. . {(0.4 7t o 4% 3kg...» . .. . r .. r 4. . .2 .. :6. .uwunn . .. . v u... ”CH3? .4. 2.‘ ‘ I: t‘. .r tH'lllu‘ .> -114) .. .14. U1 II 3 .a Ill: 3' .v .. n. I?" V. n l ‘yf‘ Jawxnauflfimmwr 5.5‘ ‘ is? 1P2... u. ,. 2.,w..«...‘.9u 53m, in“- .mwén . ‘ I‘I; n .rglg ‘. «a émcfimfifiu V.!\y . . u. THEStS Illlllllllllml lllll‘l‘lll LIBRARY :mchigan State University This is to certify that the thesis entitled Relation Between Road Time and Vibration Table Simulation Time Based on Compression Strength of Corrugated Boxes presented by Sara Hilda Languet has been accepted towards fulfillment of the requirements for MS degree in Packagig 7}!)aui gum»: we JMajor professor Date February 6, 1997 0.7639 MS U is an Affirmative Action/Equal Opportunity Institution PLACE It RETURN BOX,” remove We checkout trom your record. TO AVOID FINES return on or‘bdore d'ete due. DATE DUE DATE DUE DATE DUE M%0W|l H l -L__] l l . l _____l _—l L__lL_l WET—T T% MSU le An Affirmative Action/Equal Opportunity Inflittllon RELATION BETWEEN ROAD TIME AND VIBRATION TABLE SIMULATION TIME BASED ON COMPRESSION STRENGTH OF CORRUGATED BOXES By Sara Hilda Languet A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTERS IN PACKAGING SCHOOL OF PACKAGING 1997 ABSTRACT RELATION BETWEEN ROAD TIME AND VIBRATION TABLE SIMULATION TIME BASED ON COMPRESSION STRENGTH OF CORRUGATED BOXES By Sara Hilda Languet Recently, there has been a great deal of concern among manufacturers regarding the latest procedures used to simulate the distribution environment and its effect on product damage. Package damage caused by vibration during three hours of actual road time on a twenty-four foot delivery truck and that caused by an equal amount of laboratory vibration time are not necessarily the same. The two were compared by measuring the compression strength of “C” flute corrugated boxes that have been subjected to both treatments. The laboratory vibration tests caused between 15% and 25% more damage than the actual road trip. The compression strength of boxes exposed to a three hour road trip was similar enough to the compression strength of boxes exposed to one hour of laboratory vibration, but not significant enough to establish a genuine correlation. These results suggest that the existing analysis, interpretation and use of vibration recordings for simulation might lead to product overpaclcaging. This study also intends to compare the effect of different data recorder control parameters on distribution simulation. Two Power Spectral Density profiles were developed using different acceleration sampling rates, acceleration trigger levels, pre-trigger lengths and post-trigger lengths for the same road trip. A vibration table was driven with these two plots using identical pallet loads of test boxes. The compression strength of these boxes in each pallet load was measured after a one hour vibration simulation. The results showed no significant difference in the PSD plots and no significant difference in the compression strength implying that recording parameters are not a deciding factor in determining the simulation environment. To my husband, Bert to my son, Jose to my grand parents, Dor‘ia Celsa and Don Carmelo to my parents in law, Aland Joyce and specially to my mom, Sarita I could not have done it without you. Thank you for your support, your understanding and your love. iii ACKNOWLEDGMENTS One page would not be enough to honor all the persons that contributed to the completion of this research. Immediately after my first day of graduate classes at the Packaging School, I started to look for a thesis topic. I appreciate the guidance of the Michigan State University Packaging faculty in helping during the selection of a topic that matched my academic background and future career goals. Dr. Gary Burgess, my major professor, presented to me the starting “idea” with its theoretical foundation. He is an extraordinary teacher. Thank you for your time, your effectiveness and your optimistic attitude through the logistics setup and documentation of this research. I appreciate the participation of Dr. Paul Singh, Dr. Diana Twede and Dr. Brian Feeny as a members of my committee. Thank you for your time and support through the development of the experiment methods and procedures. I submitted proposals (Appendix A) to many companies. Only a few responded positively. I wish to thank the Plant Manager of the International Paper Converter facility (Howell, MI), Mr. Mat Mattea, and his Packaging consultant, Dave Sullivan, for supplying over 300 “C” flute Regular Slotted Container corrugated boxes for this research. Also, I want to express my gratitude to Larry Bymes, Errol Bachelder, and Ken Grimes for facilitating the use of a delivery truck though one of their contract freighters. Additionally, I greatly value the assistance of Mr. Greg Hoshal, president of Instrument Sensor Technology (181'), and of his development engineer, Mat Busdiecker. IST provided free of charge the Electronic Data Recorders (EDRs), the Dyna Max software and technical support for this study. Also, I like to thank Dr. Hugh Lockhart for whom I performed pharmaceutical packaging research that allowed me to finance part of my tuition. I appreciated the opportunity even when it was not related to my major research topic. Last, but not least, I treasure the time and support of Dr. James Jay, Dr. Susan Selke and Dr. Arlis Wiggins. “Thank you” is not enough to express my gratitude for the honest advice in the realization of my academic, professional and personal goals. iv TABLE OF CONTENTS LIST OF TAB LES ............................................................... vi LIST OF FIGURES ................................................................ vii LIST OF SYMBOLS, ABBREVIATIONS, OR NOMENCLATURE ......................... viii CHAPTER 1 - INTRODUCTION AND LITERATURE REVIEW .......................... 1 1.1 An Overview ................................................ 1 1.2 Philosophy Behind Simulation ................................... 2 1.3 Nature Of The Distribution Environment ........................... 3 1.4 Established Packaging Test Methods For Vibration ................... 4 1.5 Results Of Previous Studies ..................................... 9 1.6 Objectives Of This Study ....................................... 9 CHAPTER 2 - MATERIALS, EQUIPMENT AND EXPERIMENTAL DESIGN ............. 11 2.1 Product/Package System Under Investigation. ...................... 11 2.2 Mode And Route Of Transportation. ............................. 14 2.3 Programming The Environmental Data Recorders ................... 15 2.4 The Vibration Table .......................................... 16 2.5 Damage Criteria ............................................. 18 CHAPTER 3 - DATA AND RESULTS .............................................. 19 3.1 The PSD Plots For The Road Trip ............................... 19 3.2 Compression Strength Results For The “Control” Pallet Load ......... 22 3.3 Compression Strength Results From The Actual Transportation Environment (Road Trip) ................ 24 3.4 Compression Strength Results From The Simulations on the Vibration Table ........................... 24 3.5 Compression Strength Results For The 60 Minute Simulations Comparing Panther And EDR—l Data ............................ 30 CHAPTER 4 - CONCLUSIONS ................................................... 33 4.1 Data from the Vibration Table .................................. 33 4.2 Correlation Between The Panther And The EDR-I For The Same Simulation Time ................................. 36 4.3 Final Remarks ............................................... 36 CHAPTER 5 —APPENDICES ........................................................ 38 , Appendix A - Research Proposal To Companies ........................... 39 Appendix B - Raw Data For The Environmental Data Recorders .............. 43 Panther Data ................................................ 44 EDR-l Data ................................................ 72 LIST OF REFERENCES .......................................................... 87 Table 1. Table 2. Table 3. Table 4. Table 5. Table 6. Table 7. Table 8. Table 9. Table 10. Table 11. Table 12. Table 13. Table 14. Table 15. Table 16. LIST OF TABLES Frequency ranges associated with various sources ................................ 4 Specifications for the sand .................................................. 11 Specification for the boxes ................................................. 11 Route of transportation .................................................... 15 “Panther" Recording Control Parameters ...................................... 17 EDR-l Recording Control Parameters ........................................ 17 PSD data used ........................................................... 22 Compression test results for boxes on the control pallet load ....................... 23 Compression test results for boxes on the actual road trip ......................... 25 Compression test results for the 30 minute simulation ............................ 26 Compression test results for the 60 minute simulation ............................ 27 Compression test results for the 90 minute simulation ............................ 28 Compression test results for the 180 minute simulation ........................... 29 Compression test results for the 60 minute simulation with Panther data. ............. 3O Compression test results for the 60 minute simulation with EDR-l data .............. 31 T-test results ............................................................ 34 vi Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Figure 10. Figure 11. Figure 12. Figure 13. Figure 14. Figure 15. Figure 16. Figure 17. Figure 18. Figure 19. Figure 20. Figure 21. LIST OF FIGURES PSD plot with le bandwidth tuner ........................................... 8 The box’s contents and arrangement .......................................... 12 The test pallet load ....................................................... 14 The twenty four foot truck used during study ................................... 14 The Environmental Data Recorders .......................................... 15 (a) Panther used during study ............................................... 15 (b) EDR-l used during study ................................................ 15 Lansmont vibration table (series 1800) ........................................ 18 PSD plot from the Panther .................................................. 20 PSD plot from the EDR-l .................................................. 21 Average compression strength by layer for the control pallet load ................... 23 Average compression strength by layer for the actual road trip ..................... 25 Average compression strength by layer for the 30 minute simulation ................. 26 Average compression strength by layer for the 60 minute simulation ................. 27 Average compression strength by layer for the 90 minute simulation ................. 28 Average compression strength by layer for the 180 minute simulation ................ 29 Average compression strength by layer for the 60 minute simulation with Panther data . . 31 Average compression strength by layer for the 60 minute simulation with EDR-l data. . . 32 Average compression strength by layer for the actual and the four simulated environments 33 Average compression strength versus vibration time for the bottom layer ............. 34 Average compression strength versus vibration time for the second layer ............. 35 Average compression strength versus vibration time for the third layer ............... 35 Average compression strength versus vibration time for the top layer ................ 35 vii AB.C.D.E,F ASTM EDR IP IST PSD RCP RMS G RSC LIST OF SYMBOLS, ABBREVIATIONS, OR NOMENCLATURE Letter assigned to box location on the pallet American Society for Testing and Materials Environmental Data Recorder Acceleration Hertz International Paper Company Insuumented Sensor Technology, Inc. Number of samples Number of layers Power Spectral Density Recording Control Parameter Root of Mean Square acceleration Regular Slotted Container Standard deviation viii Chapter 1 INTRODUCTION AND LITERATURE REVIEW 1.1 An Overview An essential part of any package design is the testing and evaluation of the complete package, as well as the various components [1]. A wise cost-savings practice will eventually lead to the optimum package design. Accelerated testing aims to reach robust package design in a short period of time while providing lean manufacturing costs, uniform experimentation guidelines, better chances for reproducibility, a match to industry competitive standards and satisfaction to customers’ specification requirements. Unfortunately, accelerated testing has its disadvantages. The primary disadvantage is the possibility of not achieving a perfect match between the experimental conditions and the field environment. This research is concerned with the accelerated testing of the vibration environment. Distribution packaging takes a close look at the integrated package and product handling from factory to point of sale. It includes the actual package, processing systems, conveyor belts, forklifts, and in general any transportation forces that will be in contact with the product. Pre shipment experimentation of new package designs helps in the determination of their resistance to the hazards of transportation and storage [23.4.5.6]. Over-packaging reduces damage considerably but results in higher than necessary manufacturing costs. Under-packaging increases the risk for product damage and for customer dissatisfaction but costs less. In the United States, products used to be generally under packaged with the manufacturers willing to accept the results of some amount of damage during transit. In the late 80’s and early 90’s, a shift in the American manufacturing attitude took place. A new demand for faster, better, cheaper and 'greener' goods and services (or “lean” design and production) ruled the priorities among competitors. Some analysts called it the era of “down sizing”. Others called it “re-enginnering”. American manufacturers restructured to meet the new challenges. Searching for easy to target cost saving items like packaging has brought about the need to question the accuracy of present packaging methods and test protocol. 1.2 Philosophy Behind Simulation There are many different kinds of distribution tests. The “staircase technique” constitutes the most primitive procedure for pre shipment evaluation. It is founded on the idea that a package tumbling down a flight of stairs will disclose the primary weakness of it. It is surprising that even though this procedure lacks accuracy, it is still being used. In an effort to reduce customer claims, railroads insisted on better packaging guidelines and requested the implementation of packaging evaluation methods. Distribution testing machinery was then developed with the intent of simulating shipping hazards in an accelerated and inexpensive way [7]. The data obtained through this kind of apparatus presented reproducible results. Such equipment includes: (1) drop testers which allow accurate positioning and drop heights for edge and comer drops. In the drop tester, the package is placed on the drop table, either flat or on an angle, at a desired height. A latch is usually released to allow the table to swing out of the way, permitting the package to fall freely onto a surface. (2) a revolving drum which is a slow moving device, 14 ft in diameter, which carries a package part way up on the inside and allows it to tumble down to the bottom, striking various bars and points that are secured to the inside of the drum. A common test series would include up to 30 falls in the drum tester. (3) an incline-impact tester which was developed to simulate the effects of rail-car coupling. The results are comparable to those obtained in a drop test, but the shocks are applied only to the flat surfaces of a package. A package is positioned on a dolly on a sloping runway and allowed to freely ride until it hits a solid wall. A typical test would be one impact on each of the six sides of the package from a length of 5 feet up the incline (or about 6 mi/hr at impact). (4) a compression tester which is basically a press which applies a top force to the package to simulate stacking. A typical test would be to compress the package to several times the anticipated dead load on it in a stack for a short time. (5) a shock machine, which is much like a drop tester, but with the ability to vary the impact surface to produce both soft and hard shocks from any drop height Shock machines are typically used to determine how fragile products will behave during distribution prior to designing the package. (6) a vibration tester, or shaker, which is a table that can be made to oscillate vertically at various frequencies and amplitudes. The frequency of vibration is typically increased until the package starts to leave the table, as determined by sliding a piece of paperboard under the package as it bounces. Testing is performed at this frequency for a specified time. This last device had been modified to accept inputs from recorded information that represent any road condition desired. 1.3 Nature Of The Distribution Environment Extensive studies on both trucks and rail cars have been undertaken to measure the severity of the environment [5]. Piezoelectric accelerometers (devices used to measure acceleration) and recording equipment have been deveIOped to monitor and analyze the dynamics at different locations inside these vehicles. The signal obtained is random in the sense that no two segments of the recorded continuous waveform will match when overlaid on top of each other. Close examination of these studies reveal the following characteristics: (1) There are four different types of movement which occur simultaneously and continuously: vertical (up down) motion, lateral (side to side) motion, longitudinal (front to back) motion, and rocking (or reeling) motion. (2) The highest accelerations occur in the vertical direction. (3) The magnitude of the vertical accelerations depend on the road conditions. Bumpy roads will produce higher accelerations than freeways. (4) The highest vertical accelerations are generally found at the rear end of the vehicle. The rear end of most trailers provides a springboard effect related to the back end of the trailer hanging out over the rear axle. However, the floor accelerations directly over the axles can also be just as severe at times. (5) Different transportation profiles (road, sea, or air) and vehicles with different suspension characteristics will generate different vibration signals. The vehicle's weight and the loading can be influential as well. (6) Even though the recorded signal is random. the same characteristic frequencies are present at all times [6]. For truck travel, the sources of vibration which contribute to the characteristic frequencies mentioned are shown in Table 1. Table 1. Frequency ranges associated with various sources 1.4 Established Packaging Test Methods For Vibration Since vertical acceleration is known to be the most severe, simulated vibration tests on product/package systems are normally conducted only in the vertical mode. This is where the electrohydraulic vibration table plays a key part. High pressure hydraulic fluid, and a piston and cylinder arrangement are responsible for the vertical movement of the vibration table. The fluid is set in motion by a valve which is controlled by an electric signal. The valve releases the fluid in step with the signal. The frequency can be varied as well as the amplitude There are three different standards from the American Society for Testing and Materials related to package testing on a vibration table. The oldest one is designated as ASTM D-999, Standard Methods for Vibration Testing of Shipping Containers (initially published in 1948 and with a latest revision in 1991). ASTM D-4169, Standard Practice for Performance Testing of Shipping Containers and Systems was published in 1982 and recently revised in 1993. The newest one is denoted as ASTM D-4728-91, Standard Test Method for Random Vibration Testing of Shipping Containers (originally published in 1986 and recently revised in 1993). In the ASTM D-999 [2], the vibration table is set to move up and down sinusoidally with a fixed acceleration level. Sine testing is usually done in two phases: resonance search and endurance dwells. The product/package system is placed unrestrained on the table, either alone or in stack configuration, in order to simulate loose-load vibration. Also, it can be secured to the table platform using straps in order to simulate blocking and bracing. During the resonance search phase, the frequency of vibration is slowly swept from about 2 to 200 Hz. The intent is to identify the main resonant frequencies related to the product and/or the product/package system. The test specimen, whether the product/package system or the product itself, will violently react to one or more specific frequencies known as resonant frequencies. The endurance dwell phase is conducted by subjecting the specimen to each resonant frequency encountered during the search test phase for a period of time, usually between fifteen minutes to one hour. It is then inspected for damage to either the product or the container. The major criticism of ASTM D-999 is that it is too severe in regards to the dwell phase and yet too limited in relation to the generation of a single frequency. During actual distribution, the specimen is not likely to experience continuous vibration at any particular frequency for an extended period of time. Also, in random vibration situations, many frequencies are present simultaneously. It is possible that two components within the product that are in close proximity to each other can both be made to resonate and hit each other. In the sine test, only one component would resonate at a time, eliminating the possibility of damage due to simultaneous resonant frequencies. That is one reason that random vibration tests as described in ASTM D-4169 and ASTM D-4728 were developed. ASTM D-4169 and ASTM D-4728 provide a random vibration test that simulates actual conditions better than sinusoidal vibration because the signal driving the shaker is a statistical recreation of an actual signal collected from the floor of a particular disuibution vehicle (trailer, rail car, etc.) [3.4]. ASTM D-4169, besides proposing guidelines for drop and compression testing, presents predetermined sets of frequencies and power densities (which will be explained later) sorted by transportation mode and assurance level. ASTM D-4728 is the specific test standard for random vibration aimed at providing more flexibility by allowing the investigator to decide on the selection of the frequencies and power densities used. To record the actual motion, the transportation vehicle is insu'umented with an accelerometer that is connected to a signal recorder. Later, the recorded signal is processed and then fed into the vibration table. which will reproduce the motion of the trailer floor (at least in theory). Unfortunately, and as result of the mass of the vibration table and the lag time between the fluid, the valve, and the signal. the motion of the table will not exactly match the motion of the instrumented vehicle. That is the main reason for processing the signal first. This is done by a controller. In a controller, three different signals are handled: (a) the demand signal which is original acceleration pattern recorded off the truck, (b) the drive signal used to operate the machine, and (c) the response signal coming from an accelerometer attached to the table. The purpose of the controller is to constantly monitor the response signal. compare it to the demand signal, and decide how to adjust the drive signal in order to get a response that matches the demand signal. The drive signal must be constantly changed so that the response signal is the same as the demand signal. This system constitutes a closed loop controller and is considered to be far more practical than an open loop one in which the drive and demand signals are the same but not necessarily the same as the response signals. The closed-loop system is also used for sine testing. Since the intent is to repeat the same motion cycle after cycle, this kind of controller is able to easily home-in on the correct drive signal which makes the response the same as the demand. If the intent is to reproduce random motion from an actual road trip, a closed loop system will not work as well because the signal is not repeated. In the case of the sine testing, the drive signal can be corrected only after few table response signals have been monitored and compared to the same demand signal. In random vibration, the demand signal is different at different instants. So, it is virtually impossible to match the drive signal to the demand one. Therefore, the most common approach is to first analyze the random vibration pattern for frequency and amplitude content and then make up a random signal containing these same frequencies and amplitude to later drive the table with. This is done by taking a look at the disuibution of randomly sampled instantaneous acceleration values taken off the truck floor. A digital “bandpass filter” or tuner. that selects a specific frequency range out of a random signal, is then used to sort this information by frequency range. If a random signal is sampled at many instances, the average of the distribution of the instantaneous accelerations will be zero because there are likely to be as many positive acceleration values as negative ones [12]. The standard deviations , of this distribution of accelerationng , for number of samples. N . is calculated using the following equation: S: J2[ (Cg) Hz-(mcan) (1) A?_I Since N is large and the mean is zero, the standard deviation becomes: = fir) * 20? (2) This last result is also known as the “RMS G” or Root of Means Square acceleration. This relationship between the RMS G and the standard deviation is use to adjust the drive signal so that the table reproduces the same disuibution of accelerations for each frequency range as the demand signal. Because different analyzers employ tuners with different bandwidths, the square of the RMS G is divided by the frequency range or bandwidth to remove the effect of bandwidth. The result is known as the “power density” (PD). Power densities for respective frequency bands are graphed to obtained what is known as a “Power Spectral Density” (PSD) plot. The PSD plot is meant to be interpreted as a statistical picture of random vibration. Figure I shows a typical PSD plot. (RMSG)2 PD = W (3) Figure I. PSD plot with 1 Hz bandwidth tuner PSD plot obtained with a 1 Hz bandwidth tuner E "5 i Al— e N g3“ DBargraph w 3 2 _ +Curve contour 31b 50 71 it t t, r r r 12345678910111213 Frequency(Hz) The vibration table is driven with a random signal generated from the PSD plot which has these frequencies and amplitudes but in a different order than the original recording. The response signal will also show the same frequencies and amplitudes as the original recorded signal, but in a different order. The advantage in controlling the table with this technique is that the drive signal can be prepared in advance to produce the correct response through a simple pretest warm-up before the test begins. At that time the drive signal is ready when it is requested. This method of simulating the dynamics of the transportation vehicle is covered in ASTM D-4728. Sophisticated equipment is usually required to run the vibration table using PSD plots. Also, an accurate recording of the environment inside the particular carrier must be obtained in advance. For such reasons, smaller packaging laboratories still do vibration testing using the earlier method described in the ASTM D999. However, even if random vibration testing according to ASTM D4728 could be done, the question of how long to leave the test package on the vibration table to simulate actual road time remains. It is no longer obvious how the processing of the actual recorded signal to produce the PSD plot which drives the table affects damage potential, and hence, table time. 1.5 Results Of Previous Studies Different studies have tried to correlate road time with simulation time on the table. Two of the most recent of those studies are briefly discussed next. This first one is an abrasion resistance study [16]. Package abrasion caused by vibration during transportation and that caused in a laboratory using horizontal vibration tests were compared, using sample packages of white corrugated fiberboard printed with checkered patterns. The results of this experiment showed that an average transport trial lasting an undisclosed amount of time was comparable to five-minute laboratory vibration test at an acceleration of 1 g. The maximum abrasion found in the road trips was almost equivalent to a 20 minute laboratory vibration test. which can therefore be used as a reference for package abrasion in transport. Another study is related to “settling of cereal in cased cartons” was conducted to compare road and table time [10]. The data reported demonstrates that about one-third of the actual transportation time was required to cause the same amount of settling of cereal on the vibration table. A PSD plot was generated from an actual recording off the floor of a truck and then used to recreate the actual ride on the vibration table. The present study follows the above approach but uses compression strength of corrugated boxes instead of the amount that cereal has settled. In addition, it compared different ways to record the original signal using commercial hardware: specifimlly, the “acceleration trigger leve ”, “acceleration sampling rate”, “pre-trigger length”, and “post trigger length” recording parameters of the two different Environmental Data Recorders [8,9] used in this study. 1.6 Objectives Of This Study Even with standardized methods to conduct vibration simulation geared towards the development of acceptable packaging design, the actual distribution environment can not be exactly recreated during laboratory testing. In both of the most recent vibration testing standards, ASTM D-4169 and the ASTM D-4728, no specific correlation between table time and road time is specified. The objective of this thesis investigation was to explore the possibility of establishing a correlation between actual transportation time and the laboratory simulation time. A secondary objective was to study the effect of using different values for certain recording parameters. It was presumed that 10 when the damage resulting from real transportation equals that of an experimental simulation, a correlation between the actual and simulation time required to produce such damage can be obtained. In general, there are several causes of dynamic excitation which must be taken into consideration for any packaging vibration study. The first one would be caused by the mode of transportation. For this investigation. the trucking environment was chosen. A second cause of excitation is the dynamic input caused by the roughness of the transportation route. The transportation path used here was a circuit of Michigan highways that will be discussed later. The last significant factor to consider is the product/package system involved. There are numerous product/package systems and each one has distinct criteria by which damage is determined. In this investigation, a system was chosen so that damage was easy to observe and measure. The product/package unit was a plastic bag with 40 lbs of sand packaged inside a C flute Regular Slotted Container style of corrugated box. A wooden pallet was built to hold 24 of these units in a six tier arrangement. After distribution , the compression strength of each box can be used to establish a measure of damage. Two Environmental Data Recorders (EDRs) were used to record the vibration levels at the same time that a pallet load of boxes was being transported. This pallet was located at the rear end of the truck and the EDRs were placed alongside it. The Dyna Max software was used to analyze the data collected by the EDRs. This software has the capability of instantaneous generation of PSD plots. The PSD plot, which serves as the demand signal during vibration testing, was manually entered into the computer software that operates a Lansmont vibration table. Four identical pallet loads with same pallet configuration as the road trip load were tested on the vibration table. Each load was subjected to different simulation times. For all the pallet loads, the compression strength of each box was determined with the intention of using them to find the correlation time between the road trip trial and the vibration table simulation. Chapter 2 MATERIALS, EQUIPMENT AND EXPERIMENT DESIGN 2.1 Product/Package System Under Investigation The product/package system used to evaluate damage consisted of plastic bags containing 40le of sand (product) packed in “C” flute, single walled, Regular Slotted Container (RSC) style corrugated boxes (package). The sand and the box specifications are given in the following tables: Table 2. S ifications for the sand F Speclfication Requirement - Type Washed. screened, dried play sand Brand “Kiddies Fun” Manufacturer Gibraltar Play Sand (National Corporation) Manufacturer’s address 8951 Schaefer Highway Building Number 4 Detroit, MI 48228-2515 Tel. 313-491-3500 for the boxes Table 3. S T Material Manufacturer Manufacturer’s address Crush Test Size Unit Gross W Limit A A Deflection Width User Distributor Builders Square, Lansing, MI Density 99.4 1%compacted 96.1 ff? loose Net wei ht of the ori inal b s 50 lbs or 22.68 K Sin wall - Re “C” Flute International 1450 McPherson Park Drive Ml 48823 Tel. 517-546-1220 29 Win 60 in 50 lbs 9 in 385 lb .24 in 16in 21 in JC Penney Catalogs Fond WI 54935 Slotted Container 11 The boxes used were produced on May 15, 1996 and rejected by the manufacturer after inspection on the same day due to flap defects (a feeder ripped the sheets). They were stored for a month in a dry conditioned warehouse at International Paper (Howell, MI) before they were relocated to the vibration laboratory at the School of Packaging at Michigan State University. The amount of sand per box was chosen by taking into consideration the manufacturer’s recommended weight limit on the box, the box‘s capacity and ergonomic factors. One inch head space between the top surface of the sand bag and the bottom surface of the box flaps was allowed. Figure 2, below, shows the arrangement inside the boxes. Figure 2. The box's contents and arrangement The number of layers (NL) of boxes on the pallet was selected to provide a dead load on the bottom box equal to a significant fraction of the compression strength of the box. For the purpose of this research. measurable damage means a compression strength significantly different from the average compression strength of a box that has not been exposed to any type of loading, bending or physical damage. In this report, such a compression strength is referred to as “virgin compression strength.” The first step in determining NL was to measure the average virgin compression strength, which was used as a benchmark. The bottom box in a pallet should not be exposed to a weight higher than the virgin compression strength. Otherwise, it would completely lose all of its compression strength as soon as the pallet load was made up. If such a box was put through distribution, damage due to vibration could not be accounted for since the box would be completely damaged right from the beginning . 80, based on the results of the previous study conducted on cereal [10], this investigator and her research committee decided that about one-third of the virgin compression strength was enough initial loading. Based on this and the l3 box’s capacity, it was determined that the boxes should not hold more than forty pounds of sand. Since, the virgin compression strength was found to be about 385 lbs, one third of which is approximately 128 lbs, the resultant NL was approximately four. Six boxes were used per layer. After the pallet load of twenty-four boxes was constructed, it was shrink wrapped to provide support and to keep the boxes’ moisture content relatively constant. The pallet load sent on the actual road trip experienced about the same temperature and relative humidity as the laboratory where vibration and compression testing were done. All the simulation trials as well as the compression testing were performed in the vibration laboratory, which was conditioned at 50% RH and 75°F. For all the phases of the experiment, compression testing was done immediately after the run, whether actual or simulated, was completed. Therefore, timing and weather conditions should not be considered as factors in this experiment. Prior to testing, all the boxes used were erected from the “knocked down” configuration and conditioned at fifty percent relative humidity and seventy five degrees Fahrenheit for 24 hours. Each box had one inch thick foam cushion pads surrounding the sand bags on the inside box walls. The cushions covered only part of the height of the box about one inch head space was left on each box so that they did not support the box in compression. They simply served the purpose of keeping the bag of sand toward the center of it. This guaranteed equal and homogeneous distribution of the sand weight for all the boxes. A modified 50” x 44” pallet loaded with twenty-four boxes of sand in four layers, six boxes per layer, was used as the test unit. Each box location in a layer had an assigned letter, from A to F, which was the same for identical locations in all the four layers. Each layer had an assigned number, from one to four. The number ‘one’ was assigned to the bottom layer in the pallet and the number four to the top layer. Each box was labeled according to its position in the pallet using the described number/letter assignment during all the phases of the experiment. Each pallet load was shrink wrapped immediately after construction as shown in Figure 3 on next page. In the first phase of the experiment, a pallet load sat still in the lab for the total time equal to the actual road trip (195 minutes for this research). A second pallet load was taken on the actual road trip, Figure 3. The test pallet load Figure 4. The twenty four foot truck used during the study which will be described in the “experimental procedures.” The final phase of the experiment consisted of four simulation trials on the vibration table in order to find a correlation to the original transportation time. A separate experiment involved two simulation trials to compare the effect of vibration recording parameters. One shipping container was used per simulation trial. 2.2 Mode And Route Of Transportation A twenty-four foot delivery truck from EL. Hollingsworth & Co (3039 Airpark Drive North, Flint, Michigan 48507) was used, courtesy of the Transportation and Packaging Department of the General Motor Corporation in Lansing, Michigan. A picture of the truck is shown in Figure 4, above. The second phase of the experiment was to study and analyze the actual transportation environment that the shipping container was being exposed to. Once the delivery truck was ready for loading, the shipping container was placed in the rear end of the trailer. The bottom of the pallet was nailed to the floor of the trailer. The EDRs were also nailed to the floor as well, as near as possible to the pallet. There was one EDR located on each side of the pallet. During the trip, the velocity of the truck was almost at all times that specified as the speed limit for the road. The entire trip was 190 miles and took 195 minutes to travel. The trip took place on Friday, June 24, 1996, which was a fair weather and sunny day. Small sections of the road were under construction at this time. Immediately after the trip was completed, the pallet was disassembled and the sand bags were 15 The route taken for the actual trip is described in the Table 4 below: Table 4. Route of Direction Location or street 9 onto Wilson Road onto Harrison onto to 127 North Clare 75 to Detroit 696 to 96 to northeast 496 to Downtown east Exit 9 to Road left north Onto Harrison Road ri east Onto Wilson Road south B ’s loadin STOP carefully removed from the boxes. Compression testing was performed on each box. 2.3 Programming The Environmental Data Recorders The recording device used to capture the vibration data from the actual road trip was an Environmental Data Recorder (EDR), which is a battery powered device with a built in ui-axial accelerometer. One “Panther”(or EDR-4) and one EDR-1 were utilized. This equipment was borrowed from Instrumented Sensor Technology (IST) . Inc., 4704 Moore Street, Okemos, Ml 48864. Other equipment provided from IST included the RS-232 interface cable and the Dyna Max Module DM-6 software used to interpret the recorded vibration levels and generate the PSD plot. The EDRs were mounted on different wood mounts (provided by the School of Packaging) so that they could be nailed to the trailer floor as shown in Figure 5. (a) Panther used during study (b) EDR] used during study Figure 5. The Environmental Data Recorders 16 Prior to mounting the EDRs, they were programmed using the Dyna Max Module 6 program. Several recording control parameters (RCPs) had to be specified. The RCPs must be carefully determined in order to assure that the data recorded represents the entire trip. The memory capability of the EDR was the limiting factor. An engineer from IST [11] offered technical support in determining the critical RCPs. which included maximum event length (number of samples), the acceleration sampling rate (in samples/second) and the acceleration trigger level. First. the memory capacity available by each one of the EDRs was determined. The anticipated length of the trip then determines the maximum allowable sampling rate so that full memory capacity was used. Both EDRs were set in the overwrite mode. which allowed the EDR to substitute lower vibration readings with higher ones in the case that it used up its memory space. This means that the device will discard lower vibration readings in order to record higher ones when the length of the recording has been exceeded. Consequently, it was imperative to use a length of time believed to be the maximum to complete the trip while at the same time taking into consideration that excess time would be a waste of recording space. It was also necessary to set the trigger level, which is the lowest acceleration (G) level that will be recorded. Because of the additional memory, the Panther was set to trigger nearly every 6.6 milliseconds, while the EDR-1 was adjusted to record only accelerations higher than 2 Gs. Comparing the two previous tables, it can be noticed that the “Panther” has more capabilities and much more memory capacity than the EDR-1. So, the “Panther” should provide a better representation of the vibration levels during the actual road trip since it will pick up about seven time more data than the EDR-1. The parameters that were italicized in Tables 5 and 6 are exclusive to that particular EDR. 2.4 The Vibration Table The simulation trials using the data recorded by the EDRs in the truck were conducted on a Lansmont vibration table and system station located at the School of the Packaging at Michigan State University, Lansing, MI. This vibration table is a model 10000-10 vibration test system with a 60” x 60” platform. The “system station” was a touch-screen console computer. Method A, closed loop-automatic Tables 5 and Table 6 present the RCP settings for the “Panther” and for the EDR-1, respectively. T “Panther" Parameters Event 288 4,096 (samples) 4,096 Length 4,096 Overwrite 0 Pre-trigger rigger rigger Parameters window Number N/A At last event Acceleration event . ************** WllldOW 2,048 2,048 4,096 Overwrite ****************** equalization as in ASTM D-4728 (Random Vibration Testing of Shipping Containers), was followed. The “system station” was used to drive the vibration table. A similar setup is shown in Figure 6. Figure 6. Lansmont vibration table (series 1800) As mentioned previously, another four pallet loads identical to the one put through the actual road trip were constructed. Restraining devices that held the pallet in place on the vibration table were used to mimic the pallet load nailed to the truck floor. The simulation times chosen were based on the actual trip time. Since the actual trip was about three hours (195 minutes), the first simulation trial was done for 180 minutes. A second simulation was also conducted for only 30 minutes to see what would happen with the box compression strength under test times that were very much different from the actual road trip. Analysis of the compression strengths obtained in the first two simulations lead to a third simulation of 60 minutes and then a 90 minute simulation later on. 2.5 Damage Criteria ASTM D642 was followed to obtain compression strengths. A pre load of 50 lbs on each box was used as a reference point for box deflection. The lower platen of the Lansmont compression tester was marked with tape to provide a reference position for placement of the boxes on the tester. A floating top platen was used because it more closely mimics the loading on of stacked boxes than a fixed platen. During compression, it was noticed that when the compression strength was reached, the walls of the box began to buckle. If the test was continued after this point, the load carried by the box dropped drastically. Both the compression strength and the respective deflection for each box in each pallet were measured and recorded. Chapter 3 DATA AND RESULTS 3.1 The PSD Plots For The Road Trip The PSD plots for both EDRs were generated using the Dyna Max Module 6. The instructions on how to download the information from the EDR to the computer are well described in the EDR Manual provided by IST. The PSD plot for the “Panther” and for the EDR-1 are shown in Figure 7 and Figure 8, respectively. Although there was no significant difference between the PSD plot from the Panther and the EDR-I, the Panther PSD plot was used to run the vibration table for the purpose of comparing road time to table time because it recorded more events than did the EDR-1 for the same trip. Certain frequencies with their respective power densities were retrieved from the PSD plot in order to program the vibration table. Table 7 shows the data obtained from the EDR-1 and the Panther, which was the only data used to program the vibration table. The data was entered into the Touch Test Vibration System Station manually. The raw data for both EDRs can be found in Appendix B. 19 NmFm aco>o Lea moEEom me “out oEEom com 6.: c_ 35>“. ANIV Jucwjowed 00— o— “.0 be: ._ m m g III! II wenmfi mOnw_ "000.0 flood “0.0 ".0 (ZH/zb> 2 08d Figure 7. PSD plot from the “Panther” mmov "Ego ..eo mcEEom com “we: eanm Pm 8.: c. mEo>m ANIV Jucojomed 00— ofi L p .u .4 . . ...... .4 ..4 .. U .4 .. 4 .. .4 . . . . . . . . H ... .. n , .H ....... m... _ . h H .. H. ., . . p .. .4. a. . :... ... .. . ... 4 a .L 4 v 4 n .- a. . ... .. .. .. . ....... ... .. .. . .._. .V: m .4 .u I l... I... I ........ . t 4 . .4 : .. : .. : .. ...? ..l 4: ,4 .4 ., ..4: : . .... .. .. : .... ...:;.:.. a ... .... ...4 : . .:.. 1.“. ... ...4. .4. ... .m. ..... ......: ..... .......... 4 WOImfi .. . .. a. .. .. w . n 4 .4. .. . .4 . 4 ., .. .H ... .. . .w. ,h ; ..nl . . . .. n. . .. . h . .. ...... _ w .. ..4 ..4 n 4 .1 : , , .v 4 . 4 . .- .. ..4 .. . .. ... . . .... . . 4 ..H. . . .... 4... . ..4 . .. . ... . .. T... I .. 2...... ..I ,. s . , . . w. w. H. H h .u .n ,. U. r ... - .. .. v ..4 .4 r ..... . .w .H. S _ . . .. i m m a .H .. u g. n . .4. ..... H G . . . . ... n . m, r. m. . . ,m ,m 3. 2 ~ A ..... .. . . .... 4 ... 4 . .. ..r m _ . .4 _ A. ,4 . i . 4 I . 4 I” . 1..” . .4 e . 4 ..4 v Z.. r .4 4 i ..t ......... e n 4 ..4 #0 ) . . a. h . .. . / . r. . .... .. , . . _ x. ... . e . . . . .. . a . . .. .. 4. .u , .....f.... .....f . .....4 #0000 . . a . . . Q. _ n . . H a . . . .v . ..... 4 I, s . 4 I. L. 4 u ..4 Iv U . ... . . m. a .H. . n . .. ... 4. A a ,n no . L - . . .... , . . . ... t .4... v 4.. I e u I e e .................................. ”414.44.444.44....nr ................. . . “O . 4. 4 . .o .w . m ...... ..4 n. m . a . 2w . . m a m . - . m - H. .H .. . u . L . . ..4. ............ . a . ... .. .v .... .. a w . . ... . ...4 4.: .H. :.r. t”. . . .m ..m r ...: . .0 ..1 . n. I ...4 ...t. L .F. .h. 4 4. . I .I. t .... h .c .4 e e 4 .o. 4 M44 .m. ..b... u 4 .m. o ..w \ I“: n. a. “wad" . . w, 2v. ...? .4.. 4 4 ..4. 1... . ...4 4... . . “DO . . .. .... .4. . 4, . ... r 4 .4. .. .............. Figure 8. PSD plot from the EDR-1 22 Table 7. PSD data used FREQUENCY (Hz) POWER DENSITY (G’le) EDR-1 PANTHER 1 2 3 4 5 6 7 8 9 0.000606 0.004000 0.015187 0.003000 0.001053 0.000588 0.000385 0.000689 0.001 131 0.000865 0.000234 0.000040 0.000123 0.001036 0.008506 0.000900 0.000120 0.000270 0.000098 0.000031 0. 00021 6 0. 001 509 0. 005954 0. 00361 7 0. 001 1 91 0.000526 0.000250 0. 0002 71 0. 00042 7 0.000473 0.000437 0. 000088 0.000066 0.000206 0.00021 8 0. 000549 0.003784 0. 001 939 0. 007562 0.000587 The power density values charted in Table 7 are lower than the ones suggested in the ASTM D4169. 3.2 Compression Strength Results For The “Control” Pallet Load In order to have a control to measure against. a pallet load was built and left alone for three hours (roughly the duration of the actual trip). The pallet load was then disassembled, and the boxes emptied and compression tested. The compression data obtained is presented in Table 8 and again in Figure 9 graphically. 23 Table 8. Compression test results for boxes on the control pallet load lAYER STANDARD LOCATION IMEASUREME cm 1 2 3 Top 4 VERAGE DEVI ATI ON Compression A Stre (lb) 303 350 405 387 361 45 Deflect'nn (in) 0.19 0.17 0.22 030 0.22 0.06 Compression B Strength (lb) 393 415 437 371 404 28 Deflection (in) :H 0.23 0.23 0.24 030 0.25 0.03 Compress'nn C Strength (b) 351 332 448 265 349 76 Deflect'nn (in) 0.20 0.14 021 028 0.21 0.06 Compression D Strength (“3) 345 307 366 238 314 56 Deflect'nn (in) 0.19 027 0.22 0.29 0.24 0.05 Compress'nn E Strength (lb) 371 436 435 251 373 87 Deflection (in) 0.22 0.16 0.23 030 0.23 0.06 Compression F Strength (1b) 375 383 452 362 393 40 Deflecthn (in) 0.24 0.24 0.25 0.31 0.26 0.03 Comession AVERAGE . Stre I (b) 356 371 424 312 366 46 Deflect'nn (in) 0.21 0.20 0.23 030 0.23 0.04 _ — STANDARD Compression DEVIATION . Strength (lb) 31 50 33 68 Deflect'nn (in) 0.02 0.05 0.01 0.01 Figure 9. Average compression strength by layer for the control pallet load Average Compression Strength per Pallet Load Layer 550 450 Oonpression Strength (Lbs) Bottom 1 2 3 ' Top 4 Layer Location in the Palet Load 24 The average compression strengths for the control pallet load show an ascending trend starting with the bottom layer up to the third layer. The top layer for the control case seems to be weaker than any of the other layers. In theory, that top layer should exhibit the highest compression strength results since the boxes in that layer bear no load. This contradictory situation can be understood by explaining the box storage conditions prior to the experiment. Before any testing was performed, all the 300 boxes were stacked flat (in the knocked down configuration), one on top of the other, on a pallet The boxes used to construct the “control” pallet load were closer to the bottom of that stack. There was a high probability that the flutes of the bottom boxes were compressed and damaged. This same phenomena was experienced with the pallet loads used during the simulations to compare the Panther and the EDR] data. 3.3 Compression Strength Results From The Actual Transportation Environment (Road Trip) After the actual road trip was completed, the pallet load was disassembled and the boxes were again emptied and compression tested. The trip took 195 minutes to complete. The truck traveled a distance of 190 miles for an average velocity of 58 miles/hr. The compression data obtained is presented in Table 9. The highest standard deviation for the compression strength was 7% of its average value. For the deflection, the highest standard deviation was 12% of its average value. The average compression strengths per layer is plotted in Figure 10. 3.4 Compression Strength Results From The Simulations on The Vibration Table For the simulation phase. the same physical testing procedure as in the actual road trip was used. Compression data was obtained at four different times on the vibration table. The four identical pallet loads were exposed to 30 minutes, 60 minutes, 90 minutes and 180 minutes on the vibration table and the compression test results are presented in Tables 10 through 13, and Figures 11 through 14, respectively. 25 Table 9.Compression test results for boxes on the actual road trip IAYER STANDARD IDCATION IVIEASUREME an I 2 3 Tq) 4 VERAGE DEVI ATI ON Compression A ‘ Stre (1b) 265 255 382 458 340 98 Deflect'nn (in) 0.14 0.05 022 035 0.19 0.13 Compress'nn B Strength (lb) 318 409 551 344 406 104 Deflection (in) 0.17 0.12 0.26 022 0.19 0.06 Compression C Strength (1b) 343 248 389 430 353 78 Deflection (in) 020 0.11 029 0.27 022 0.08 Compress'nn D Strength (1)) 279 412 540 332 391 114 Deflection (in) 0.18 0.17 025 0.32 0.23 0.07 Compress'nn E Strength (1b) 277 306 379 410 343 62 Deflect'nn (in) 0.16 032 025 025 025 0.07 Compression F Strength (1b) 346 331 491 415 396 73 Deflect'nn (in) 023 0.13 024 0.33 0.23 0.08 Comesshn AVERAGE Strength (1)) 305 327 455 398 371 69 Deflect'nn (in) 0.18 0.15 025 029 022 0.06 STANDARD Compression DEVIAT‘ION Strength (lb) 36 72 8‘ 5° Deflect'nn (in) 0.03 0.09 0.02 0.05 Figure 10. Average compression strength by layer for the actual road trip Average Compression Strength per Pallet Load Layer 550 450 250 § Conpreuion Strength (Lbs) Bottom 1 2 3 Top 4 Layer Location 'n the Palet Load I180 Minute Road Trip} 29 Table 13. Compression test results for the 180 minute simulation LAYER 3 STANDARD 53 Deflection Deflection Deflection Deflect'nn Deflection AVERAGE Deflect'nn DEVIATTON Figure 14. Average compression strength by layer for the 180 minute simulation Average Compression Strength per Pallet Load Layer 550 450 400 350 300 250 Oon'pression Strength (Lbs) § Bottom 1 2 3 Top 4 Layer Location in the Pallet Load [180 Minute Simulation} 3.5 Compression Strength Results For The 60 Minute Simulations Comparing Panther and EDR-1 Data This research also intended to compare the effect of using different EDR recording control parameters data gathering. Two Power Spectral Density profiles were developed using different acceleration sampling rates, acceleration trigger levels, pre-nigger and post-trigger lengths for the same road trip. The compression strength after a one hour vibration simulation was used to evaluate any differences. Table 14 and Figure 15 show the results of the compression tests using Panther data. Table 15 and Figure 16 show the results using the EDR-1 data. Table 14. Compression test results for the 60 minute simulation with Panther data STANDARD Compression 83 Deflection Defkct'nn Deflect'nn Deflection Dethction Deflection AVERAGE “3’0“ Deflect'nn DEVIATION 31 Figure 15. Average compression strength by layer for the 60 minute simulation with Panther data Average Compression Strength per Pallet Load Layer 550 500 400 350 300 250 200 150 Conpression Strength (Lbs) Bottom 1 2 3 Top 4 Layer Location in the Pallet Load [60 Minute Pantherj Table 15. Compression test results for the 60 minute simulation with EDR-1 data STANDARD 109 Deflect'nn 0.07 122 Detkct'nn 0.04 87 Deflect'nn Deflection Deflection AVERAGE DEVIATION 32 Figure 16. Average compression strength by layer for the 60 minute simulation with EDR-1 data I Average Compression Strength per Pallet Load Layer 550 500 450 400 350 300 250 200 1 50 Conpresslon Strength (Lbs) Bottom 1 2 3 Top 4 Layer Location in the Pallet Load [60 Minute EDR1 E Comparing the compression strengths of the two different 60 minute simulations on the vibration table, the results showed significant differences in the standard deviations but not in the means, implying that variations in the boxes, as explained previously with the control case results, were responsible for this behavior. Chapter 4 CONCLUSIONS As explained previously, only one of the total of eight pallet loads was exposed to the actual transportation environment. One of them was used as control and the other six were used for simulation testing on the vibration table. Because the total time of the actual road trip was 195 minutes, vibration testing began with extreme values that were arbitrarily selected as 30 minutes and 180 minutes. After analysis of the compression test data from this two simulations, the next simulation was done for 60 minutes and then 90 minutes where it was decided to end the experiment. 4.1 Data From The Vibration Table In Figure 17, the effect of table time on the average compression strength for boxes in a given layer can be observed. As table time increased, the boxes in general lost compression strength. The compression strength in each layer for each simulation was statistically compared to the actual transportation environment using a Meet [12,13]. The results are shown in Table 16. Figure 17. Average Compression Strength by layer for the actual and the four simulated environments Average Compression Strength per Pallet Load layefl 500 commas 180 ...eoMNurEs +REAL-3m somes 450 400 350 Conpression Strength (Lbs) Bottom 1 2 3 4 Top Layer Mrrber it the Palet Load 33 34 The T-test results in Table 16 do not point to a strong correlation between anyone of the simulations and the actual road trip. The numbers shown in this table are the percent probabilities that the means of the compared samples come from the same population. With the control pallet out of the picture, and considering each layer independently, the most probable matches are highlighted in Table 16. For the bottom layer, the 90 minute simulation had the highest correlation to the three hour road trip. For the second and third layers, the 60 minute simulation correlates best to the actual trip. If the top layer is considered alone, then the 30 minute simulation is the best choice. Since the results pointed in different directions depending on the layer considered, no correlation between the simulation time and actual road time am be reached. Table 16. T-test results Percent ‘ with the Road T LAYER Versus Value Road Trip ve rsus versus T-Test T-Test T-Tcst T-Test The next four figures show the average compression strengths versus time within a specific layer. The results are presented in Figures 18 through 21. Figure 18. Average compression strength versus vibration time for the bottom layer Average Compression Strength versus vibration time for the 1st Layer 3 380 g, 360 — $2 340 _ Mm“ «a- Vibration table results 5 no" Fbad Tr'p restlts .3 320 '- + 'Control' - no vibration 9 300 - ° 2 8 280 L ‘ so so 90 180 Trne in m’nutes 35 Figure 19. Average compression strength versus vibration time for the second layer 400 380 h- 360 - to a C) l Q N O b Oonpression Strength (lb) Time in m'nutes ”......M 1".” “Mn-u... arr” ... ‘ l J ? 3O 60 90 180 m."- Vibration Table results .... Road Trip results + 'Control' - no vbration Average Compression Strength versus vibration time for the 2nd Layer I Figure 20. Average compression strength versus vibration time for the third layer Average Compression Strength versus vibration time for 3er layer 2; 460 ‘0 g. 440 - ° ‘ ...... Vbration Table results 420 -.r— ‘2 K «on Road Trip results .3 400 b \\ + 'Control' - no vbration ° 380 — E; 360 l l l 4:1!_J 30 60 90 180 Time in minutes Figure 21. Average compression strength versus vibration time for the top layer 500 Compression Strength (b) 450 - 400 '- 350 - L Y’ M 1 4 30 60 90 180 Time in ninutes Average Compression Strength versus vibration time for the 4th Layer } «Q. Vibration Table results *9» Road Tr'p results + 'Control' - no vibration 36 Figures 18 and 19 reveal that the two bottom layers in the pallet load generally experienced an increase in compression strength as simulation time increased. This is just the opposite for the two upper layers in the pallet load. This unexpected behavior for the two bottom layers may be related to the high standard deviation in the compression strength results. 4.2 Correlation Between The Panther And The EDR-l For The Same Simulation Time This investigator arbitrary decided to test another two identical pallet loads in 60 minutes simulations and compare the damage produced in each case. Table 14 and Table 15 show the data obtained after simulation using the Panther and the EDR-1 PSDs, respectively. There was no significant difference between the two different pallet loads used on the vibration table for a period of 60 minutes, which was expected because there was no signifieent difference between the PSD plots from the different EDRs. Based on this, it can be concluded that the choice of RCP’s when using any EDR does not significantly influence the PSD plot (for a three hour trip) and hence the compression test results. Any differences between the simulations using the Panther and the EDR-1 PSDs comes from the previously explained fact that the boxes used for this two simulation came from the bottom boxes of the “knocked down” stack. 4.3 Final Remarks The purpose of this research was to explore the possibility of establishing a correlation between vibration table simulation time and actual road trip time using the compression strength of “C” flute RSC corrugated boxes as a measure of damage. Another objective was to investigate whether or not the recording parameters can influence vibration table performance. The compression strength results were inconclusive in establishing any possible correlation between vibration table simulation time and the road trip time. Box compression strength is probably not a good indicator of damage and should not be used to compare road time to table time due to: (1) the huge variation in compression strength values (big standard deviations). (2) the fact that the virgin compression strength (385 lbs) was somewhat low to begin with. 37 Previous research demonstrated that it took one third of the transportation time to simulate the settling of cereal in cases on pallets. No such correlation was found here. Substantial reduction in test time should motivate test laboratories to use vibration tables because of the savings involved in the reduction of test time. Unfortunately, the amount reduction still remains in question. On the positive side, these results do demonstrate that there is no need to standarize recording parameter settings, since for a short duration trip the results appear t be insensitive to them. It has become quite evident that determining a correlation between actual distribution time and simulation laboratory time is quite a complex task. A limiting factor is the ability to select a product for which it is easy to quantify damage. This research clearly shows that compression strength of boxes should not be selected. There are literally hundreds of different experiments which could be done to follow up on this study. Additional research should be focus on studying other product/package systems in combination with other transportation modes and routes, all aimed at reducing simulation time. APPENDICES 38 Appendix A - Research Proposal To Companies 39 SUBIVII'ITED TO: BAXTER HEALTHCARE CORPORATION Route 120 and Wilson Road Round Lake, Illinois 60073 ANALYSIS AND COMPARISON OF LABORATORY AND REAL TINIE DISTRIBUTION VIBRATION LEVELS AND TEST DURATIONS SUBMITTED BY: Sara H. Languet, EIT. MBA Gary Burgess, Ph.D, Major Professor School of Packaging Michigan State University East Lansing, MI 48824-1223 Tel. 517-355-9580 Fax. 517-355-8999 March 11, 1996. 40 W School of Packaging, Michigan State University. East 11min: MI 48824-1223 Tel. 517-355-9580 INTRODUCTION One interest of the School of Packaging is the study and investigation of shock and vibration levels during distribution. Recently. there has been a great deal of concern among manufacrurers regarding the procedures used in the measurement of product damage. which is a costly problem. The proposed thesis intends to set standard methods for selecting the appropriate test parameters for vibration simulation. OBJECTIVE 1) To analyze and compare the basic laboratory methods and equipment used to determine the distribution vibration levels. 2) Formulation of a possible ASTM standard based on the outcomes of the study. 3) Possible publication based on the outcome of the study. PERSONNEL, MATERIALS AND EQUIPMENT ‘ Two pallet loads of the same product-provided by the Baxter ‘ A delivery truck on its way to a destination with n'uck driver provided by Baxter ‘ Three Electronic Data Recorders (EDRs) -available from the School of Packaging. ‘ Proper EDR software ~availablc from the School of Packaging TEST METHODS PHASE 1- Preparation steps: 1) Determination of the product to be used. Product suppliers must be contacted. Two identical product pallets loads must be obtained. One pallet load will be road tested and the other will stay at the School of Packaging for later Vibration testing. 2) Determination of actual qualitative and quantitative damage after delivery. 3) Determination of the needed parameters for programming the EDRs. 4) Program and load the EDRs with the proper information for the vibration data collection. PHASE 2 -Data collection steps: 1) A loaded delivery truck in route to destination stops at the School of Packaging. 2) Take out two pallets loads of the product from delivery truck into the Packaging loading dock. 3) Instrumentation and wiring of the delivery truck and of one of the product-pallets to be delivered (perhaps can be done at Baxter facilities'. 4) Vibration data collection from the delivery truck riding to its final destination using EDRs. The graduate student 15 going to follow the truck till the TWO DAY journey is completed. 5) Qualitatively and quantitatively measure product damage level with specifications of PHASE 1 Step 2. 6) Remove instrumentation from the delivery buck and grad student returns to Lansing. PHASE 3 -Data analysis steps: 1) Download data from the EDRs to the computer. 2) Amlysis of data with proper software to supply Baxter with vibration information (6 levels. identification of the critical component of the product) that will help in the packaging design and specifications of future products. 3) Further analysis and comparison of the data collected to determine the vibration levels to be used in a possible ASTM standard for vibration simulation. PHASE 4 -Paperwork steps: 1) Submission of first thesis' drafl and possible publication's first draft to student's committee. 2) Submission of second thesis draft and possible publication's first draft to student's committee. 3) Submission of final copy and possible publication copy to Baxter and the students committee. PROJECTED OUTCOMES 1) Graduate student thesis. 2) Possible ASTM standard and publication. 4] 3:95.— : 3.5 H»: ”nose: :3 oteoiflm mm<2d D szwwd wxmmi mzcz Eons.» 22820 new 33,891 . .. i HUN ” v mmcil mxmmi 2w>mm Eaten Snatch new 33891 - i i : i um: i c i t: i i _.IIIWI.I_ “ 1.5203 mzo .332” £2.35 new senescent i i i 4 i i r .. i i ..m. i . i um. i . i t N —. mxmmkg ..:oesu83333308881 -iiuiiwiiiiwi. limit. ....Wii .18 anm4 ZO-Hém; 203391.55 m2; 1mm .m> m2; >MOP4