I‘M l “1‘le 1 IIHHIWIHIWI‘llllfllmllllllmll!W ms IHIUHIIHHIIIlllllllllllllllllllllllllllllllllHllllllllll . 02074 2148 f :33?) LEBRARY l Mflhhu’ul Sifiifl University This is to certify that the thesis entitled —AN INSTRUMENTED BOTTLE— The use of strain gage technology to measure the torsional forces applied to a package while online. presented by William George Kilhridge has been accepted towards fulfillment of the requirements for M. S. degree in Packaging Ma r professor Date 'Z/’§/fl 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. Dug org“ DATE DUE DATE DUE 11/00 chIRODateDmpBS—p.“ -AN INSTRUMENTED BOTTLE - THE USE OF STRAIN GAGE TECHNOLOGY TO MEASURE THE TORSIONAL FORCES APPLIED TO A PACKAGE WHILE ONLINE By William George Kilbridge A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE School of Packaging 1999 ABSTRACT -AN INSTRUMENTED BOTTLE - The Use of Strain Gage Technology to Measure the Torsional Forces Applied to a Package while Online. By William George Kilbridge The purpose of this research was to design, fabricate and test an instrumented package which, when placed in a production line could measure selected forces. This study focused on the measurement of torque exerted on a bottle during the application of a closure. Strain gages, adhered to an elastic member were used to measure these forces. The study demonstrated that an instrumented bottle, once calibrated, can provide a stable and repeatable method for the measurement of application torque. The vision for this instrumented bottle is to un-tether it, thus, enabling it to wirelessly transmit collected data to a system that will monitor or track machine function. The intent is to provide operations and maintenance staff with a tool that can provide information pertaining to machine performance. To Lucy ACKNOWLEDGMENTS I would like to extend a sincere thanks to my thesis committee: Dr. Robb Clarke, Dr. Burgess, and Dr. Soutas-Little. Thank you for the open doors and many hours of guidance though this project. An additional thanks goes out to those whose resourcefulness assisted me in attaining this goal: Bob Hurwitz, Dave Keller, and Tom Pellazaro. Finally, I would like to thank my wonderful wife Lucy, my family, and friends for their unwavering support. TABLE OF CONTENTS LIST OF TABLES ............................................................................. vi LIST OF FIGURES ........................................................................... vii INTRODUCTION ................................................................................. 1 CHAPTERI LITERATURE REVIEW .................................................... 3 CHAPTER II BACKGROUND INFORMATION ....................................... 7 CHAPTER III DESIGN AND FABRICATION .......................................... 15 CHAPTER IV METHODS .................................................................. 22 CHAPTER V RESULTS .................................................................... 27 CHAPTER VI CONCLUSIONS ........................................................... 34 APPENDIX A ................................................................................... 37 APPENDIX B ................................................................................... 38 APPENDIX C ................................................................................... 43 REFERENCES ................................................................................. 57 9‘99”!" “39".“? 11. 12. 13. 14. LIST OF TABLES . ANOVA for Initial Dynamic Torque Test ............................................. 30 ANOVA for the Revised Dynamic Torque Test .................................... 31 Static Torque Test Data ................................................................. 38 Static Torque Linear Regression ...................................................... 39 Calculation of Modulus of Elasticity Based on Static Torque Data ........... 4O 95% Confidence Interval for Static Torque ......................................... 42 Dynamic Torque Data(90 RPM) ....................................................... 43 Data Collected from Dynamic Torque Trials ....................................... 45 Single Factor ANOVA for Dynamic Torque Test Setting 1 ..................... 46 . Single Factor ANOVA for Dynamic Torque Test Setting 2 ................... 46 Single Factor ANOVA for Dynamic Torque Test Setting 3 ................... 47 Single Factor ANOVA for Dynamic Torque Test Setting 4 ................... 47 Linear Regression for Dynamic Torque Test Data .............................. 48 95% Confidence Interval Calculation for Dynamic Torque Test ............. 49 vi LIST OF FIGURES 1. Foil Type Strain Gage .................................................................... 10 2. Wheatstone Bridge Circuit .............................................................. 11 3. Simple Illustration of Shear Strain .................................................... 12 4. Test Bottle End Caps ..................................................................... 16 5. Test Bottle Body ........................................................................... 16 6. Static Test Finish .......................................................................... 17 7. Dynamic Test Finish ..................................................................... 18 8. The lnstrumented Shaft and Plugs ................................................... 19 9. lnstrumented Shaft Gage Orientation ................................................ 20 10. Wheatstone Bridge Schematic ....................................................... 20 11. The lnstrumented Bottle ............................................................... 21 12. Dynamic Torque Test Setup on Fowler Capper ................................. 25 13. Calibration Curve from Static Torque Test ........................................ 27 14. 95% Confidence Interval for Static Torque Test ................................. 28 15. Calibration Curve For Dynamic Toque Test — 90 RPM ........................ 31 16. Validation of Calibration Curve for Dynamic Torque Test ..................... 32 vii INTRODUCTION Currently, there is a push in the production arena to move from the widely accepted preventative maintenance (PM) methods into a predictive maintenance (PdM) mode. Predictive maintenance moves away from task scheduling at specific intervals to the periodic measurement of machine function. The goal is to anticipate machine failure by data analysis, and provide the appropriate maintenance activity prior to failure. The United States Navy, an advocate of PdM, discovered in the submarine fleet that large amounts of time and resources were lost in performing unneeded preventive maintenance.[1] It is estimated that “US. industry needlessly squanders an excess of $200 billion each year on inadequate or unnecessary maintenance procedures.”[2] With advances in computer technology and controls, much of the packaging industry has moved to sophisticated control systems which, when properly applied, can improve machinery operation and product quality. These same controls have provided manufacturers with very powerful tools to improve machine maintenance through the collection of data. Error logs and down time histories are examples of collected data which, when analyzed, can be used by maintenance staff to improve machine performance through informed maintenance decisions. There are, however, some machines or processes that cannot be reasonably integrated with these high-tech monitoring systems. In general, the limiting factor is price. Take for instance a large high-speed liquid filling line equipped with a 100 head filling and closing carousel. To equip each head to measure a force applied to the package would be prohibitively expensive. Packaging and production personnel currently have a few alternatives in these situations when online monitoring is not feasible. One alternative is to establish and perform a manual sampling plan to inspect product after it has been processed. A second alternative is to install an ancillary system that can divert product for automatic sampling. In either case, the system is functioning in a reactive mode, and generally only one attribute is monitored. Depending on the sampling frequency used by these systems, a machine may have the opportunity to produce a large quantity of out of tolerance product — in other words scrap — before the line personnel realize that there is a problem. The vision for this research evolved from the observation that the above mentioned ancillary systems only provide a few pixels in the picture of overall machine performance. Upon this realization, the goal was set to develop an instrumented or “rigged” package that could be introduced into a packaging operation with the intent of measuring the forces experienced in the capping process. In theory, decisions regarding line operation and maintenance could be made proactively by comparing recorded measurements to a pre-established machine baseline. Chapter 1 LITERATURE REVIEW Many sources were used to find information for this research. The databases used included the following: PIRA, Dissertation Abstracts, and the United States Patent Office. There was a plethora of resources available in the Engineering and Materials disciplines to aid in the design of the “rigged” model. Unfortunately little information was found in journals or theses citing the use of an instrumented package intended to measure packaging machine/line performance. In 1991, California State University—Long Beach utilized strain gauge technology to measure the torque counterforce in automobiles. The purpose of the study was to “design, fabricate, and evaluate a solid state torque measurement device that would operate in real time and be completely contained within a specific vehicle.”[3] In this study, an instrumented torque rod was tested statically and dynamically. The collected torque data was used to detenninelestimate engine horsepower without the use of a chassis dynamometer. Among the conclusions formulated, it was stated that “a solid state torque measurement device that operates in real time and can be completely contained within a vehicle can be made for a reasonable cost, and still maintain a high degree of correlation to true horsepower.”[3] 2ND In a 1996 presentation at the 4 annual Institute of Electrical and Electronics Engineers Pulp and Paper Conference several methods of measuring torque in an elastic member were discusses. The methods included Bonded Strain Gage Direct Torque Sensors, Phase Shift Direct Torque Measurement Sensors, and Magneto-Elastic Direct Torque Sensors. It was stated that strain gauges “can be bonded to the surface of a machine element in such a way as to be sensitive to the principle strains and thereby the principle stresses. Knowing the calibrated measure of the principal stresses and the section area of the machine element, the reaction force in an element can be calculated.”[4] The relationship between an applied moment (torque) and the stress/strain experienced by a shaft was discussed. The article states that “ both (combined strain and shear stress) can be seen to be proportional to T, the applied torque.”[4] “All practical torque sensors use the measurement of one or both of these elements (strain and or shear stress) as their means of deriving the torque transmitted by a machine element.”[4] A search of patents issued in the United States between 1979 and 1998 was performed. The keywords “packaging machinery” and “application torque” proved to be fruitful. The search revealed that, in at least two instances, commercial development has occurred in the area of online application torque measurement using an instrumented package. Humphries and Sangster of England were granted patent number 5,319,984 on June 14‘”, 1994. The patent states that the aim of the invention was “to provide a remedy to the problem of loose caps or insufficiently tightened caps, and/or to the problem of over-tightened caps.”[5] The device is described as follows: “a device comprising a dummy body, shaped to perform as a container body in such machinery, an externally screw-threaded neck part having the form of the neck of a container which is capped by such machinery, a substantially rigid stem secured rigidly to the neck part and to the dummy body, at least one strain gauge which is bonded along the outside surface of the stem to provide an indication of a torque applied to the neck part, and electrical-signal generating means within the dummy body, in which the strain gauge is connected to the electrical-signal generating means to generate an electrical signal when the device is in use, which is indicative of the torque applied by such capping machinery to the neck part, whereby such an indication is obtained without rotational displacement of the neck part relative to the dummy body."[5] The patent explained several details of the system, but did not provide any specifics regarding the strain gage positioning or placement. Conceptually, the method presented in the patent for measuring torque with a strain gage is the same as that discussed by Beihoff and many engineering texts - the strain exhibited by an elastic member is proportional to the applied force. On May 16‘“, 1995 Trendel and Spencer were issued patent number 5,415,050 for their invention. The patent was assigned to Owens-Brockway Glass Container Inc. of Toledo, Ohio. The device was described as a “self- contained apparatus for measuring threaded closure application torque that includes a container-shaped enclosure having a threaded finish portion configured to receive a threaded closure.”[6] Although the concept used by Trendel and Spencer utilizes strain gages, it is quite different from the method used by Humphries and Sangster. The patent states: “In the preferred embodiment of the invention, the finish portion of the enclosure is mounted for rotation with respect to the remainder of the enclosure. A load sensor assembly is mounted within the enclosure and coupled to the finish portion of the enclosure for developing the electrical measurement signal as a function of torque applied to the finish portion of the enclosure with respect to the remainder of the enclosure. The electronic circuitry stores the maximum or peak value of the sensor signal as indicative of the torque applied to the closure. The stored signal is scaled for direct reading in units of torque, and may be reset by an operator for reinsertion through the closure application system.”[6] The above mentioned load sensor assembly is known in engineering texts as a transmission dynamometer. In short, thin bars, which are equipped with strain gages, couple the freely rotating finish shaft to the fixed base of the unit. The thin bars deflect under the applied force. The strain gages stretch or compress creating a change in resistance. The strain gage output is proportional to the applied load. Although the above patents do not provide any data, they are important to note. A search of dissertations, theses, and journals regarding the use of instrumented packages intended to measure production line forces proved to be unproductive. The existence of patents, however, demonstrates that the premise is sound, and that there is potentially a need commercially for the type of online force measurement device proposed in this research. Chapter 2 BACKGROUND INFORMATION While developing the “rigged” bottle, many concepts and methods had to be learned. This chapter focuses on the concepts used in the development and testing of the test bottle. Measurement What is measurement? In it’s most simple form; measurement is the process of “obtaining a quantitative comparison between a predefined standard and a measurand.”[7] Measurement can be broken down into two methods: indirect and direct comparisons. Direct measurement uses a standard to determine the quantity present and indirect measurement methods utilize an input from the entity being measured. Indirect methods use a system that takes the input and “processes" it into a format that can be recognized or compared.[7] This study utilized an indirect method of measurement using resistance strain gages. mm The calibration of a measuring system is critical. If the data coming out of a system cannot be correlated to a known value, the system is useless. Another very important aspect of calibration is not only applying known values to a measurement, but also in demonstrating that the system can perform reliably. “At some point during the preparation of a measuring system, known magnitudes ofpinput quantity must be fed into a the sensor-transducer, and the system’s output behavior must be observed... this calibration procedure establishes the correct output scale for the measuring system.”[7] Streg—StrgiLRemnsjlip During axial loading, stress (0) is defined as the ratio of force (F) to cross- sectional area (A): 0': f. A Strain (s) is the ratio of a materials change in length (Al) to original length (I): Al 8 = — I The stress-strain diagram describes the behavior of a material. The proportional limit of a material represents the point at which the deformation of a material makes the transition from being elastic (capable of returning to its’ original length) to plastic (permanent deformation of the material). The slope of the linear portion of the stress-strain diagram is called the modulus of elasticity (represented by the symbol E). The modulus of elasticity is an indication of a materials stiffness. Robert Hooke in 1676 introduced a relationship between stress, strain, and the modulus of elasticity.[8] For a uniaxial state of stress the relationship, now called Hooke’s Law, states that: mlq E = Modulus of Elasticity o = Stress e = Strain Strain Gages Strain gages are devices used to measure the change in length of a material as a load is applied. The underlying principle behind strain gages states that a change in electrical resistance will occur when an electrical conductor changes in length. Lord Kelvin first demonstrated this phenomenon in 1856 “Electrically conductive materials possess a strain/resistance relationship defined as the ratio of relative electrical resistance change of a conductor to the relative change in its length.” Resistance for any given electrical conductor is represented by the following relationship[9]: R = Resistance p = Resistivity l = Length A = Cross Sectional Area When a conductor is placed in tension, the above relationship states that the resistance will increase due to an increasing length and a reduced cross sectional area due to Poisson’s effect. The relationship is opposite when a conductor is placed in compression — the conductor is shortened and the cross sectional area increases.[10] Using a “grid of fine resistance wire bonded to a test surface as a means of measuring strain“ was described in the late 1930s by Edward Simmons at the California Institute of Technology.[10] Today, strain gages come in many shapes and sizes, but the most prevalent is the foil type resistance strain gage. The foil type gage is a laminated structure of very thin foil on a backing material, usually polyamide. A photo etching process is used to create the fine gage grid. Figure 1 is an illustration of a typical general-purpose strain gage. Figure 1. Foil Type S‘train Gage.[10] Due to various material properties and test conditions, several gage characteristics must be considered during the strain gage selection process. Some of these characteristics include gage factor, resistance, temperature effects, operating temperature range, and fatigue Iife.[10] Wheatstone Bridge The majority of applications utilizing strain gages incorporate a circuit called a Wheatstone bridge. S.H. Christie devised the circuit in 1833.[7] The purpose of a bridge circuit is to convert very small resistance changes — as commonly seen in strain gages — into a voltage that can be amplified and read on a volt meter or other data acquisition device. The bridge is made up of four arms each containing a resistor. The resistors are labeled R1, R2, R3, and R4. The resistors can take the form of actual resistors, or resistive elements such as 10 strain gages and thennistors. A power source (Vin) supplies the bridge with power, and the change in voltage (Vom) feeds the data acquisition device. Figure 2 is a simple illustration of a Wheatstone Bridge circuit. a, ’ ‘ R; V out R4 ‘ ’ R3 viri Figure 2.Wheatstone BridgeTSircuit.[10] In a normal state, a bridge circuit is balanced, i.e., V0... is zero when R1/R4 = R2/R3. A change in resistance in one of the arms, for instance R1, will create an unbalanced system producing a voltage across Vom.[10] If the change in voltage is displayed on a voltmeter, the strain can be calculated. If the bridge circuit is connected to an instrument specifically intended for the measurement of strain, the instrument circuitry calculates and displays the strain. The bridge circuit can be configured in several ways when using strain gages: quarter bridge, half bridge, and full bridge. A quarter bridge places a strain gage in one of the four bridge arms. The half bridge places two strain gages in two of the four 11 bridge arms. The full bridge circuit replaces all four of the resistors in the bridge circuit with strain gages. The application to be studied will dictate which circuit configuration is appropriate. To_r_sion Formula for a Thin W_a_fled Tube Hooke's law can be used to calculate stress and strain in a material when torsional forces are applied. A cylindrical body in torsion experiences shear stress (1) and shear strain (7). Shear stress is considerably different than normal stress. Unlike normal stress, the acting force on a body is not perpendicular to its’ plane. The shear force orientation is tangential to the bodies’ plane resulting in a “sliding” of material on one side of the plane across the material on the other side.[8] Figure 3 is a simple illustration of shear strain. Figure 3. Simple Illustration of Shear Strain.[11] 12 The maximum shear stress (tmax) in a body is determined by the following equation: tmax=T-I=—CE J T = Applied torque (Force x Lever Arm) J = Polar moment of inertia C = Cylinder/Tube Radius The polar moment of inertia (J) is determined by the following equation: J— 11:=I‘(do4 —dl4) " 32 do = Ouside Diameter di = Inside Diameter When the properties of a material (E and v) are known, it is possible to calculate both shear stress and shear strain based on an applied torsional load. These theoretical calculations can be used to indicate whether a measured strain value is appropriate, or expected based on material properties. Shear strain (y) can be calculated using the following formula: 2*(1+u)*r max: I E v = Poissons Ratio E = Modulus of Elasticity 1’ = Shear Stress 13 Due to the positioning of the strain gages on the transducer shaft, the theoretical value for shear strain will be half as large as the strain indicated by a full bridge circuit. This is due to the 45 degree positioning of the gages on the transducer shaft. At 45 degrees the tensile and compressive strains are half of the theoretical shear strain value. Thus, with four active strain gages in the full bridge circuit, the actual output ends up being twice that of the theoretical value (4 gages x 0.5).[12] 14 Chapter 3 DESIGN AND FABRICATION A variety of engineering texts were used to aid in the design of the test bottle. Several models were designed and built during this project. Each model generation made improvements on the previous. Several objectives guided the design of the test bottle. The design objectives included the following: the test bottle must be capable of being instrumented; the design must measure application torque, the bottle must be stable, i.e., it must provide repeatable results; the design must be capable of being integrated into different package geometries; and it must be compatible with the available machinery in the Michigan State University School of Packaging production lab. The model consists of three components: the bottle housing, the finish, and the instrumented shaft. The Bottle Housing The bottle housing consists of three basic parts: two end disks, and the cylindrical body. The end disks are 3/8-inch thick Lucite fiat stock. Figure 4 is an illustration of the two end disks. The disks were turned to a diameter that produced a press fit (approximately 2.5 inches) into the cylindrical body. Both and disks had a 7/8-inch diameter hole bored through the center. The bottom disk had two 8-32 setscrew holes drilled and tapped through the side of the disk 90 degrees apart. The purpose of the setscrews was to hold the end of the transducer shaft in a fixed position. 15 2. IN. OD — — .875 IN. ID 50 \\ // /' _/ 1”” _ \ k Q:9 D Q:9 \ , / \\\-~-~-V:_JA/j/I/ / \\\.:___\v q , *:9 —_.9// \\\~‘__f -‘_/.//'/ \\--...,:,L // TOP END CAP BOTTOM END CAP Figure 4. Test Bottle End Caps. The cylindrical body consisted of a 3-inch diameter Lucite tube with a %- inch wall thickness. Figure 5 is an illustration of the Lucite test bottle body. 2.50 IN. ID ——\\ /~j;j::;il..—- \\\ \ ‘1" CYLINDRICAL BODY Figure 5. est Bottle Body. The tube was 5 inches long. The function of the cylindrical body was to provide a structure for the model and to protect the otherwise vulnerable strain gages on the transducer shaft. The size of the tube was arbitrarily chosen; a variety of sizes and shapes could have been selected to produce equivalent results. 16 The Bottle Finfleg The intent of the finish design was to provide a method of testing both static and dynamic forces. Thus, 2 separate finishes were developed; one for static and one for dynamic testing. The design allowed the finishes to be easily interchangeable. The static test finish had a clearance hole for a ‘/4-20 flat head bolt through its center, and two diametrically opposed 1I8-inch pinholes on a 0.600-inch diameter. The static test finish had a square profile. Figure 6 illustrates the static test finish. The static finish mated with a 4 inch Lucite disk intended to provide a consistent lever arm length. DIMENSIONS IN INCHES .265 THRU /\ ff? )1 1l8 DOWEL PIN HOLE <1: :7 PRESS FIT X2 .200 DP. I 0.587 \\ ”.7? , I : I:L-‘»;I:j I: r 0.100 ROTATED 45 DEGREES MATERIAL' ALUMINUM 6061 -T6 igure 6. Static Test Firfih. 17 The dynamic test finish had a clearance hole for a 1/4-20 flat head bolt through its center, and two diametrically opposed 1/8-inch pinholes on a 0.600- inch diameter. The profile of the dynamic test finish was that of a 28-400 bottle. Figure 7 is an illustration of the dynamic test finish. This finish allowed for the application of actual 28-400 continuous thread (CT) closures using capping machinery. DIMENSIONS IN INCHES 28-400 FINISH ——\ F3??? ,4 9:; NOTE: APPROXIMATELY 6 THREADS PER INCH 1.075 1/8 DOWEL 0.987 \ PIN HOLE \v/‘is\. 0.860 ”’4 ODD; / 9i, .f ?’ VIEW OF BOTTOM MATERIAL: ALUMINUM 6061-T6 Igure 7. Dynamic Test Finish. The lnstrumenteds_ha_ft_ The heart of the test bottle was the transducer shaft. Transducers operate by using “a spring element which is simply a piece of suitable metal designed to deform elastically and linearly when the desired parameter (force) is applied to it.” It “has strain gauges attached to it so that the deformation is converted into an electrical output.”[13] The electrical output is calibrated using known applied loads. 18 The elastic element was a 7/8-inch diameter aluminum tube with a wall thickness of 0.062 inches. The tube had a flanged plug made for each end. The plugs were epoxied and pinned into place. The top plug had a %—20—threaded hole through its center, and two diametrically opposed 1/8-inch pinholes on a 0.600-inch diameter. The total length of tube (plugs included) was 5.500 inches. Figure 8 is an illustration of the instrumented shaft and plugs. \-- TOP PLUG I C) \ \-- WALL THICKNESS = .0625 IN. xi CHEVRON STRAIN GAGES MOUNTED TO MEASURE TORSION (FRONT AND BACK) 5.50 IN. :9 “*l .875 IN. Ik ”) /- BOTTOM PLUG l \‘— -/ I \-_/ Figure 8. The lnstrumented ShafT and Plugs. The intent of the instrumented shaft was to measure both the torque and top load compression experienced by the test bottle during filling and closing. To accomplish the measurement of both forces, two independent circuits were used; one sensitive only to torsional loading, and one sensitive to compression. 19 The circuit for measuring torsion utilized special 350-ohm strain gauges designed for that purpose. The circuit formed a full bridge with two strain gages containing a total of four gage elements mounted in a chevron pattern. In a chevron pattern, the gage elements were positioned at 45 degrees to the axis of the tube, thus taking advantage of the principle axis of strain on the surface of the tube. The two pair of gages were mounted on opposite sides of the tube. Figure 9 below provides a simplified illustration of the strain gage orientation, and Figure 10 is a schematic of the Wheatstone Bridge circuit. Pi- S- P- 8+ Figure 10. Wheatstone ridge Schematic. 20 The gages were positioned in this manner because “when four gages are mounted on the shaft in pairs directly opposite each other, and in the directions of both principle axes, the bridge output will be: directly proportional to the applied torque, independent of axial thrust, independent of bending, and independent of temperature changes."[15] The gages were mounted on the shaft 4.5 inches from the bottom of the test bottle. Figure 11, below, is a photograph of the final bottle design. Figure 11. e nstrumente- 'otte. 21 Chapter 4 METHODS To accurately simulate realistic operation, the instrumented bottle had to be tested both statically and dynamically. The static tests were required to demonstrate that the strain gages performed properly. The dynamic tests provided the opportunity to see how the instrumented bottle performed while being loaded at realistic strain rates. Static Torgue Testing There were three objectives for static testing of the test bottle. The first objective was to demonstrate that the installation of the strain gages was correct. The second objective was to develop a mathematical relationship (the equation of a line) between a known applied load and the transducer output. The final objective was to demonstrate that the mathematical relationship was accurate. The static torque testing was performed using a method similar to that used in ASTM D 3474-80 “Standard Practice for Checking the Calibration of Torque Meters used in Packaging Applications”. A fixture was built to hold the bottle in a fixed vertical position. A special finish for the test bottle was designed for this test (Figure 6). The finish mounted directly onto the transducer shaft and a four-inch diameter Lucite disk mated with the finish. A length of wire rope was fixed to, and wrapped around the outside diameter of the disk. The wire passed over a pulley and had a loop in the end. A dead load was hung from the loop in 22 one pound increments. The following equation was used to determine the torque applied to the bottle: T = Torque F= Magnitude of Force (pounds) L = Lever Arm (2 inches) A Measurements Group P-3500 portable strain indicator was used to display the strain experienced by the transducer shaft. The strain indicator was wired as a full bridge circuit and balanced at zero when unloaded. The output of the strain indicator was in microstrain (p x 10 ‘5) units; the standard measure of strain when using strain gages. Weight was hung from the test bottle starting at one pound and increasing in one pound increments up to twelve pounds resulting in a range from 2 to 24 torque inch pounds (the four-inch diameter Lucite disk produced a two-inch lever arm). The strain data was collected at each step. Five trials at each weight increment were performed. The data were plotted, a linear regression was performed, and an equation of the line was determined. To validate the equation, thirty unknown weights were hung from the bottle. The strain values were applied to the equation to determine the unknown applied torque. The unknown weight was then weighed on a bench-top scale and compared to the calculated measure. Dynamic TorgueTesting The objective of the dynamic testing was to determine a relationship between an application torque and the output of the test bottle. 23 A bench top Zalkin capping machine was used to apply the 28-400 CT closures to the test bottle. The Zalkin capper featured a magnetic torque-limiting head which would “slip” once the set torque was reached. The machine was fitted with variable speed controls for both head rotation and vertical displacement. An LED output displayed capper spindle speed in revolutions per minute. A fixture was built to hold the test bottle securely on the machine base. The static testing finish on the test bottle was replaced with the 28-400 threaded finish. The machine cycle — spindle speed and vertical displacement — was set up to ensure that the capper head would not slip on the cap as the maximum torque was reached. It was initially noted that, at high spindle speeds and high torque settings, the head could slip on the cap, resulting in erroneous data. The spindle speed was set at 90 RPM (2.0 on the potentiometer dial), and the vertical displacement speed was set at 2.5 on the potentiometer. The vertical position of the head was set at 9.25 inches from the base. These settings proved to be acceptable across the torque range being measured. Figure 12 is a photograph of the dynamic torque test setup. As mentioned, the capper head was fitted with a magnetic clutch to limit torque. The applied torque was adjusted by increasing the penetration between the magnetic field in the fixed head body and a movable magnetic ring. The application torque increased as the penetration of the inner ring into the head body increased. The manufacturer supplied a setup gage consisting of four 24 torque settings; 12, 17, 22, and 27 torque inch pounds. The dynamic tests were performed using these settings. / . Figure 12. Dynamic Torque Test Setup on FowleGCper. The Measurements Group M—3500 Strain indicator was wired as a full bridge. A Measurements Group M-3650 Peak Read Indicator was wired to the M-3500 to record and hold the peak strain values reached during the capping cycle. 25 A sample of 50 measurements was taken at each torque setting and a mathematical relationship (the equation of a line) was determined. To validate the relationship, the machine was run at randomly selected torque settings using ten caps at each setting. 26 Chapter 5 RESULTS Static Torque Testing The calibration process for static torque demonstrated that the test bottle performed in a very linear fashion within the range of interest. Figure 13 illustrates the calibration curve produced from the collected strain data. I Static Torque Test : Calibration Curve 250 9 is‘ g 200 9 ‘9. m _//0/ g 150 _ ’91 .2 ,,,e/’/ g 100 -l V9V’ .E ./-»’/ ,9 5° “ 9/ y = 8.7258x R2 = 0.9959 ID V 0 l T T I l j 0 5 10 15 20 25 30 Applied Static Torque (in/lb) I Figure 13. Calibration Curve from the Static Torque Test. To validate that the recorded strain values were appropriate the Modulus of Elasticity (E) of the transducer shaft was calculated using the collected strain data and shaft geometry. Assuming a Poisson’s ratio of 0.3 the calculation yielded an average E of 9,501,024 psi. This calculation demonstrated that the 27 shaft performance was characteristic of aluminum, which has a documented E of approximately 10,000,000 psi. A 95% confidence interval was determined using the calibration data. A test of thirty random torque values was performed to validate the calibration curve. The collected strain data were applied to the calibration equation to predict the applied torque. Figure 14 illustrates that the predicted torque data from the calibration curve fell within the 95% confidence interval. Static Torque Trial Predicted Torque vs. Actual Torque 25.0 , l .» ‘ 9,90” g 20.0 - ,. ./° ' f I ,, 9/ 3 15.0 — ,. "91" ,, 9 E ,,9 ' .2 w" i - e: / ’ .8 10'0 I. ,, )2"? ’ -- 95% Confidence Level .‘é I 9.; “912'?" " '5 .9 an / 0 9 " '99/r— ’ 0- l sky"? / y = 0.9971x R2 = 0.9973 0 0 I '1? j I? 0.0 5.0 10.0 15.0 20.0 25.0 Actual Torque (In/lb) Figure 14. 95% Confidence Interval for the Static Torque Test. Dynamic Torque Testlig Several dynamic torque trials were run prior to collecting the final data. The early trials provided the opportunity to gain a further understanding of the machine operation and adjustments. The most appropriate machine configuration was determined based on the information gathered in the trials. 28 A single factor Analysis of Variance (ANOVA) was used to determine if the data collected at a given application torque came from the same population. ANOVA is a statistical method which allows the researcher to determine whether there is “equality or nonequality” between means.[16] A single factor ANOVA has a single variable (in this case the set application torque) “which may influence the dependant variable Y.”[16] The null hypotheses for the dynamic torque testing stated the following: H0: X sample 0: X samplel = X sample 2 = X sample 3 = X sample4 = X sample 5 H,: The six means are not equal. If the p-value is less than 0.05, the null hypothesis must be rejected. If the p- value is greater than 0.05, the null hypothesis must be accepted — suggesting that the samples are statistically equivalent based on the criteria and methods used. Table 1 summarizes the ANOVA for the initial dynamic torque trial. In this trial, the P-value for torque settings 17 and 22 torque inch pounds is less than 0.05 indicating a statistical difference between sample means. The test methods used in the trial were evaluated, and two potential causes of variability were identified: the changeover between torque settings, and spindle orientation. Upon examination of the changeover process, it was determined that variation in setting the torque head (due possibly to backlash between the inner and outer magnets) could have attributed to the differences between samples. The setup process was modified to ensure that the changes between torque settings were more consistent. The orientation of the spindle at 29 cycle start was questioned as well. To eliminate variation, the spindle was placed in the same position prior to the start of each machine cycle. Table 1. ANOVA for Initial Dynamic Torque Test (or = .05) Variation orque Setting 1 2 orque Setting 17 orque Setting 22 orque Setting 27 Table 2 summarizes the ANOVA for the data collected after the changes to the setup procedure had been made. In this experiment, all of the P-values were above 0.05, indicating that all of the sample means were equivalent based on the methods and criteria used for the test. The calibration curve from the data collected in the revised dynamic torque test is illustrated in Figure 15. As seen in the static torque tests, the relationship between the dynamic torque setting and recorded strain was very linear. A 95% confidence interval for the calibration data was determined to illustrate the consistency between it, the initial calibration data, and the test data. The calibration curve was used to calculate a predicted torque based on the 30 recorded strain from fifty samples taken from each of the four available torque settings. Table 2. ANOVA for the Dynamic Torque Test (a = .05) - Revised Procedure. Variation Dynamic Torque Calibration Curve (at 90 RPM) 300 I 250 9 /' E 5 g 200 _ 2 E. g 150 - '>' / E 5 100 9 3 y = 8.9384x R2 a 0.9919 50 . . . . . . . . 9 10 12 14 16 18 20 22 24 26 28 Torque Settlng (In/lb) Figure 15. Calibration Curve for Dynamic Torque Test — 90 RPM 31 Figure 16 illustrates that the predicted torque - calculated using the calibration curve — fell well within the 95% confidence interval. Valldatlon of Calibration Curve for Dynamic Torque Test (90 RPM) 30 — a 25 - 5 8 . 20 J 3 E .— 15 I - - - 95% Confidence Interval “- 10 9 L / y . 0.9999x 5 T f T T I I T T l 10 12 14 16 18 20 22 24 26 28 Actual Torque Setting (in/lb) Figure 16. Validation of Calibration Curve for Dynamic Torque Test. A calculation was performed to determine if the rate of loading had the potential to produce elevated recorded strain in the shaft due to wave propagation. The following calcualtion determines wave speed through aluminum: _ Exg CL C = wave speed in (in/sec) E = Modulus of Elasticity 9 = gravity or = material density (lbfina) 32 According to the wave speed equation it was determined that waves travel through aluminum at a rate of approximately 195,595 inches per second. This value is orders of magnitudes higher than the rate at which the shaft is loaded in practice. Based on this calculation, wave propagation through the instrumented shaft should not have any effect on the recorded strain values. 33 Chapter 6 CONCLUSION This research demonstrated that strain gages can be incorporated into an instrumented package with the intention of measuring the forces experienced when introduced to a packaging machine. The application and selection of an elastic member (the component that experiences the applied force) must be carefully considered. In this project, a circuit was applied to an elastic member (the instrumented shaft) with the assumption that it could provide an accurate strain measurement under an applied force. The force under study was the torque experienced during the application of a continuous thread (CT) closure. The circuit performed in a very stable fashion. The behavior of aluminum at elevated strain rates was examined. The relationship between applied torque and indicated strain was linear and repeatable in both static and dynamic situations. Several trials were run at various torque settings. With the aid of statistical analysis, it was determined that the collected data from the various trials were statistically equivalent, thus, they demonstrated stability. The primary objective of this research was satisfied. A method to measure a force online with an instrumented package was successfully developed and tested. Limitations of the Research Throughout this study, attempts were made to ensure that the designs and tests were representative of actual production issues. The simple cylinder 34 selected as the bottle body was chosen because a real package system or geometry was not available. In choosing an elastic member, an effort was made to ensure that it, with some reinforcement, could be applied to an actual package geometry. The Fowler capper used in this study is only one of many methods available to apply CT closures. The results from this test can only be applied to this type of machine. The instrumented bottle should be tested on other capping devices to determine if it can provide useful information regarding the application of closures in other machines. Are_as for Future Study Further study should be performed to identify additional methods of force measurement that can be incorporated into the structure of a rigged bottle. Load cells, thermocouples, and accelerometers are just a few of the possible measurement devices that could be included in an instrumented package. Additional work is required to identify methods that can be used to un- tether the instrumented bottle, thus, allowing wireless transmission of strain data. Radio telemetry and infrared data transmission are two potential technologies. By removing the constraint of wires, an instrumented package could be conveyed freely through the various stations of a packaging line to measure forces as it moved. 35 APPENDICES 36 APPENDIX A EQUATIONS Equation for Torque: T = Torque F= Magnitude of Force L = Lever arm Equation for Polar Moment of Inertia (J): J— 113*604 —dl4) ‘ 32 do = Ouside Diameter di = Inside Diameter 1: = 3.1487 Equation for Shear Strain (y): ymax=2*(1+u)*t v = Poissons Ratio E = Modulus of Elasticity r = Shear Stress Equation for Modulus of Elasticity in Torsion of Thin Wall Cylinder: I32*(1+v)*(T)*(-n_ 99w“ 93:30; 956580 2:38..” N: .98 8a.: 829 S .8289. Swan 9.98 6.23: 95.23: 9 858.8m EBESEBE u. 22 mm a <>oz< N_. mcozmaomno $3. Lem. 2888 83 988 m 89.0.32 3.9.0 99939. . moans: mozwzflw co_mm9mcm .98.: watch 265 Lo“. :23“:me Emma: .v can... 39 Table 5. Calculation of Modulus of Elasticity Based on Static Torque Data. Observed Observed Calculated Torque Strain Strain Modulus (in/lb) (microstrain) (in/in) (psi) 2 19 0.000019 9,046,400 4 40 0.00004 8,594,080 6 58 0.000058 8,890,428 8 75 0.000075 9,167,019 10 91 0.000091 9,444,044 12 108 0.000108 9,548,978 14 124 0.000124 9,702,994 16 141 0.000141 9,752,148 18 157 0.000157 9,853,086 20 174 0.000174 9,878,253 22 190 0.00019 9,951,040 24 207 0.000207 9,964,151 2 20 0.00002 8,594,080 4 41 0.000041 8,384,468 6 58 0.000058 8,890,428 8 75 0.000075 9,167,019 10 91 0.000091 9,444,044 12 107 0.000107 9,638,221 14 124 0.000124 9,702,994 16 140 0.00014 9,821,806 18 156 0.000156 9,916,246 20 173 0.000173 9,935,353 22 190 0.00019 9,951 ,040 24 207 0.000207 9,964,151 2 20 0.00002 8,594,080 4 41 0.000041 8,384,468 6 57 0.000057 9,046,400 8 74 0.000074 9,290,897 10 91 0.000091 9,444,044 12 107 0.000107 9,638,221 14 123 0.000123 9,781,880 16 140 0.00014 9,821,806 18 156 0.000156 9,916,246 20 172 0.000172 9,993,116 22 189 0.000189 10,003,691 24 206 0.000206 10,012,521 2 20 0.00002 8,594,080 4 40 0.00004 8,594,080 6 57 0.000057 9,046,400 40 8 74 0.000074 9,290,897 10 90 0.00009 9,548,978 12 107 0.000107 9,638,221 14 123 0.000123 9,781,880 16 140 0.00014 9,821 ,806 18 156 0.000156 9,916,246 20 173 0.000173 9,935,353 22 189 0.000189 10,003,691 24 206 0.000206 10,012,521 2 19 0.000019 9,046,400 4 40 0.00004 8,594,080 6 57 0.000057 9,046,400 8 74 0.000074 9,290,897 10 90 0.00009 9,548,978 12 107 0.000107 9,638,221 14 123 0.000123 9,781,880 16 140 0.00014 9,821,806 18 156 0.000156 9,916,246 20 172 0.000172 9,993,116 22 188 0.000188 10,056,902 24 206 0.000206 10,012,521 Avera e E 9,501,024 41 xmm c Ifiiéiméfm < 0.35; : $-32 03581;" D N e... < l _ I mimmiilofiic Eisaduo 9&1 xVanww ”macaw—sow was: _ 8.9va xwm 00.v~ F0.- _.0.»000 00. _.a _ov.0ow a0.» 30a» av.~0 3.0» 00. I. vw Emma v0.00 00.~000 00¢» om.a0p a0.» 0000 av.m0 3.0» Ba Nu 00.0w 0».0F 3.0000 0a.»0 000»: am.» 0a.0o0 av.N0 3.0» cc.» om N00: 000: 00.0.»: 00. 3. 00.00: am.» or. :00 av.~0 3.0» 8.0 0: a».0F 00.0 r 5.30 00.0w cad: a0.» av. FNN av.N0 3.0» 8.0 0: »0.3 o_..0_. »0.0» 00.0 _ov.0mv a0.» 0a....» av.N0 3.0» 00.: 3 5.? 0m. 2. 3.00 N0»- _ow.»oF a0.» 0¢.00. 0vN0 3.0» cc. _.- N: 5. 3 Eva 0v.000 Nada- _o0.oa a0.» :0. F E... 0v.~0 3.0» 00.0- or 0N.a av.» 90.000: Nvdv. Eve» _a0.» »0.N00- 0v.m0 3.0» 00.0.. 0 cm.» 0.0 »0.0aN0 NV.»0- _ov.»0 a0.» 5. 30- 0¢N0 3 .0» 00.»- 0 00.0 a0.0 V0.»000 N39»- _ov.ov am.» 00.90. 0v.~0 3.0» 80. v »a.~ Fae :N.00ola| N~.0a- 300: am.» 00.000- 0YNO 3.0» cc. Fm: N :E: :E: 99...: 3-5 _ 9 o o m < 97.9: x Long: 526.. _ £90 £000 awe... cacao... 0030 Lo: 503330 .0295 8:08:00 $0a .0 0.22. 42 APPENDIX C DYNAMIC TORQUE DATA Table 7. Dynamic Torque Data (90 RPM). Units in Microstrain. Torque Torque Torque Torque Setting Setting Setting Setting 1 2 3 4 99 147 196 244 99 149 199 243 99 148 200 243 99 148 198 245 100 150 200 241 97 146 201 249 101 150 198 245 96 150 198 242 100 148 196 243 100 149 199 243 97 153 197 245 98 149 200 246 102 149 197 240 100 147 201 245 100 149 199 244 99 150 201 245 100 150 200 243 99 152 195 243 99 150 198 245 97 149 199 245 99 147 200 243 97 150 201 243 98 150 199 244 98 154 199 242 97 153 198 243 97 154 196 242 101 150 204 243 98 148 199 243 102 151 197 243 99 151 199 244 97 151 198 243 99 149 198 247 100 149 197 248 43 100 1 50 1 99 245 1 00 1 50 201 245 96 149 200 244 1 02 1 54 201 247 1 02 148 202 244 98 1 50 201 244 101 1 51 201 246 1 02 1 51 201 244 98 1 51 200 243 1 01 1 53 200 244 98 149 1 98 245 99 149 201 243 99 1 50 200 243 1 00 1 51 201 245 99 1 53 198 246 103 149 201 244 101 1 52 1 96 245 AVG 99 1 50 199 244 STD 1.70 1.88 1.84 1.66 N 50 50 50 50 MAX 1 03 1 54 204 249 MIN 96 146 1 95 240 3N 3N a3 aa 03 0a3 30.. Now owN »a3 ..0.. NE. 0VN 0a? V03 3.. 0vN aa.. 003 aa 0VN aaF N03 N03 3N »av 303 car »VN 03 00.. For 0VN aap 003 aa 0VN aaF 03 SF 3N 03 03 car 3N 0a.. N03 0a 0¢N 08 »3 aa 3N 0a.. .03 0a 9N av 03 0a NVN aaw a3 0a 0VN aaF »3 »a NVN SN »3 0a 03 NON a3 For 3N aaF 003 aa v 0 N F v 0 N 3 9.030 9.530 9.030 9.330 9.030 9530 9530 9.030 26.0... 26.8 26.0... 26.0.. 95.0.. 2.90.. witch 39o... 0 3...... v 3...... SN aaF 303 »a 9N aaF N03 aa 3N 0a? 303 For 0VN EN 003 5.. 9N aaF 003 No_. 0VN aaw 003 0a 0VN 0a3 N03 2: MEN 8.. N0.. N2. »VN 8N N03 For 0¢N aaF 303 aa 0VN aaF N03 Nor 0VN SN 00.. aa 0vN »aF 003 For »VN »aP N03 N03 NVN aaF N03 »a NVN »a.. 03 003 0¢N aap 003 03 3N aaw N03 0a 0VN 0a? 303 3.. 0vN 0a.. a3 aa 3N »av 003 aa 0VN »a3 a3 Nov 0VN av 03 5.. mwN 0a? »3 aa NVN 0a? a3 0a NVN 0a3 »3 on: NVN »a3 »3 »a 3N »a3 .03 »a 3N aaP 03 aa 3N »av 003 8.. v 0 N _. v 0 N F v 0 N 3 9.030 9.590 9.030 9530 9.030 9530 9.030 9.030 9.030 9.030 acfiom 9.300 2.96... 266.3 witch 26.0.3 266... 390... 2.0.0.3 966... 266... watch 266... 26.0... n 3...... N 3...» _. 3...... 63.30.25. 5 0...: 33:... 2.90... 2:395 Eot 028.30 330 .0 Q3... 45 Table 9. Single Factor ANOVA for Dynamic Torque Test Setting 1 Groups Count Sum Average Variance Sample 0 50 4962 99.24 2.88 Sample 1 10 989 98.9 2.1 Sample 2 10 1006 100.6 1 .82222 Sample 3 10 999 99.9 4.32222 Sample 4 10 996 99.6 4.26667 Sample 5 10 994 99.4 2.04444 ANOVA Source of SS df MS F P-value F cn't Variation Between 20.72 5 4.144 1.43149 0.220138 2.31127 Groups Within 272.12 94 2.89489 Groups Total 292.84 99 Table 10. Single Factor ANOVA for Dynamic Torque Test Setting 2 Groups Count Sum Average Van'ance Sample 0 50 7500 150 3.55102 Sample 1 10 1510 151 7.333333 Sample 2 10 1501 150.1 6.1 Sample 3 10 1507 150.7 2.9 Sample 4 10 1504 150.4 4.266667 Sample 5 10 1491 149.1 4.766667 ANOVA Source of SS df MS F P-value F cn't Variation Between 23.01 5 4.602 1.075287 0.379134 2.311268 Groups Within 402.3 94 4.279787 Groups Total 425.31 99 46 Table 11. Single Factor ANOVA for Dynamic Torque Test Setting 3. 47 Groups Count Sum Average Variance Sample 0 50 9958 199.16 3.402449 Sample 1 10 1985 198.5 1.833333 Sample 2 10 1982 198.2 1.733333 Sample 3 10 1977 197.7 2.455556 Sample 4 10 1986 198.6 0.488889 Sample 5 10 1994 199.4 4.044444 ANOVA Source of SS df MS F P-value F crit Variation Between 27.04 5 5.408 1.942351 0.094571 2.31 1268 Groups Within 261 .72 94 2.784255 Groups Total 288.76 99 Table 12. Single Factor ANOVA for Dynamic Torque Test Setting 4. Groups Count Sum Average Variance Sample 0 50 12204 244.08 2.76898 Sample 1 10 2443 244.3 3.788889 Sample 2 10 2443 244.3 3.788889 Sample 3 10 2442 244.2 2.177778 Sample 4 10 2439 243.9 6.322222 Sample 5 10 2444 244.4 4.266667 ANOVA . Source of SS df MS F P-value F crit Variation Between 1.97 5 0.394 0.11618 0.988533 2.311268 Groups Within 318.78 94 3.391277 Groups Total 320.75 99 88... 88... 88... 88... o 83$ 83... one... Foaga>x $2.. $2.. $2.. $2.. $2.. $2... .32.. o .8925 .3031 $931.98.. 3%: 003.128.. .90 .ami. Em Samuem .080 $089.3 8m EB :03 308.... 8. .828... :38... ~33 amimowm 33.58% F 8.898”. n. 0.5 n. Ms. mm .3 <>02< 08. 22.9.88 at... am 9333 .8... .3 m 8.8.2 Name 295 m 08... m 22.32 8.3.3.0 5.3903. .330 3o... 6:90... 2:356 .00 co_mmo.aom 39.... .03 0.3... 48 3.3 5.3 3.53 3.3- oo. 33 end 3.33. no. 3 3.3 emee 3 3.3 3.3 3.neoe 3. 3- oo.N3 end odene. no. 3 3.3 3.3 3 3.3 no.3 nod3e 3.3- 8.3 end 3.3e. 3.3 3.3 emee 3 no.3 «dd enoone 3.3- 8.3 end ne.one- - 3. 3 3.3 3.8 3 e~.3 3.3 Kddve 3.3- oodd end 3Nee- 3. 3 3.3 e~.ee 3 3.3 3.3 3.neoe 3.3- oo.~3 end od.ene- no.3 3.3 made 3 3.3 3.3 3.neoe 3. 3- oo.~3 end od.ene- no. 3 3.3 3.8 3 3.3 3.3 3.neoe 3. 3- oo.~3 end od.ene- no. 3 3.3 nude 3 3.3 3.3 3.neoe 3. 3- oo.~3 end od.ene- no. 3 3.3 nude 3 5.3 e~.3 dndene 3.3- oo.o3 end tunne- 3. 3 3.3 emee 3 3.3 3.3 Kddve 3.3- oodd end ohmee- 3. 3 3.3 enee 3 3.3 5.3 bosome 3.3- oo. 33 end 3.33. no. 3 3.3 emee 3 5.3 e~.3 dndene 3.3- oo.o3 end Knee. 3. 3 3.3 ewee 3 e~.3 3.3 3.33 3.3- oodd end 3.~ee- 3. 3 3.3 emee 3 3.3 5.3 nosowe 3.3- oo. 33 end nnéve. no. 3 3.3 3.8 3 3.3 no.3 nod3e 3.3- 8.3 end 3. 3e- 3. 3 3.3 ewee 3 3.3 5.3 bosome 3.3- oo. 33 end nnéwe. no. 3 3.3 emee 3 em.3 3.3 Eddve 3.3- oodd end 3Nee- 3.3 3.3 3.3 3 no.3 Ndd enoone 3.3- 8.3 end neoee- 3. 3 3.3 3.3 3 3.3 no.3 nodvee 3.3- 8.3 end 3. 3e. 3. 3 3.3 emee 3 e~.3 3.3 Eddve 3.3- oodd end 3.~ee- 3. 3 3.3 emee 3 3.3 3.3 3.33 3.3- oodd end 3Nee- 3. 3 3.3 emee 3 no.3 Ndd enoone 3.3- 8.3 end 933- 3. 3 3.3 omen 3 5.3 em.3 dndene 3.3- oo.o3 end Knee. 3. 3 3.3 emee 3 ._E_._ :5: Li... 3-..... > o o m < v.3“. x 533 526-. £85 dcfimw Sam. on .m 268... 0.8930 8.. 5.3.3.8 329:. 8:03:00 $8 .3 flow... 49 No.5. No.9. moéoo crow- ooKE. omo whom”. oo.oo Voon mmo 5 oo.: oo.? oo.oom o...o~- oo.omv moo oo.oom- oo.oo 3.2. mmo D. oN.N_. .1on :6on 9.65- oo.oo ooo onNoo- N..._.o vodx. mNom NF mod? ovow mndoom o... E- oo.No_. omo oo.omo. oo. _.o woo“ mmom NF oNNF Snow {down 03K- oo.oo moo oNNoo. N... _.o voon mwom NP no.3 oN.o_. omdmmm 94K- oo.oo—. moo undmo- o_.._.o vodn omom NP ONNF v_..o_. 95¢on 92:. oo.oo moo onNoo. N... 5 woos mmom NF nod? om.o_. oodmmm 9.3... oo.oo_. omo unomo. of; woos mwom NP oN.N_. Sow {down 9.65- oo.oo omo ohdoo. Nvéo woos mwom NF Sump oo.ow oo.ovom ofmn- oo.oo omo vo. go. 3.5 voon owom N? ooNv Nod oo.ooom eon- ooNo moo oo.ooo. taro woos mmom NF oYNF no.9. nokomm gums- oo. For omo modem. no. 5 woos mmom NP omdw owor mhdoom or. E- oo.No_. omo oo.omo. mo. 5 woos mmom N_. ova no.9. noNon 9N5- oo. For omd ooévo. no. Po woos mmom N_. om.m_. ovov mnooom o... E- oo.NoF moo oo.omo- moéo took mNom NF ovNF no.9. noNon ofmn- oo. _.o_. moo mo.v¢o. no. 5 Vooh mmom NP oer boo? woNon ofimn- oo.Fo_. moo ooévo. no.5 3.2. mwom NF Sam? oo.ow oo.ofiom ofmn- oo.oo moo vo. who. Euro coon mmom NF S..N_. mo.o_. oo.ovom 2.2.- oo.oo moo 3.2.0. 3. Po woos mwom N... 3.9 oo.ov oo.ovom 92mm- oo.oo mod #0. who. 3.. _.o 3.9m mmom N_. No. 3 Foo hoooom 9.2.- oo.oo moo Foooo- or. 5 vodn mNoo NP oN.N_. 3.3 :oovm 92¢»- oo.oo moo Eamoo- NV. Po voow mmom NP ooNF woo oo.ooom Son- ooNo omo oo.ooo- 2.. 5 wooh mmom N_. ova? no.3 woNon oFNN- oo. For omo moévo- no. _.o Vooh mmom NF no.3. owor omdmmm 9.4K- oo.oow omo unooo. o_.._.o cook. mmom NF oN.~_. 3.2. {down ofivn- oo.oo omo ondoo. N_.. _.o woos mmom NP ovNF moor noNon ofmm- oo._.o_. moo ooévo. no.5 woos omom NF oNNP oo.ov mvNNov ofion- oo.mo_. moo oo.owo. oo.—o woos mwom NF 50 no.9. no.9. no._.ov 9..NN- ooé9. on.o mono—H Nnon finon own n_. oN.n_. ow.9. nvoov nn.on- oo.n9. nn.o 9..on F- on.on vn.on mun nw 9..nF no.9. mn.n3~ 9.. FN. oo.n9. on.o oo.onT Fnon wnon own nw no.9. no.3 no. row 9..NN- oo. For nn.o mono —... nn.on Vnon nun nv oo.nw on.9. onnnn 9..nN- oo.o9. on.o oo.ooN- nn.on vn.on mud n_. nn.n_. no.9. no.nno ofimm- oo.nE. ono nnémm- ¢n.on Vnon own nw no.9. no.9. no. no? 9..NN.. oo. _.9. on.o mono _.- nn.on vnon own n_. nn.nw nn.9. _.n.nnn nwéw- oo.o3. on.o oo.9.N.. nn.on vnon mun n_. mnne nn.9. _.n.nnn 9..vm- oo.o3 on.o oo.nFN- nn.on vn.on nun n_. oN.n_. o_..nw nvoov ofom- oo.n9. on.o 9..on _.- on.on Vnon own nw nfinw no.0? nn.n$~ oréw- oo.noF on.o oo.on _.- _.n.on vnon mud n_. oo.nr ov.9. nvnofi 9..oN- oo.n9. on.o ofon _.- on.on vnon own nw n_..nv no.9. nn.n¢¢ 9.. FN. oo.nov nno oo.on T. _.n.on vn.on nun nF n_..n_. noor nn.nvv 9....N- oo.nov on.o oo.on _.- nn.on vnon mud n_. nfinw no.9. nn.nv¢ or. FN- oo.nmw ono oo.on T _.n.on Vnon mud n_. oo.nr on.9. ononn 9..nN- oo.omw on.o oo.noN- nn.on vnon own n_. mn.nF nnnw _.n.nnn owéw- oo.o3. ono ooovN- nn.on vn.on own nF Fn.n_. 3.9. no.nnn 9..nN- oo.nE. ono _.n.N¢N.. nn.on vnon own n—. Nn.n_. no.9. nnéno 9..nN- oo.nS. on.o nn.nnN- nn.on vn.on mud n_. nn.n_. no.9. no.nno 92mm- oo.ng on.o nnéNN- wnon vn.on mud nw no.nw no.9. noéov nFNN- oo. _.9. ono mono _.- nn.on ¢n.on own n—. oN.n—. ow.9. niooe. 9..o~- oo.n9. on.o 9..on —... on.on vn.on own nw nfinv no.9. nn.nvv or. N. oo.n9. on.o oo.on _.- nn.on vnon own nv no.3 no.9. nvéom own..- oo.n9. on.o vnnm T oNon vn.on mud n—. 929. no.9. on.n$~ or. _.N. oo.n9. ono oo.o9.- _.n.on vnon own n_. n_..nw no.9 mn.n3. orém- oo.n9. nno oo.on _.- nn.on ¢n.on own n—. oo.n_. nnor on.nno 9..nN- oo.o9. on.o oo.oom- nn.on Vnon mud n_. No.n_. nn.on onéno 9..oN.. oo.nE. on.o nn.nnN.. nn.on vnon own n—. 51 oo.nN ow. Fm on.own vnoN oo.ooN on.o on.onN on.on coon on.o NN no.nN oo. PN no.onn vn.nN oo. row on.o nn.n¢N on.on ¢n.on on.o NN Sunm n. _.N wn.noo finoN oo.oop on.o PoonN on.on vnon on.o NN vino vn. FN _.n.noo vnoN oo.o9. on.o FoonN on.on Vnon on.o NN NNnN N... rm on.noo vnnN oo.noF on.o Von—.N nn.on vnon on.o NN nn.nN nm. _.N nonvo vnéu oo.now on.o no. FNN vnon wnon oNo mm anm NFQN on.ooo vnnm oo.n9. on.o Von—.N nn.on wnon on.o NN NNnN Nvém on.noo wnnN oo.no_. on.o vonFN nn.on ¢n.on on.o NN oo.nw on.o? _.n.nno owém- oo.o3. on.o oo.orN- nn.on vnon oNo n_. non? no.9. no. wov oFNN- oo. _.9. on.o mono? nn.on Vnon on.o n_. oN.n_. on.o? nvoow on.oN- oo.nov on.o 9..o9... on.on ¢n.on on.o nF no.nw no.9. no. wow owNN- oo. F9. on.o No.no_.- Noon ¢n.on on.o nv oo.n_. on.o? on.ono oan- oo.o9. on.o oo.ooN- nn.on vnon on.o nw nn.n_. no.9. oo.nno 9..o~- oo.nS. on.o nnémm- vnon vnon on.o n_. won? moor onéno on.on- oo.nE. on.o nn.nnN- on.on vnon on.o n_. Fonw Sow no.nnn 9..n~- oo.o3. on.o rnNVN- on.on wnon on.o nw mo.n_. no.9. onéno 9..oN.. oo.nvw on.o nn.onN- on.on vnon on.o nw on.nw on.on _.n.nno ofivm- oo.o3. on.o oo.oFN- nn.on ¢n.on on.o nw no.9. no.9. no. no». oFNN- oo. F9. on.o mono _.- Nnon wnon oNo nw 3.9. on.9. :non 9..oT oo.».ow on.o NNén—H on.on vn.on on.o n_. nfinw no.9. on.n$~ or. K. oo.now on.o oo.on _.- wnon vnon on.o n_. oo.n_. on.9. on.ono ofnm- oo.o9. on.o oo.ooN- nn.on Vnon on.o n_. nn.n—. no.9. oo.nno ofoN- oo.nS. on.o nnéNN- Vnon vnon oNo nF n_..9. no.op on.n¢¢ or. K. oo.n9. on.o oo.on? —.n.on Vnon oNo n_. no.9. no.9. no. no? oFNN- oo. For on.o Nono T nn.on wnon on.o nw oo.nw on.9. _.n.nno ofivm- oo.o3 on.o oo.9.N.. nn.on vnon oNo n_. mo.n—. no.9. onéno on.on- oo.ni on.o nn.nnN- on.on Vnon oNo n_. oo.nw on.ow on.ono ofnm- oo.oow on.o oo.ooN- nn.on ¢n.on on.o n_. 52 fivnm vn. Pm —.n.noo vnom oo.oow on.o FoonN on.on vn.on on.o NN nn.on no. —.N nonvo vnéw oo.n9. on.o no. PNN ¢n.on vnon on.o NN nn.nN nN. PN no.n_.o vném oo.now on.o no. VNN Vnon vnon on.o NN Ixnn vn. _.N _.n.noo vnow oo.o9. on.o FoonN on.on ¢n.on oNo NN oo.nN 9V. Pu on.omn vnow oo.ooN on .o on.onm on.on vnon on.o NN gnu vn. Fm rnnoo vnoN oo.oo_. on.o FoonN on.on vnon on.o NN wvnm vn. Fm _.n.noo vnoN oo.oo_. on.o FoonN on.on vnon on.o NN no.nN oo. Pm no.onn vnnN oo. row on .o nn.nvw on.on vnon on.o NN Snow Po. FN no. Poo vnNN oo.oo_. on.o ofivoN nn.on Vnon on.o NN SVnN wn. FN _.n.noo vnom oo.oow on.o _.o.onN on.on vnon on.o NN nn.on NF. Fm on.ooo vnnN oo.now on.o Von—.N nn.on «non on.o NN NNnN NF. rm on.ooo vn.nN oo.noF on.o on _.N nn.on vnon on.o NN From 3. PN no. So VnNN oo.oow on.o ofvom nn.on vnon oNo NN NNnN NF. _.N on.ooo vnnN oo.now on.o Von Fm nn.on wnon on.o NN nn.on nm. ..N oo.nro vnéN oo.n9. on.o no. FNN ¢n.on wnon on.o NN NNnN NPéN on.noo wnnN oo.no_. on.o Von Fm nn.on Vnon on.o NN Yvon Vn. Pm Fn.noo vnou oo.oor on.o _.o.onN on.on «non on.o NN Sunm wn. PN _.n.noo vnom oo.oop on.o Foonm on.on vnon on.o NN Sinw Vn. FN Pn.noo vnom oo.oo_. on.o _.o.onN on.on ¢n.on on.o NN Yvon vn. Fm _.n.noo ¢n.oN oo.oo_. on.o PoonN on.on vnon on.o NN NNnN NF. Pm on.ooo vnnN oo.now on.o Von FN nn.on vn.on on.o NN gnu wn. _.N _.n.noo vnoN oo.oow on.o Poonm on.on vn.on on.o NN _....nN 5. PN no. FNo vnNN oo.oow on.o ofivom nn.on ¢n.on oNo NN gnu Vn. PN Fn.noo vn.oN oo.oo_. on.o _.o.onN on.on Vnon on.o NN Snnw 3. FN no. «No vnNN oo.oo_. on.o o_..voN nn.on vnon on.o NN Venn vn. Fm _.n.noo vnom oo.oow on.o Poona on.on Vnon on.o NN nn.on nm. _.N no.n_.o vném oo.n9. on.o no. FNN ¢n.on vnon on.o NN SVnN Vn. FN _.n.noo ¢n.oN oo.oor on.o FoonN on.on Vnon on.o NN 53 oo.nN ovoN oo.o9.o 3.: oo.ovm on.o noévo no._.n 3.on on.oo nN oo.nN ovoN oo.o9.o 3. _.n oo.ovm on.o no. F3 no. 3 3.on on.oo nN oNnN ofom oo.nnnv 3.no oo.nwm on.o 9..o _.o oo. Pn 3.on owoo nN nv.n~ nn.oN _.n.nwoo 3.on ooéwm on.o 3.nno 3. _.n 3.on oNoo nN nvnN nn.on _.n.nwoo 3.on ooévm on.o vonno 3. Po 3.on oNoo nN oo.nN owom oo.oowo 3. _.n oo.ovm on.o no. _.3 no. Pn 3.on on.on nN wnnw _.n.oN on.No3 3.nn oo.nVN on.o 3.ooo _.F. Fn 3.on on.oo nN oo.nN nn.on no. _.ooo 3.vn oo.nvm on.o nn.ooo 3.. _.n 3.on on.oo nN oN.nN 3.on oo.nnnv 3.no oo.nvm on.o on.9.o oo. 3 3.on oNoo nN nvnN nn.oN 3.9.oo 3.on oodcn on.o 3.nno 3. wn 3.on oNoo nN nvnN nn.oN 3n _.oo 3.on oo.3~ on.o 3.nno 3. Fn 3.on on.oo nN nn.on on.on no.nnnv 3.oo oo.nvm on.o o_..¢No no. 3 3.on on.oo nN oN.nN 9..oN oo.nnnv 3.no oo.nvw on.o ofiowo oo. 3 3.on on.oo nN nVnN nn.oN _.n.nvoo 3.on oo.3~N on.o 3.nno 3. _.n 3.on oNoo nN no.nN oo. _.N no.onn 3.nN oo. 3N on.o nn.n¢~ on .on 3.on on.o NN NNnN N... 3 oo.noo 3.nN oo.n9. on.o 3.9.N nn.on 3.on on.o NN 3.on 3. _.N _.n.noo 3.on oo.o9. on.o _.o.onN on.on 3.on on.o NN NNnN N... _.N on.noo 3.nN oo.n9. on.o VonFN nn.on 3.on on.o NN 3.on 3. PN _.n.noo 3.on oo.o9. on.o PoonN on.on 3.on on.o NN nn.nN nm. _.N nonwo 343 oo.n9. on.o no. FNN 3.on 3.on on.o NN on.on on. _.N nvoon 3.oN oo.noN on.o oo.ooN nn.on 3.on oNo NN 3.on 3. FN _.n.noo 3.oN oo.o9. on.o _.o.onN on.on 3.on on.o NN 3.nm 3. 3 _.n.noo 3.on oo.oow on.o Fo.onN on.on 3.on on.o NN nn.nN no. _.N on. _.nn 3.on oo.noN on.o unnom nn.on 3.on on.o NN nn.on nm. 3 no.n_.o 3.v~ oo.no_. on.o no. _.NN 3.on 3.on oNo NN nn.on nm. 3 no.n_.o 3.v~ oo.no_. on.o no. _.NN 3.on 3.on on.o NN 3.on 3. FN _.n.noo 3.on oo.o9. on.o vo.onN on.on 3.on on.o NN nn.nN nm. 3 no.n_.o 34am oo.no_. on.o no. 3N 3.on 3.on on.o NN oo.nN o¢.oN oo.oo—.o 3. rn oo.oVN on.o no. :3 no. 3 3.on on.on nN oo.nN ovoN oo.oo 3 3. wn oo.ovu on.o no. _.3 no. 3 3.on on.oo nN no.nN nn.on no. _.ooo 3.vn oo.nvm on .o nn.noo 3.. 3 3.on on.on nN _.n.nN oo.oN no.oono 3.nn oo.o¢~ on.o _.o.ooo oo. 3 3.on on.oo nN nn.nN on.on no.nnnv 3.oo oo.nvm on.o oficmo mo. 3 3.on on.on nN nn.nN nn.on on.mo3 3.nn oo.nvw on.o 3.ooo _.F. 3 3.on on.on nN oo.nN ovoN oo.oo—.o 3. _.n oo.ovm on.o no. _.3 no. 3 3.on oNoo nN nwnw nn.oN nn.nroo 3.on oo.3~ on.o 3.nno 3. 3 3.on on.on nN ownw 3.oN nn.noov 3.no oo. :3 on.o nn.ooo noon 3.on on.oo nN on.on 9..oN oo.nnnv 3.no 09va on.o 9..o 3 oo. 3 3.on on.oo nN oN.nN o_..oN oo.nnnv 3.no oo.nvm on.o 9..o3 oo. 3 3.on on.on nN o _..nN 3.on nn.noov 3.no oo. 3N on.o nN.ooo oo.on 3.on on.on nN ninm nn.oN _.n.n_.oo 3.on ooévm on.o vonno 3. 3 3.on on.oo nN oo.nN ovom oo.oo_.o 3.nn oo.ovm on.o no._.3 no.3 3.on oo.oo nN _.n.nN oo.oN no.oono 3.nn 00.93 on .o 3.ooo oo. 3 3.on on.on nN _.n.nN oo.oN no.oono 3.nn oo.ovw on.o Fo.ooo oo. 3 3.on on.oo nN nn.nN on.on no.nnnv 3.oo oo.nvm on.o o_..vno mo. 3 3.on on.oo nN oNnN on.on oo.nnnw 3.no ooNvm on.o 93 3 oo. 3 3.on on.oo nN oo.nN ovom oo.oo_.o 3._.n oo.ovm on.o noévo no.3 3.on on.on nN oo.nN ovoN oo.oo3 3._.n oo.ovm on.o no._.3 no.3 3.on on.oo nN oNnN 9..o~ oo.nnnV 3.no oo.N¢N on.o ofio 3 oo. 3 3.on oNoo nN ninm nn.oN _.n.nvoo 3.on oo.3~ on.o 3.nno 3. 3 3.on on.on nN nn.nN on.on no.nnnv 3.oo oo.nvm on.o owémo No. 3 3.on on.on nN nn.nN on.oN no.nnnv 3.oo oo.nvm on.o o —. .vmo mo. 3 3.on on.on nN nn.nN on.on no.nnnv 3.oo oo.nvm on .o o _. .vmo No. 3 3.on oNoo nN BEIN‘QWN no. room 3.vn oo.nVN on.o nn.noo 3.. 3 3.on on.oo nN NnnN enoN onNo3 3.nn oo.nvm on.o 3.ooo 2.. 3 3.on oNoo nm a on.on no.nnn¢ 3.oo oo.nvm on.o o_..VNo No. 3 3.on on.oo nm 55 xww c I j + 4 + :< n m < I xmm . . mwo+o+mu §|~m|< 035?.“ .I. D < wu_ l + x n I _ u MWDIO I Elo>fipmno «CM 5% me ”mcozmaam 9..nn_. ooNo xww nv.nN nn.oN _.n.nwoo 3.on oo.vvm on.o 3.nno 3. 3 3.on oNoo nN 3.nN no.oN oo.noSV 3.oo oo.ovm on.o oN.noo oo.on 3.on oNoo nN _.n.nN oo.oN no.oono 3.nn oo.ovm on.o _.oooo oo. 3 3.on oNoo nN nv.nN nn.oN _.n.nwoo 3.on oo.3VN on.o 3.nno 3. 3 3.on oNoo nN no.nN NnoN no. _.ooo 3.vn oo.nvm on.o nn.noo 3..3 3.on oNoo nN nvnN nn.oN _.n.nnoo 3.on oo.3- on.o 3.nno 3. 3 3.on oNoo nN oo.nN ovoN oo.oo—.o 3. _.n oo.ovm on.o no. V3 no. 3 3.on on.oo nN nn.nN oN.oN no.nnn.v 3.oo oo.nvm on.o ofivwo No. 3 3.on oNoo nN 56 REFERENCES 57 Bibliography 1. Robinson, Charles J.,Ginder, Andrew P., Implementing TPM: The North American Experience, Productivity Press, Portland, OR, 197 pp. (1995). 2. Fogel, 6., “Improve Output Through Reliability-Based Maintenance,” Packaging Technology and Engineerigg, 2, pp. 22-27 (1997). 3. Grain, Lanoe J., Strain Gauge Measurement of Torque Counterfoce jg Automobiles, California State University, Long Beach, CA, 55 pp. (1991). 4. Beihoff, B., “ A Survey of Torque Transduction Methodologies for Industrial Applications”, in Pulp and nger Industry Technicgl 1996. Proceedings, IEEE, Annual Technical Conference, pp. 220 - 229 (1996). 5. Humphries, Glyn A., and Sangster, Ronald, “Device and Method of Monitoring Toque or Force,” United States Patent Office. USA (1994). 6. Trendal, Alois F., and Spencer, Albert T, “Method and Apparatus for Measuring Threaded Closure Application Torque,” United States Patent Office. USA (1995). 7. Beckwith, Thomas G. and Marangoni, Roy D. and Lienhard, John H., Mechanical Meafiurements, 5th ed., Addison-Wesley Publishing, New York, NY, 876 pp. (1993). 8. Bowes, William H. and Russell, Leslie T. and Suter, Gerhard T., Mechanics of Engineering Materials, John Wiley & Sons, Inc., New York, NY, 610 pp. (1984). 9, Chalmers, G,F., “Materials, Construction, Performance, and Characteristics,” in AL \Nil‘ldOW, ed., Strain Gauge Technology, Elsevier Applied Sciences, New York, NY, pp. 1-38 (1992). 10. Scott, K. and Owens, A., “Instrumentation,” in AL. Window, ed., Strain Gauge Technology, Elsevier Applied Sciences, New York, NY, pp. 151- 216(1992) 11. Cook, Nathen H., Mechanics and Materials for Design, McGraw-Hill Book Company., New York, NY, pp.479 (1984). 12. Pople, J., BSSM Stgin Measurement Reference Book, in AL. The British Society of Strain Measurement, Newcastle upon Tyne, England, 208 pp. (1 979). 58 13. VVIndow, A.L., “Strain Gauging Techniques For Transducers,” in AL. Window, ed., Slain Gaug Technology, Elsevier Applied Sciences, New York, NY. pp. 85-95 (1992). 14 . Lineback, L.D., “Experimental Stress Analysis Notebook,” Measurements Group, Inc, 6, pp. 6-15 (1987) 15. Murray, William M. and Stein, Peter K., S_train Gage Techniques, Massachusetts Institute of Technology, Cambridge, MA, 588 pp. (1956). 16. Hamett, Donald L. and Horrell, James F., Data. Statistics. and Decision Models with Excel, John Wiley & Sons, Inc., New York, NY, 602 pp. (1998). 59 TRT UNIV. LI E BRRRIES III III IllIIIIIIIlIIIlllIIII 22 17 iZ‘IIIEB .I nzcuxcnn s IIIIIIIIIIIIIIIIIIII III 3129 0 13