacking pattern shows the maximum compression strength loss at the heaviest.package weight. The column type pallet pattern produces the greatest loss in compression strength at higher static loads compared to the interlock and pinwheel patterns most likely because the bottom boxes are loaded.primarily'on their corners during vibration. Many of the corners of the bottom boxes in the pinwheel and interlock patterns contact the:middle:of the boxes above them (not the corners as in the column type pattern). This means that the corners of the bottom boxes are not abused as much which then leaves them stronger after vibration. Figures 7, 8, and 9 illustrate this. Since compression strength is primarily related to the strength of the corners(Maltenfort, 1988, 1989), greater loading of the corners during vibration is expected to result in a lower compression strength after vibration. 4. Because the dynamic forces on the bottom boxes never exceeded the compression strength of the box for any of the box types (SW-A, SW-B, and DW), no conclusion can be reached regarding the effect of dynamic forces on box failure. In other words, it is not known whether periodic dynamic forces 61 in excess of the static compression strength will cause a box to fail. 5. The use of stack.resonance tests as outlined in ASTM D 999 resulted in dynamic forces up to about 2.2 times the static loads at resonance. See Tables 6, 7, and 8. Earlier studies using ASTM D 4728 (Singh and Leinberger, 1992) produced dynamic forces up to 5 times the static loads using the composite truck spectrum. Therefore, ASTM.D 999 is not as severe a test as ASTM D 4728. 6. The Rule 41 recommendations on gross weight were found to be valid. The greatest static load placed on any of the boxes was 213 lbs. This was due to the three 71 lb boxes on top of it. According to Table 4, this box weight exceeded the Rule 41 recommendation of 65 lbs only for the SW-A box. The SW-A.boxes with 213 lb static load on top should therefore have shown the greatest amount of damage and presumably the greatest loss.in.compression strength" According to Figure 12, this was in fact the case. 7. Based on Newton's law, dynamic forces can be accurately predicted by measuring accelerations provided that there is no separation between the stack of boxes during vibration. The dynamic compression levels at resonance where the boxes separate and bounce can not be predicted. Therefore, the ISB can not be replaced by accelerometers during transit if resonance conditions are expected. List of References Antle, John Rodney. 1989. Measurement of Lateral and Longitudinal Vibration in Commercial truck Shipments. MSU: School of Packaging. ASTM. 1991. Selected ASTM Standards on Packaging. Philadelphia, PA: ASTM. Blatt, Frank . 1986. Principles of Physics. Massachusetts: Ally and Bacon Inc. Brandenburg, Richard K., Lee, Julian June-Ling. 1985. Fundamentals of Packaging Dynamics. USA: MTS System Co. Dority, Jack. 1989. Solving Distribution Damage Problems in the Laboratory. Journal of Packaging Technology. Vol. 3. No. 1:pp13-16. Fibre Box Association. 1989. Fibre Box Handbook. Rollong Meadows,IL: Fibre Box Association. Harris, Cyril M and Crede, Charles E. 1976. Shock and Vibration Handbook. USA: Mcgraw-Hill Inc. Maltenfort, George G. 1988. Corrugated Shipping Containers. Plainview, NY: Jelmar Publishing Co., Inc. Maltenfort, George G. 1989. Performance and Evaluation of Shipping Containers. Plainview, NY: Jelmar Publishing Co., Inc. Marcondes, Jorge Aparecido. 1988. Dynamic Analysis of a Less Than Truckload Shipment. MSU: School of Packaging. Singh, Paul and Leinberge, Dave. 1993. Identification of Methodology, Equipment, and Software necessary to Perform Environmental Data Acquisition of Dynamic Compression Values on Individual Packages in Shipment. MSU: School of packaging. Timoshenko, 8., Young, D.H., and Weaver, JR. W. 1974. Vibration Problems in Engineering (Forth Edition). New York: 62 63 John Wiley & Sons. Wiley. 1986. Mechanical Engineers' Handbook. New York: Wiley. Wilson, T Douglas. 1988. How Cushion Packaging Design Can Control Vibration Damage During Transport. Journal of Packaging Technology. Vol. 2. No. 3:pp106-108. APPENDIX 64 INSTRUMENTED SHIPPING BOX Technical Description of Instrumented Shipping Box : The instrumented shipping box consisted of the following main items, (these items were installed in a 13"width, 16" length by 10" tall box sized to approximate a standard shipping container). 1. One, 8-bit controller style microprocessor containing 64K of non-volatile static ram, a real-time clock calendar, both a synchronous and asynchronous serial port, and a lithium backup battery. 2. Four, 8-megabyte dynamic ram storage (mass storage) memory banks. These memory devices used synchronous serial to communicate with the microprocessor. 3. Two, 8-channel multiplexed twelve-bit analog to digital converters (A/D). The A/D used synchronous serial to communicate with the microprocessor. 4. Four, instrumentation amplifiers to process the signal from the sheer beams. 5. Four, 0-50001b. shear beam style load cells. These load cells were temperature compensated from 0 to 150 degrees fahrenheit, were 3 mV/V signal at full output, and used a 350 ohm bridge resistance. 6. One, 0-r50g monolithic capacitive accelerometer. The sensitivity was approximately 20 mV/V signal with a full frequency response from 0 Hz. to 500 Hz (3db down). The range of the accelerometer was limited to i 20gs. 65 7. Two, 4amp/hour 6 volt lead acid, gel type batteries. 8. A hexademical rotary input switch to control the mode of operation. Two push button switches to increment, decrement, or commence the function of a given mode. 9. A RS232 style serial port (9600 baud) to send or receive information from a PC computer. 10. Three power suppliers, one five volt supply for the digital circuitry, one eight volt supply for the analog circuitry, and one precision five volt supply for the A/D. A crowbar was used to protect the circuitry against overvoltage. 11. A built-in battery charge for the gel batteries. The batteries can.be recharged in approximately five hours. A two color led indicates the condition of charging (red=charging, green = full charge). The box could run for approximately 20 hours on a full charge. 12. A 16x2 full alpha-numeric liquid crystal display (LCD) was used to provide information to the box user. Explanation of Features : The four load cells were placed at each corner of the boxes lid. The lid was undercut so that it could float free of the box sides and was supported by the load cells. With this type of arrangement the forces acting on individual corners could be measured (0-50001b.) as well as the sum of the forces on the entire lid (200001b.). The box was constructed of 5/8th" plywood with the lid from 3/4" plywood. The maximum test force placed on the box (to date) was 3000+lb. The load 66 cells were supported on the internal steel box-beam frame that provided rigid support.and overload protection. The lid could be removed and a smaller lid version placed on an individual corner to test the accuracy of a given load cell. The microprocessor used an 11.0592 Mega-Hz. crystal resulting in a machine cycle time of 1.1Us. This crystal frequency provided both the highest processor speed and allowed the development of special timing for the asynchronous serial port. The real-time clock calendar provided a means of time/date stamping test runs. The lithium backup battery provided by the retention of microprocessors programs and data without the need for EPROM or EEPROM type memory. The bulk of the test data was stored in the mass storage devices. A means of testing the mass storage for stuck bits was provided. The LCD provided an independent means of viewing the lid weights, box temperatures, battery charge, and amount of mass storage filled. The two A/Ds provided a total of sixteen channels of 12- bit data acquisition. A/D #1 provided information from the four sheer beams, monitored the analog voltage supply, the 12V battery pack, and two temperatures (sheer beam and box ambient). A/D #2 provided information from the single axis accelerometer and would allow for future expansion. Due to the amount of time it took to access the A/Ds are saved the information in mass storage:a:maximum sampling time of 200 Hz. was achieved. For convenience and due to different testing 67 criteria the sample rate could be incrementally varied from 0.1 Hz to 200 Hz. To read and save data from one channel of an A/D required approximately 0.275 ms. To provide the most "instantaneous" reading of the A/D channels, they were red in a burst mode at the beginning of a sample cycle saving the data to the microprocessor to the mass storage. At 200 Hz the mass storage could be filled in approximately 21 minutes. A typical data gathering session would proceed as follows. The user turns the box on and lets it "warm-up" for five to ten minutes.During this time the sampling rate is set and at the end of this time a tare weight is taken. When the box is put in the data gathering mode a file heading is created that records the start time and date, the sample rate, the zero value for the accelerometer. As data is gathered it is saved, along with a check sum, in the mess storage. One sample of all the A/D channels is one packet and has a one-byte check sum. At the end of a data gathering session the number of packets is added to the header information. The instrumented box is turned.on a serial download.mode and connected to a PC to transfer the data. The header information provides the program.in the PC with the number of packets of information to receive, and necessary consultants to recreate the actual information. Storing raw data in the instrumented box allows for faster sampling time and more efficient use:of the mass storage. Providing a check sum with 68 each packet provided a means of checking data integrity. s .. RI 0 .. ...: . m 7 .. . a H m 9 . .. .NI8 ... . . . U m 0 ... m 3 _. H . .. sIIQ _, .:. ,7 HI3 . . .. .. m .t... \ .:.. 5:...