Queensland University Of Technology School Of Engineering Systems Faculty Of Built Environment Engineering
DEVELOPMENT OF A HIGH-SPEED SENSING AND DETECTION SYSTEM FOR AUTOMATIC REMOVAL OF PACKAGES WITH LEAKY SEALS FROM A HIGH-SPEED FOOD PROCESSING AND PACKAGING LINE
Gary Raymond Gibson – Student No. 00924725 B.Eng (Mechanical), AD Mechanical Eng RPEQ 7587, MIEAust CPEng NPER 1302262 (Mechanical and Structural)
Principal Supervisor: Professor Doug Hargreaves
Associate Supervisor: Associate Professor Gopinath Chattopadhyay
Submitted to Queensland University of Technology for the degree of: BN72 MASTER OF ENGINEERING (by Research)
JULY 2009
ABSTRACT Contamination of packaged foods due to micro-organisms entering through air leaks can cause serious public health issues and cost companies large amounts of money due to product recalls, consumer impact and subsequent loss of market share. The main source of contamination is leaks in packaging which allow air, moisture and microorganisms to enter the package. In the food processing and packaging industry worldwide, there is an increasing demand for cost effective state of the art inspection technologies that are capable of reliably detecting leaky seals and delivering products at six-sigma. The new technology will develop non-destructive testing technology using digital imaging and sensing combined with a differential vacuum technique to assess seal integrity of food packages on a high-speed production line. The cost of leaky packages in Australian food industries is estimated close to AUD $35 Million per year. Contamination of packaged foods due to micro-organisms entering through air leaks can cause serious public health issues and cost companies large sums of money due to product recalls, compensation claims and loss of market share. The main source of contamination is leaks in packaging which allow air, moisture and micro-organisms to enter the package. Flexible plastic packages are widely used, and are the least expensive form of retaining the quality of the product. These packets can be used to seal, and therefore maximise, the shelf life of both dry and moist products.
The seals of food packages need to be
airtight so that the food content is not contaminated due to contact with microorganisms that enter as a result of air leakage. Airtight seals also extend the shelf life of packaged foods, and manufacturers attempt to prevent food products with leaky seals being sold to consumers. There are many current NDT (non-destructive testing) methods of testing the seal of flexible packages best suited to random sampling, and for laboratory purposes. The three most commonly used methods are vacuum/pressure decay, bubble test, and
i
helium leak detection. Although these methods can detect very fine leaks, they are limited by their high processing time and are not viable in a production line. Two nondestructive in-line packaging inspection machines are currently available and are discussed in the literature review. The detailed design and development of the High-Speed Sensing and Detection System (HSDS) is the fundamental requirement of this project and the future prototype and production unit. Successful laboratory testing was completed and a methodical design procedure was needed for a successful concept.
The Mechanical tests
confirmed the vacuum hypothesis and seal integrity with good consistent results. Electrically, the testing also provided solid results to enable the researcher to move the project forward with a certain amount of confidence. The laboratory design testing allowed the researcher to confirm theoretical assumptions before moving into the detailed design phase. Discussion on the development of the alternative concepts in both mechanical and electrical disciplines enables the researcher to make an informed decision.
Each major mechanical and electrical
component is detailed through the research and design process. The design procedure methodically works through the various major functions both from a mechanical and electrical perspective. It opens up alternative ideas for the major components that although are sometimes not practical in this application, show that the researcher has exhausted all engineering and functionality thoughts. Further concepts were then designed and developed for the entire HSDS unit based on previous practice and theory.
In the future, it would be envisaged that both the
Prototype and Production version of the HSDS would utilise standard industry available components, manufactured and distributed locally. Future research and testing of the prototype unit could result in a successful trial unit being
incorporated
in
a
working
food
processing
production
environment.
Recommendations and future works are discussed, along with options in other food processing and packaging disciplines, and other areas in the non-food processing industry.
ii
ACKNOWLEDGEMENT The preparation of a Thesis is not possible without the direct and indirect support for a large number of people. I would like to take this opportunity to acknowledge everyone who has contributed to complete this project: •
Associate Supervisor and Previous Principal Supervisor – Associate Professor Gopinath Chattopadhyay, former QUT Senior Lecturer and current CQU Director Engineering Postgraduate Studies.
Thank you for introducing the
project to me and assisting in commencing and continuing my journey through the Masters program, and also for your sincere and tireless support, assistance, encouragement and guidance throughout the entire project. •
Principal Supervisor and Head of School, School of Engineering Systems, QUT – Professor Doug Hargreaves.
Thank you for your guidance, support and
mentoring through the later stages of this project. I am very grateful for your assistance in the direction of project and during publication preparation. •
Associate Professor Wageeh Boles, QUT Nathan Newman’s electrical engineering supervisor, who provided inspirational electrical background data and vital research assistance.
•
Peter Fell, Engineering Librarian, QUT Gardens Point for his valuable assistance in preparing my Literature Review and also teaching the advanced information retrieval skills subject which greatly assisted my research ability.
•
Delwar Hossain Khan, QUT coursework in Masters of Engineering Management who provided original contribution and background support documentation.
•
Wesley Mergard and Nathan Newman, QUT Undergraduate Students who assisted me through their introductory work linked to the research of this project.
•
QUT Research administration staff including Elaine Reyes, Christine Percy, Diane Kolomeitz and Kate McKee for their valuable support with administration and documentation preparation and procedures during my Masters program.
•
GRG Consulting Engineers Pty Ltd staff including Andrew Hartley and James Storer for assisting me in completing the engineering drawings under my supervision.
•
Finally, my wife Michelle and children Marcus and Sarah for their support and sacrifice and encourage throughout this Masters program.
iii
STATEMENT OF ORIGINALITY I declare that to the best of m knowledge the work presented in this thesis has not been previously submitted for degree or higher at any other tertiary education institution. To the best of my knowledge and belief, the project report contains no material previously published or written by any other person except where due reference is made.
Signed:…………………………………….. Date:………………………………………..
iv
LIST OF RESEARCH PUBLICATIONS
PUBLICATIONS RESULTING FROM THIS THESIS
Referred National Conference Paper Gibson, G.R., Chattopadhyay, G, and Hargreaves, D (2008) “Non-destructive testing technology to assess seal integrity of packaged products”, 9th Global Congress on Manufacturing & Management, GCMM 2008, Holiday Inn, Gold Coast, Queensland. (Based on Chapters 1 and 2). RM No. 2009002148
Referred International Conference Paper Gibson, G.R., Chattopadhyay, G, and Hargreaves, D (2009) “High-Speed Sensing and Detection System for Automatic Removal of Packages with Seal Quality Issues from a Food Packaging Line”, 22nd International Congress on Condition Monitoring and Diagnostic Engineering Management, COMADEM 2009, Miramar Palace, San Sebastian, Spain. (Based on Chapters 1 to 5). Referred International Conference Paper (in progress) Gibson, G.R., Chattopadhyay, G, and Hargreaves, D (2010) “Cost Effective Solution for Detection of Leaky Food Packaging Seals using Innovative Non-Contact Testing”, 23rd International Congress on Condition Monitoring and Diagnostic Engineering Management, COMADEM 2010, Nara Prefectural New Public Hall , Nara, Japan. (Based on Chapters 3 to 5).
v
NOMENCLATURE A
Area cross section
Cd
Co-efficient of Discharge
∆
Delta, Deflection
ƒ,F
Force
|G|
Gradient
H
Head
L
Length between start of chamber and sensor
m³
Cubic Metres
m/s
Metres per second
n
Number of molecules of gas
N
Speed, Tension
σ
Sigma, Stress
Pa
Pascals
ρ
Rho, Density (kg/m³)
π
Pi
p
Pressure
P
Air Pressure, Misread, False Alarm
pcd
Pitch Circle Diameter
PRS
Production Run Size
Q
Discharge, Flow Rate
R
Universal gas constant
^
To the power of
T
Absolute Temperature
µ
Mu, Co-efficient of Friction
V
Air Volume
vi
CONTENTS ABSTRACT ...........................................................................................................i ACKNOWLEDGEMENT ......................................................................................iii STATEMENT OF ORIGINALITY .........................................................................iv LIST OF PUBLICATIONS AND EVENTS ............................................................v NOMENCLATURE ..............................................................................................vi LIST OF TABLES.................................................................................................4 LIST OF FIGURES ...............................................................................................5 CHAPTER 1 SCOPE AND OUTLINE OF THESIS ....................................................................7 1.1 INTRODUCTION .......................................................................................7 1.2 BACKGROUND OF PROBLEM ................................................................7 1.3 PREVIOUS THESIS - 2001 ..................................................................... 10 1.4 SCOPE OF STUDY ................................................................................. 12 1.5 LAYOUT OF THESIS .............................................................................. 13 CHAPTER 2 INTRODUCTION AND OVERVIEW ................................................................... 14 2.1 INTRODUCTION ..................................................................................... 14 2.2 PROJECT AIMS ...................................................................................... 14 2.3 LITERATURE REVIEW ........................................................................... 15 2.4 CON-CURRENT RESEARCH WORK ..................................................... 21 2.4.1 MECHANICAL ...................................................................................... 22 2.4.1.1 Introduction .................................................................................. 22 2.4.1.2 Food Packaging ............................................................................ 22 2.4.2 ELECTRICAL ....................................................................................... 23 2.4.2.1 Problem Description ...................................................................... 24 2.4.2.2 Related Problems and Solutions ..................................................... 24 2.4.2.3 Project Plan .................................................................................. 24 2.4.2.4 Expected Outcomes ...................................................................... 26
2.5 CONCLUSION......................................................................................... 26
1
CHAPTER 3 DESIGN AND DEVELOPMENT ......................................................................... 27 3.1 INTRODUCTION ..................................................................................... 27 3.2 LABORATORY DESIGN TESTING ......................................................... 27 3.2.1 MECHANICAL TESTING ....................................................................... 27 3.2.2 ELECTRICAL TESTING ........................................................................ 32
3.3 DESIGN PROCEDURE ........................................................................... 36 3.3.1 MECHANICAL AND ELECTRICAL COMPONENTS ................................ 36
3.4 DEVELOPMENT & ANALYSIS OF ALTERNATIVE CONCEPTS ........... 50 3.4.1 MECHANICAL CONCEPTS ................................................................... 50 3.4.2 ELECTRICAL CONCEPTS .................................................................... 54
3.5 RESEARCH & DESIGN OF COMPONENT EQUIPMENT ...................... 59 3.5.1 MECHANICAL DETAIL DESIGN ............................................................ 60 3.5.1.1 Conveyor Belt and Cleats .............................................................. 60 3.5.1.2 Vacuum Tunnels ........................................................................... 70 3.5.1.3 Vacuum Chamber ......................................................................... 74 3.5.1.4 Gaskets ....................................................................................... 84 3.5.1.5 Regenerative Circuit ...................................................................... 85 3.5.1.6 Vacuum Pump and Control ............................................................ 87 3.5.1.7 Packet Rejection ........................................................................... 92 3.5.1.8 Packet Leveller ............................................................................. 94 3.5.2 ELECTRICAL DETAIL DESIGN ............................................................. 98 3.5.2.1 Sensing and Decision Making ........................................................ 98 3.5.2.2 Camera ...................................................................................... 104
3.6 CONCLUSION....................................................................................... 104 CHAPTER 4 SELECTION AND PROCUREMENT................................................................ 105 4.1 INTRODUCTION ................................................................................... 105 4.2 SELECTION & PROCUREMENT OF EQUIPMENT.............................. 105 4.2.1 PROTOTYPE UNIT ............................................................................ 106 4.2.2 PRODUCTION UNIT ........................................................................... 108
2
4.3 ASSEMBLY & TESTING ....................................................................... 110 4.3.1 MECHANICAL COMPONENTS ........................................................... 110 4.3.2 ELECTRICAL COMPONENTS ............................................................. 113
4.4 COST ANALYSIS .................................................................................. 115 4.4.1 PRODUCT COST ............................................................................... 115 4.4.2 RESEARCH AND DEVELOPMENT COST............................................ 116 4.4.3 EQUIPMENT COST ............................................................................ 117
4.5 CONCLUSION....................................................................................... 119 CHAPTER 5 CONCLUSIONS, RECOMMENDATIONS AND FUTURE WORKS ................. 120 5.1 INTRODUCTION ................................................................................... 120 5.2 CONCLUSIONS .................................................................................... 120 5.3 DISCUSSIONS AND LIMITATIONS ...................................................... 121 5.4 RECOMMENDATIONS ......................................................................... 122 5.5 FUTURE WORKS ................................................................................. 122
REFERENCES ................................................................................................. 124
APPENDICES................................................................................................... 127 A - DRAWINGS ........................................................................................... 127 B - DESIGN CALCULATIONS..................................................................... 129 C - COST OF EQUIPMENT......................................................................... 135 D - LABORATORY TESTING DATA ........................................................... 136 E - EQUIPMENT BROCHURES.................................................................. 143
3
LIST OF TABLES Table 3.1: Heights of Twisties Packets at Various Vacuum Levels .......................... 137 Table 3.2: Heights of Doritos Packets at Various Vacuum Levels ........................... 138 Table 3.3: Test Case Summary ............................................................................. 33 Table 3.4: Idea Combination Excel Spreadsheet ..................................................... 38
4
LIST OF FIGURES Figure 1.1: Chip Packet with Improper Seal ..............................................................8 Figure 1.2: Sealed Snackfood Packet at Atmospheric Pressure and Partial Vacuum .. 10 Figure 1.3: Concept Proposed in 2001 ................................................................... 11 Figure 2.1: NPC-501 Leak Seal Inspection System ................................................. 17 Figure 2.2: PTI 550 Seal Scan .............................................................................. 18 Figure 3.1: Laboratory Testing of Multipack Doritos packet displaying packet under vacuum .............................................................................................. 28
Figure 3.2: Photograph showing a Sealed Doritos™ Packet inside a –20kPa Vacuum Chamber ............................................................................................ 29
Figure 3.3: Graph showing Change in Heights of Twisties Packets ........................... 29 Figure 3.4: Graph showing Change in Heights of Doritos Packets ............................ 30 Figure 3.5: Variation of packet height under differential vacuum conditions ............... 34 Figure 3.6: Typical Application of Thru-Beam Sensors............................................. 39 Figure 3.7: Key components of the Omron Z4LC Parallel Beam Sensor .................... 40 Figure 3.8: Preliminary results of using Sobel Edge Detection .................................. 45 Figure 3.9: Preliminary results of using Thresholding and Region Growing ................ 46 Figure 3.10: Sample Polygon Approximation from our preliminary results ................. 50 Figure 3.11: Single Packet Discrete System ........................................................... 51 Figure 3.12: Continuous Conveyor ........................................................................ 52 Figure 3.13: Continuous Cleat Belt ........................................................................ 53 Figure 3.14: Sample Edged Profile Function from our preliminary results .................. 56 Figure 3.15: Model of Major Component Equipment ................................................ 59 Figure 3.16: Diagram Showing the Lip Seal Profile to be cut into Cleat ..................... 61 Figure 3.17: Flexco HD Tatch-a-Cleat .................................................................... 62 Figure 3.18: Flighted Modular Belt ......................................................................... 63 Figure 3.19: Two-Piece Cleat Concept................................................................... 64 Figure 3.20: Remacleat Concept ........................................................................... 66 Figure 3.21: Assumed Dimensions for Lip Seal....................................................... 68 Figure 3.22: Seal Dislodgement Point .................................................................... 69 Figure 3.23: Folded Sheet-metal Concept .............................................................. 71
5
Figure 3.24: UHMWPE Assembly Concept ............................................................. 73 Figure 3.25: Diagram of the Layout of the Walls to Ensure an Airlock is created ........ 77 Figure 3.26: Air Pressure Distribution in Assembly before Equalisation ..................... 78 Figure 3.27: Air Pressure Distribution in Assembly after Equalisation ........................ 79 Figure 3.28: Compact Vacuum Chamber Design .................................................... 80 Figure 3.29: Chamber Top Plate ........................................................................... 81 Figure 3.30: Screw-in Plug used to Seal Tapped Holes ........................................... 82 Figure 3.31: Section View of Chamber ................................................................... 83 Figure 3.32: Fillet on Underside of Top Plate .......................................................... 83 Figure 3.33: Gasket Arrangement for Chamber ...................................................... 84 Figure 3.34: Schematic Diagram Representing the Expansion of Air Entering the Vacuum Chamber ............................................................................. 85
Figure 3.35: Chamber Pressure Diagram ............................................................... 86 Figure 3.36: Pressure of Air Entering Vacuum Chamber .......................................... 86 Figure 3.37: Rotary Vane Pump ............................................................................ 89 Figure 3.38: Rotary Piston Pump........................................................................... 90 Figure 3.39: Pneumatic Push Rod on Conveyor...................................................... 93 Figure 3.40: Brush Concept .................................................................................. 95 Figure 3.41: Foam Roller Concept ......................................................................... 96 Figure 3.42: Flexible Pad Concept......................................................................... 97 Figure 3.43: Test Performed on Packets with Triangulation Sensor .......................... 99 Figure 3.44: Recommended Ultrasonic Displacement Sensor ................................ 101 Figure 3.45: Parallel Laser Concept ..................................................................... 102 Figure 3.46: Digital Imaging Concept ................................................................... 103 Figure 4.1: Proposed Final Design of Prototype HSDS .......................................... 106 Figure 4.2: Proposed Final Design of Production HSDS ........................................ 108 Figure 4.3: Proposed Food-Grade Industrial Castor (Rotarola) ............................... 109 Figure 4.4: Test Rig Configuration for Testing the Leakage Rate over One Cleat ..... 112 Figure 4.5: Detailed View of Tailpiece Configuration for Testing Leakage over One Cleat .................................................................................................................. 113
6
CHAPTER 1 SCOPE AND OUTLINE OF THESIS 1.1
INTRODUCTION
Quality assurance is becoming an increasing issue in the snack food industry in an age where consumers are aware of the risks associated with spoiled food. The Smith’s Snackfood Company (TSSC) experienced such issues in their potato chip products because of unsealed packets. In previous work of this project, it was found that an economically suitable on-line testing machine was not available and it was therefore proposed to develop a new design. This Thesis investigated current literature regarding current package testing methods and defined the design functions and constraints of the problem. A laboratory experiment was completed and presented regarding the testing of the concept of packet change in height in a vacuum. After exploring various ideas and concepts, a detailed design was proposed. The Thesis also detailed the progress of the project throughout 2004 and made several recommendations for future work. Chapter 1 details the background of the current problem, and previous preliminary research on solving the leak detection problem common to many food processing companies throughout the word. 1.2
BACKGROUND OF PROBLEM
Potato chips are generally processed by continuous or batch methods, seasoned and packaged into individual packets, then placed in cartons for distribution to wholesale or retail outlets.
The packaging process at TSSC Tingalpa plant generally involves
automatic weighing and “bagging” through a vertical form, fill and seal (VFFS) machine. This involves the chips falling into a formed film package and sealing of both ends via thermal sealing. Historically, some chips may be still free falling when the thermal
7
sealing process is taking place and some residual chips may become lodged in the seal. This issue results in improper and non air tight incomplete seals (see Figure 1.1).
Figure 1.1: Chip packet with improper seal Although most packets are sealed properly, previous quality audits at TSSC Tingalpa plant indicated, “a fresh batch of potato chip packets in a carton was observed and 1.5 % were found with leaks” (Khan, 2001a). Although the obvious solution to this problem at TSSC would incorporate a redesign or new VFFS machines, this would incur significant plant and equipment capital costs, as there are many such machines at the Tingalpa plant. Operators also adjust the VFFS machines, which can affect the quality of the film sealing process. Another variable in the solution is film quality, and although lower film quality results in significant cost reduction per packet produced, it can have negative effects on seal integrity. Through all this, some companies consider the cheaper option is to discard defective packets than to invest in process redesign or additional capital investment. For most sizes of chip packets, defective packets become rejected when hand packed into boxes, as the packets will deflate when handled. Although this method obviously does not test for extremely small leaks, it does provide 100% product testing. As packaging operations become more automated, manual packing of potato chip packets will increasingly become redundant. Food processing companies worldwide are continually working through methods to reduce labour for Repetitive Strain Injury (RSI) type tasks. Good Internal Rate of Return (IRR) on productivity capital proposals 8
is attractive to company executives.
The introduction of automatic case packing
machines directly after the VFFS machines will result in human intervention being reduced, and identifying of faulty seals subsequently reduced. Many snackfood and baking companies using VFFS machines will heavily rely on automated inspection points to reject faulty packets.
An opportunity is available to introduce a Hazard
Analysis Critical Control Point (HACCP) prior to the packets being placed in a carton for direct distribution to the market. Another issue is the small Multipack chip packets at TSSC that have only small revenue per item, and are cost prohibitive to justify manual packing. The unchecked or uninspected packets are transported directly to an automatic outer bag VFFS machine or Cartoner machine to complete the packaging process. During this process, the individual multipack packets are not tested or inspected for leaks. Although recent advances in VFFS machine have resulted in increased ability to detect chips trapped in the packet seal, considerable capital investment is required to introduce these machines across an entire packaging line. TSSC and other snackfood and baking companies, are therefore faced with a decision between the following two options: Option 1: Continue to have chip packets entering the market with potential leaks. There is a direct financial advantage with this option, with no significant capital cost and therefore little financial risk. However, there is potential disadvantage in terms of continuing loss of revenue which is large enough to consider, and does not strive towards product improvement.
Many modern snackfood and baking manufacturing
companies strive to improve their quality and consumer complaints records, with Continuous Improvement (CI) being a major Key Performance Indicator (KPI) for all employees. Option 2: To reject defective packets using an in-line non-destructive testing machine. A major advantage of this option is an increase in revenue from enhanced consumer confidence and reduced consumer complaints. However, for this option to be justified, the payback period or IRR must be within the life of the machine or better. Some companies IRR can be a high as 20-30% of capital investment.
9
The second option was explored and the associated problem statement was “to design a non-contact, in-line (100% NDT) package seal testing apparatus initially for TSSC”. The proposed method was to use the change of height of packets in partial vacuum to determine seal integrity. 1.3
PREVIOUS THESIS - 2001
In 2001, it was found that there were not any economic in-line seal testing methods applicable to TSSC potato chips available in industry. Therefore, TSSC raised the issue to In Motion Engineering for finding out an effective solution (Khan, 2001b).
The
concept that was developed was based on the following primary statement or theory: "If a sealed flexible package is subjected to a partial vacuum, then the air inside the packet will expand. " (see Figure 1.2)
Figure 1.2: Sealed snackfood packet at atmospheric pressure and at partial vacuum Tests performed by In-Motion Engineering found that this statement was true. This package swelling occurs because of the air inside the packet being at higher pressure than the air surrounding the packet. The package material is forced to deflect outwards due to the following holding correct: Forceon membrane = ∆ Pressure x Area .................................................................... (1.1) Force on a Small Element of the Packet when Subjected to Partial Vacuum (Douglas et al, 1995) The secondary theoretical statement of the concept was: "If the package has a gross or fine leak, the package will not puff up". Testing was conducted and it was found that this statement is true for most leaks encountered by TSSC. This occurs due to the air inside the packet leaking through the hole, and the subsequent packet not expanding.
10
This allows pressure equalisation on either side of the packet membrane and there is no net force on packet. In-Motion Engineering attempted to build an in-line testing apparatus for TSSC that determined seal integrity and made a decision (i.e. accept/reject), based on the height of packet in a partial vacuum. Although very limited literature was available on this machine, it was found that the machine was unsuccessful because: •
packets are not smoothly transferred to the vacuum chamber by a rotor feeder;
•
air leakage through the rotor feeder is extremely high;
•
mechanical spring loaded sensor device does not give accurate result of packet heights; and
•
packets were not passing through the delivery feeder smoothly.
The machine was deemed not reliable for uniform flow of packets and 100% accuracy in sorting out the leaky or unsealed packets from the sealed ones (Khan, 2001c). The project was then handed to Queensland University of Technology (QUT) as a Cooperative Education for Enterprise Development (CEED) project in 2001. Khan in 2001 developed the concept of measuring the change in packet height with a non-contact displacement sensor. This helped account for various size packets entering the system. Khan developed the concept for a discrete multi-batch testing machine (Figure 1.3), where a lid covers packets on a conveyor belt and the air is evacuated.
Figure 1.3: Concept proposed in 2001 (Khan, 2001d) 11
The change in height measurement of the packet was completed with an optical proximity sensor, and a logic decision was made. The perceived problems with this concept included: •
large effective sealing area potentially leading to a high leakage rate and inaccurate results;
•
discrete process therefore dynamic speeds of components could limit the rate of operation;
•
high level of control required to operate; and
•
proposed sensor relied heavily on package positioning and package colour consistency.
It was noticed that the fundamental problems with the previous two concepts were vacuum generation and package measurement, therefore these two areas of engineering warranted detailed review of literature. 1.4
SCOPE OF STUDY
Further develop the theory of vacuum generation to accurately test leaky seals on specific potato chip packets through a high speed food processing and packaging line. Develop both mechanical and electrical theory to enable future commencement of working prototype machine suitable for testing within a food processing environment. Develop designs and consider machine options to incorporate this technology into allied food processing industries including baking, dairy and meat and smallgoods. Chapter 5 further discusses other future opportunities and potential works in other areas of nonfood processing.
12
1.5
LAYOUT OF THESIS
The main body of the Thesis consists of five chapters and five appendices. Chapter 2 details the project aims, literature and con-current research work completed in both mechanical and electrical disciplines. Chapter 3 discusses the design and development phase of the Thesis and includes laboratory testing completed, design procedure, consideration of alternative concepts and their development and analysis, and research and development of the component equipment in both mechanical and electrical design. Chapter 4 details the selection and procurement of both the prototype and production unit, and includes a detailed cost analysis. Mechanical and Electrical assembly and testing are also included. Chapter 5 summarises the Thesis and reflects on the results and achievements. It also discusses recommendations and future works.
The Appendices include design
drawings and calculations, equipment costs, laboratory testing data and equipment brochures.
13
CHAPTER 2 INTRODUCTION AND OVERVIEW 2.1
INTRODUCTION
Chapter 2 discusses the current project aims, and provides a detailed Literature Review. The Chapter also discusses background of the con-current undergraduate research work completed and the expected project outcomes. 2.2
PROJECT AIMS
GRG Engineering Services Pty Ltd in conjunction with QUT set up on a joint venture and IP agreement to progress the original concept into a final detail design for production of a prototype unit. There are three main criteria or aims to consider in the design of the High-Speed Sensing and Detection System (HSDS) and these are discussed as follows: 1. Performance The performance criteria will look at the performance of the design, for example, the concept or design adequately conforms to expectations.
The design shall contain
engineering features and components that enable the HSDS to perform adequately in a high-speed food processing and packaging line, ensuring Good Manufacturing Practice (GMP), food safety/quality practices and advanced engineering practice as well. It is planned to design a prototype machine that will perform at a minimum of 100 packets per minute. During prototyping the machines drive motor and pump could be fitted with variable speed drives (VSD) to allow speed increases incrementally over prototyping period.
Ideally a production machine with a speed of 200 packets per
minute would account for the majority of current VFFS and other packaging machines available in the packaging industry market.
14
Physical size and infeed limitations are specified and are required to be maintained. The performance also relates to achieving the maximum packets per minute during peak production times. 2. Manufacture Manufacture of a prototype unit and production trials have been excluded, an additional Masters or PHD study would be required to progress the design into a fully operational unit suitable for production testing within the food processing and packaging industry. The author does not initially plan to progress this next step. The project aims to provide design details that will enable manufacture of the plant, firstly as a prototype unit, then in production quantities. 3. Operation and Maintenance Ease of operation of the HSDS during both normal and peak periods of production is of paramount importance. The project aims to provide operation and quality controls at a convenient place for both viewing and operating. Preventative and routine maintenance of the unit are important to its life span. The design aims to allow the production operator or maintenance personnel easy access to clean and maintain the plant. 2.3
LITERATURE REVIEW
Contamination of packaged foods due to micro-organisms entering through air leaks can cause serious public health issues and cost companies large sums of money due to product recalls, compensation claims and loss of market share. It can potentially have a serious impact on consumers as a result of medical expenses, hospital bills and lost earnings due to health problems. In a worst case scenario, the impact on governments can also be significant in terms of overloading the public health system and hospital resources. The main source of contamination is leaks in packaging which allow air, moisture and micro-organisms to enter the package. Quality assurance is becoming an increasing issue in the snack food industry in an age where consumers are aware of the risks associated with spoiled food.
TSSC
experienced such issues in production of potato chips because of unsealed packets. In
15
previous work of this project, it was found that an economically suitable on-line testing machine was not available and it was therefore proposed to develop a new design. An extensive review of previous theses was conducted to ascertain a commencement level for the literature review relating to package seal testing.
Previous practical
experience in various food processing environments highlighted the importance of the project and requirements of food packaging. Current methods of package seal testing and inspection were investigated through research on various databases, World Wide Web, journals and traditional library catalogue searches. The most common types of food sealing include cans, bottles, containers, flexible plastic packets and vacuum sealing. Flexible plastic packets are widely used, and are the least expensive form of retaining the quality of the product. These packets can be used to seal, and therefore maximise, the shelf life of both dry and moist products. Examples of foods contained in flexible packets in supermarkets include sealed pre-cut salad (moist), potato chips (dry), dried apricots (moist), pasta (dry), baking products (dry) and flavour sachets (moist). The seals of food packets need to be airtight so that the food content is not contaminated due to contact with micro-organisms that enter as a result of air leakage (Jos Huis in't Veld, 1996). Airtight seals also extend the shelf life of packaged foods, and manufacturers attempt to prevent food products with leaky seals being sold to consumers. Despite their best efforts however, this problem still occurs because of failure to detect leaky seals during packaging, with potentially dangerous consequences from moisture, air and micro-organisms entering through air leaks.
Given the
implications for consumers, the manufacturers and government of food products, there is an increasing demand for cost effective state of the art inspection technologies that are capable of reliably detecting leaky seals. There are many current Non-Destructive Testing (NDT) methods of testing the seal of flexible packets best suited to random sampling, and for laboratory purposes. The three most commonly used methods are vacuum/pressure decay, bubble test, and helium leak detection. Although these methods can detect very fine leaks, they are limited by their high processing time and are not viable in a production line. Two non-destructive
16
in-line packaging inspection machines are currently available. These are the “NPC-501 leak seal inspection system” and the “PTI 550 Seal Scan”
Figure 2.1: NPC-501 leak seal inspection system Source: (www.jzw.com.au/page65.html) The NPC 501 distributed through JZW Control Systems provides a continuous seal inspection of every single packet and the system checks for pin hole leaks by passing the product through a region of high electric field at a maximum rate of 220 packets per minute. This machine is mainly used in medical, pharmaceutical and for high cost food processing where the quality of product can be severely compromised by poor quality of seal, or package.
The stand alone machine cost is between AUD50,000 and
AUD70,000 excluding infeed and outfeed conveyor systems. The PTI 550 Seal Scan is a product introduced by Package Technologies and Inspection to offer continuous on-line inspection of pouches at up to 150/minute. This system utilizes ultrasonic technology to scan the integrity of seals of pouches with accuracy levels towards six-sigma. It is again a high end machine again targeting medical, pharmaceutical and for high cost food processing. 17
Figure 2.2: PTI 550 seal scan Source: http://www.packexpo.com/ve/37638/products.html Another machine, the PAC Guard 400 by Mocon Inc. completes off-line testing and is suitable for snackfood products; a major disadvantage is it relies on detecting carbon dioxide leakage from inside the packet to verify a leaky packet. TSSC use nitrogen to purge oxygen from the packets prior to sealing; the PAC Guard would require a redesign to effectively test the chip packets. Another limitation is the machine only completes off-line testing, and is therefore only suitable for periodic quality control inspections. This project will develop a cost effective, state of the art technology for accurate evaluation of seal integrity in a high speed food packaging production line, so that accurate decisions on product safety can be made. The technology will be based on real time imaging and sensing technology combined with an automated differential vacuum system. Such a system is not available for the food processing industry in general. Most of the existing methods are unreliable at high production line speeds, and are unable to inspect both clear and opaque packaging materials with the same accuracy. Machines that are available for high production line speed, are not cost effective for real-time, high speed, non-destructive testing of food package. A review of current work in this field to date is presented here. Matsuno (1995) discussed quality control and testing
18
technologies in food processing. The cost of conventional testing of food packets is very high. It is over US$10000 per million packets (Sizer, 1994) and can not be used for testing all packets due to time and cost constraints. A study by the National Centre for Food Safety and Technology (NCFST) on nondestructive testing for seal integrity (Harper, 995) has outlined existing technologies. Frazier et al (2000) discussed use of ultrasound images for detecting defects in food package seals. Penman (1993) and (Penman et al, 1992) discussed X-ray detection of foreign material in food production. Graves et al (1994) extended this to a 3Dimensional model. Bowman (1993) worked on imaging and machine vision. (Davies et al, 1995) addressed issues related to design of real-time contaminant detection. However, very few researchers have addressed the current and future needs of the industry for a cost effective solution to the seal integrity problem. The main reasons are as follows: •
visual inspection used by companies has limitations due to limited resolution of human eye, operator skills and human errors;
•
machine vision imaging has a limitation when packets are opaque;
•
infrared imaging is sensitive to environmental conditions;
•
capacitance test for leaks can miss areas outside the cover of conductive plates;
•
ultrasound is a costly method with low accuracy at high speeds and for high package content variation;
•
eddy current probes for testing seal integrity do not work for clear plastic packets;
•
magnetic resonance imaging is costly and does not detect voids in the seal; and
•
x-ray detection technology has limitations when packaging material is opaque and introduces a radiation hazard into the environment under repetitive use.
The innovative outcomes of this proposed research will significantly advance the body of knowledge in several fields, including Non-Destructive Testing, real time imaging and sensing, differential vacuum, and development of algorithms for automatic decisionmaking. The innovative and integrated framework developed in this project has tremendous potential in other food processing/packaging organisations in Australia and
19
internationally. Food packaging and subsequent export is a major contributor to the Australian economy. The cost of 0.5% leaky seals has a potential impact of AU$35 million per annum losses in this sector. The innovations of the proposed research lie in the following areas: •
use of a differential vacuum technique for expanding food packets to detect leaky seals;
•
innovation is required to design mechanisms and develop sealing technology for a vacuum chamber that can allow repeated testing of potato crisps packets at 80 per minute. The developed vacuum must be retained and measurements of height and volume differences taken within a limited test time in order to detect leaks. This is a cost effective innovative contribution to decision support for nondestructive testing of food packets in high speed production lines;
•
novel image capturing and processing techniques for identification of leaky seals; and
•
a new approach to non-destructive testing by developing a hybrid system combining image processing and sensing for higher accuracy in predicting change of height and volume of packets tested in a differential vacuum system; and
•
development of robust algorithms to accommodate different package sizes, packaging materials, contents and size of leaks.
The excellent time-frequency localisation of the wavelet transform (WT) is proposed to be used here. By decomposing the representation of a package into multi-scales, an accurate comparison between two entities (one before vacuum and the other after vacuum) will be constructed. The WT will describe packets in both low and high resolution (Jaggi et al, 1994). The basic geometry of packets will be determined from the low frequency information and high frequency information will be utilized to perform more specific recognition tasks that detect packets as either properly sealed or leaky. The proposed system has the potential to be robust even in the presence of noise due to variations in package shape, weight of film and material content. It is anticipated that
20
it will be able to perform extremely well even in complex cases where the change can easily be missed using traditional techniques. In this project, sophisticated processing techniques will be used to combine both image data and sensed heights. State of the art sensors will measure the height changes of packets under differential vacuum. New algorithms will be developed to achieve high accuracy while the packets are in motion on a high-speed conveyor. This research will advance the traditional knowledge verification approach using image processing or sensing to create a hybrid sensing and imaging technology for real time decision making on seal integrity. The developed algorithm is expected to be more accurate and cost effective than previously proposed and trialled methods. The novel system components developed in this research will be: •
a differential vacuum testing chamber;
•
an image capturing and processing system;
•
a height sensing system;
•
a hybrid prediction module;
•
an optimal quality control decision support module to reduce errors of rejecting good packets and accepting leaky packets; and
•
robust algorithms will be developed to achieve high accuracy in detecting leaky packets in high speed production line.
2.4
CON-CURRENT RESEARCH WORK
Two undergraduate theses were completed under my supervision during the completion of my Master’s thesis preparation.
Nathan Newman (Newman, 2004) completed
electrical research on industrial image processing applications and their relevance to the proposed design.
Wesley Mergard (Mergard, 2004) completed mechanical
research in my office under constant direction and supervision to develop essential mechanical equipment and components.
21
2.4.1
MECHANICAL
2.4.1.1 Introduction An important stage of the project was the review of current literature relating to the problem of package seal testing. For the undergraduate student to understand the importance of the project, the requirements of food packaging and the current methods of package seal testing were investigated. 2.4.1.2 Food Packaging Sealed food packaging is often required to prevent the two main forms of food degradation- oxidation and microbial activity. These forms of food spoilage are a major problem to the food industry, as consumers are increasingly aware of the dangers of food spoilage to their health. Oxidation of food causes the food to become stale and unappealing, and presents a very real health hazard. This form of food spoilage occurs when the integrity of a packaging seal is compromised, and air is allowed to flow into a previously controlled environment. The same is true of microbial degradation of food. Loss of integrity of food packaging allows these microorganisms into an ideal environment for their growth, leaving them to multiply exponentially and cause serious health problems. Properly sealed packaging allows companies to guarantee the quality of their products when they reach consumers. The most common types of food sealing include: •
cans;
•
bottles;
•
containers;
•
flexible plastic packages; and
•
vacuum sealing.
Flexible plastic packages are becoming widely used, as they are the least expensive form of retaining the quality of the product. These packets can be used to seal, and therefore maximise, the shelf life of both dry and moist products.
22
However, the main disadvantage of this type of packaging is that it is quite prone to the following types of leaks, which lead to product degradation: •
porous leaks: - This is a result of matter, including air and moisture, being able to pass through the fine pores of the plastic;
•
fine leaks: - Fine leaks are generally regarded as leaks up to the size of a pinprick. Although this type of leak is not obvious to the consumer, microbial activity and oxidation will most certainly occur, and product quality can be significantly diminished;
•
gross leaks: - These commonly occur because of improper heat sealing, or a tear created in the packet during packaging/distribution. Not only does the product degrade, but also the consumer is well aware of the leak. This often leads to customer dissatisfaction and reluctance to buy from the company again. This in turn causes loss of revenue for the company.
TSSC use flexible packages to seal the freshness of their chip products. These include Twisties, Doritos, and many others. However, the packages can sometimes have gross leaks because of improper heat sealing. Therefore, TSSC and other food processing companies require an in-line testing apparatus to test 100% of product. This will minimize customer dissatisfaction, as all products are being tested and faulty products will be rejected. Before development of this machine could commence, market research was required to investigate current methods to check the seals of flexible packaging. 2.4.2
ELECTRICAL
An initial search for literature describing similar solutions was undertaken. While there is an abundance of literature on the topic of Image Processing, there was found to be little relating specifically to this problem. The focus of the project is a new image processing technique for detecting changes in packets (such as height, shape and volume) from images taken before and after the packet being subjected to a vacuum. The project is multi-disciplinary in nature, and as such makes significant contributions to several research areas:
23
•
innovative image processing algorithms capable of identifying seal integrity of packets from their visual characteristics under vacuum;
• development of software for optimal detection of seal integrity and decision making or process control logic; • integration of physical components to construct a testing apparatus for such a system with consideration of the likely operation environment. The algorithms and techniques developed are reasonably robust, even under several difficult cases identified during the formal testing process. 2.4.2.1 Problem Description This project is focused on using image capture and processing techniques to solve this problem. There are several unique challenges of this project that need to be addressed: •
this system needs to be robust and handle many random variations of the packets as they pass through the production line;
•
this system must operate at a speed at least equal to current production line speeds;
•
the system must be able to be integrated with the mechanical vacuum chamber; and
•
the system needs to be supported by comprehensive test results and shown to be able to accurately and reliably identify the integrity of chip packet seals.
2.4.2.2 Related Problems and Solutions Photoelectric and laser sensors are commonly used in food processing and industrial plants worldwide to take various measurements of products. Such a solution for this project may be viable but would provide limited data to assess the condition of the packet. Image processing techniques are commonly used in industrial plants to identify and measure products. This project has several unique aspects. While the image processing techniques applied here are commonly used; a similar solution in not known to exist.
24
2.4.2.3 Project Plan The initial phase of the project will include a broad research of digital image processing techniques. As the project progresses, the research will become more focused on relevant areas of digital image processing. The following specifications regarding the mechanical vacuum system and the packets will need to be investigated: •
visual characteristics of the packets (patterns etc);
•
visual characteristics of the likely mechanical vacuum system, i.e. the background of the captured images;
•
position of the packets relative to the camera;
•
consistency of packets position; and
•
movement of packets during vacuum process.
The important conditions or specifications of the system will affect the accuracy, reliability and load of the system.
Accuracy it vital is the ability of the system to
correctly diagnose each packet. Reliability is the ability of the system to consistently apply the diagnostic to each packet.
The load is the amount of computational
processing required and will affect the speed of the system. The goal is to choose the conditions or specifications that will maximise the accuracy and reliability and minimise the load. The following conditions or specifications of the system will need to be investigated: •
the resolution of the captured images;
•
the placement of camera(s) in relation to mechanical vacuum systems;
•
lighting characteristics of typical operating environments (the system may require its own light source);
•
noise characteristics of typical operating environments ( the system may need to be optically isolated);
•
speed required of the system to be viable;
•
calibration procedures; and
•
limitation of the accuracy of the system. 25
Various image processing methods will be investigated and tested for suitability. These will include edge detection, thresholding, region growing and other methods to be identified. A core-system will be initially developed based on the results of the previous investigations. This will be an un-optimised, formed system used for preliminary testing consisting of low volumes of packets with typical variations. A standard system will then be developed integrating optimised image capture and image processing algorithms.
This system will possibly be written in programming
language “C”. This system will be subjected to formal testing of actual packets, and the results of this testing will be analysed. The result of this analysis will form an integral part of the project, as is will provide information on the accuracy and reliability of the system that the end-user will need to consider upon installation. Full development and optimisation of the standard-system may not be achievable in the time available. Some advanced functionality may need to be postponed for an extended-system. 2.4.2.4 Expected Outcomes The system will be able to accurately and reliably classify packets according to the quality of their seal. The system will operate at a speed equal to or greater than current production line speeds. The system will be robust, that is, it will be resistant to noise due to variations in packets and will perform well in all cases. 2.5
CONCLUSION
The literature review provided detail of the importance of leak seal prevention and current technology available. Project aims in relation to performance, manufacture, operation and maintenance were also summarised.
Con-current research by both
mechanical and electrical undergraduate students provided valuable background development and options.
26
CHAPTER 3 DESIGN AND DEVELOPMENT 3.1
INTRODUCTION
Detailed design and development of the HSDS is the fundamental requirement of this project and the future prototype and production unit.
This chapter details all the
laboratory testing completed and methodical design procedure needed for a successful concept. Discussion on the development of the alternative concepts in both mechanical and electrical disciplines enables the researcher to make an informed decision. The final detailed research and design of the mechanical and electrical major components close out the chapter 3.2
LABORATORY DESIGN TESTING
The laboratory design testing allowed the researcher to confirm theoretical assumptions before moving into the detailed design phase. The mechanical tests confirmed the vacuum hypothesis and seal integrity with good consistent results. Electrically, the testing also provides solid results to enable the researcher to move the project forward with confidence. 3.2.1
MECHANICAL TESTING
The concept of a differential vacuum technique for expanding food packets to detect leaky seals was trialed in the laboratory using various TSSC multipack packets. The test used both fully sealed and packets containing a known leak.
27
Figure 3.1 Laboratory testing of multipack Doritos packet displaying packet under vacuum Data was recorded to determine a change in height versus vacuum pressure of sealed and leaking packets. This data was then presented in a graph format to diagrammatically display the relationship between fully sealed and leaky packets. The data will also provide assistance and guidance during both the design activity and also future testing of prototype equipment. Results: Throughout the experiment, the following general observations were made: •
packets without a leak “puffed up” with a dome shape (see Figure 3.1 and 3.2), as the surrounding pressure was lowered;
•
the profile of the packet became more uniform at lower surrounding pressures e.g. at pressures lower than -24 kPa (gauge), the packets generally appeared quite smooth;
•
in some instances, defective packets still “puffed up” because the opening would self-seal. Self-sealing occurred where the leak was smaller than usual (due to experimental errors), and
•
self-sealing was not a problem if the leak was not in the heat sealed area of the packet.
28
Figure 3.2 Photograph showing a sealed Doritos™ packet inside a –20kPa vacuum chamber
Twisties Results The measurements recorded at the time of the experiment for the Chicken Twisties™ packets are detailed in Table 3.1 in Appendix D. The height change of the Twisties™ packets at various vacuum levels is shown in Figure 3.3.
Figure 3.3 Graph showing change in heights of Twisties packets
29
Doritos Results The measurements recorded at the time of the experiment for the Doritos packets is detailed in Table 3.2 Appendix D. The height change of the Doritos packets at various vacuum levels is shown in Figure 3.4.
Change in Height [mm]
Graph Showing the Change in Height of Doritos Packet vs Vacuum Level Change in Height of Sealed Packet
50.00 40.00
Change in Height of Packet with Leak
30.00 20.00
Expon. (Change in Height of Sealed Packet )
10.00 0.00 0
10
20
30
40
Vacuum level of Chamber [kPa]
Figure 3.4 Graph showing change in heights of Doritos packets
Discussion: The experiment confirmed the hypothesis that sealed chip packets experience a change in height when they are subjected to a vacuum. It was found that the change in packet height increased exponentially with vacuum level between differential pressures of 10 kPa to 32 kPa. Generally, the Doritos packets experienced a greater change in height than the Twisties packets, possibly because of thinner and more flexible packaging (plastic without foil) and a higher initial air volume. The maximum heights of the packets occurred at approximately 32 kPa and were found to be 60.3mm for Twisties and 74mm for Doritos. For differential pressures greater than 30 kPa, the bag walls appeared in very high tension and at risk of bursting or localised seal tearing.
30
Most unsealed packets experienced very little change in height because the leak allowed pressure equalisation between internal and external air. However, some packets experienced self-sealing and would puff up not quite as much as if the packet was fully sealed. Self-sealing may only be a problem with smaller leaks but further testing will be needed to be conducted when the sensing method has been designed. Despite the high level of care that was taken throughout the experiment, the experiment still contained errors. Potential sources of error included: •
the maximum height was used despite the fact that this point may have possibly been because of a crease in the packet;
•
no significant consideration was given to inflation time;
•
chips were initially flattened out to overcome any major packet creases or chips standing up. This would not be performed on packets in the in-line testing machine because of the damage to the product;
•
two types of chip packets were tested;
•
difficulty in making consistent 5mm hole in the heat seal;
•
difficulty was experienced in achieving exact pressures.
Result Tolerances
were within 150 Pa of specified pressure. The following recommendations can be made for the design of the prototype testing apparatus: •
the system must accommodate for packet inflation heights of up to 74mm;
•
higher differential pressures may lead to results that are more accurate;
•
self-sealing is less of a problem at higher differential pressures. Once the sensing system has been designed, future testing will be required to determine what pressure is required. At a problem definition stage, the system must be designed to withstand an internal gauge pressure of -30 kPa.
31
3.2.2
ELECTRICAL TESTING
Due to the prototype not being available for formal electrical testing, a mock-up test chamber or rig was constructed to enable preliminary testing to be completed. Lighting and camera was set-up against a white paper background to simulate a production environment. A Logitech Quickcam was used to capture the images and a formal test program was developed by Nathan Newman (Electrical Undergraduate Student). The test program was written with Borland Visual Delphi 6.0 and the Logitech Quickcam Software Development Kit. This program has the following functionality: •
connect to Logitech Quickcam;
•
capture before and after images (references and vacuum images);
•
apply the optimised “Profile” algorithm to the images; and
•
output test statistic (proportional to the height of the packet).
Test Cases: Testing was conducted using five 50g Smiths potato chip packets. Each packet was a different flavour and had different coloured packaging and slightly different surface visual designs. Each physical packet was used as two test packets: firstly with the seal intact and again after the seal was compromised. Each test packet was tested ten times with random orientations. The random orientations were used to simulate the likely operation of the production HSDS unit that will be fed packets at high speed by dropping packets onto the infeed conveyor belt. Each test case is composed of two images, a reference image and a vacuum image. The reference image is captured after the packet is placed in the test rig, but before a vacuum is applied. The vacuum image is captured after the appropriate vacuum is applied to the dry box. A summary of the test cases is shown in Table 3.3:
32
Test Case #
Physical Packet
Test packet #
Seal Condition
1-10
A
1
Good
11-20
A
2
Bad
21-30
B
3
Good
31-40
B
4
Bad
41-50
C
5
Good
51-60
C
6
Bad
61-70
D
7
Good
71-80
D
8
Bad
81-90
E
9
Good
91-100
E
10
Bad
Table 3.3 Test case summary
Note a seal condition of “good” denotes the seal of the packet is intact. A seal condition of “bad” denotes that the seal has been manually broken. A tabulation of the test results can be found in Appendix D. A graphical representation of the above results can be seen in Figure 3.5.
33
60.00%
% Change in Height
50.00% 40.00% 30.00% 20.00% 10.00% 0.00% -10.00% Test Cases
Good Packets
Threshold
Bad Packets
Figure 3.5: Variation of packet height under differential vacuum conditions (Source: Unpublished final year undergraduate thesis experiment with differential vacuum at QUT 2004) The top function is the relative change in height of packets known to be “good”, and the bottom function is the relative change in height of packets known to be “bad”. No correlation exists between the top and bottom functions, and are shown on the same graph for easy reference. A threshold value was selected, and was based on the relative height change of 5%. During this test, this threshold value represented the best result for this set of data. During prototyping and subsequent setting up of the final system, this value would need calibrating from a large amount of test data. The test results indicate that packet orientation significantly affect the percentage change in height detected by the system. In this set of test results there are three false alarms or readings, i.e. three test cases that were good seals, were implied to be bad seals.
34
These cases, where the relative change in height drops below the threshold are shown in Figure 3.5 and in the tabulated results in Appendix D. The test numbers are 50, 64 and 81. The cause of this error was identified to be an orientation error when placed in the test rig. Other test cases (6, 8, 9, 27, 61, 62 and 68) also displayed a low relative height change of between 5 and 10%. Packet orientation was also identified to be the cause of these readings. From illustrations, it appears that errors and low readings tended to occur when the packet s sitting on one end. A mechanical device, Packet Leveller as discussed and detailed in Section 3.5.1.8 would used to counteract this issue. During the testing, no “bad” misreads were recorded; this was due to the threshold for the relative change height being chosen above all the bad test cases.
During
prototyping this approach would be adhered to, but may result in increasing the number of false alarms occurring. Two main reasons for relative height variations have been identified, i.e. Self Sealing and Packet Movement. During the packet puffing up stage, the broken seal can close securely on itself. Packet Movement can occur due to the packet being ‘kicked” around by the sudden application of negative pressure. During the future prototype stage, these issues would be closely monitored and solutions developed to reduce or eliminate these height variation constraints.
35
3.3
DESIGN PROCEDURE
The design procedure methodically works through the various major functions both from a mechanical and electrical perspective. It opens up alternative ideas for the major components that although are sometimes not practical in this application, show that the researcher has exhausted all engineering and functionality thoughts. 3.3.1
MECHANICAL AND ELECTRICAL COMPONENTS
The mechanical system consists of a number of components incorporated in the overall design.
Various concepts for each component were considered, and are detailed
below: A number of options or functions were considered during the initial concept development, including horizontal, vertical and polar (circular) transportation. The conveyor system was broken down into the following five functions and ideas were generated for each individual function: Conveyor Transportation Functions (A) Logic: The packets will be loaded from the existing conveyor line into the HSDS efficiently without product damage and correct orientation.
The packets must also
progress through the HSDS without internal product or package damage. •
Horizontal Function:
A1: Conveyor Belt (continuous) A2: Rollers (continuous) A3: Air Draft ((continuous or discrete) A4: Pneumatic Ram (discrete) A5: Loaded Spring (discrete)
•
Vertical Function:
A6: Cleat Conveyor (continuous)
•
Polar Function:
A7: Large Barrel A8: Star-wheel
36
Vacuum Generation Functions (B) Logic: The HSDS will subject the packet to a partial vacuum to ensure the sealed packets expand precisely. •
B1: Double Gate Airlock
•
B2: Star Wheel
•
B3: Dribble Valve
Defect Sensing Functions (C) Logic: The height/profile of the packet will be assessed at both atmospheric pressure and partial vacuum without contact •
C1: Maximum Height
•
C2: Maximum Width
•
C3: Change in Height
•
C4: Change in Profile from Side
•
C5: Change in Volume (3d scanning)
Rejection Functions (D) Logic: The signal from the decision-making algorithm will be used to physically reject unsealed packets. •
D1: Ram to Push Packets
•
D2: Air Jet
•
D3: Selector Gate
37
Control and Decision Making Algorithm Functions (E) Logic: The HSDS will be controlled to ensure leaky packets are rejected. •
E1: Discrete Progression and Sensing
•
E2: Continuous Progression and Sensing
•
E3: Discrete Progression and Continuous Sensing
•
E4: Continuous Progression and Discrete Sensing
The above functions stimulated new ideas from the author, although the overall system ideas were still limited. An Excel spreadsheet table was created to quantify the ideas and work through to a logical solution.
Although some idea combinations were
considered impractical, it did force the author to generate new and creative ideas. The best were selected for further concept development in later Chapters. Function Function
Function Function Function
A
B
C
D
E
Number of Ideas 8
3
5
3
4
Combination 1
7
1
3
2
3
Combination 2
1
3
5
2
2
Combination 3
2
2
2
2
3
Combination 4
4
1
4
3
3
Combination 5
7
2
3
1
2
Combination 6
7
3
4
2
2
Combination 7
6
1
5
3
4
Combination 8
8
1
3
1
3
Combination 9
7
3
4
1
3
Combination 10
8
2
3
1
2
Table 3.4: Idea combination Excel spreadsheet
38
Non-Contact Sensing and Decision Making Functions (F) Logic: The HSDS will require non-contact sensing to determine decision making protocol: •
F1: Through-beam (Photoelectric sensors)
•
F2: Parallel Beam Laser Sensor
•
F2: Reflex Sensor
•
F3: Proximity Sensor
It was proposed that the chip packets should be measured with a non-contact sensing method.
Non-contact sensing will complete the operation without damage to the
integrity of the packet. Although many sensors are available, only those that can provide package height measurement were explored. It was found that the most common type of non-contact sensors are photoelectric sensors. Photoelectric sensors are optically based and are widely used throughout industry. There are "three basic methods of sensing are used in photoelectric: thru-beam, reflex and proximity"(Juds, 1988). Through-beam (Photoelectric Sensor): The Through-beam is the oldest and most familiar type of photoelectric sensing. The through-beam sensor essentially consists of a source that transmits a pulsed light to a directly facing detector (see Figure 3.6). If the beam is interrupted, the output of the sensor changes, and object detection is achieved.
Figure 3.6 Typical Application of Thru-Beam Sensors 39
The main advantages of this type of sensor include: •
longest range;
•
highest possible signal strength; and
•
little effect from surface colour and reflectivity.
Parallel Beam Laser Sensor: Parallel Beam Laser Sensors have been recently developed and achieve continuous height/thickness.
Figure 3.7 shows, when the beam of light is interrupted by the
measured target, the CCD (charge coupled device) within the receiver registers areas of light from emitter.
Figure 3.7: Key components of the Omron Z4LC parallel beam sensor Source: http://www.omron.com.au/product_info/Z4LC/index.asp The main advantage of this sensor is that it provides continuous measurement of an object with a resolution of typically 10 microns. Reflex Sensor: Reflex sensors are similar to through-beam sensors, except that they use retro reflective targets to return the light along the path it came from. This allows the emitter and receiver to be located together in one housing, reducing installation costs. For this reason, this type of sensor is generally used in shops to alert the attendant of a new customer. Proximity Sensor: Proximity sensing is achieved by sensing the light directly reflecting from the target. Although target proximity distances can be measured, photoelectric proximity sensing is limited by sensing distance and accuracy, and is affected by colour and reflectivity of the target (Juds, 1988).
40
Camera Function (G) Logic: The camera type, placement, numbers and position angle relative to the packet is important in consistent measuring and quality control •
G1: Camera Type – Retail or Industrial
•
G2: Camera Quantity
•
G3: Camera Positioning
Details: The placement of the camera in the prototype system and hence angle relative to the packet was carefully considered. It was found that the best angle to view a chip packet from is the side when the packet is lying flat on a surface. It is planned in the prototype system that the packets will tend to fall and stabilise in this manner. This angle of viewing provides the greatest visual difference of a packet between normal and vacuum pressure. In addition the possibility of using additional cameras was investigated. This would provide an orthogonal view that may contain information not present in the side view. The only option for an alternative view in the prototype vacuum chambers being considered would be a plan view. A plan view of a packet when lying flat on a surface would provide little improved information and would not register significant visual changes between normal and vacuum pressure. The HSDS should incorporate two cameras in the final design. The first one should capture a reference (before vacuum) image when the packet enters the vacuum chamber (after the packet has come to rest). The second will capture during vacuum image when the packet is subjected to maximum negative pressure. Resolution Functions (H) Logic: The image resolution size will determine the picture quality and repeatability. •
H1: Resolution Size
41
•
H2: Picture Quality
•
H3: Repeatability
Details: Several image resolutions were considered in the design. The most common resolution among retail and industrial cameras is 640x480 pixels. The computational processing required in the HSDS is not overly intense and operates relatively quickly on images of this size. Test images revealed that this resolution maintained sufficient detail. It was decided to use resolution of 640x480 pixels.
Grey Scale Conversion Functions (I) Logic: Images require conversion from RGB to grey scale for packet imaging •
I1: Geometric mean of component colours
•
I2: Arithmetic mean of component colours
Details: Two methods of converting an ‘RGB’ image to grey scale were identified. It was found that the use of either method was suitable for the purposes of this project. Functions I to K cover visual environment and noise variables and are summarised below: Background Noise Functions (J) Logic: A suitable background is required to capture sample images, and when distinguishing the packet profile •
J1: Colour (white)
•
J2: Contrast (bright and consistent)
•
J3: Sheen (low sheen or non-reflective)
•
J4: Finish (plain, not glossy)
•
J5: Background Surfaces (light coloured and consistent, not busy)
42
Details: Several consistent backgrounds were tested when capturing sample images. The best work background to work was white plain paper (not glossy). Similarly light coloured, consistent and not-reflective (low sheen) surfaces were found to be an effective background. Various tests were conducted to evaluate the effects of background noise in sample images. This was achieved by capturing sample images of a packet with a variety of visually busy backgrounds. It was found to very difficult to distinguish the packet from the background in such circumstances.
Tests confirmed that the background be
consistent white of light in colour and non-reflective (low sheen). Lighting Functions (K) Logic: Suitable background lighting is required to allow a suitable vision of the captured sample images. This lighting also improves the packet profile image.
Testing was
conducted using several different light sources. •
K1: Lighting Type (Halogen or similar)
•
K2: Number of lights (Two Minimum)
•
K3: Positioning of lights (Above and either side of camera, directed towards the centre of the image)
Details: Testing was conducted using several light sources. Ambient light (from ceiling lights such as fluorescents) was found to be insufficient for capturing a ‘clean’ workshop image.
It was decided to use a directed light source and during testing, the best
available light source was found to be halogen desk lamps, which emit a bright white directed light. During the next set of tests, two halogen desk lamps were used and placed just above and either side of the camera directed toward the centre of the image. The HSDS visual environment would have dedicated light sources.
43
Packet Noise Functions (L) Logic: Packet film could be a source of disruptive noise. Testing was conducted using several different packet patterns or logos for sources of disruptive packet noise. •
L1: Packet film surface (all smooth surface)
•
L2: Filter Operations (patterns had minimal effect on distinguishing the packet)
Details: Testing of a variety of different packets revealed that these patterns had minimal effect on distinguishing the packet or filtering operations. In addition the use of these patterns to assist in identifying changes in the packets was considered. It was found that no additional information was obtained from patterns on the packets and as such was not considered as part of the solution. Spatial Domain Functions (M) •
M1: Edge Detection
•
M2: Thresholding and Region Growing
Edge Detection: Edge detection was identified early in the research as an effective way of identifying the external contour of a packet. Several methods of edge detection were investigated and the 3x3 Sobel edge detection filter was found to be the best, in addition, this filter is efficient in the spatial domain and operates relatively fast on the size of images in this design. As can be seen in Figure 3.8, the Sobel edge detection filter is good at identifying the external contour of the sample packet.
Some disruptive noise is generated at the
bottom of the image due to shadows. These shadows can be reduced with additional light sources but are difficult to completely eliminate. The most important part of the contour to consider was towards the top of the packet.
44
Figure 3.8: Preliminary results of using Sobel Edge Detection (Source: Unpublished final year undergraduate thesis experiment with differential vacuum at QUT 2005)
It was decided that the Sobel edge detection filter should play a role in the solution. Thresholding and Region Growing: Thresholding is the process of accepting or rejecting pixels based on their intensity. It is commonly applied to ‘edge’ images since the intensity of a point represents the strength of the edge at that point. In addition, the process of image region growing is used to accumulate pixels in a common object or area. The combination of these processes is an effective way of distinguishing objects in an image. By region growing the background using a suitable threshold, the packet could be identified and the contour of the packet could be found. The result of this operation is commonly referred to as a binarised image since there are only two distinct intensities. Figure 3.9 shows the ‘edge’ image of a packet and to the right an image that has undergone region growing of the background using a suitable threshold. The contour of the packet in the binarised image tented to be distorted around the bottom of the image and this is mainly due to shadow of the packet.
45
20
20
40
40
60
60
80
80
100
100
120
120
140
140
160
160
180
180 50
100
150
200
250
300
350
400
450
500
550
50
100
150
200
250
300
350
400
450
500
550
Figure 3.9: Preliminary results of using Thresholding and Region Growing (Source: Unpublished final year undergraduate thesis experiment with differential vacuum at QUT 2005)
The value of the threshold will need to be automatically calibrated in the final system. A possible method for doing this is by using a histogram of the ‘edged’ image. It was decided that thresholding and region growing should form part of the solution. In addition, histograms should be considered as a method of calibration the threshold operation. Summary During the design development process, constraints and considerations became apparent. The six main constraints are discussed below: Safety: The HSDS must conform to Workplace Health and Safety regulations and legislation of relevant states and territories throughout Australia, and also any specific company technical specifications. The machine shall also not impose undue hazard to either consumers or employees.
In important consideration in relation to food
processing, the machine shall not encourage the growth of bacteria or other microorganism’s., and must be easily cleaned during production sanitation periods. Product Damage: The HSDS shall not damage the chips or packaging incurring any quality issues or potential loss of revenue.
46
Accuracy and Precision: The HSDS shall have the capability of rejecting leaky packets at a high level of accuracy and also work within a tight range or limit without wasting too many sealed packets. The accuracy and precision parameter settings shall adjustable to assist in maximising the cost benefits of installing the machine. Cost: The initial, operational and maintenance costs of the machine must be economically viable for TSSC and other food processing and packaging companies. Size: The HSDS shall be reasonably compact in design to ensure fitment within an existing production/packaging line. Some companies such as TSSC Tingalpa plant have space constraints and therefore require careful consideration. Leak Size: Plant staff indicated, “The most common leak is approximately 5mm in the sealing area of the packet”. Although quantified statistical data regarding leak sizes was not initially available, testing occurred at the TSSC Plant during July 2004, and such data may currently be available. It was decided to consider both a short- term and long-term problem definition as potentially different assumptions would be determined between the two. The major advantage of defining short-term problem was that certain assumptions could be simplified. The constraint variances between the two problem definition terms are discussed: Safety: The HSDS must conform to Workplace Health and Safety regulations and legislation, specifically considering QUT and TSSC environments and working practices. Product Damage: No variances Accuracy and Precision: No variances Cost: A prototype machine build cost would be less than AUD 10,500 excluding cameras and PLC equipment, and engineering costs
47
Size: To suit QUT or off-site testing facilities, and compact enough to fit into TSSC current facilities. Leak Size: No variances The electrical system consists of a number of components incorporated in the overall design.
Functions F to M cover the various concepts of image capture and
representation and are detailed below: During the research phase, several other transformation domains were also considered as part of the solution. Image subtraction was investigated as a partial method for detecting changes in the packet shape. Two sample images were obtained in which the packets’ shape differed but position remained the same. The difference image was obtained by subtracting corresponding pixel values of the sample images. It was found that this operation had some advantages, in that background detail was removed/attenuated; and the image became less sensitive to thresholding operations. However, the image difference was negligible, when trying to find any changes in shape of the packets. This was mainly due to shadows in the image and the patterns shifting significantly between sample images. It was decided that the image subtraction would not be utilised as part of the solution. Frequency or Fourier and Discrete Cosine domain which is used primarily to speed up an intensive filtering operation and was considered too complicated for this solution. The Hough domain, which finds lines in images and each point in the domain corresponds to a line of particular angle and the intercept in the image domain was also decided to be not a suitable solution to the problem. Trials with potato chip packets were completed using the Hough domain, and due to the complex shape of the packet, mapped points are difficult to distinguish.
48
The image processing task will focus on capturing images of a package before applying vacuum and at the peak value of vacuum. Features of the “before” images and those extracted from the “after” images will be compared and evaluated. In the spatial domain, operations such as image subtraction, edge detection, thresholding and region growing, will be considered. Image subtraction will be investigated as a partial method for detecting changes in packet shape. The difference image will be obtained by subtracting corresponding pixel values of the sample images. This will provide useful pointers for seal integrity. This method will have the advantage of removal/attenuation of background detail that might clutter the images and cause inaccuracies in the evaluation. Practical issues need to be addressed include the existence of shadows, due to the setup of lights at the testing location. Edge detection would be used for an effective way of identifying the external contour of a packet. The effects of noise due to shadows will be addressed in the algorithm. The size changes of the package under vacuum/pressure would be detected at the top portion of the package, to reduce the effects of shadows. Thresholding will be applied to ‘edge’ images since the intensity of a point represents the strength of the edge at that point. This can be incorporated with image region growing to accumulate pixels in a common object or area for an effective way of distinguishing objects in an image. Frequency domain processing would be used for provide valuable improvements in feature extraction. Techniques for processing in the Hough domain would also be explored. The Hough transform for finding lines in images would be used. Each point in the Hough domain will correspond to a line of particular angle and intercept in the image domain. The intensity of a point in the Hough domain will represent the density of that line in the image domain. A threshold value will be set in the Hough domain to characterize shape changes approximated by straight lines. Further, polygon approximation and segmentation techniques will be used to identify the changes of shape. After the image has been binarised, it can be identified as the largest grouping of pixels. An iterative algorithm would then be used to find the longest straight line (major axis) that can fit into the packet region. From this segmented image the longest fitting line at a right angle to the major axis would be drawn and the top half
49
of the packet would be modelled as a triangle. The image can be further segmented for a more accurate model. Figure 3.10 shows two different sample packets with polygon approximation indicating that this approach can be used to characterize changes in the packet shape as a result of applying vacuum.
Figure 3.10: Sample polygon approximation from our preliminary results (Source: Unpublished final year undergraduate thesis experiment with differential vacuum at QUT 2005)
3.4
DEVELOPMENT & ANALYSIS OF ALTERNATIVE CONCEPTS
Based on details from the design procedure in the previous section, the section sets out to amalgamate the process both a mechanical and electrical perspective. This section addresses concepts for the entire process. 3.4.1
MECHANICAL CONCEPTS
Consideration of moving packets continuously through a vacuum chamber was investigated and a preliminary design completed. The proposed design included a belt conveyor complete with cleats operating through an airlock tunnel and into a dedicated vacuum chamber. An image processing and sensing system to determine relevant height change would be positioned directly at the outfeed of the vacuum chamber. A reject algorithm system will be developed at the conveyor exit to reject leaky packets based on a pre-determined packet height.
50
Operational data was obtained from TSSC in regard to the current range of multipack packets.
Information such as current conveyor system speeds. packet per minute
output and con-current processing lines operating, etc are vital to establish a system criteria and manufacture of a suitable prototype. Various site visits to TSSC Tingalpa were completed to enable the researcher to fully understand the current leakage problem, observe the operation of the various procesing lines and different set-ups. During the mechanical design and development stage a number of concepts were considered, from single packet discrete systems to continuous multiple packet systems. The discrete system as shown in Figure 3.11 would check one packet at a time, and is generally suitable for an off-line Quality Control or HACCP inspection. In some food processing business’s with tight or closely controlled HACCP inspections upstream from this point, a single and regular quality check may be deemed to meet their Food Safety procedure and policy.
IR Sensor
Top Door Actuator
Bottom Door Actuator
To Main Vacuum Chamber
To Reject Line
To Packing Conveyor
Figure 3.11: Single packet discrete system
51
The advantage of this concept is the unit is cost effective and quite compact, and can be placed in a convenient location within the packaging area or remotely within the quality laboratory. The main disadvantage is number of checked packets per minute; it is estimated that the maximum would be 20 packets per minute based on careful placement of the packet directly above the IR sensor. A second concept was a continuous system either retro-fitted to an existing conveyor, or manufactured new with an integral conveying system (see Figure 3.12). The system relies on a series of pneumatic rams to provide airlocks and a vacuum chamber of the same height. A reject mechanism would be installed after the chamber, and would comprise of either a pneumatic ram or pneumatic “air” knife to physically blow the rejected packet off the conveyor. This rejected packet would then be inspected by either a production or quality personnel, who would record and verify the fault.
Sensor Sensor Rejection Ram Airlock 2 Gate 2 Airlock 1 Gate 1 Chip Packets Loaded Onto Conveyor
Figure 3.12: Continuous conveyor
52
The advantage of this system is its ability to process more packets simultaneously and can be retro-fitted onto an existing conveyor system. Disadvantages include, limitation on belt speed due to the nature of the inspection, and potentially high wear rates on critical components. The number of pneumatic stokes per production shift would incur high maintenance costs. In this system the packets would require an entry conveyor and mechanism that would line up all the packets in a predetermined fixed spacing to ensure no damage to the packets. The third concept shown in Figure 3.13 is a proposed continuous cleat belt conveyor system through a custom designed vacuum chamber.
Figure 3.13: Continuous cleat belt The theory behind this principle was, “if a region between the adjacent cleats, that contain a packet, can progress through the system without ever being exposed to both atmosphere and internal vacuum at any one time then an airlock is created”. The basis to this theory was achieved by having the cleats pass through a rectangular sealed
53
tunnel longer than the cleat spacing. The cleats theoretically created a seal on all four internal faces of the tunnel. The prototype machine would incorporate the following process stages: •
the packets are manually loaded onto the moving two adjacent cleats. At this stage, the packets are to be “flattened out” by the operators' hand. This will “smooth out” any package irregularities in the chip packets to ensure the accuracy of the prototype results. The production version would incorporate a servo-driven packet leveller prior to the Vacuum Tunnel entry;
•
using a non-contact electrical sensor, the height of the packet will be measured as it passes by;
•
rubber cleats and a “tunnel” act as an airlock. This ensures that the vacuum chamber is never directly exposed to atmospheric pressure;
•
the packet travels through a vacuum chamber in order to allow the bag to “puff up”;
•
using a non-contact electrical sensor, the height of the packet will be measured as it is in the “puffed up” state. This height profile is compared to the original and a reject/accept decision is made;
•
the package passes through another airlock;
•
the packet is rejected if it is found to be faulty using a quick jet of air, and the accepted packets continue along the conveyor line.
The system also features a regenerative circuit that 'recycles' some of the vacuum exiting the system. The re-use of the vacuum could potentially save energy costs and affect the Vacuum Pump size. 3.4.2
ELECTRICAL CONCEPTS
The design of the “profile” algorithm is completed by a relative simple and effective method of detecting changes in the profile, shape or “puffiness” of the packets (Newman, 2004). The process is measured by the change in relative height before and during vacuum. It was determined by research that this could be achieved by finding the
54
top profile of the packet. There are numerous ways of achieving this, and based on the research conducted, the following algorithm was developed: •
the visual environment must be prepared so that the background is white or light coloured, consistent and non-reflective. Sufficient light source is also required to achieve the correct visual environment;
•
the threshold to be calibrated, were sample images should be taken using the established background and characteristic packets. A suitable threshold can be found by viewing the histograms of the ‘edged’ sample images;
•
two images are captured: a reference image (before vacuum), and a vacuum image (during a calibrated vacuum pressure);
•
the images are converted to grey scale;
•
the images are passed through a 3x3 spatial Sobel edge detection filter;
•
a simplified method of region growing and thresholding is achieved by, commencing at the top of the left-most column of pixels, scanning down the column until a pixel is found with intensity greater than the threshold. The position of this pixel relative to the bottom of the image is recorded as the height of the profile for the particular column;
•
repeat the previous step of region growing and thresholding for each column in the image to produce a profile function;
•
the profiles of the reference and vacuum images can be compared to find any changes in height and shape. A simple test is to find the maximum of each profile function and find the relative change in height ([vacuum height – reference height] / reference height). Any significant relative change in height (e.g. greater than 5%) denotes that the packet has ‘puffed up’ and therefore is well sealed.
55
For the ‘edged’ image shown in Figure 3.14, the profile function is shown to be right:
Figure 3.14: Sample Edged Profile Function from our preliminary results (Source: Unpublished final year undergraduate thesis experiment with differential vacuum at QUT 2004)
The ability of the algorithm to handle “real time” speed of the HSDS is an important consideration in the design process; therefore careful consideration was needed to determine how fast the system must run. The prototype and final HSDS will travel at approximately 100 packets per minute or 1.67 packet per second. The design requires that two images per packet travel through the algorithm, therefore the time allowed for image transfer and processing is 0.835 seconds. In the design, it is assumed that the image transfer must be less than 0.3 seconds and the processing must be less than 0.5 seconds. An uncompressed 640x480 ‘RGB’ image is approximately 8 Mbits in size, and the image must be transferred in less than 0.3 seconds requiring the transfer rate to be greater than approximately 26 Mbits/s. The transfer rate is exceeded by the following common technologies: 100Mb+Ethernet, USB 2.0 and Firewire. The older technology of 10Mb Ethernet and USB1.0/1.1 are not acceptable.
56
During the design process, it was found necessary to optimise the ‘Profile’ algorithm to meet the speed requirements of the final system. The following steps were taken to reduce the number of processor operations required by the algorithm: •
All operations were combined into a “one-pass” algorithm, in lieu of iterating over the images’ pixels several times to complete all the required operations;
•
The grey scale conversion process was optimised by using an arithmetic mean rather than a geometric mean. In calculation the mean, dividing the sum of the red, green and blue components by three was excluded. To compensate this, the value of the chosen threshold was multiplied by three due to a single performance run, as opposed to an operation for each pixel. The grey scale conversation process was then integrated into the “one-pass” algorithm.
•
The Sobel edge detection operator was then optimised by approximating the magnitude of the gradient as lGl = lGxl + lGyl. In addition, this process was also integrated into the “one-pass” algorithm.
The “one-pass” algorithm was designed to allow thresholding of each column of pixel in the image to be easily be integrated.
During the process, once the threshold is
exceeded in a column, the algorithm then moves to the next column thus eliminating any unnecessary operations. In the design process, coding of the optimised ‘Profile’ algorithm in programming language ‘C’ displayed an actual time range of approximately 0.03 seconds. This time exceeded the permissible speed requirement or range of 0.5 second by an approximate factor of ten. Further testing during the prototype development stage would be required to ensure limited movement in the packets during their travel through the HSDS. The above algorithm relies on obtaining a reasonably accurate image of the packet side view. Further testing would reveal changes in height threshold and relevant calibrating would be completed for various packet shapes, types and heights.
57
Another method is the investigation and application of wavelet transform techniques, as mentioned in the Literature Review, “the basic geometry of packets will be determined from the low frequency information and high frequency information will be utilized to perform more specific recognition tasks that detect packets as either being properly sealed or leaky”. The dyadic wavelet transform can also be explored as a technique for effectively recognizing and characterizing package shape changes (Khalil and Bayoumi, 2000) and (Tieng and Boles, 1994). By using the wavelet transform of the affine arc length or enclosed area of a package contour. A dissimilarity function comparing the extreme of the resulting coefficients would be used to compare leaky and sealed packets at a number of scales (Tieng and Boles, 1997). The zero-crossings of the wavelet transform can be used for package identification (Cook, and Boles, 2002). In a factory environment, captured images are expected to be noisy and further research would address removal of the noise prior to object recognition and classification. By representing the signal in multiple resolutions, the wavelet transform will allow noise to be effectively isolated, regardless of the temporal location. Carefully calculated threshold decomposition will effectively remove additive White Gaussian noise from a wide range of signals (Lang et al, 1996) and (Donoho, 1995). NonGaussian noise, such as spikes and unwanted harmonics will be removed using Bspline wavelets (Ainsleigh and Chui, 1996). The multi-resolution properties of the wavelet transform would be used to detect, isolate and remove noise in a similar manner for improving the quality of hearing (Whitmal et al, 1996) and removing phase noise from satellite SAR imagery (Martinez et al, 2001). These noise removal techniques would be adapted in the future to provide a high possibility of enhancing the performance of the image processing. The proposed system has the potential to be robust even in the presence of noise due to variations in package shape, weight and material content. It is anticipated that it will be able to perform extremely well even in complex cases where the leak size is less than 3 mm. In future works a number of electrical techniques could be trialled to determine the most effective method for production use.
58
3.5
RESEARCH & DESIGN OF COMPONENT EQUIPMENT
The selection of critical equipment within the Vacuum Chamber and Tunnel depended on the results of the laboatory tests and relevant engineering calculations. Both were completed in these areas to directly determine a selection criteria of suitable materials of contstruction and to assist with design tendencies and preferences.
The Vacuum Chamber and Tunnel system consist of both a static shell and a moving cleat conveyor belt within the shell to provide a suitable airlock chamber.
Vacuum Chamber Airlock Chamber Cleat Conveyor Belt
Image Processing & Sensing System Reject System
Figure 3.15: Model of Major Component Equipment Within the food industry, stainless steel is a typical material of construction. It provides both suitable strength and rigidity, and a hygienic surface suitable for either direct or indirect food contact. Specific polymers displaying advanced mechanical properties may also be suitable.
59
Standard conveyor belting and cleats are readily available and suitable to food processing requirements. These types of systems typically operate within an atmospheric pressure environment, and may not be suitable under a vacuum or negative atmosphere pressure. Relevant research and engineering calculations were completed on the Conveyor support Frame to verify suitability of standard materials of construction. Refer to Appendix B for FEA Analysis of Conveyor Frame and Vacuum Chamber and Tunnels for both prototype and production units. During this research, it was highlighted that a critical relationship existed between the static airlock chamber and inertia cleats. Standard flexible cleats were deemed not suitable to maintain a suitable vacuum within the airlock chamber. Therefore, a rigid type of cleat having superior mechanical properties would be required to achieve these criteria. 3.5.1
MECHANICAL DETAIL DESIGN
The HSDS requires detailed mechanical design in the following areas: •
conveyor belting,
•
airlock tunnels,
•
vacuum chamber,
•
vacuum pump and control,
•
rejection.
3.5.1.1 Conveyor Belt and Cleats The horizontal conveyor belt was selected as the preferred method of conveying the packets through the HSDS. The conveyor belt plays a critical role in the HSDS for both packet travel and assisting in retaining vacuum as the packet progresses through the system. To properly design the belt the problem was defined, various concepts were developed and the specific details were finalised. The functional statement was to design a conveyor belt and cleat configuration to travel through the HSDS that creates a physical air seal based on the following parameters: •
Move the packet through system;
60
•
develop a high efficiency air seal within a airlock tunnel or chamber;
•
allow the packet to exit and enter the HSDS via suitable means;
•
minimise air flow leakage rate to ensure not excessive, vacuum pumping rates and therefore costs, and
•
ensure belts and cleats are easily replaced and maintenance friendly
Detailed below are various concepts for the belt and cleat configuration for the prototype HSDS; the majority of the concepts focused on different cleat designs: Polyvinyl chloride (PVC) Belt and Cleat: The PVC belt and cleat concept would use an industry standard 75mm high flexible PVC cleat vulcanised onto an approximately 2mm thin PVC belt. To maintain a seal on the top and side faces of the tunnel, a lip seal profile would be cut into the top and sides (see Figure for top profile).
Figure 3.16: Diagram showing the lip seal profile to be cut into cleat
The direct contact between the bottom surface of the belt and the bottom of the tunnel or chamber ensured a seal was created. A local belt supplier (Andrew Cohurst from Precision Belting, Salisbury) indicated that this type of cleat will travel around the head and tail roller readily, providing the roller diameters are at least 150mm (for a T75 cleat). The PVC belt and cleat concept displayed the following advantages: •
the belt and cleat are inexpensive, e.g. the T75 cleat is available from the local supplier for around $20 per metre (excluding labour). The whole conveyor belt is relatively inexpensive, at a cost of only $900 for a 5.6m endless belt x 300 mm wide 2mm belt with cleat spacing of 200mm;
61
•
white belt available, which is used in food industry to help identify foreign objects e.g. metal and timber, and
•
a wide variety of cleats are available in many sizes including 30mm, 50mm, 75mm and 100mm.
However, the primary disadvantage of this concept was that the PVC cleat was too flexible. Based on the assumption that each cleat must seal against a pressure of approximate - 30 kPa, it was theoretically determined that the deflection would create excessive leakage. In addition, when the forces due to friction acting on the cleat are in the same direction to the force due to the change in pressure (when cleats are exiting the system), the deflection would potentially be significantly increased. A secondary disadvantage was that the lip seal profile was extremely hard to cut during manufacture. Flexco Tatch-a-Cleat ™: This concept was similar to the PVC cleat concept, in relation to the sealing with the exception that the Tatch-a-Cleat was far more rigid. The high rigidity was accredited to the thick design and the use of rigid Styrene Butadine Rubber (SBR) compound (a rubber type material similar to car tyres). The cleat is mechanically fastened to the belt, and relies on the use of bolts through the belt.
Figure 3.17: Flexco HD Tatch-a-Cleat High loading on the cleat and the stress concentrations around the holes in the belt, will potentially cause localised elongation around the boltholes. The disadvantage is a cleat could become dislodged from the conveyor belt and lose contact with the belt. This would cause lost contact with the top of the belt and a significant air leakage at each revolution, resulting in increasing operational costs. Further to this, and after many
62
cycles of fatigue loading, the bolts may tear through the belt and lead to unexpected costly production down time. Despite the increased rigidity, the difficulty of cutting the lip seal profile with this cleat is still present. Modular Belt: The Modular Belt concept involved the use of a standard flight modular belt to produce a vertical seal through the tunnel or chamber. A tolerance normal running fit between the tunnel and cleat will produce the seal for this concept. Although leakage will occur, the maximum gap width will be limited to the sum of the fundamental deviation plus tolerance on tunnel plus tolerance on cleat.
Figure 3.18: Flighted modular belt Source: http://www.intralox.com/products.asp?page=800ohirf The presence of a linking system in modular belts by plastic pins did not impair the viability of this concept. Each cleat must provide a seal on all four-tunnel faces in one vertical plane. Therefore, the presence of holes in the belt between cleats is independent of the ability for the cleat to create a seal. The Modular belt and cleat concept displayed the following advantages: •
no special design required and standard modular belting could be utilised;
•
hygienic for direct food contact, particularly in a dry or packaged food environment;
•
flights are available in high impact form that can readily withstand the high loading due to pressure or vacuum;
63
•
minimal or no belt slippage problems with the use of modular belt.
The base of the belt may not be able to create a seal because of the presence of teeth and /or small air gaps. This concept may result in an excessive amount of leakage that it would become impractical to run. A potential option to consider during prototyping would be to include a solid base fro the belt to travel on; this would reduce and/or eliminate the air gap issue. The belt would also need to be checked for jamming issues occurring. This may impact in a small piece of the flight becoming dislodged and reaching the finished product. Two-Piece Cleat: Two-piece cleats are often used in heavy-duty applications, e.g. bulk materials handling. This set-up consist of a urethane insert bolted into a rubber base, and can be imported through Sandvik Materials Handling. The proposed concept would use the base part of the cleat cold bonded onto a rubber belt with a specially designed sealing insert as shown in Figure 3.19.
Rubber Seals
Cleat
Hard Plastic Conveyor Belt Bolt Holes
Figure 3.19: Two-piece cleat concept
64
The Two-piece cleat and bonded belt concept displayed the following advantages: •
the replaceable rubber seals are reasonably inexpensive and simple to manufacture and could be replaced by maintenance staff without removing belt;
•
the rigid plastic insert allows the cleat to be able to withstand high loading with minimal bending;
•
the cleat insert height can be designed to suit the maximum inflation height of packets. The advantage is, it minimises the amount of air introduced into the system and reduces pumping and energy costs. In addition, the leakage over the cleat may be minimised because the effective length of the seal is reduced.
•
multiple seals can be used to decrease leakage. However, too many seals cause spacing reduction and potentially congested with debris and therefore encouraging bacterial growth.
Although the cleats are not in direct contact with raw product, this concept may not be acceptable in a production environment due to the increased interfaces and crevices, and the potential for bacteria to grow. In addition, the use of complicated assembly components increases the risk of a component to break off, become loose and enter the production line, causing the breakdown of another machine, or the component reaching the consumer. Remacleat: This concept featured a T75 Remacleat cold-bonded to a rubber belt. This is similar to the PVC concept as a standard T cleat is bonded to a flexible belt.
The major
advantage of the Remacleat range is the extensive rigidity when compared to the PVC concept. In order to create a seal, sections of flexible rubber were fastened on both vertical faces of the cleat (Figure 3.20).
65
Rubber Sheet Seals Bolts
Cleat Plate
T75 Remacleat
Figure 3.20: Remacleat concept This type of seal is inexpensive and easy to manufacture and can be replaced by maintenance staff without replacing the whole belt. The stainless steel cleat plate holds the seals in place and adds to the rigidity of the cleat to withstand forces from pressure and friction. Major advantages of this concept include: •
dual sealing, to assist creating amore effective vacuum seal;
•
the cleat plate is an integral mating surface during the cleats travelling around the head and tail rollers.
Major disadvantages include: •
the stainless steel cleat plate provides additional mass to each cleat. This additional mass will increase forces due to centripetal acceleration as the cleat travels around the rollers;
•
the concept contains four major mating faces per cleat where bacteria can harbour;
•
a large pulley diameter is recommended by the manufacturer to prevent cleat tearing;
66
•
the nature of the assembly is a potential food safety risk due to machine components loosening and the risk of contamination occurring downstream of the production environment;
•
the supplier (Sandvik Materials Handling) only stocks black-coloured belts and the Remacleat range is only available in black.
Prototype Conveyor Belt and Cleat Design Criteria and Selection: During future prototyping, it is recommended to trial two types of belt and cleat arrangement to determine fit for purpose: •
the cleat spacing and belt width would be determined by the packet size. Based on using a TSSC 28 gram packet of approximately 150mm x 150mm, the cleat centres would be 200mm with a belt width of 200mm. This would allow the packets to be easily loaded onto the belt and progress through the machine without excess air leakage due to large compartment size;
•
belt selection: Two options would be considered and trialled, Remacleat from Sandvik Materials Handling and Modular Belting from Intralox. With Remacleat, Sandvik recommended a 2-ply 6.5mm or a 3-ply 8mm rubber belt with a low friction base for the T75 cleat. Although the 6.5mm belt was thicker than the initial proposed belt thickness of 2mm, the selection would be compatible with the T75 cleat. In using the Intralox Modular belt and cleat system, the prototype machine will have a solid bed to eliminate air leaks;
•
rubber Seal: Clark Rubber was contacted regarding available rubber sheets to use for the seal, and 0.8mm 1.5 mm and 3mm rolls were found to be commercial and industry standard. The 1.5mm thick natural rubber was chosen as it was considered to provide reasonable wear life and not inhibit excessive force required for seal deflection. The thickness was selected by the writer based on previous practical engineering knowledge.
Natural rubber was
selected over neoprene due to superior resistant to wear; •
cleat plate: The cleat plate would be manufactured from 304 S/S to adhere to the health and safety regulations of TSSC and other food processing companies. The cleat plate criteria would be of sufficient thickness to prevent excessive cleat bending when vacuum is applied but not excessively thick to cause the cleat to disengage from the belt. A 2mm thick plate was initially selected based on practical experience and the thickness selection was made
67
difficult by the unknown pressure distribution through the tunnels and 2mm thickness was assumed for the purpose of the test rig. Once the prototype unit has been built, the pressure distribution will be known and this thickness can be minimised; •
speed: The speed of the belt was limited to the maximum rate of manual loading onto the prototype, which was assumed 2 seconds per packet. Therefore, the speed of the belt when manual loading = 0.1m/s. The belt speed for a future inline testing or production machine would be 0.33m/s (200mm cleat spacing and 100 packets per minute).
Limited literature was available regarding the design of lip seals to be used in the proposed fashion. Therefore, the dimensions for sealing were assumed as per Figure 3.22. These assumptions need to be confirmed in future work before building prototype. It was decided to use a robust, “belts and braces” approach on the cleats during future prototyping to ensure minimal deflection of the cleat under vacuum. During prototyping it would be also recommended to trial less rigid cleats to provide a less complicated solution.
Top of Tunnel
Figure 3.21: Assumed dimensions for lip seal
68
Recommendations for Future Work: Late in the research, a new type of cleat was found during research, it is manufactured from urethane and was developed in Europe. This would be investigated further before a prototype unit is manufactured. The production HSDS would also incorporate a solid stainless steel bed to eliminate air leaks. A FEA model has been completed in Appendix B to confirm correct stainless steel bed thickness. In order to minimise excess friction between the conveyor belt and bed from occurring, it would be recommended to food grade Teflon tape or thin UHMW strips. The Teflon tape has the advantage of being applied across the full width to reduce any air gaps, and has been successfully used throughout the food processing industry for packaging purposes. In addition, the fastening system must be modified to ensure the bottom of the seal does not pull out of cleat (Figure 3.23). This may be achieved by the introduction of another bolt or a rubber cement.
Seal may be pulled out from Cleat
Figure 3.22: Seal dislodgement point
69
As previously mentioned, one of the disadvantages of this concept was that it required assembly which may come apart and become a consumer health hazard. Therefore, testing should be conducted to determine if the cleat plate is necessary for stiffening purposes. If it is not required, it can be eliminated from the cleat assembly and the rubber seal could be cold bonded directly onto the cleat using a rubber cement. However, doing so may require the whole belt to be replaced when seals wear out and a cost/risk analysis should be performed to determine the best outcome.
Rubber
cements will require research to find possible cements that can be dissolved when seal replacement is required. These components and their continued development determine the effectiveness of the entire system. 3.5.1.2 Vacuum Tunnels The vacuum tunnels are designed to develop the vacuum or sealing effect prior to the packet entering the vacuum chamber. The vacuum tunnels act an intermediate air-lock between outside air and the chamber as a seal between the vacuum chamber. The following section explores two possible concepts and details the areas of specific design. The functional statement was to design the vacuum tunnels to permit packet travel through the HSDS that creates a physical air seal based on the following parameters: •
construct a continuous rectangular sealing surface to suit cleat and seal;
•
develop an airlock to ensure vacuum is not lost or loses are minimal;
•
incorporate a regenerative circuit to “recycle” a proportion of the vacuum exiting the system;
•
design a sealing surface with a low co-efficient of friction to reduce seal and wall wear;
•
design of tunnel to withstand an internal pressure differential of approximately 30 kPa vacuum;
•
all components easily replaced for maintenance and access purposes;
•
the sealing surface must be low friction to reduce seal and wall wear, and
•
the tunnel must be 84mm (+ 0.5mm - 0mm) high and 200 +/- 0.2 wide to suit belt.
70
Detailed below are various concepts for the Vacuum Tunnels and configuration for the prototype HSDS: Pressed or Bent Stainless Steel Sheet Metal: This concept involved the use of a single section of Stainless Steel pressed into the shape of a tunnel that is designed for bolting to the conveyor main frame.
The
conveyor belt travel on the inside of the sheet metal vertical wall (Figure 3.24).
Figure 3.23: Folded sheet-metal concept
The stainless steel sheet metal concept has the following advantages: •
continuous welding along the length of the tunnel was not required due to the majority of work being completed by a brake press; thus minimal or no welding is required. Welding disadvantages are heat created and potential shrinkage and deformation or bending of the stainless steel.
Additional post weld
treatment such as pickling and passivation is also required, plus grinding or shaping the metal profile; •
a Teflon coating could be applied in strips or baked onto the inside of the tunnel to reduce the need for lubrication;
•
minimal joints and interfaces were required, reducing the risk of bacteria growth, and
•
stainless steel is wear resistant.
71
This concept was discarded due to potential excessive leakage occurring in the corners of the tunnel due to the radius created in sheet metal bending. From the writers experience in previous fabrication workshops, and by the nature of the pressing operation, a minimum radius is produced during the process. This minimum radius is a least the same as the thickness of the material. Assuming a material thickness is 2.5mm, then the internal bend radius would be 2.5mm and hence the external bend radius would be 5mm. An extra variable to consider is the accuracy repeatability in the pressing process, with respect to tolerances, parallelism and perpendicularity of the sheet metal’s different faces. Teflon: Teflon was first considered due to its low co-efficient of friction properties resulting in long cleat seal life. However, Teflon does not possess durable wear properties required for this application. This would relate to frequent material replacement, and therefore high operational and maintenance costs.
An industrial plastics supplier (E-Plas)
confirmed this statement and advised other plastic solutions. UHMWPE (Ultra High Molecular Weight Polyethylene), Polycarbonate (Lexan®): For the prototype HSDS, sheets of UHMWPE or Polycarbonate would be fastened together to form the vacuum tunnel with a stainless steel base (Figure 3.24).
72
Top Plate
Walls
Bottom Plate Figure 3.24: UHMWPE assembly concept
This concept would facilitate the requirement of accurate manufacturing, and the tunnel walls can be precisely milled to the correct height. The distance between the walls of the tunnel could be adjusted to account for any slight manufacturing inaccuracies of belt width by manoeuvring the walls firmly against the belt before tightening fasteners. The UHMWPE or Polycarbonate concept has the following advantages: •
the low co-efficient of friction of UHMWPE walls reduce the need for lubrication;
•
standard manufacturing processes and practices can be used, and
•
easy replaceable walls for maintenance purposes.
The UHMWPE concept has the following disadvantages: •
walls may potentially experience excessive wear due to constant rubbing with cleat and belt, and
•
increase manufacturing costs due to large quantities of tapped holes.
73
Prototype Vacuum Tunnels Design Criteria and Selection: Local industrial plastics supplier, E-Plas, recommended using UHMWPE due to it's low coefficient of friction of around 0.15 and high wear-resistant properties. This material was chosen for the walls due to the excellent wear/friction properties and the cost effectiveness when compared to Teflon. Other advantages of using UHMWPE include a standard of white for this material, which was a major advantage to food processing manufacturers.
White is considered the preferred colour in both a processing and
packaging environment, and it the desirable colour for digital imaging. Finite Element Analysis (FEA) using Femap with NX Nastran software was completed to determine material thickness as shown in Appendix B.
The FEA confirmed a
thickness of 12mm for the walls. Recommendations for Future Work: Future manufacturing research is recommended to produce a UHMWPE or Polycarbonate tunnel with reduced fasteners, and possibly moulded as one or two large sections. When proceeding to full mass production of HSDS units, it would become economically viable to produce the tunnels in one piece and have the items listed on a spare parts list for periodic replacement. 3.5.1.3 Vacuum Chamber The vacuum chamber is designed to ensure the packet is held at a lower pressure during critical measuring to assist in accurate decision-making. Similarly to the vacuum tunnels, the functional statement was to design the vacuum chamber to permit packet travel through the HSDS that creates a physical air seal based on the following parameters: •
construct a continuous rectangular sealing surface to suit cleat and seal;
•
develop an airlock that maintained a vacuum pressure environment of approximately 30 kPa vacuum;
•
tolerance on pressure differential = +/-2% of differential pressure;
74
•
suitable for digital imaging requirements and other external electronic requirements;
•
packet equilibrium time of approximately 2 seconds maximum;
•
allow the packet to reach equilibrium before final image capture;
•
packet speed 30 packets per minute for prototype unit and 100 packets per minute for production HSDS;
•
incorporate a regenerative circuit to “recycle” a proportion of the vacuum exiting the system;
•
design a sealing surface with a low co-efficient of friction to reduce seal and wall wear;
•
design of tunnel to withstand an internal pressure differential of approximately 30 kPa vacuum;
•
all components easily replaced for maintenance and access purposes;
•
the sealing surface must be low friction to reduce seal and wall wear, and
•
the tunnel must be 84mm (+ 0.5mm - 0mm) high and 200 +/- 0.2 wide to suit belt.
The material used was clear polycarbonate because it is transparent, which will assist in visual performance assessment of leaks (using Bubble test) and packet viewing. In addition, it possessed high resistance to brittle failure, making it a safe material to use for a pressure vessel. Polycarbonate of 12mm thickness was assumed because this was the largest standard size available and FEA confirmed the thickness as adequate. The design concepts are the same for both the Vacuum Tunnels and Chamber therefore the concept selection has not been repeated. Further discussions on material selection and fastening systems for the Vacuum Chamber are detailed below: Material Selection: Important design criteria in the selection of the material are listed below; •
low coefficient of friction while maintaining high wear properties;
•
material cost, consider initial and whole of life cost;
•
material availability (consider local versus overseas) and procurement period to receive spare parts and components;
75
•
ease of manufacture including machinability of the materials which can impact on cost, and
•
suitability for food grade purposes.
Fastening Connection System: A well designed fastening connection system is essential to the operations of the HSDS unit. The fasteners, and their selection, ensure correct assembly, provide structural integrity and reduce tunnel air leakage. Food processing manufacturers prefer the use of a minimum of 304 Grade stainless steel (S/S) fasteners to meet strict hygienic standards. These fasteners are corrosive resistant and do not require any protective coatings, unless used in an aggressive environment. In packaging areas full “wet sanitation” is generally not required, and equipment is usually wiped down. The use of 304 Grade Stainless Steel fasteners to Grade A2-70 to ISO3506 would suffice in this type of environment. In the event of a more aggressive environment where strong cleaning chemicals are used, or the general area is highly corrosive, then 316 or higher alloy is recommended. It was decided to use 304 S/S bolts to meet the strict hygiene requirements of the TSSC Company. The shape of the bolt head would vary according to the operating environment. Hexagon heads provide no hidden areas for debris or bacterial growth, but have sharper areas that act as a catch point. Domed heads are curved in nature but have hidden areas for potential growth. See design calculations in Appendix B for further details. Prototype Vacuum Chamber Design Criteria and Selection: The chamber design has a number of important design features.
Discussions are
based around the following: •
vacuum Tunnels,
•
vacuum Chamber sizing,
•
top plate design,
•
vacuum generation and control (will be discussed as future work).
76
Vacuum Tunnels: The tunnel entrance and exit is determined by a number of factors. The minimum length is calculated according to the size of packet compartment. In the prototype the length is proposed to be 200mm. The design also requires a vacuum regenerative cycle, which will be discussed in a later section. The regenerative port also requires one compartment, therefore a minimum of two entry and two exit compartments will be designed into the prototype unit. The layout shown in Figure shows the proposed prototype solution.
Figure 3.25: Diagram of the layout of the walls to ensure an airlock is created
This design ensures a tunnel seal will be covered over a minimum of two cleats at any one time. The value of “x” was originally chosen to be 216mm, which is the radius of the port (16mm) plus the spacing of the cleats (200mm). However, the distance between the ports needed to be a multiple of 200mm and “x” was chosen to be 300mm (based on the assumption that the vacuum chamber length is a multiple of 200mm). To ensure at least a minimum of two cleats are acting as a seal at any time the total length of the tunnels would be nominally 600mm.
Calculations and assumptions below
indicate the theoretical optimum number of cleats in the tunnel. It was decided to use Torricelli's theorem, formula 3.1 below which states that the velocity of the issuing jet, or in this case, an air leak, is proportional to the square root of the head producing flow (Douglas et al 1995). Therefore, if the leakage past a cleat can be approximated as an orifice plate with a small restricting hole, the flow rate Q can be given by the following:. Discharge e, Q = CdA√(2p/r) ............................................................................... (3.1) Flow rate of vacuum leak past conveyor cleat (Douglas et al, 1995)
77
where A is analogous to the leakage area of a cleat and p is the pressure difference over a cleat. This theoretical analysis is not sufficient to evaluate the optimum number cleats because the area is very hard to determine and is dependant on the following factors:
seal effectiveness,
seal deflection,
cleat deflection,
accuracy of manufacture.
Chamber sizing: The length of the chamber is important in assuring that the packet has undergone suitable vacuum for measuring comparison purposes. The packet requires being held at a low pressure for at least 2 seconds until measurements were taken, the length (L1) between the start of the chamber and the sensor was determined as follows: L1 = minimum equilibrium time x maximum belt speed = 2 x 0.333 m/s = 0.666m = 3.33 cleat spaces (this was rounded up to 4 cleat spaces for design simplicity) Therefore L1 = 800mm To avoid moving air from cleat leakage blowing past package during measurement, the sensor was mounted 200mm (i.e. 1 cleat space) from the end of the chamber. Thus, the total length of the chamber was 1000mm. Every time a packet advanced into the chamber a volume of air also entered (shown in red in Figure 3.26), causing pressure fluctuations.
P1
P2
V1
V2
Figure 3.26: Air pressure distribution in Assembly before equalisation
78
P3
V3
Figure 3.27: Air pressure distribution in assembly after equalisation
Where P is the air pressure and V is the air volume.
p1v1 + p2 v2 = p3 v3 Where, v3 = v1 + v2 ∴ p3 =
p1v1 + p2 v2 v1 + v2
Re arranging gives, v2 = v1 (
p2 − p1 ) p 2 − p3
where : v1 ≈ 3.1×10 −3 m 3 p1 = 71.3kpa p2 ≈ 86.3kpa ( from regenerative circuit ) p3 = 101.3 − 30 × 0.98 = 71.9kpa ( 2% fluctuation) ∴ v2 = 3.1× 10 −3 (
86.3 − 71.3 ) = 74.4 × 10 −3 m 3 86.3 − 101.3
79
It was decided to achieve this volume by designing the vacuum chamber the full length of the tunnel walls
Regenerative Circuit Pipes Vacuum Chamber
Slotted Vacuum Plate (Section 0)
Figure 3.28: Compact vacuum chamber design Volume to be contained by cover:
Vcov er = 74.4 ×10 −3 + 2 × Vregenerative pipe − 5 × vcell −3
Vcov er = 74.4 ×10 + 2 ×1.7 × ( Height cov er =
π × 0.032 2 4
) − 5 × 3.1×10 −3 = 0.0616m 3
Vcov er 0.0616 = = 0.131m ≈ 0.135mm Widthcov er × lengthcov er 0.2 × 2.350
FEA was completed to verify the design thickness of the conveyor top and vacuum chamber base. A vacuum pressure of 30 kPa was applied to the Stainless Steel conveyor top, which revealed a minimum permissible thickness of 2.8mm. The next thickness stock material available is 3mm 304 s/s.
80
Detailed Top Plate Design: First design concepts excluded the provision of a top plate covering the Vacuum Chamber cleats. However, digital imaging may have been adversely affected by having a top cover in the first picture (at atmospheric pressure) but not in the second because light cannot reflect off a top surface. This suggested the use of a top plate in the vacuum chamber. The three main features of this concept that required detailed design is shown in Figure 3.29.
Vacuum Chamber Slots Regenerative Circuit Ports
Fillet
Figure 3.29: Chamber top plate The Top Plate has a number of design features as detailed below:
•
the Top Plate end has a 5mm fillet ground out to assist in the transition of the seals at the entry to the Vacuum Tunnel.
This feature assists the seals in
bending over, to reduce any shock loading and hence increase fatigue life (Figure 3.32);
•
incorporate regenerative circuit pipes inside the chamber cover (see Figure 3.31) as it facilitates compact design and virtually eliminates the risk of operators leaning on the pipes and applying undue force.
81
Although the
connection method shown in Figure 3.29 does not permit assembly, the manufactures discretion would be used to decide what connections to be used so that the regenerative pipes can be directly screwed into the top plate. A solution to this was to introduce of a screw-in “Plug” (see Figure 3.30) manufactured from UHMWPE. This optional plug can be used in the prototype phase to determine if longer tunnels with more cleats are preferential to using the regenerative concept, by sealing off the tapped holes in the top plate. If it is found to be beneficial to use the regenerative concept, a 5mm fillet will be required on the underside of the plate around the portholes.
Figure 3.30: Screw-in plug used to seal tapped holes •
These slots were required to apply a vacuum to the packets as they entered the vacuum chamber. The vacuum chamber effectively started and ended at the start and ends of these slots. A 5mm fillet was required on the underside of the slots.
82
Regenerative Circuit Pipes
Vacuum Chamber Slots
Figure 3.31: Section view of chamber
Fillet to Assist in Seal Bending
Figure 3.32: Fillet on underside of top plate
83
3.5.1.4 Gaskets The functional statement is for all gaskets to provide adequate sealing for its intended purpose and be a food grade gasket.
In a corrosive environment, additional
considerations are required in the selection of appropriate gaskets. Suitable gaskets are required to ensure sealing between the walls and the top and bottom plates. The gasket arrangement shown in Figure was proposed to allow some minor compression adjustment from changing of bolt tension to provide a minor tunnel height adjustment. This feature would be useful during future testing and commissioning phases of the prototype unit to find optimal chamber height. The proposed gasket material is a natural rubber of 0.8mm thickness.
Gaskets
Figure 3.33: Gasket arrangement for chamber
84
3.5.1.5 Regenerative Circuit The functional statement for the Regenerative Circuit is to reduce pressure fluctuations, assist in creating a vacuum environment in both the entry and exit Vacuum Tunnels. Pressure fluctuations in the vacuum chamber can cause inaccuracies in measurement. This can be minimised by reducing the amount of air entering the system with each package. It is achieved by decreasing the amount of expansion of the air as it enters the vacuum chamber by lowering its pressure.
Figure 3.34: Schematic diagram representing the expansion of air entering the vacuum chamber
Design calculations in Appendix B show pressure and volume within the vacuum chamber. An original concept of reducing P1 was to use a “roughing pump” (high flow but minimal pressure drop across pump) with the input port mounted to the first tunnel. The use of this pump would increase both manufacturing and operational costs. It became evident that this pressure drop could be obtained by equalising the pressure of the air exiting the system with the pressure of the air entering the system. Figure 3.35 explains diagrammatically the theory relative to the Appendix B regenerative circuit design calculations.
85
Figure 3.35: Chamber pressure diagram
This iterative process was modelled in Microsoft Excel and it was calculated that the air entering into the system would be lowered to 86.3 kPa after nine cleat spaces have progressed. Figure 3.36 displays a graphical format of the iterative calculation process.
Pressure of Air Entering Vacuum Chamber vs Cleat Spaces Moved 105
95 90 85 80
Cleat Spaces moved
Figure 3.36: Pressure of air entering vacuum chamber
86
31
29
27
25
23
21
19
17
15
13
11
9
7
5
3
75 1
Pressure [kpa]
100
3.5.1.6 Vacuum Pump and Control The functional statement for the Vacuum Pump and Control is to maintain a suitable vacuum pressure with minimal pressure fluctuations, and have provisions for variable vacuum pressure with the use of a vacuum control system. Vacuum creation (Mergard, 2004): To select a specific pump, the head H and flow rate Q required are needed to find the operating point. The head required is dependant on the differential pressure required to inflate bag and will be determined by testing of the prototype unit. Q is the flow rate through the system and can be arbitrarily represented using the continuity equation. O = E flow into system - E flow out of system In the case of the vacuum chamber: 0 = (Q leakage + Q cell + Q cell expansion) - Q pump The Q
cell
term can be eliminated because the rate the cells (air pocket between
adjacent cleats) enter the system is the same as they exit. Therefore, Q pump =Q leakage + Q cell expansion Q pump needs to be measured experimentally once the prototype has been built because Q leakage is affected by:
the number of cleats in the tunnels; as the number of cleats in the tunnels in increased, the total pressure change occurs over more seals and hence the leakage rate is decreased;
the effectiveness of the cleats/seals;
87
the friction and speed between seal/tunnel because this will influence the level of cleat bending.
Vacuum Systems (Mergard, 2004): Vacuum systems are used in many different applications throughout various industries. However, vacuum systems fundamentally consist of "a vacuum pump and vacuum vessel, together with the necessary plumbing and vacuum-measuring equipment" (Roth, 1990). These systems can be categorised as either static or dynamic. In a static system, the pump is isolated from the chamber after it is appropriately evacuated, whereas the pump of a dynamic system continues to work on the system during the process after differential pressure is reached. Static systems are more suited to discrete or batch applications and dynamic systems are generally more suited continuous applications. Given the fact that the chip-testing problem is in a continuous processing line, dynamic systems will be the focus of literature research. However, the research was limited to the following two aspects of a continuous vacuum system: •
vacuum pumping;
•
maintaining internal vacuum despite introduction and extraction of product.
Vacuum pumping: Vacuum systems can be classified according to the vacuum range they operate into the following Categories: •
rough or Low Vacuum: 1013 mbar- a 'few' mbar,
•
medium Vacuum: a few mbar - 10-3 mbar,
•
high Vacuum: 10^-3 mbar-10-7 mbar,
•
ultra-high Vacuum: below 10-7 mbar.
(Roth, 1990)
88
Mechanical rotary and "Dry" pumps were researched in further detail, as they operate in the range of around 700mbar pressure that was required for package testing. Mechanical Rotary Pumps: Mechanical rotary vacuum pumps are the most commonly used type of rough vacuum pump, because they can be used continuously, and a "quick pump down cycle can be repeated immediately" (Roth, 1990). These pumps seal minor leaks using oil, and the vessel can be contaminated with oil vapour for pressures lower than 0.1mbar. The two main types of mechanical rotary pumps are the rotary vane and the rotary piston pumps. The rotary vane pump consists of a rotor, with seal blades, offset inside a stator (Figure 3.37). A cycle consists of induction, isolation, compression and exhaust stages.
Figure 3.37: Rotary vane pump Source: www.lesker.com This design has the advantage of minimal vibration, as the only eccentric rotation would be a result of the relatively light blades. The rotary piston pump is similar to the vane pump, in that it goes through the same four stages, except that a cylindrical piston rotates around the stator (see Figure 3.38). Although this pump has higher vibration forces, far higher flow rates can be achieved.
89
Figure 3.38: Rotary piston pump Source: www.lesker.com Dry Pumps: Dry pumps are a type of mechanical pumps that operate without the use of oil. The main advantage of this is that there is an increased resistance to particles and chemicals in the environment. For example, the absence of oil reduces debris build-up in pump lubrication channels. Dry pumps include the following types: •
carbon vane rotors,
•
screw rotors,
•
roots stages,
•
claw rotors.
(Roth (1990)) Vacuum Containment: Maintaining an internal vacuum/pressure, despite introduction and extraction of product, was found to be a less extensive and less diverse area of study than vacuum pumping. However, pneumatic conveying systems were found to use this concept to introduce bulk material into the conveying system. Here the objective is to introduce the product with minimal loss of differential pressure.
90
There are three common methods to achieve this goal: •
rotary airlock feeders: This concept is essentially an air-sealed star wheel where an airlock is created in each cavity of the continuously rotating rotor;
•
the double door discharge gate: Only one gate can be opened at any one time to ensure an airlock. Gates are opened by means of gravity/spring, pneumatic ram or rotating cam; and
•
the dribble valve: As the product accumulates in a hopper, the weight of the product acting on the valve overcomes the required seating force created from a differential in pressure. This allows product to "dribble" in.
Vacuum control To ensure the chamber is maintained at the correct pressure, vacuum control methods need be employed. These methods must ensure a pressure differential regardless of belt speed and to maintain a constant pressure differential after potentially two 8 hour shifts of continuous operations. Control for Belt Speed: As the belt speed is increased, so does the leakage rate. For a given vacuum pump, an increase in Q leads to a decrease in H, leading to inaccurate decision-making regarding the chip packets. The mathematical relationship between Qleakage and belt speed is not known because the effect of friction is not known. At different belt speeds, the system has differing system characteristic curves because the flow rate required (to produce a given head) is greater for the system with a high belt speed than that of the system with low belt speed. The same pump can be used in conjunction with a Variable Speed Drive (VSD) to allow the prototype to be adaptable to various belt speeds because Q/N = constant (Douglas et al. 1998). During commissioning and in production, the operator can simply alter the speed of the pump until the correct pressure differential is obtained. This concept will also be useful in altering the pressure differential as different packets may have various vacuum requirements, especially in different packaging industries.
91
Control for Continuous Running: The vacuum chamber must stay at a constant pressure to ensure the legitimacy of the results.
This can be achieved by simply attaching a vacuum regulator between
surrounding air and the vacuum chamber. Although this method may be appropriate to the prototype, this would increase the operations cost of the inline testing apparatus by constantly introducing air into the system.
An alternate and more efficient method
would include an electrical feedback loop between a pressure transducer in the vacuum chamber and the VSD controlling the pump. 3.5.1.7 Packet Rejection The functional statement of the Rejection System is to provide a precise and accurate method of rejecting all unsealed packet from the HSDS unit. This operation would be completed directly after the packet exits the Vacuum Tunnel. A number of standard food processing options are available to perform with operation and include the following: •
compressed Air Knife Nozzle,
•
mechanical Reject Sweeping Arm, and
•
compressed Air Push Rod.
As shown above compressed air is a common denominator in two of the options. Compressed air is used extensively in food process and packaging machine and service applications. It is considered a clean method, involving little or no oil present in the compressed air system, particularly if the correct ancillary equipment is incorporated in the entire system. The disadvantage of the Mechanical Reject Sweeping Arm and the Compressed Air Push Rod (see Figure 3.39) is these operations require a mechanical device to enter the path of the conveyor system, and could potentially damage the conveyor cleats. Synchronisation would need to be very accurate in order to not clash with the conveyor system.
92
Figure 3.39: Pneumatic push rod on conveyor The Compressed Air Knife Nozzle operation will not enter the path of the conveyor system as shown on Drawing BN72-001 and BN72-101. It relies on a jet of air to blow or push the rejected packet off the side of the conveyor. The system incorporates a solenoid valve to activate the compressed air knife nozzle. Correct positioning and orientation of the air knife is vital to the success of the reject operation; this is normally completed during prototyping or commissioning phase of the project. Valve Selection: The pneumatic solenoid valve receives a signal from the PLC to reject an unsealed packet. The solenoid valve shall require the following operational functions: •
two port,
•
functionality: Operate a “normally closed” positional valve,
•
material: Stainless Steel main body,
•
size: 3/8 port size,
•
operating Pressure: Nominal 700 kPa or 7 Bar compressed air pressure.
The SMC VCW41-6-D-4-03 met the above requirements and is locally available from SMC Pneumatics.
Other manufacturing brands including Festo and Norgren also
supply similar types of pneumatic valves which are readily available.
93
The food
manufacturing company will normally stipulate their preferred brand in specification document. Air Knife Nozzle: Air Knife Nozzles are a standard purchase item from local Australian suppliers as Spray Systems in Melbourne. Varying widths are available to suit physical space limitations; air knife hole sizes also affect the compressed air flow rate of the nozzle. Air Knife Nozzle Positioning: It is proposed to position the compressed air knife nozzle centre about 5mm nominally above the top of the belt to ensure that the moving air does not flow over the top of an empty packet. The nozzle would be mounted, perpendicular to the belt to maximise its effectiveness in discarding rejected packets. As discussed previously, final positioning will be determined during actual operation.
During prototyping, and to simplify the
automation process, the air knife would be positioned at a specific multiple of the cleat spacing from the sensors to ensure it is ‘in-phase’. 3.5.1.8 Packet Leveller The functional statement for the Packet Leveller device is to incorporate the following within the HSDS unit: •
provide packet height consistency by equalising the packet height prior to first camera digital imaging;
•
stimulating the packet and reducing the “self-seal” issue, therefore better preparing the packet prior to first camera digital imaging;
•
full and incremental height adjustment to suit varying packet heights and varying calibrated applied pressure.
The Packet Leveller device will greatly assist the digital imaging process by providing a consistent packet height for comparison purposes before and after the vacuum process. Precise packet height is essential to the success of accurately diagnosing sealed or unsealed packets and will enable the HSDS unit to perform with less packet reject errors.
94
The packet stimulation process will also assist in rectifying the issue of “self-sealing” at the site of the leak, as the process will attempt to “deflate” of self-sealed packets by applying an even pressure. Detailed below are various concepts for the packet levelling for the prototype HSDS: Brush: A basic design, for prototype use only, is a stationary flexible brush as shown in Figure 3.40. The travelling packet would be pressed down by the stationary brush. Although this concept would be simple and inexpensive, the packet may slide and twist with respect to the belt, causing potential measurement errors.
Brush
Cleat
Conveyor Belt Direction
Chip Packet Figure 3.40: Brush concept
Foam Roller: The foam roller would again be predominately utilised and trialled at prototyping phase, but would generally be not suitable in a high volume production line. It does provide a
95
cost effective solution and could potentially used for low production volumes on a budget type production machine. Figure 3.41 shows a simplistic version of the process, excluding any drive and oscillating mechanism.
In practice, the roller mechanism would be linked to the
conveyor drive system using a cam action and possibly belt driven. This concept is not highly recommended due to the load being applied over a small area which may cause product damage. A large diameter roller may reduce this issue.
Compression Roller
Chip Packet Cleat Conveyor
Figure 3.41: Foam Roller concept
Flexible Pads: The Flexible Pad concept consisted of pads mounted to an independent conveyor mounted directly above the main belt (see Figure 3.42). Both conveyors would operate at the same speed (velocity), and the Flexible Pad would gently apply an even load on 96
the packets as they passed by. The pads are manufactured from Ethylene Propylene Diene Monomer (EPDM) for flexibility around the head and tail rollers.
Velocity x m/s
Velocity x m/s
Figure 3.42: Flexible pad concept
The major advantage of this concept was that the pads provided even pressure to the packets for a relatively long time, allowing the packets to be deflated before measurements are taken. The supplier advised that EPDM sponge is impermeable to most oils and does not encourage bacteria growth like most conventional sponges. The disadvantage of this concept was that it would be difficult to ensure exact timing between pads and cleats because slight inaccuracies in roller machining. Production Pneumatic Bag Leveller: The Production Pneumatic Bag Leveller would consist of a height adjustment control to different packet heights; this could be by manual operation or servo controlled mount option. The integral flexible leveller could either be cam operated synchronising with the conveyor cleats or belt speed, or be servo controlled and operate through a common PLC program in sync with the conveyor system. This option would depend on the type of packet being feed through the HSDS unit and design accordingly.
97
3.5.2
ELECTRICAL DETAIL DESIGN
The HSDS requires detailed electrical design in the following areas: •
sensing and Decision Making,
•
camera.
3.5.2.1 Sensing and Decision Making The sensing and decision making system is crucial to the future prototype unit. It is critical in allowing the leaky or unsealed packets to be discriminated against sealed packets.
The following section explores this area of the system and recommends
suitable methods. As previously discussed, the air inside the packets expands as it is subjected to a lower pressure; this was seen as a change in height/profile. Therefore, the first function of this system was to “determine the height/profile of a chip packet”.
Secondly, “the
system must automatically determine if the packet contains a significant leak by using the height/profile data. Although a ‘significant leak’ was originally defined as ‘a round 5mm hole’, it would be of significant benefit to accommodate for smaller holes. The functional statement was to design a conveyor belt and cleat configuration to travel through the HSDS that creates a physical air seal based on the following parameters: •
the decision-making system requires adjustable parameters for varying levels of accuracy and precision;
•
increased numbers of rejected sealed packets determine cost effectiveness of HSDS;
•
fine tuning of settings to ensure a balance between productivity and quality functions;
•
non-contact system;
•
production operational speeds a maximum 0.33 m/s, and
•
calibration variant for various types of packets.
98
Detailed below are various concepts for the sensing and decision-making system based on diffuse, capacitive, ultrasonic, digital imaging, and parallel-beam sensors.
The
features, advantages and disadvantages of each will be discussed in relation to the prototype HSDS unit. Diffuse Height Detection: This concept featured two triangulation laser sensors, mounted above the conveyer belt to measure the profile of the packet as it passes by. One sensor would be mounted in atmospheric pressure before the first Vacuum Tunnel and the second inside the Vacuum Chamber. The continuous analogue output of these two sensors would be compared using a PLC, and a decision could be determined regarding the change in packet height.
The primary advantage of this concept is the process can be fully
automated using standard components. However, this concept was attempted in the 2001 CEED project (Khan, 2001) and it was found that the sensor was affected by multiple packet colours so much that the output would be unusable in decision making (see Figure 3.43). Wave Form Test No.7 with 5 packets without belt on top 1.20E+01
1.00E+01
Packet height in volt
8.00E+00
6.00E+00
4.00E+00
2.00E+00
2449
2377
2305
2233
2161
2089
2017
1945
1873
1801
1729
1657
1585
1513
1441
1369
1297
1225
1153
1081
937
1009
865
793
721
649
577
505
433
361
289
217
145
1
73
0.00E+00
Reading at 2500 points in 2 seconds
Figure 3.43: Test performed on packets with triangulation sensor The second major disadvantage of this type of sensor was that if the packet were not loaded exactly in the centre of the belt, the laser would not be measuring the highest 99
point of the packet. Therefore, the decision was made to disregard this type of sensor for this application. Capacitive Transducer: The capacitive transducer was to measure the height of the packet as a packet passed by. The transducer would treat the chip package as a ‘parallel plate’ and the subsequent capacitance would be measured according to the displacement.
The
analogue output of the sensor would give a continuous profile of the packet to assess if the packet has puffed up. A decision would be made by comparing the profile at partial vacuum. The major advantage of this concept is the sensor has a relatively large sensing area, which assists in ‘averaging out’ errors due to loading or creases in the packet. For example, the Omron E2K-X15MY1 has an effective diameter of approximately 25mm. In addition, the measurement of displacement was independent of the optical properties of the packets to be assessed. The major disadvantages of this concept are: •
the measurement was dependent of package material and thickness;
•
the contents level of the packet affects the measurement, and
•
issues arise such as, chip lying parallel with the sensor or chips “bunching up”, and the system may interpret the package as “puffed up”.
Ultrasonic: The Ultrasonic concept is similar to the first two concepts discussed above as the analogue proximity probes are used to determine the package height at atmospheric pressure and at a partial vacuum.
The advantage of the concept is that unlike
capacitive transducers, ultrasonic sensors are ideal for detecting thin films (including chip packets), regardless of optical properties.
Discussions with Control Logic
recommended the Pepperl+Fuchs UB500-18GM75-EO1-V15 see Figure 3.44.
100
Figure 3.44: Recommended ultrasonic displacement sensor
Although this sensor was inexpensive, Pepperl+Fuchs indicated that it would be ‘dramatically affected’ by change in surrounding air pressure. Ultrasonic sensors are dependant on the speed of sound through the medium and the speed of sound is related to density, which is dependant on pressure. Therefore, this concept was not considered, as slight vacuum chamber pressure fluctuations would incur unnecessary measurement errors. Parallel Laser: The Parallel Laser concept uses a through beam laser to measure the profile of the packet at atmospheric and at a lowered pressure. As the package passes between the emitter and receiver, some beams of light are blocked and the receiver measures the amount of light that was not blocked see Figure 3.45.
101
Figure 3.45: parallel laser concept The analogue voltage output of the receiver is proportional to the area of interference, and the package height is measured. The major advantages of this type of measurement are: •
the measurement is independent of packet colour as it is in no way retro reflective;
•
minimal measurement errors, if the package is not positioned in the centre of the belt due to loading errors.
This is a significant advantage over previous
concepts; •
fast response time, e.g. the minimum response time of the OMRON Z4LBS30V2 is 0.3ms;
•
the distance between emitter and receiver can typically be up to 300mm, and
•
very accurate measuring control.
There are three main types of decision making tests that could have been used:
maximum height of package in vacuum,
change in packet height profile,
change in cross-sectional area of package.
Digital Imaging: The Digital Imaging concept was recommended by Dr Wageeh Boles from QUT and as discussed in detail in earlier sections, an electrical engineering undergraduate student assisted the project with design concepts. This concept worked by taking a photograph of the packet at atmosphere pressure and a photograph inside the vacuum chamber. The two photographs were then automatically compared using image processing and decision was made to reject/accept package (see Figure 3.46).
102
Figure 3.46: Digital imaging concept The test was based on the change in the number of pixels (i.e. Change in area), containing the packet between atmospheric pressure and vacuum.
The main
advantage of this concept was that it was not restricted to 30mm sensing height (like the parallel-beam sensor). This may reduce manufacturing costs and assist in making the machine more applicable to different industries with widely varying packet sizes. The disadvantage of this concept is that difficulty may be experienced in determining which pixels do contain areas of packet and which contain an image of the background. In addition, focusing issues may arise because packets may not necessarily be in the centre of the belt. To explore this option further an Electrical Engineering student was introduced to this concept as his final year project. Modern Imaging and Vision equipment is been currently developed through a number of advanced imaging company’s, and although the current technology is cost prohibitive, increased demand will easy pricing. Appendix D shows on type of imaging system complete with integral cameras. Future research should compare technologies and cost advantages of each to determine the most suitable equipment for the HSDS production unit.
103
3.5.2.2 Camera The prototype cameras would be industrial type as previously discussed. When undertaking the production unit it may be possible to integrate the complete imaging system with an internal PLC. This would need further research and understand the advantages and disadvantages of this proposal. The Vacuum Chamber camera position has two options, namely, propose to incorporate the entire camera within the Chamber, and verify the camera’s suitability to work in a negative pressure zone, or mount externally and incorporate a “special” transparent area for accurate imaging. Both have positive attributes and would require resolving during prototyping. A third option is to provide a vacuum seal at the Chamber and camera lens interface. SKF do supply standard seal types can normally withstand a pressure differential of the order of 70 kPa, and a seal housing would be introduced to prevent seal slippage. 3.6
CONCLUSION
Mechanical and electrical laboratory testing confirmed the concept and theory of the differential vacuum and imaging process.
All individual mechanical and electrical
components were discussed in detail through the ideas generation process. Further development and analysis of alternative concepts both mechanically and electrically were discussed and shortlisted. Mechanical and electrical detailed design was then completed all component equipment to ensure material and component type selection was suitable for both prototype and production units. Drawings in Appendix A and design calculations in Appendix B detail and highlight the design philosophy. Original equipment manufacturers’ for components were also discussed and introduced throughout the chapter, along with theoretical explanation of the logic and reasoning behind the selection rational.
104
CHAPTER 4 SELECTION AND PROCUREMENT 4.1
INTRODUCTION
Different selection criteria are required between procuring equipment and components for the prototype versus the full production unit. Material selection also varies between applications when comparing a food processing environment to heavy industry and a corrosive fertiliser manufacturing facility. The chapter discusses procurement comparisons between the prototype and production versions of the High-Speed Sensing and Detection System (HSDS). Various costs disciplines are also discussed in this Chapter.
4.2
SELECTION & PROCUREMENT OF EQUIPMENT
In the future, it would be envisaged that both the prototype and production version of the HSDS would be utilise standard industry available components, manufactured and distributed locally. There can be reasonable cost advantages in procuring locally manufactured components. This would generally allow quick access to spare parts and also improve the delivery lead-time during production.
105
4.2.1
PROTOTYPE UNIT
Design Drawings in Appendix A show the prototype version of the High-Speed Sensing and Detection System (HSDS).
Vacuum Chamber
Cleat Conveyor Belt
Regenerative Circuit Pipes
Reject System
Industrial Digital Cameras
Figure 4.1 Proposed final design of prototype HSDS
During the future prototype manufacture, it would be envisaged to utlilise standard industry available and/or secondhand components to ensure cost effectiveness. The prototype would enable the researcher to complete exhaustive trials on a variety of food packets and gather essential statistical data. The manufacture of the prototype, by operation, would confirm any inherent design or manufacture difficiencies within the system, and will highlight any reliability and maintenance issues before trialling within a “live” food processing environment. Procurement time and cost will depend on availability of either purchasing a good quality second-hand conveyor system versus a complete new manufacture.
106
Procurement lead-times can be reduced by purchasing various components concurrently. From previous design information in Chapter 3, certain physical parameters can be refined to determine prototype overall dimensions. The prototype conveyor would be 3000mm roller centres (i.e. between the drive and tail roller). The belt width would be developed @ 300mm, and cleat centres @ 200mm. The height of the prototype unit purchased would be nominally 940mm from ground level to top of belt. This height is within the range of human ergonomics work practices and is typically an industry standard height within the food processing. The conveyor length is currently determined from the following parameters: •
entry - two cleat spacing (400mm) including camera position and manual bag leveller;
•
vacuum Tunnels – three cleat entry and exit spacing (2 x 600mm);
•
vacuum Chamber – four cleat (800mm) including camera position;
•
exit – two cleat spacing (400mm) including Air Knife Reject system; and
•
feed In and Feed Out – 2 x100mm allowance to transfer packet from upstream to downstream of packaging line.
Before “live” trialling, the height would be adjusted to suit the operations requirements. For prototyping purposes, the castors accompanying the second-hand conveyor would suffice for trialling purposes.
Quality checks with the purchaser would be required
before trialling within a food processing environment.
107
4.2.2
PRODUCTION UNIT
Design Drawings in Appendix A show the production version of the High-Speed Sensing and Detection System (HSDS).
Regenerative Circuit Pipes Vacuum Chamber Cleat Conveyor Belt
Industrial Digital Cameras Reject System
Figure 4.2 Proposed final design of production HSDS
From previous design information in Chapter 3, certain physical parameters can also be refined to determine production overall dimensions. The production conveyor would again be 3000mm roller centres (i.e. between the drive and tail roller). The length could be adjusted longer if required to suit the purchaser’s production line requirements. Lengthening would typically occur in 200mm increments. The conveyor length at this stage could not be reduced, until the number of chamber requirements is confirmed through trials during prototyping. Again, the belt width would be developed @ 300mm, and cleat centres @ 200mm; the width and cleat centres can be modified to suit the purchaser’s requirements within limits.
108
The height of the production unit purchased would be nominally 900mm from ground level to top of belt. This height complies with Workplace Health and Safety guidelines for human ergonomics work practices and is typically an industry standard height within the food processing and packaging industry. The production version would have adjustable castor heights through an integrated nut and thread within the conveyor support legs. This option gives full flexibility to fine tune each end based on uneven floors. The castors would also include a lock to restrict lateral and longitudinal movement. It is proposed to use the locally produced Rotarola industrial duty castor (see Figure 4.3) which has been specifically designed for food processing environments by utilising bacteria resistant materials and stainless steel body.
Figure 4.3 Proposed Rotarola food-grade industrial castor Source: (Rotarola technical brochure)
109
4.3
ASSEMBLY AND TESTING
Assembly operational procedures or work instructions will vary between the prototype and production units, and different quality control practices would be implemented. The details below set-out required design, documentation and policies in the assembly and testing process. 4.3.1
MECHANICAL COMPONENTS
The prototype unit would be modified from a second-hand conveyor using local tradespeople to complete. Workshop or Manufacturing Drawings can be produced from the model using the Autodesk Inventor software program. The production unit would be manufactured initially in “made to order” or small batch quantities depending on market demand. To streamline production lead-time various components could be manufactured con-currently at different manufacturers workshops if required. Again, detailed workshop and manufacturing drawings would allow a certain amount of computer controlled manufacturing to be completed by using parts drawings and converting into “dxf” file format. The majority of parts would be manufactured from stainless steel, as this is industry standard within a food processing and packaging environment with direct contact to food surfaces. All non-metal components e.g. polycarbonate, would be food grade type materials. Food grade materials, means that the component or part is not poisonous if accidently consumed.
Again industry standards including food safety national
standards and Good Manufacturing Practices (GMP) dictate this requirement. The prototype and production units would require a Risk Assessment complying with AS4360 to be completed prior to either unit entering a production or trial facility. From a Workplace Health and Safety perspective, the production unit would be submitted to a rigorous inspection by engineering, production and maintenance personnel prior to commissioning and subsequent operations. A thorough Food Safety audit would also be completed prior to becoming operational. Food manufacturers have slightly different policies in relation to installation, commissioning and production start-up, but all are
110
guided by Workplace Health and Safety and Food Safety regulations and legislation. Purchasers normally have specific “Scope of Works” or “Specifications” that are issued during the tender process. Prototype Development and Testing: As a part of future research and development, a prototype unit would be able to verify the concept of vacuum sealing with conveyor cleats before a large investment was incurred into the production unit. The initial testing would allow optimisation of the prototype design by evaluating how many cleats should be used in both the Vacuum Tunnel and Chamber. As previously discussed, the higher the number of cleats develops a seal, the lower the flow rate from leakage and hence lower the operational costs. The prototype would be required to perform the following functions: •
permit calculation of the leakage rates across varying number of cleats, in both the Vacuum Tunnel and Chamber;
•
identifies and highlight major leakage areas by visual inspection;
•
adjustable Belt Tensioning to investigate the effects of varying tensions;
•
trial a manual packet leveller to verify is this rectifies “self sealing” issue; and
•
trial different digital imaging processes.
The HSDS must conform to Workplace Health and Safety regulations and legislation, specifically considering QUT and TSSC environments and working practices.
The
Vacuum testing chamber and Tunnel must be able to withstand an internal pressure of 30KPa (gauge). Incorporated in the Vacuum Tunnel and Chamber prototype unit would be a number of tapping points.
The barbed tailpieces are locked into various positions along the
chamber as shown in Figures 4.4 & 4.5. The Vacuum Pump is connected to one to the tapping points, and a calibrated Vacuum Gauge connected to the other.
111
The flow rate would then be measured through the pipe connecting the chamber to the vacuum pump. This process would be repeated for up to five cleats by unscrewing the barbed tailpieces and swapping them with the brass nuts.
Barbed Tailpiece Steel Vacuum Chamber
Figure 4.4 Test rig configuration for testing the leakage rate over one cleat
112
Brass Nut
Barbed Tailpiece
Figure 4.5 Detailed view of tailpiece configuration for testing leakage over one cleat
Tests completed during this procedure would confirm the following: •
conveyor Cleat rigidity and suitable materials;
•
friction and wear on Cleat materials;
•
optimum number of cleats to create a reliable vacuum pressure drop;
•
adequacy of Regenerative Vacuum Circuit; and
•
leakage areas e.g. through cleats and chamber itself.
4.3.2
ELECTRICAL COMPONENTS
Further statistical data analysis would be gathered and analysed during the testing phase to ensure correct thresholds and ranges are fine tuned. As specified in the ‘Profile’ algorithm, it is necessary to carefully prepare the visual environment of the system in order for it to be effective. The background must be consistent, white of light in colour and non-reflective. The environment must have a dedicated light source and shadows and reflections should be minimised. 113
The PLC program will require developing and continually modified during the later prototyping phase and early full production testing. Testing of imaging speed would also be confirmed during prototyping for verification. For production use, the Panel View screen would visually show the pass/fail analysis of the imaging program; this program would also be incorporated and written to suit the PLC’s brand’s language. The exact camera positions would be confirmed during prototype testing of the unit and the effects of the camera being inside the Vacuum Chamber versus the camera being positioned external to the Vacuum Chamber. Both scenarios have advantages in their positioning; the exposure of a camera can be detrimental, and sealing issues exist when the camera is positioned externally to the Vacuum Chamber. It is envisaged that a sealing device be trialled for the external camera application, or a clear fixed section resist to vacuum pressure. The positioning of the camera inside the Vacuum Chamber may also complicate the chamber design. A light source is also required to capture the passing image accurately. The same logic will also be needed with the positioning of the light either inside or external to the Vacuum Chamber. Modern technology allow for compact, bright lighting which would only provide minimal space constraint issue either within or outside the Vacuum Chamber. Additional testing for both external and packet noise will need verifying by the trialling of a large variety of packet configurations.
External electronics noise from outside
equipment would be analysed and verified for suitability.
114
4.4
COST ANALYSIS
The cost analysis considers the total cost which includes Research and Development, Engineering, Equipment, Installation and Commissioning.
The purchaser will also
consider “whole life costs” which include Product or Productivity, Quality, Operations and ongoing Maintenance. 4.4.1
PRODUCT COST
The Quality cost to business’s producing leaky packets and sending to the consumer is not often directly quantifiable to the food processing business.
As mentioned
previously, there is a potential for loss in business and market share due to contaminated or stale food products. The accuracy and precision of the production HSDS is paramount, as up to 100 packets per minutes may be testing through the machine. This equates to 6000 packets per hour and 240,000 packets per 40 hour shift. A small percentage in reject errors results in a large number of good packets being rejected. Although this is not preferable, the food processing company may accept this reject error, as being more preferential than sending through leaky packets through to the consumer. The selection of operating point should ultimately be based on a large amount of test data and a cost analysis. A discussion of the proposed cost analysis follows. Consider the total cost (TC) for production run of size PRS: TC = P (false alarm) x PRS x $(rejecting good packet) +P (misread) x PRS x $(accepting bad packet) + Q x $(testing) By selecting a ‘high’ threshold, the probability of false alarms (P(false alarm)) will be relatively high and the probability of misreads (P(misread)) will be relatively low.
115
The cost of rejecting a good packet ($(accepting good packet)) will be low. If the packets are simply discarded then the cost per packet will be manufacturing cost of the packet. Small additional costs may arise if these packets are retested. The cost of accepting a bad packet ($(accepting good packet)) will be high. This will include the probabilistic cost of a badly sealed packet becoming contaminated and leading to litigation. It will also contain an amount based on potential product returns and degradation of the customer goodwill. The cost of testing ($(testing)) needs to be economically viable for companies to implement. This will include the initial cost of the system, maintenance costs and any speed implications on the overall production line. It is believed that selection of an operation point described above will reduce the overall cost of the system. The optimum operation point should be derived from a comprehensive cost analysis utilising a large amount of test data. 4.4.2
RESEARCH AND DEVELOPMENT COST
A sometimes hidden cost is the establishment of a new or prototype machine is the research and development cost. Even though the Australian Government has very generous tax concessions for researchers, the design and development does require a significant investment by Government indirectly and/or private/company investors. GRG Engineering Services initially contributed $7,500 to develop a prototype machine. As the research further developed, it become obvious that this amount of money would be insufficient to cover the equipment costs, let alone any engineering labour costs to manufacture. Our company has contributed “in kind” for engineering drafting costs to the value of approximately $10,000; workshop and manufacturing drawings would evaluate to another $10,000 in drafting costs, excluding any cost for engineering and drafting software.
116
Fortunately, for the HSDS project, much of the engineering design and procurement costs have been completed, resulting in significant savings of Engineering and Research and Development consulting costs. 4.4.3
EQUIPMENT COST
Prototype and production equipments costs are detailed in Appendix C with individual Microsoft Excel spreadsheets. The costs include the following items: •
Conveyor Support Structure and Frame: As discussed earlier in the Chapter, the prototype conveyor frame and support structure would be purchased second-hand at an estimated cost of $1,000. This would also include a basic conveyor deck, which is an anticipated requiring modification for prototyping purposes.
•
Castors: Two fixed and two swivel food grade castors are required for the production unit. Costs were procured from a local castor manufacturer who specialises in supplying castors to the food processing industry. Other castor manufacturers supply similar types, but the Rotarola wheel offers the advantage of suitability over a number of food processing industries and are cost effective.
•
Conveyor Belting: Intralox quotation is attached in Appendix C
•
Airlock Tunnels and Vacuum Chamber: Budget Estimate received from local plastics manufacturer
•
Vacuum Pump and Control: Budget estimate received from Busch Pumps and specifications
•
Rejection: Budget estimate received from local supplier
117
•
Digital Imaging Equipment: Budget estimate and technical brochures received from local suppliers
•
Camera: Two industrial cameras are required for the both the prototype and production units.
•
PLC Panel View Screen: For the production unit, the writer would highly recommend Allen Bradley products for PLC and Panel View Screens.
The majority of larger food
processing manufacturers use this brand of equipment and standardising would be favourable. Various companies are agents for Allen Bradley equipment, a budget cost screens Supplier: Various Allen Bradley distributors Nominal Cost: $3,500 + Workshop Installation and Programming •
Installation Costs: With the HSDS unit, direct installation costs are minimal, due to the machines “wheel in wheel out” design through the use of castors. Other installation costs may prevail, if existing equipment requires modification or replacement due to the introduction of the HSDS production unit. Therefore Installation Costs are a variable to each different food processing or packaging facility.
•
Commissioning Costs: The original equipment manufacturer (OEM) of the HSDS unit would complete most of the commissioning at the factory, and also invite the purchaser to complete Factory Acceptance Tests (FAT) prior to deliver and installation. The factory commissioning allows the OEM to complete the majority of testing in the factory with sample food packets.
This initiative reduces onsite costs, and
allows the OEM to work through any small start-up or “teething” issues within a low pressure environment.
118
The unit can then be delivered to site with less production downtime risk and to the purchaser’s satisfaction. Minor commissioning at start-up should be minimal, resulting in the purchasers being able to ramp up to full production quickly and efficiently. Typically after the installation of new equipment the purchaser budgets a certain cost to ramp up production over a designated time period, and any improvement to maximum production time will save the purchaser operational costs. 4.5
CONCLUSION
Selection and procurement of components required for both the prototype and production version of the HSDS were discussed and compared. Assembly and testing processes of both the prototype and production units were explained.
Important functionality check requiring verification during prototype
development and testing were also detailed. Finally, a cost analysis was completed including “whole of life costs”.
The costs
included research and development, engineering, equipment, installation and commissioning, product, quality and operations and maintenance.
119
CHAPTER 5 CONCLUSIONS, RECOMMENDATIONS AND FUTURE WORKS 5.1
INTRODUCTION
Contamination of packaged foods due to micro-organisms entering through air leaks can cause serious public health issues and cost companies large amounts of money due to product recalls, consumer impact and subsequent loss of market share. The main source of contamination is leaks in packaging which allow air, moisture and microorganisms to enter the package. In the food processing and packaging industry worldwide, there is an increasing demand for cost effective state of the art inspection technologies that are capable of reliably detecting leaky seals and delivering products at six-sigma. This project was initiated to extensively develop an original concept of vacuum technology in relation to sealed food products, and the conclusion summarises the entire process accordingly. The recommendation and future works discussions should allow future researchers to further develop the process to at least full prototyping. 5.2
CONCLUSIONS/SUMARY ORIGINAL CONTRIBUTION
There are many current NDT (non-destructive testing) methods of testing the seal of flexible packages best suited to random sampling, and for laboratory purposes. The three most commonly used methods are vacuum/pressure decay, bubble test, and helium leak detection. Although these methods can detect very fine leaks, they are limited by their high processing time and are not viable in a production line. Two nondestructive in-line packaging inspection machines are currently available. The detailed design and development of the High-Speed Sensing and Detection System (HSDS) is the fundamental requirement of this project and the future prototype and production unit. Successful laboratory testing was completed and a methodical design procedure was needed for a successful concept.
120
A literature review and search indicated suitable on-line testing machines were limited and not suitable to test the seals of TSSC potato chip packets. An intense design and development process commenced both mechanically and electrically to provide solid design solutions by introducing the High-Speed Sensing and Detection System unit. The design would inspect food processing packets in ambient static pressure air, and compare the height change which occurs in a Vacuum Chamber. The digital imaging algorithm would make a decision on the seal integrity, and accept or reject the packet. During this design process, laboratory tests were completed in both areas of mechanical and electrical discipline.
Favourable results in both engineering areas
displayed a need to progress the project into detailed design phase. The proposed algorithm has several advantages over mechanical methods of detection. It has the potential to identify shape changes in packets as opposed to the absolute height obtained from the mechanical sensors. In addition there is the ability to conduct more calculations on the data available and store results for quality control purposes. The detailed design procedure analysed concepts for each major component, and logically worked to a well engineered result. Development and analysis of various unit concepts were also completed for both the prototype and production versions of the HSDS unit. Major components were selected and cost analysis completed on both the prototype and production units, along with proposed methods of testing the prototype in both areas of engineering. The costs confirm the cost effectiveness of the design and the ability to be competitive in the packaging equipment market. 5.3
DISCUSSIONS AND LIMITATIONS
The digital imaging system has been shown to be effective under controlled tests. There will be many more challenges to overcome when integrating with the mechanical equipment and Vacuum Chamber and Tunnels. Testing of the final system may produce different results and require review of the algorithm.
121
The system is sensitive to the orientations of the packets, but the packet leveller could be further developed to ensure the packet orientation remains constant to ensure imaging accuracy algorithm. A future working prototype unit would verify mechanical and electrical constraints, but also confirm current design procedures and assumptions. The current design has been significantly improved and advanced for prototyping purposes. Some of the design advances could also be directly utilised in the production unit. 5.4
RECOMMENDATIONS
The potential technology could also be utilised in other food processing areas such as, Diary, Seafood, Smallgoods, Meat and Beverages to confirm if other opportunities exist. All areas of food processing offer many future applications with packaging technology constantly changing. The HSDS concept and future production unit will also need to keep breast of changing technology and advances in packaging. Other non-food processing areas include the aviation industry, where the inclusion of this type of machine would simulate the environment inside an aircraft. All potential dangerous and harmful goods could be put through the conveyor system to check leak detention and prevent potentially dangerous events from occurring. The application should also be considered in other areas of industry where vacuum testing works effectively. A future undergraduate Thesis could be completed on using this technology in aircraft operations and research other possible opportunities in the future. 5.5
FUTURE WORKS
The author is prepared to supervise or guide a future Masters student to complete the prototype machine. One of my companies, GRG Engineering Services Pty Ltd currently has funding available through previous cash deposits in trust.
122
Future works should progress and extensively test the prototype HSDS to ensure all electrical and mechanical components are functioning within their design parameters. All components should be closely monitored for wear and damage, and replace with enhanced or different designed components until all components are deemed reliable. Food processing and packaging company’s can not afford major production downtime during operation, and the extensive up-front trialling will reduce of eliminated the majority of these engineering issues. Mechanically, a primary area that requires specific attention is the operation of the Vacuum Chamber and Tunnels, to ensure that the selection of cleat spacing numbers will provide the desired result. Increasing the number of cleats would lead to a lower pressure change per seal and a lower leakage rate In relation to the algorithm development, if during future prototyping, the algorithm proposed is not sufficiently precise and/or accurate then further processing of the profile function may yield better results. For example, the curvature of the profile could be used to assess shape change. It would also be possible to apply a polygon approximation algorithm to the profile function to obtain more information about the shape of the packet. For this, the ends of the packet can be found from the gradient of the profile function and hence the major axis. In summary, the final system can be applied to a variety of packaged food items where a high level of food quality control and GMP is required. The production unit will reduce the risk of contamination and reduced product life to virtually all packaged food items with similar packaging material.
123
REFERENCES Khan, D.H. 2001. Prototype Integrated Seal Testing Machine, Unpublished Thesis, QUT Brisbane. Mergard, W.J.
2004. Vacuum Testing of Potato Chips Packets for Seal Integrity,
Unpublished Thesis, QUT Brisbane. Newman, N. 2005. Industrial Image Processing Application, Unpublished Thesis, QUT Brisbane. Douglas, J.F., Gasiorek, J.M, & Swaffield, J.A. 2001. Fluid Mechanics, Essex : Pearson Education Limited Hibbler, R.C. 2000. Mechanics of Materials, New Jersey: Prentice Hall. Juds, S.M. 1988. Photoelectric Sensors and Controls, New York: Marcel Dekker Incorporated. Roth, A. 1990. Vacuum Technology, Amsterdam: Elsevier Science Publishers Incorporated. Jos H. J. Huis in't Veld, 1996.
Microbial and biochemical spoilage of foods: an
overview, International Journal of Food Microbiology, Volume 33, Issue 1, November 1996, Pages 1-18 Matsuno R.
1995. The Japanese overview: quality control and its supporting
technology, Food Control, Volume 6, Issue2, 1995, Pages115-120 Sizer, C. 1994. Packaging cost report, Tetra Pack Research Centre, USA, Sept. 28, 1994. Harper C L., Blakistone, B A., Litchfield J. B, Morris S A, 1995. Developments in food packaging integrity testing, Trends in Food Science & Technology, Volume 6, Issue 10, October 1995, Pages 336-340
124
Frazier, C H., Tian, Q; Ozguler, A; Morris, S A.; O'Brien, W D. Jr., 2000. High contrast ultrasound images of defects in food package seals, IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, Volume 47, Issue 3, 2000, Pages 530-539 Penman, D, 1993. “Automated x-ray detection of contaminants in continuous food streams”, in Proc. Of 1st IVCNZ conference. Penman, D. Olsson, D, Beach, D,
1992.
“Automatic x-ray inspection of canned
products for foreign material”, Machine Vision Applications, Proc of SPIE 1823, 1992. Graves, M, Batchelor, B, and Palmer, S. 1994 “3D X-ray inspection of food products”, applications of image processing, Proc SPIE 2298. Bowman, C, 1993.
“Scientific and industrial imaging: machine vision, the eyes of
automation, in Proc. Of 1st IVCNZ conference. Davies, E, Patel, D, Johnstone, A. 1995. “Crucial issues in the design of real-time contaminant detection systems for food products”, real-time imaging, vol. 1, no. 6,1995 Jaggi, S, Willsky A.S., Karl W.C., and Mallat S., “Multiscale geometrical feature extraction and object recognition with wavelets and morphology," in Proceedings of International Conference on Image Processing, vol. 3, pp. 372-375, 1995. Khalil, M. I. and Bayoumi, M.M. “Affine invariant object recognition using dyadic wavelet transform," in Proceedings of Canadian Conference on Electrical and Computer Engineering, vol. 1, pp. 421-425, 2000. Tieng, Q.M. and Boles, W.W. “Object recognition using an affine invariant wavelet representation," in Proceedings of the 1994Second Australian and New Zealand Conference on Intelligent Information Systems pp. 307-311, 1994. Quang, T and Boles, W.W. “Wavelet Based Affine Invariant Representation: A Tool for Recognising Planar Objects in 3D Space”, IEEE Transactions on Pattern Analysis and Machine Intelligence, PAMI, Vol. 19, No. 8, pp. 846-857, August 1997.
125
Cook, J. and Boles, W.W. “Iris Based Human Identification: Enhancing Performance through Pre-processing and Feature Reduction,” Proceedings of the 4th Australasian Workshop on Signal Processing and Applications, WoSPA’02, September 2002. Lang, M.,Guo, Odegard, H.J.E., Burrrus, C.S. and Wells, R.O. “Noise reduction using an undecimated discrete wavelet transform," IEEE Signal Processing Letters, vol. 3, no. 1, pp. 10-12, 1996. Donoho, D.I. “De-noising by soft thresholding," IEEE Transactions on Information Theory, vol. 41, no. 3, pp. 613-627, 1995. Ainsleigh, P.L. and Chui, C.K. “A b-wavelet-based noise-reduction algorithm," IEEE Transactions on Signal Processing, vol. 44, no. 5, pp. 1279-1284, 1996. Whitmal, N.A. Rutledge, J.C. and Cohen, J. “Reducing correlated noise in digital hearing aids," IEEE Engineering in Medicine and Biology Magazine, vol. 15, no. 5, pp. 88-96, 1996. Martinez, C.L., Canovas, X.F. and Chandra, M “Sar interferometric phase noise reduction using wavelet transform," Electronics Letters, vol. 37, no. 10, pp. 649-651, 2001. Standards Australia 1996. Australian Standards: ISO Metric Hexagon Commercial Bolts and Screws (AS1111-1996). Standards Australia 1996. Australian Standards: ISO Metric Hexagon Nuts, including Thin Nuts, Slotted Nuts and Castle Nuts (AS1112-1996). Standards Australia 2004. Australian Standards: Risk Management (AS4360-2004) ISO 3506-1:1997 Mechanical properties of corrosion-resistant stainless steel fasteners Part 1: Bolts, screws and studs
126
APPENDIX A DRAWINGS
127
128
APPENDIX B DESIGN CALCULATIONS
129
130
FASTENING SYSTEM BOLT CALCULATIONS (Section 3.5.1.3) The minimum bolt diameter was dictated by the load capacity of the bolt. The tensile load subjected to the bolts was the pretension load required to prevent the walls from slipping with respect to the chamber top and bottom (Mergard, 2004). To maintain static equilibrium,
∑F
x
= 0 = ∆ Pr essure × Area − 2 friction
Where,
f = µN ∴ Bolt Tension = N =
∆ Pr essure × Height × Length 2µ
Assuming, ∆Pressure = 30,000 N/m2 Height ≈ 0.085m Length = 1m (for the purpose of bolt analysis) µ = 0.15 (worst case scenario if gasket was accidentally not installed)
∴ Bolt Tension per metre = N = 30000x0.085x1/0.3 = 8500 N/m However, this was the total load taken up by the pretension of many bolts. Empirical formulae recommend the spacing of flange bolts is to be less than or equal to 10 times the bolt diameter to maintain seal clamping, but greater than 5 times bolt diameter to ensure standard tools can be used. It was decided to use a bolt spacing of 10 x because the differential pressure was relatively low (only 30 kPa).
131
Assuming a bolt diameter of 3mm, Bolts per metre = 1/10xBolt Major Diameter = 1/0.03 = 33.3 = 34 bolts per metre
∴ Force Per Bolt = 8500/34 = 250N Proof load = Stress Area x σut = 0.00000503m2 x 400 x 106 = 201212N 2012 > 250 … 3mm is adequate (Grade 4.6) The maximum length of the bolt inside the walls was limited to half the height of the wall (41.25mm). The length of engagement was based on the assumption that it is required that the bolt fails before the external thread strips. Therefore, Length of Engagement = 2At/0.5πdpcd x σy(bolt)/σy(external thread) Source: http://www.roymech.co.uk/Useful_Tables/Screws/Thread_Calcs.html Where, At = Tensile Area = 5.03mm dpcd = 2.675mm σy(bolt) = 400 MPa σy(external thread)= σy(UHMWPE) = 27.6 MPa Source: www.matweb.com/search/SpecificMaterial.asp?bassnum=O4009 Length of Engagement = 2x5.03/0.5πx2.675 x 400/27.6 = 35mm Another solution method is to use the Machinery’s Handbook calculation for minimum thread engagement. This was less than 41.25mm and the assumption of 3mm bolt was still valid.
132
REGENERATIVE CIRCUIT CALCULATIONS (Section 3.5.1.5) (Mergard, 2004) Assumptions: Ideal Gas; PV =nRT Fixed Temperature Closed system n1=n2 V1 =Cell Volume = Cleat Height x Belt Width x Cleat Spacing (approximation) =0.075x0.2x0.2= 0.003 m3 P2= 71.3kPa= Vacuum chamber pressure (30kPa below atmospheric pressure of 101.3kPa) Therefore: P1V1=P2V2 Rearranging gives: V2=P1V1 /P2 Where V2= the volume of air introduced into vacuum chamber after expansion. Upon start-up, the vacuum pump is started and the system has the following pressure distribution: •
P1=101.3 kPa
•
P2=101.3 kPa
•
P3=101.3kPa
•
P4= 101.3kPa
•
Chamber Pressure = 71.3kPa (assumed)
First Iteration After the cleats have moved one space, the system has the following pressure distribution: •
P1=101.3 kPa
•
P2=101.3 kPa
•
P3=71.3kPa
•
P4= 101.3kPa
133
When the cleats travel past the regenerative circuit ports, the pressure equalises across all four cells and the pressure can be found as:
Pafter equalisation =
P1v1 + P2 v 2 + P3 v3 + P4 v 4 v1 + v 2 + v3 + v 41
Assuming all volumes are the same, Pafter Equalisation =(P1+P2+P3+P4)/4 Pafter Equalisation =(3x101.3+71.3)/4=93.8kPa Second Iteration After the cleats have moved two spaces, the system has the following pressure distribution: •
P1=101.3 kPa
•
P2=93.8 kPa (from previous iteration)
•
P3=71.3kPa
•
P4= 93.8kPa (from previous iteration)
Hence, Pafter Equalisation =(101.3+93.8+71.3+93.8)/4=90.05 kPa
134
GRG CONSULTING ENGINEERS PTY LTD RECTANGULAR PLATE THEORY CALCULATIONS Fully Supported Conditions
Created by: Checked by:
C. Walker G. Gibson
Project No: Project Title: Client: Designed:
BN72 High-Speed Sensing & Detection System QUT G. Gibson
02/02/2009
User Input
02/03/2009
Calculated Output
Comments: (If any)
Stainless Steel Vacuum Chamber Top Section sbujected to 30 kPa Vacuum Plate Thickness Plate length Plate Width Load on plate
5 800 320 30.00
mm mm mm kPa
Machinery's Handbook 20th Edition page 444 Condition 6 - Recangle Plate with all edges supported at top and bottom and a uniformly distributed load over the surface of the plate W Total Load 30.00 kPa L Long Side 800 mm l Short Side 320 mm mm 2 A Area of Plate 256000 t PL. Thickness 5 mm N/mm 2 E Modulus of Elasticity 2.00E+05 W Total Load 7680.0 N S Maximum Tensile Stress 83.55 MPa d Deflection 1.568 mm Simply Supported Result OK Machinery's Handbook 20th Edition page 444 Condition 7 - Recangle Plate with all edges fixed and a uniformly distributed load over the surface of the plate W Total Load 30.00 kPa L Long Side 800 mm l Short Side 320 mm mm 2 A Area of Plate 256000 t PL. Thickness 5 mm N/mm 2 E Modulus of Elasticity 2.00E+05 W Total Load 7680.0 N S Maximum Tensile Stress 61.28 MPa d Deflection 0.354 mm Fixed Condition Result OK Rectangle Flat Plates - Uniform - Simply Supported AISC 5th Edition - Recangle Plate with all edges simply supported and uniformly distributed load q Unifrom Load 30.00 kPa a Long Side 800 mm b Short Side 320 mm t PL. Thickness 5 mm a/b Length Ratios 2.50 β Stress Constant 0.662 α Deflection Constant 0.123 N/mm 2 E Modulus of Elasticity 2.00E+05 fb Maximum Tensile Stress 81.29 MPa d Deflection 1.541 mm Simply Supported Result OK Rectangle Flat Plates - Uniform - Fixed Supports AISC 5th Edition - Recangle Plate with all edges fixed and uniformly distributed load q Unifrom Load 30.00 kPa a Long Side 800 mm b Short Side 320 mm t PL. Thickness 5 mm a/b Length Ratios 2.50 β Stress Constant 0.4980 α Deflection Constant 0.0281 N/mm 2 E Modulus of Elasticity 2.00E+05 fb Maximum Tensile Stress 61.19 MPa d Deflection 0.353 mm Fixed Condition Result OK Allowable Load Fy 200 Factor 0.75 Maximum Safe Load 150
MPa MPa
AISC 5th Edition page 213
Minimum Thread Engagement Calculations Thickness of nut required for bolt failure.
ISO Metric profile
External (bolt thread)
Size mm
Thread Designation
Simple Thread Designation
3
M3x0.5
M3
Pitch mm 0.5
0.11811
2
Major Dia
Pitch Dia
Minor Dia
d=D
d2=D2
d3
Class
max.
6g
2.98
6g
min.
max. 2.874
0.117323
2.655
min. 2.58
Basic mm Minor Dia
Pitch Dia
Major Dia
D1
d2=D2
d=D
min.
6H
2.459
max.
min. 2.599
min.
max.
2.675
2.775
0.096811 0.102323 0.105315 0.109252
3
max.
Tap Drill 3.172
2.5
0.11811 0.125 0.0984
From Machinery's Handbook, 28th Edition. p.1443 Le = 2 x At / [Π x Knmax x (½ + 0.57735n x (Es min ‐ Kn max))] where: Le =
Engagement Length (inches)
Tensile Stress Area of the bolt (inches2) At = n= Threads per Inch Knmax = Maximum Minor diameter of internal thread Esmin =
Minimum pitch diameter of external thread
Therefore :‐ Le = Le =
[2 x (Π x ((3/2)/25.4)2)] / [Π x (2.599/25.4) x (½ + 0.57735 x 2 x ((2.58 ‐ 2.599 / 25.4)))] 0.136569 inches 3.4689 mm
min.
2.439
2.272
0.11315 0.105 0.1016 0.096 0.089449
Internal (nut thread)
Class
max.
Page ________ of ________
GRG CONSULTING ENGINEERS PTY LTD FEA Design Report
Created by: Checked by: Version: Revised by:
G. Gibson
Project No: Project Title: Client: Designed:
BN72 High-Speed Sensing & Detection System QUT G. Gibson
31/03/2009
Comments: (If any)
Conveyor Support Top is fastened directly to Conveyor Structural Support frame. FEA Analysis assumed a concentrated load over support areas.
FEA Design Analysis Settings 25 TET 3 NX Nastran Static Design Load Analysis N FEMAP 9.3.1
Mesh Size Mesh Type Minimum Elements Analysis Program Analysis Type Analysis Name Evidence of Convergence Software Package Loads Surface Load Self Weight Connections
-30 kPa Y
Contact
Nil
Glued
Nil
Boundary Conditions Constraints Load
Conveyor Structural Support Frame Vacuum Pressure
Filename: FEA Design Report - Production Conveyor Support Frame.xls
1 of 1
Page ________ of ________
GRG CONSULTING ENGINEERS PTY LTD FEA Design Report
Created by: Checked by: Version: Revised by:
G. Gibson
Project No: Project Title: Client: Designed:
BN72 High-Speed Sensing & Detection System QUT G. Gibson
31/03/2009
Comments: (If any)
Vacuum Chamber Top Section is fastened directly to Main Vacuum Chamber Body Support. FEA Analysis assumed a distributed load over entire bottom flange
FEA Design Analysis Settings 25 TET 3 NX Nastran Static Design Load Analysis N FEMAP 9.3.1
Mesh Size Mesh Type Minimum Elements Analysis Program Analysis Type Analysis Name Evidence of Convergence Software Package Loads Surface Load Self Weight Connections
-30 kPa Y
Contact
Nil
Glued
Nil
Boundary Conditions Constraints Load
Vacuum Chamber Support Vacuum Pressure
Filename: FEA Design Report - Production Vacuum Chamber Top Section.xls
1 of 1
Page ________ of ________
GRG CONSULTING ENGINEERS PTY LTD FEA Design Report
Created by: Checked by: Version: Revised by:
G. Gibson
Project No: Project Title: Client: Designed:
BN72 High-Speed Sensing & Detection System QUT G. Gibson
31/03/2009
Comments: (If any)
Vacuum Chamber is fastened directly to conveyor top and support frame. FEA Analysis assumed a distributed load over entire bottom flange
FEA Design Analysis Settings 25 TET 3 NX Nastran Static Design Load Analysis N FEMAP 9.3.1
Mesh Size Mesh Type Minimum Elements Analysis Program Analysis Type Analysis Name Evidence of Convergence Software Package Loads Surface Load Self Weight Connections
-30 kPa Y
Contact
Nil
Glued
Nil
Boundary Conditions Constraints Load
Conveyor Top and Support Frame Vacuum Pressure
Filename: FEA Design Report - Prototype Vacuum Chamber.xls
1 of 1
APPENDIX C COST OF EQUIPMENT
135
BN72 - High Speed Sensing & Detection System - Prototype Machine BUDGET ESTIMATE - 30/03/09 ITEM SUPPLIER 1. ENGINEERING INVESTIGATION & DESIGN (incl. Drafting) 1.1 Mechanical Design and Development 1.2 Design and Development Drawings
GRG Engineers GRG Engineers
Prepared by: G. Gibson REF: DWG No. BN72-001
UNIT PRICE
QTY
LINE PRICE
COMMENTS
$10,000 $10,000
1 1
$10,000 $10,000
Not Chargable Not Chargable
sub-total
$20,000
2. CONVEYOR EQUIPMENT 2.1 Support Frame and Extension 2.2 Remacleat cleat 2.3 Remacleat Seal 2.4 Remacleat Bolt Plate 2.5 Conveyor Belt - Endless PVC 2.6 UHMW Vacuum Chamber 2.7 Manual Bag Leveller
Metal Fabricator Rema Tip Top Rema Tip Top Rema Tip Top Rydell Plastic Fabricator Metal Fabricator
$2,000 $20 $10 $25 $300 $1,200 $500
1 32 64 64 1 1 1
sub-total
$2,000 $640 $640 $1,600 $300 $1,200 $500
Ford Engineering - Estimate Rema Tip Top Rema Tip Top Rema Tip Top Rydell - Estimate All Type Plastic - Estimate Ford Engineering - Estimate
$6,880
3. MECHANICAL & ELECTRICAL EQUIPMENT 3.1 Vacuum Pump 3.2 Electronic Test Box & Wiring 3.3 Electrical Contol Box & Wiring 3.4 Industrial Camera & Control 3.5 Air Reject & Solenoid Valve 3.6 Gaskets 3.7 General Assembly 3.8 Vacuum Pipework 3.9 Miscellaneous Items & Brackets
Electrician Electrician Electrician Metal Fabricator Metal Fabricator Metal Fabricator Metal Fabricator Metal Fabricator
$800 $500 $500 $10,000 $500 $100 $500 $1,000 $1,000
sub-total
1 1 1 1 1 2 1 1 1
$800 $500 $500 $10,000 $500 $200 $500 $1,000 $1,000
Busch Vacuum Pumps & Systems Glec - Estimate Glec - Estimate Adept Electronic Solutions - Estimate Glec - Estimate Estimate Estimate Ford Engineering - Estimate Estimate
$15,000
SUMMARY 1. Design 2. Conveyor Equipment 3. Mechanical & Electrical Equipment
$20,000 $6,880 $15,000
TOTAL
$41,880
EXCLUSIONS A. Existing conveyor frame, castors and drives to be used.
Cost Estimate Sheet Prototype Machine 090330.xls
1
3/07/2009
BN72 - High Speed Sensing & Detection System - Production Machine BUDGET ESTIMATE - 30/03/09 ITEM SUPPLIER 1. ENGINEERING INVESTIGATION & DESIGN (incl. Drafting) 1.1 Mechanical Design and Development 1.2 Workshop and Manufacturing Drawings
GRG Engineers GRG Engineers
Prepared by: G. Gibson REF: DWG No. BN72-100
UNIT PRICE
QTY
LINE PRICE
COMMENTS
$10,000 $10,000
1 1
$10,000 $10,000
Not Chargable Not Chargable
sub-total
$20,000
2. CONVEYOR EQUIPMENT 2.1 New Conveyor Top & Support Structure 2.2 Intralox Cleated Belt 2.3 UHMWPE Vacuum Chamber 2.4 Pneumatic Bag Leveller
Metal Fabricator Intralox Plastic Fabricator Metal Fabricator
$5,000 $2,300 $5,000 $2,000
1 1 1 1
sub-total
$5,000 $2,300 $5,000 $2,000
Ford Engineering - Estimate Quotation All Type Plastic - Estimate Ford Engineering - Estimate
$14,300
3. MECHANICAL & ELECTRICAL EQUIPMENT 3.1 Vacuum Pump 3.2 Electronic Test Box & Wiring 3.3 Electrical Contol Box & Wiring 3.4 Industrial Camera & Control 3.5 PLC/Panel View Screen 3.6 Air Reject & Solenoid Valve 3.7 Gaskets 3.8 General Assembly 3.9 Vacuum Pipework 3.10 Miscellaneous Items & Brackets
Busch Electrician Electrician Electrician Electrician Electrician Metal Fabricator Metal Fabricator Metal Fabricator Metal Fabricator
$800 $500 $500 $10,000 $3,500 $500 $100 $500 $1,000 $1,000
1 1 1 1 1 1 2 1 1 1
sub-total
$800 $500 $500 $10,000 $3,500 $500 $200 $500 $1,000 $1,000
Quotation Gelec - Estimate Gelec - Estimate Adept Electronic Solutions - Estimate Allen Bradley Distributor Gelec - Estimate Estimate Estimate Ford Engineering - Estimate Estimate
$18,500
4. ASSEMBLY, INSTALLATION AND COMMISSIONING 4.1 Assembly 4.2 Installation 4.3 Commissioning
OEM OEM OEM
$5,000 $2,000 $5,000
sub-total
1 1 1
$5,000 $2,000 $5,000
Estimate Estimate Estimate
$12,000
SUMMARY 1. Design 2. Conveyor Equipment 3. Mechanical & Electrical Equipment 4. Assembly, Installation and Commissioning
$20,000 $14,300 $18,500 $12,000
TOTAL
$64,800
EXCLUSIONS A. Maintenance Contract
Cost Estimate Sheet Production Machine 090330.xls
1
3/07/2009
QUOTATION #: DU1879
INTRALOX AUSTRALIA PTY. LTD. ALL-PLASTIC CONVEYOR BELTING SUBSIDIARY OF LAITRAM, L.L.C. P.O BOX 155 SOMERTON MELBOURNE, VICTORIA AUSTRALIA 3062 Toll Free Australia: Tel: 1800-128 742 Fax: 1800-120705 Toll Free New Zealand: Tel: 0800-449 173 Fax: 0800-449 446 ABN: 59-077-781-992
PAGE # 1
Original Version
Quote Date:
Sold-To: Account # 830144 GRG CONSULTING ENGINEERS PTY. LTD. P.O. BOX 1456 SUNNYBANK HILLS QUEENSLAND AUSTRALIA 4109
31/3/09
Prices are valid thru: 30/4/09 Prices are valid for 30 days
Phone #: 01161732484151 Fax #: 01161732484187 Contact: ANDREW HARTLEY Bill-To: GRG CONSULTING ENGINEERS PTY. LTD. P.O. BOX 1456 SUNNYBANK HILLS QUEENSLAND AUSTRALIA 4109 -
Prices include ground delivery within Australia and New Zealand.
Ship-To: GRG CONSULTING ENGINEERS PTY. LTD. P.O. BOX 1456 SUNNYBANK HILLS QUEENSLAND AUSTRALIA 4109
Account Representative: MINDY HELMER Quotation Entered By: Wesley Hoppe Line Seq
Intralox Part # Quantity
Description
[ Customer Part # ]
Application: 1
1
2
6
S3D8 G5CPK1 NG
Net Unit
Extended
Price
Net Price
SNACK FOODS
BELT: Series 800 Flush Edge Flat Top Polypropylene White FLIGHTS: 3"/76mm Streamline SPACING: 203 MM RODS: Natural Polypropylene LENGTH: 7 M ( For A Total Of 2.14 Square Metres) WIDTH: 306 MM +/- 1 (18 Links) Lead Time: 10 Working Day(s)
2,000.00
2,000.00
SERIES 800 5.2"/132 MM (8T) NATURAL ACETAL EZ CLEAN SPROCKET WITH 40 MM SQUARE BORE Lead Time: Stock
47.06
282.36
PLEASE NOTE: Mould-to-order (MTO) products, belts with non-standard configurations, and all specially-machined items are not returnable. This transaction reflects Volume Based Pricing. Prices are subject to change if the overall footage or quantity of a belt line item changes. Some belt widths may not be returnable. Please contact your Customer Service Representative for more information.
Subtotal: BELT WIDTH TOLERANCE AT 70 DEGREE F
GST Amount: TOTAL -----> AUD
TAX EXEMPT NUMBER IS REQUIRED UPON PLACEMENT OF THE ORDER. IF TECHNICAL RECOMMENDATIONS ARE REQUIRED, CONTACT OUR SALES ENGINEERING DEPARTMENT. Intralox, Australia's General Terms and Conditions of Sale are provided on the following page. In the event you did not receive these terms and conditions, please call Intralox for a copy.
2,282.36 228.24 2,510.60
Panel View Component When you need an essential component, with added value and at a reduced cost, look to the new Allen Bradley Panel View component family of operator interfaces. Enjoy the convenience and effectiveness of the Panel View component which features built in programming software and integrated mounting clamps to improve productivity and maintenance. The Panel View component family are easy to install and simple to learn and operate. Offering a full line of displays form 2¶¶ to 10¶¶ they are suitable for integration with Allen-Bradley Micro Logix and SLC 500 Families.
Features of the Panel View component:
Built in programming software
Secured programming access
Serial and Ethernet communications
Optimized for Micro Logix based Systems
Unicode language switching
Alarm messages and history
Pricing as at 01 July 2008 2711CF2M
Panel View C200 2¶¶ Monochrome 4 Key Touch
$291.20
2711CT3M
Panel View C300 3¶¶ Monochrome Touch
$418.40
2711CT6C
Panel View C600 6¶¶ Colour Touch
$1104.00
2711CT10C
Panel View C1000 10¶¶ Colour Touch
$3152.00
All prices quoted are subject to change and are excluding GST.
For information on any of our products please contact us: Telephone: (08) 8733 0600 Fax: (08) 8733 0606 Email:
[email protected]
Web: www.EncompassAutomation.biz
Product Specifications Bulletin
2711C-F2M 2711C-K2M
2711C-T3M 2711C-K3M
2711C-T6M 2711C-T6C
2711C-T10C
Type
PanelView C200
PanelView C300
PanelView C600
PanelView C1000
DISPLAY Size
2¶¶ (49x14mm)
3¶¶ (67x33mm)
5.7¶¶ (115x86mm)
10.4¶¶ (211x158mm)
Resolution
122x32 pixels
128x64 pixels
320x240 pixels
640x480 pixels
Type
Monochrome Trans- Monochrome Transflective STN flective FSTN
Operator Input
4 Function key or Combination Function/Numeric Key
Analog Touch or Combination Function/Numeric Key
Monochrome Transmissive FSTN or Colour Transmissive CSTN
Colour Transmissive TFT
Analog Touch
Analog Touch
ELECTRICAL Communication Port
RS232 (9 pin D-Shell) RS422/RS485 Connector
RS232 (9 pin D-Shell) RS422/RS485 (connector) Ethernet
Programming Port USB DeviceNet Port, Ethernet also support programming Real Time Clock
No Battery Backup
Battery-backed
Power Requirements
18...30V DC @ 24V DC, C200 and C300 5W, C600 10W, C1000 18W ENVIRONMENTAL
Operating Temperature
0...50°C
Enclosure
NEMA/UL Type 4X (indoor) 12, 13 and IEC NEMA/UL Type 12, IP54, IP65 13 and IECIP54
Certifications
cULus Listed, CE marked, C-Tick
NEMA/UL Type 4X (indoor) 12,13 and IEC IP54, IP65
MECHANICAL 4-Key or Touch Numeric KeyPad
0.2kg 0.3kg
0.2kg 0.3kg
0.7kg
1.6kg
Dimensions Overall (HxWxD)
80x116x54mm 119x139x55mm
80x116x57mm 119x139x55mm
135x189mm
231x289mm
Cutout Dimensions
64x99mm 99x119mm
64x99mm 99x119mm
135x189mm
231x289mm
APPENDIX D LABORATORY TESTING DATA
136
MECHANICAL TEST DATA (Section 3.1.1) Without Leak Vacuum Level
With 5mm Leak Atmospheric Lowered Change
Atmospheric Lowered in
[kPa]
Pressure Pressure Height
(gauge)
Height at
Height at Height at Change Height at
[mm]
[mm]
[mm]
Pressure
Pressure in Height
[mm]
[mm]
[mm]
10
34.57
37.38
2.81
33.80
34.00
0.20
12
40.44
45.99
5.55
34.00
34.24
0.24
14
40.60
46.00
5.40
37.00
41.23
4.23
16
37.90
47.74
9.84
37.00
37.60
0.60
18
37.30
48.73
11.43
36.50
36.50
0.00
20
36.80
47.40
10.60
36.20
36.20
0.00
22
36.10
49.50
13.40
38.00
38.80
0.80
24
37.30
54.40
17.10
41.30
48.60
7.30
26
35.99
65.00
29.01
36.60
36.60
0.00
28
38.00
62.00
24.00
37.00
37.03
0.03
30
34.80
60.50
25.70
34.00
34.00
0.00
32
40.37
74.00
33.63
39.00
39.60
0.60
6.04
36.62
30.82
7.50
14.60
7.30
2.08
10.25
10.13
2.17
4.05
2.26
Range Standard Deviation
Table 3.1 Heights of Twisties Packets at Various Vacuum Levels
137
Without Leak Height at
With 5mm Leak Height at
Height at Change Height at
Vacuum Atmospheric Lowered in
Atmospheric Lowered Change
Level
Pressure
Pressure Height
Pressure
Pressure in Height
[kPa]
[mm]
[mm]
[mm]
[mm]
[mm]
[mm]
10
33.17
38.90
5.73
33.00
34.14
1.14
12
31.04
35.60
4.56
37.30
33.80
-3.50
14
36.50
44.30
7.80
37.42
39.60
2.18
16
35.50
43.38
7.88
37.00
39.80
2.80
18
33.70
41.86
8.16
34.00
34.00
0.00
20
34.60
47.20
12.60
34.50
34.50
0.00
22
35.14
44.58
9.44
35.56
42.60
7.04
24
33.36
48.90
15.54
33.40
45.00
11.60
26
34.00
50.60
16.60
33.80
34.00
0.20
28
38.00
56.30
18.30
36.30
35.70
-0.60
30
32.00
58.00
26.00
33.00
32.00
-1.00
32
35.40
60.30
24.90
33.40
33.40
0.00
6.96
24.70
21.44
4.42
13.00
15.10
1.92
7.67
7.21
1.73
4.15
4.03
Range Standard Deviation
Table 3.2 Heights of Doritos Packets at Various Vacuum Levels
138
ELECTRICAL TEST DATA (Section 3.1.2)
Test # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
Packet # 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4
Actual Condition good good good good good good good good good good bad bad bad bad bad bad bad bad bad bad good good good good good good good good good good bad bad bad bad bad bad bad
Reference Height 235 260 244 257 309 318 279 292 286 228 244 298 225 297 271 225 281 279 244 274 287 266 222 243 311 291 246 235 270 290 276 233 272 237 247 282 257
Vacuum Height 305 311 310 301 395 343 384 316 305 276 247 312 224 307 275 232 282 283 252 280 324 303 257 272 449 333 360 314 400 316 286 235 279 239 251 294 261
139
Height Change 70 51 66 44 86 25 105 24 19 48 3 14 -1 10 4 7 1 4 8 6 37 37 35 29 138 42 14 79 130 71 10 2 7 2 4 12 4
% Change 29.79% 19.62% 27.05% 17.12% 27.83% 7.86% 37.63% 8.22% 6.64% 27.05% 1.23% 4.70% -0.44% 3.37% 1.48% 3.11% 0.36% 1.43% 3.28% 2.19% 12.89% 13.91% 15.77% 11.93% 44.37% 14.43% 5.69% 33.62% 48.15% 24.48% 3.62% 0.86% 2.57% 0.84% 1.62% 4.26% 1.56%
Inferred Correct Condition Condition good yes good yes good yes good yes good yes good yes good yes good yes good yes good yes bad yes bad yes bad yes bad yes bad yes bad yes bad yes bad yes bad yes bad yes good yes good yes good yes good yes good yes good yes good yes good yes good yes good yes bad yes bad yes bad yes bad yes bad yes bad yes bad yes
38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78
4 4 4 5 5 5 5 5 5 5 5 5 5 6 6 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 8 8 8 8 8 8 8 8
bad bad bad good good good good good good good good good good bad bad bad bad bad bad bad bad bad bad good good good good good good good good good good bad bad bad bad bad bad bad bad
318 304 264 313 251 296 312 303 256 221 306 232 317 276 221 283 299 281 232 316 263 304 299 275 294 236 303 243 307 280 263 316 319 276 288 273 293 293 250 287 221
333 309 267 465 339 387 409 391 346 272 391 342 326 285 223 280 310 283 234 316 274 307 300 291 316 333 314 268 339 289 477 363 285 210 290 285 303 300 256 297 225
15 5 3 152 88 91 97 88 90 91 85 110 9 9 2 -3 11 2 2 0 11 3 1 16 22 97 11 25 32 79 26 161 4 9 2 12 10 7 6 10 4
140
2.72% 1.64% 1.14% 48.56% 35.06% 30.74% 31.07% 29.04% 35.16% 23.08% 27.78% 47.41% 2.84% 3.26% 0.70% -1.06% 3.68% 0.71% 0.86% 0.00% 4.18% 0.99% 0.33% 5.82% 7.48% 41.10% 3.63% 7.27% 10.42% 28.21% 9.89% 50.95% 13.79% 3.26% 0.69% 4.40% 3.41% 2.39% 2.40% 3.48% 1.81%
bad bad bad good good good good good good good good good bad bad bad bad bad bad bad bad bad bad bad good good good bad good good good good good good bad bad bad bad bad bad bad bad
yes yes yes yes yes yes yes yes yes yes yes yes no yes yes yes yes yes yes yes yes yes yes yes yes yes no yes yes yes yes yes yes yes yes yes yes yes yes yes yes
79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100
8 8 9 9 9 9 9 9 9 9 9 9 10 10 10 10 10 10 10 10 10 10
bad bad good good good good good good good good good good bad bad bad bad bad bad bad bad bad bad
225 314 230 295 302 294 269 295 291 302 224 298 223 318 314 301 318 261 271 230 313 273
228 323 241 375 384 350 334 357 405 423 290 387 223 317 315 309 317 265 275 240 327 277
3 9 11 80 82 56 65 62 114 121 66 84 0 -1 1 8 -1 4 4 10 14 4
141
1.33% 2.87% 4.78% 27.12% 27.15% 19.05% 24.16% 21.02% 39.18% 40.07% 29.46% 28.19% 0.00% -0.31% 0.32% 2.66% -0.31% 1.53% 1.48% 4.35% 4.47% 1.47%
bad bad bad good good good good good good good good good bad bad bad bad bad bad bad bad bad bad
yes yes no yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes
142
APPENDIX E EQUIPMENT BROCHURES
143
industrial
POSITION | IDENTIFICATION | VERIFICATION | MEASUREMENT | FLAW DETECTION
vision solutions
PHARMACEUTICAL LABEL WITH RSS CODE
CONSUMER PRODUCT WITH UPC CODE
2D CODE PRINTED ON PHARMACEUTICAL VIAL
2D CODE PRINTED ON UNDERSIDE OF PLASTIC BOTTLE
04 Industrial Vision Solutions
DALSA’s barcode product supports UPC, EAN, Code 39, Code 93, Code 128, Codabar, Interleaved 2 of 5, Pharmacode, BC412, Postnet, Planet, OneCode, RSS14 (Limited, Composite, Expanded)
1D barcodes are commonly used on products for traceability and sorting. Machine vision verifies that the barcode matches the product that it is printed on.
2D Matrix codes are widely used across many industries for part traceability and process control. The codes are popular for their small footprint, built-in error correction and large data capacity.
DALSA 2D matrix algorithms provide decoding and grading of ECC 000, 050, 080, 100, 140 and 200, QR. MicroQR and PDF 417 matrix codes.
1D Barcode Readers
2D Matrix Code Readers
environments.
CHARACTERS PRINTED ON CAST METAL PART
PRINTED DATE AND LOT CODE ON PRODUCT LABEL
OCR is based on pattern matching and so can be applied to a diverse range of verification applications outside of character reading. Often manufacturers will use OCR to build a library of parts that can later be identified and sorted.
DALSA products include trainable Object Character Recognition (OCR) tools that can handle the variation and diversity of most printing methods in use today.
Character or symbol recognition is common in many manufacturing or production environments.
Date codes and lot codes printed on products provide critical expiration and traceability information. Products with unreadable codes become defective as consumers cannot verify product quality.
Character and Object Readers
are designed for accurate results in the toughest of manufacturing
of product lots or grading of print codes, DALSA’s Identification tools
symbols on products. For traceability of production parts, verification
that involve reading printed characters and decoding 1D or 2D
Identification encompasses a range of machine vision applications
02. identification
LASER ETCHED 2D CODE ON METAL
BACKGROUND INTERFERENCE
CIRCULAR PRINT
POOR CONTRAST
DALSA OCR tools can read a variety of printed characters and symbols under equally challenging conditions. New font variations can be quickly trained and saved to a pattern data base. Similarity scores are provided for the character verification process to indicate match quality.
DOT PEEN 2D CODES ON PLASTIC
Detecting Print Variation
DALSA meets this challenge by providing robust identification tools that can handle the wide variation in print appearance and part position. DALSA tools also provide grading of printed codes that allows manufacturers to detect and correct deteriorating print quality.
Direct part marking of data matrix codes present many challenges for industrial identification. With a range of printing methods available, from direct etching and stamping to laser scribing and peening, direct part marking on metal, plastic and other materials offer manufacturers extensive printing flexibility together with variation in print quality.
Engineered for Industry
• Work in process inventory management – verify parts as they navigate through the fabrication process • Cradle to grave part traceability • Product verification – assure 1D or 2D code matches printed text • Product identification and sorting • Date and lot code verification • Code Verification. Detect problems with the marking system for preventive maintenance.
Identification Applications
case study
www.goipd.com 05
The application requires operator visualization for each inspected container.
Containers are indexed between pockets of a conveyor, three pieces at a time across the field of view of three 640 x 240 imagers connected to a single VA40. After moving into the field of view, the products are held and spun to ensure that an image of the 2D code is located and captured by the system. The cameras capture 25 images of each spinning container for the VA40 to process within a 150 ms allotted inspection window.
The company produces 450 containers a minute, in any orientation.
A Pharmaceutical company needs to locate and read 2D product codes on round plastic containers after packaging. The code contains a product description, batch lot, and expiry date that must be verified for each product.
HIGH SPEED 2D READING
02
06 Industrial Vision Solutions
NON-SYMMETRICAL
Teeth Verification on Gear
SYMMETRICAL
Defects found at part assembly are easier and much less expensive to fix than in the finished product. For example, a vision system prevents these two similar parts from being interchanged.
Pop-top can lids are checked to verify that they are ‘top side up’ and have the pull ring in place before they are joined to beverage cans.
The low contrast of this image might make for a difficult inspection, but DALSA’s geometric pattern tools are easily able to distinguish the pull ring from the background.
Part Verification
Aluminum Lid Verification
barcode, to render 100% product inspection.
VERIFICATION OF PACKAGE SEAL
VERIFICATION OF CORN KERNEL GRADE
Often, presence of product is detected by color as the position and extent of component foods vary too much to be reliably measured.
Machine Vision is used by the food industry to verify product content as well as processing and packaging.
Food Verification
tasks, such as measurement of part dimensions or reading of product
identification and flaw detection. Verification is often combined with other
generally so broad, they utilize the same tools for positioning, measurement,
assemblies and packaged goods. The range of verification applications are
Machine vision systems are widely used for the verification of parts,
03. verification
Ensuring that a correct type and quantity of aerator heads are correctly packed into this crate would be much more challenging without color verification tools.
Package Verification of Water Aerators
• Search and match tools to find parts and verify assemblies • Edge, corner, line, circle and line segment detection tools to find part “features” • Blob analysis tools for counting and dimensioning areas of similar color or contrast on the part • Counting tools to determine number of parts and indicate missing parts • Color tools to measure amount and location of colored elements such as automotive fuses, wire, foodstuffs, and pills • Measuring tools for further qualifying parts and assemblies
Software Capabilities
Color tools are often used to detect the presence and order of parts on an assembly, such as the blue and red plastic components on this medical instrument.
Assembly Verification Using Color
Verification has many uses in the production and packaging of products, and in automotive, electronics, pharmaceutical and medical manufacturing.
DALSA’s Vision Appliances are easy to set-up and simple to train. In the case of verification the primary concern is with presence and correctness of assemblies and parts. A trained machine vision system will evaluate a number of characteristics such as brightness, shape, dimension, orientation and color to achieve reliable inspection results.
Easy Set-Up and Trainability
• Blister pack verification • Molded part verification • Solder joint verification • Bottle cap and safety seal verification • Print verification • PCB assembly verification • Cable wiring verification • Package verification • Feature (thread, hole, notch) verification
Verification Applications
case study
www.goipd.com 07
Assembly Verification
Screw Presence Verification
Glue Bead Verification
Wheel Verification
Automotive manufacturers require verification checks at every stage of the production process. Machine vision verification tools provide reliable automation of mundane repetitive tasks:
AUTOMOTIVE
03
Using DALSA measurement solutions, production quality can be monitored at any stage in the body shop. Results can be sent to the factory enterprise and documented for step-by-step quality control.
The Automotive industry has many applications that require online and offline measuring systems.
08 Industrial Vision Solutions
CONNECTOR INSPECTION
STAPLE INSPECTION
3D PROFILE MEASUREMENT
FILL LEVEL MEASUREMENT
GLAND INSPECTION
Productivity Improvements for a Multitude of Applications
BEAD INSPECTION
Gauging for Quality Control
For general manufacturing needs, machine vision measurement provides a fast, highly accurate and cost-effective way to assure product quality and customer satisfaction
Manufacturers of medical instruments measure each part of the assembly process to strict tolerances. An incorrectly manufactured part could have dire consequences.
Critical Thresholds for Medical Imaging
provide the precision and repeatability to ensure manufacturing accuracy.
measurement tools, combined with the right optics and stable lighting,
is as important as the vision algorithms themselves. DALSA sub-pixel
tolerances. Attention to the inspection environment and image quality
verification to checking high-precision dimensional accuracy and geometrical
Manufacturing requirements for measurement range from presence
04. measurement
• Bead tool to measure thickness and uniformity of adhesive beads or similar applications
• Laser tools for measuring height on parts determined by angle of projected laser lines
• Math tools to create custom measurements that span multiple cameras
• Caliper tools to measure between edge points
• Geometric fitting tools to fit lines, angles, arcs and circles to edge points
• Shape finding tools to locate distinct shapes like corners on parts
• Edge finding tools to accurately find edge transitions on parts for gauging
• Preprocessing tools to manipulate or enhance the camera image to highlight features to measure
• Calibration tools to remove camera distortion and translate sub-pixel measurements locally or globally into real world units
• Positioning (search) tools to accurately landmark measurements on moving parts
Software Capabilities
Selecting the right optics and lighting to achieve a clear image with minimal distortion is another important consideration for noncontact measuring.
DALSA offers a range of area and line scan cameras with its Vision Appliances to suit almost any measurement application.
1024 X 1200 LINE SCAN IMAGE
Selecting the correct resolution is critical to distinguishing the smallest feature for measuring. In the application below, a DALSA 1024 pixel line scan camera is used to image different sized horse shoes. In applications where the part being gauged is large, images may be sourced and combined from multiple cameras to perform measurements.
Imaging for Measurement Accuracy
Area or Line Scan Cameras
• Presence/absence • Dimensional accuracy – geometrical tolerances • Thickness and uniformity of parts
Measurement Applications
case study
www.goipd.com 09
Measurements are being held to +/-0.0002 at speed. Each camera uses an IPD designed collimated back light. The VB interface reads out dimensions in thousandths of an inch.
Frontal Camera View
A dual camera VA41 Vision Appliance is used to inspect 2 parts at a time for an effective throughput of 1320 parts per minute.
A ribbon of coiled metal enters a highspeed press where parts are stamped, bent or folded with every punch of the press. The presses run at a top speed of 660 strokes per minute, creating 2 parts with every stroke.
METAL CONNECTOR INSPECTION
04
Connect
12 Industrial Vision Solutions
Connect either locally with a keyboard and mouse or remotely via network.
01
Sensor
Adjust sensor, select single or multiple cameras, adjust trigger, set lights and adjust exposure.
02
Set-up in 5 easy steps
Apply Tools
Using the icon driven tool palette, apply the tools needed for the inspection
03
Integrate Setup I/O and information exchange between system, network and 3rd party controllers.
04
Inspect Run your application and save the setup to be loaded on power-up.
05
quickly setup and deploy machine vision to a wide range of manufacturing tasks.
tool. For the first-time, new and experienced users alike have the capability to
the factory floor. We’ve taken all of our knowledge and developed a user interface that abstracts the complexity of machine vision to render a practical inspection
This breakthrough software represents years of experience developing solutions for
machine vision made simple
iNspect software
LABEL INSPECTION ON PACKAGE
iLabel allows users to setup fine ROI processing areas for detecting very small manufacturing defects.
iLabel can inspect labels from 1” to over 10” in size and is capable of inspections speeds in excess of 300 labels per minute.
Operator access is an important consideration in factories. iNspect provides the capability to restrict or lockout unauthorized users.
Included with iNspect is iLabel, an application designed to be used by someone familiar with a packaging line, but not necessarily machine vision. iLabel is quick to setup and rapidly verifies the placement, quality (including print) and correctness of labels on bottles, boxes, cans and other ridged, packaged goods. iLabel also provides the capability to read and verify product barcodes.
For highly controlled manufacturing environments like Pharmaceutical, it is also required to log access and any changes made to the system. iNspect offers the ability to log access and change information to a secure drive on the company network.
Administration
SURFACE FLAW DETECTION
BEAD MEASUREMENT
• Emulator for offline development • Support for custom local interfaces • Direct connect to 3rd party interfaces • Image logging and playback
Label Inspection Made Easy
POSITIONING
CAP VERIFICATION
• Multiple cameras and image sizes • Same interface for set up and runtime • Access control • Solution switching via l/O or network • FREE updates
iNspect Features
www.goipd.com 13
RESULTS
INSPECTION
IMAGES AND
LOGGING OF
PROVIDES
iNSPECT
iNspect offers a Visual Basic API for advanced users wishing to develop custom operator interfaces. The standard operator interface provided with the product is available in various languages such as English, Chinese, French, Italian, Japanese and Spanish.
iNspect supports digital I/O, serial and Ethernet communications for interfacing 3rd party equipment, operators and the factory enterprise. Compatible protocols, such as Modbus, Profibus and Ethernet/IP, provide standard languages for connecting complementary factory devices. DALSA IPD is proud to be an encompass partner of Rockwell Automation.
Factory Integration
challenging inspection tasks.
script-like formulas for solving
capabilities allow users to develop
applications and industries. Advanced
that can be applied to a wide range of
iNspect offers a complete set of tools
PRODUCT IDENTIFICATION
14 Industrial Vision Solutions
04. Image window Displays image during setup and live image at runtime. Images are acquired from cameras, files or sequence of files.
03. Program instruction toolbar Provides quick access to commonly used instructions. These include acquisition, subroutine creation, program steering, conditional statements and scripting.
02. Image window controls Load, acquire, save and zoom images. Select Region-OfInterest shapes and apply image preprocessors and algorithms.
01. Solution management Open and save solutions, start and stop inspection. Includes single-step debug operations.
05. Feedback windows Viewing windows provide immediate status of program events. They provide feedback of instruction timing, algorithm results, variables, hardware I/O, result reporting and more.
05
04
02
01
User Development Interface
06. Program The program window displays the sequence of instructions or actions that comprise an inspection. Program snippets can be copied and paste back into the program or a subroutine.
05
05
06
03
machine vision industry, Sherlock offers both power and flexibility to solve your vision applications while providing the assurance that comes with this popular product.
deployed in thousands of installations worldwide. Recognized throughout the
resolve a wide variety of automated inspection applications. This graphical design environment provides a rich suite of proven tools and capabilities that have been
Sherlock is advanced machine vision software that can be easily configured to
the choice among integrators
sherlock software
Sherlock provides tools for color correction, classification and presence. It also supports color mapping, a technique which allows you to segment the image by color in order to apply mono tools to the task.
Color Tool
CORNER FINDER TOOL
The corner finder tool generates an array of “corner points” that can be manipulated by Sherlock formulas to measure the space between “peaks and valleys” of machined parts, such as bolt threads.
Corner Finder Tool
BEAD TOOL
The bead tool algorithm inspects a bead (thin line) of material. A typical application is inspecting beads of glue that attach gaskets to automotive assemblies.
Sherlock provides interfaces to a variety of communication mediums and supports standard factory protocols such as Modbus and Ethernet/IP.
Communication
CALIBRATION TOOL
Sherlock offers several methods for translating pixel to real-world coordinates. Calibration tools also correct for lens and perspective distortion.
Calibration Tool
LASER LINE TOOL
Laser tools are used to measure the profile of parts or to detect irregularities such as the placement of protective wrapping on this high-pressure pipe. At the right, a gap in the wrapping is followed by lifting of the wrapping, as shown by the upward step in the reflected laser line points.
Sherlock tools and capabilities allow you to tackle a wide range of industrial applications. Included are a variety of specialty tools that have been specifically designed to simplify difficult inspection tasks.
Bead Tool
Laser Line Tool
Specialty Tools
Sherlock provides a comprehensive set of vision tools and capabilities that can be applied to applications across all industries. You can quickly build a solution using Sherlock’s extensive library of preprocessors and advanced algorithms or if you need something special, you can write custom scripts, import proprietary tools and develop your own custom operator interfaces.
Rich Suite of Tools for any Application
• Flexible Region of Interest selection • Extensive set of conditioning functions • Advance pattern finding tools for object alignment and robot guidance • Precise tools for computing the dimensions
Sherlock Features
www.goipd.com 15
A complete Visual Basic interface is provided for developing custom operator interfaces.
CUSTOM OPERATOR INTERFACE
Sherlock’s Java Script based scripting tool, complete with drag and drop instruction editing, allows you to develop custom formulas for in-line and background operations.
Customization
EDGE GUI TOOL
Many of the tools provide graphical feedback that allows you to tune the algorithm to match your application needs.
POINTS FROM SPOKE REGION OF INTEREST
Access Type # Inputs # Outputs
Application
I/O
Software
iNspect Lite
Direct 24V Opto 10 8
Remote Local
2 x (1.1) 10/100/1000 1 26
CE, RoHS
CE, RoHS
9.5H x16Lx5D
24V @ 1A
iNspect iLabel
Direct 24V Opto 10 8
Remote Local
2 x (1.1) 10/100/1000 1 26
Analog 2 Area No 640 x 480 1024 x 768 60f/s
256MB 128MB Flash
1X
VA20
16 Industrial Vision Solutions
MONO ANALOG CAMERAS
GENIE COLOUR CAMERAS
Full range of cameras and accessories
Approvals
9.5H x 16Lx5D
Setup GUI Operator
Display Options
Centimeters
USB Ethernet (Mbps) Serial (RS232) Visual (LEDs)
Communication
Analog 1 Area No 640 x 480 640 x 480 60f/s
Dimensions
Sensor Type # Sensors Sensor Format Color Support Sensor Size Min Sensor Size Max Sensor Speed
Image
128MB 64MB Flash
24V @ 1A
Program Storage
Memory
0.7X
VA15
Power
Relative
Processing Scale
STANDARD | VISION APPLIANCES
CE, RoHS
CE, RoHS
9.5H x 16Lx5D
24V @ 1A
iNspect iLabel
Direct 24V Opto 10 8
Remote Local
2 x (2.0) 10/100/1000 1 26
Analog 2 Area No 640 x480 1600 x 1200 60f/s
512MB 128MB Flash
2X
VA30
LINE SCAN CAMERAS
9.5H x 16Lx5D
24V @ 1A
iNspect iLabel Sherlock
Direct 24V Opto 10 8
Local/Remote Local
2 x (1.1) 10/100/1000 1 26
Analog 2 Area No 640 x 480 1024 x 768 60f/s
256MB 2GB Flash
1X
VA21
vision appliance platforms
CE, RoHS
9.5H x16Lx5D
24V @ 1A
iNspect iLabel Sherlock
Direct 24V Opto 10 8
Local/Remote Local
2 x (2.0) 10/100/1000 1 26
Analog 2 Area No 640x480 1600x 1200 60f/s
512MB 4GB Flash
2X
VA31
Access Type # Inputs # Outputs Application
I/O
Software
INDUSTRIAL ENCLOSURES
LIGHTING SOLUTIONS
CE, RoHS
7.6 x 20 x 21.5
Dimensions Approvals
24V @ 2.5A
iNspect iLabel Sherlock
Breakout 24V 8 8
Local/Remote Local
Power Centimeters
Setup GUI Operator
Display Options
2 x (2.0) 10/100 1 7
3 Area No 640 x 480 1600 x 1200 60f/s
# Sensors Sensor Format Color Support Sensor Size Min Sensor Size Max Sensor Speed USB Ethernet (Mbps) Serial (RS232) Visual (LEDs)
Analog
512MB 40GB
3X
VA40
Sensor Type
Program Storage
Relative
Communication
Image
Memory
Processing Scale
EXPANDED | VISION APPLIANCES
CE, RoHS
7.6 x 20 x 21.5
24V @ 2.5A
Sherlock
Breakout 24V Opto 12 8
Local Local
2 x (2.0) 10/100 1 7
2 Area/Line Yes 1024 x 1 8192 x1 User Defined
CamerLinkTM
512MB 40 GB
3X
VA50
CE, RoHS
7.6 x 20 x 21.5
24V @ 2.5A
iNspect iLabel Sherlock
Breakout 24V 8 8
CE, RoHS
7.6 x 20 x 21.5
24V @ 2.5A
Sherlock
Breakout 24V Opto 12 8
Local Local
2 x (2.0) 10/100/1000 1 7
2 Area/Line Yes 1024 x 1 8192x 1 User Defined
CamerLinkTM
1024MB 40 GB
10X
VA51
INDUSTRIAL ISOLATION
Local/Remote Local
2 x (2.0) 10/100/1000 1 7
6 Area Yes 640 x 480 1600 x 1200 60f/s
3 Analog 3 IEEE1394
1024MB 40 GB
10X
VA41
DALSA IPD VA61 AND VA30 VISION APPLIANCES
www.goipd.com 17
CE, RoHS
7.6 x 20 x 21.5
24V @ 2.5A
iNspect iLabel Sherlock
Breakout 24V 8 8
Local/Remote Local
2 x (2.0) 10/100/1000 1 7
2 expandable Area Yes 640 x 480 1600 x 1200 60f/s
GigE
1024MB 40GB
10X
VA61
Busch Vacuum Pumps and Systems
Page 1 of 1
BUSCH ALL OVER THE WORLD
Go
Australia
Company
Products
Applications
Service
What's new
Contact
What is vacuum?
Seco - Dry Running Vacuum Pumps 4 - 140 m3/h (50 Hz) 4,8 - 168 m3/h (60 Hz) Vacuum 100 - 150 hPa (mbar) Over Pressure 0,6 - 1 bar (g) Seco vacuum pumps are the ideal vacuum generator anywhere in industry where oil-free sealing is required. There is also an over pressure version of the Seco SD available. z
compact
z
low maintenance
z
economic
z
oil-free
Tiny SV
Miniseco SV
Technical Data [50 Hz] SV 1003
SV 1008
SV 1002
SV 1004
SV 1006
3
8
2
4
6
Ultimate pressure [hPa (mbar)]
150
150
150
150
150
Nominal motor rating [kW]
0,1
0,25
0,18
0,18
0,25
Nominal displacement [m3/h]
Seco SV Technical Data [50 Hz] SV 1010 C
SV 1016 C
SV 1025 C
SV 1040 C
SV 1063 B
SV 1080 B
SV1100 C
SV1140 C
10
16
25
40
63
80
100
140
Ultimate pressure [hPa (mbar)]
150
150
120
120
100
100
100
100
Nominal motor rating [kW]
0,37
0,55
0,9
1,25
1,7
2,2
3,0
4,0
Nominal displacement [m3/h]
To receive more information on Seco vacuum pumps - Please click here
©2000-2002 Dr.-Ing. K. Busch GmbH
HOME | COMPANY | PRODUCTS | APPLICATIONS
http://www.busch.com.au/seco.htm
|
SERVICE
|
WHAT'S NEW
|
CONTACT
31/03/2009
