DESIGN AND IMPLEMENTATION OF A MULTIPLE...

DESIGN AND IMPLEMENTATION OF A MULTIPLE FUEL

INJECTOR DRIP SENSING TEST STAND SYSTEM TO SIMULTANEOUSLY TEST

AUTOMOTIVE FUEL INJECTORS

PROJECT 4014 Jason Chekansky Hugh Campbell

Matthew Kubarek Courtney Stout

Peter Boyer Brian Crawford Richard Peffer

Rochester Institute of Technology Kate Gleason College of Engineering

87 Lomb Memorial Drive Rochester, NY 14623

Executive Summary The Critical Design Report details the steps taken by the Fuel Injector Drip Test Stand design

team to design and implement a test system to detect fuel drips from automotive fuel injectors.

Fuel injectors are designed to provide a precise volume of fuel per pulse to the intake port.

Accumulation of fuel develops on the injector tip throughout the course of operation and

eventually overcomes viscous forces allowing the drip to pass into the intake tract. These drips

are magnitudes larger than the incoming metered fuel vapor pulse. The large, unmetered drips of

fuel have a dramatic effect on engine emissions.

The design team employed methods as prescribed by the Engineering Design Planner Notebook.

The faceted approach is detailed in the following report. The first chapter, Recognize and

Quantify the Need, discusses the project mission, goals, and scope of the project. The second

chapter, Concept Development, details the approach taken by the team from brainstorming to

design conceptualization. Chapter three, Feasibility Analysis, uncovers the measures taken to

identify the best solution to meet project goals through a thorough breakdown of each design

concept regarding various parameters. The fourth chapter, entitled Design Objectives and

Specifications, provides the reader with both qualitative and quantitative measures of product

performance. Chapter five discusses analysis and synthesis methods taken to research and

develop the design concepts. In Chapter 6, conclusions from sensor system testing are weighed

and a final sensor system decision is made. Chapter 7, entitled Detailed Design, describes the

sensor package, data acquisition system, and test enclosure in full detail. Chapter 8 describes

implementation results and concludes with a look at project accomplishments. The Critical

Design Report includes a technical data package, with detailed drawings of assemblies, sub-

assemblies and components. A complete Bill of Materials is also included in the package.

The fuel injector drip sensing test stand system is composed of an array of accelerometer sensor

assemblies to detect fuel drips, a data acquisition system to interpret and report drip occurrence,

and a test enclosure to house the injector positioning system and sensor assemblies.

The accelerometer sensor system includes an accelerometer, diaphragm, base fixture, annular

clamp ring, and signal conditioning apparatus. The system converts fuel drip impulse energy

into a voltage signal read by the data acquisition system. The setup also provides a large catch

basin to ensure capture of all fuel drips, guaranteeing system performance and reliability.

Data acquisition is accomplished utilizing LabVIEW software and National Instruments

hardware. Upon drip detection, a voltage signal is passed through a BNC board to a PCI card in

the DAQ computer. LabVIEW interprets and reports the voltage input as a drip. Drip occurrence

is reported in a preformatted MS-Excel spreadsheet. The operator is also given system

flexibility, as testing and reporting operations are able to be varied utilizing a LabVIEW operator

interface.

The test enclosure consists of an explosion proof, quiescent test chamber and fuel injector

positioning system. The enclosure is capable of simultaneously testing multiple fuel injectors

and withstanding a full vacuum load. The test enclosure is also equipped with pressure relief

panel similar in design to an explosion proof room. The fuel injector positioning system is

capable of changing fuel injector orientation through a 90o range. Fuel injector adapters provide

consistent injector positioning as well as fast, clip-on attachment.

Thorough testing of the final system has been conducted to ensure that the system met project

goals and requirements. The design team has implemented an automated solution to the unique

problem of detecting fuel drips from injector tips.

Table of Contents

Executive Summary .......................................................................................................................2

Table of Contents ...........................................................................................................................4

Table of Illustrations......................................................................................................................7

1.0 Recognize and Quantify the Need ....................................................................8

1.1 Project Mission Statement ........................................................................................... 8

1.2 Product Description ...................................................................................................... 8

1.3 Scope/ Limitations......................................................................................................... 8

1.4 Stakeholders .................................................................................................................. 8

1.5 Key Business Goals ....................................................................................................... 9

1.6 Financial Analysis ......................................................................................................... 9

1.7 Preliminary Market ...................................................................................................... 9

1.8 Order Qualifiers............................................................................................................ 9

1.9 Order Winners .............................................................................................................. 9

1.10 Innovation Opportunities ............................................................................................. 9

2.0 Concept Development........................................................................................9

2.1 Accelerometer Sensor System .................................................................................... 10

2.2 Infrared Emitter/ Detector Sensor System ............................................................... 11

2.3 Strain Gauge Sensor System ...................................................................................... 12

2.4 Test Enclosure ............................................................................................................. 13

2.5 Fuel Injector Positioning System ............................................................................... 14

3.0 Feasibility Assessment .....................................................................................16

3.1 Technical Feasibility ................................................................................................... 16

3.2 Performance Feasibility.............................................................................................. 17

3.3 Economic Feasibility ................................................................................................... 17

3.4 Safety Concerns........................................................................................................... 18

3.5 Scheduling Matters ..................................................................................................... 18

4.0 Design Objectives and Specifications.............................................................19

4.1 Design Objectives ........................................................................................................ 19

4.2 Performance Specifications........................................................................................ 20

4.3 Design Practices .......................................................................................................... 21

4.4 Code Compliance ........................................................................................................ 22

5.0 Problem Analysis and Design Synthesis ........................................................24

5.1 Accelerometer Sensor System Design ....................................................................... 24

5.1.1 Safety Concerns................................................................................................... 24

5.1.2 Diaphragm Material Selection........................................................................... 24

5.1.3 Diaphragm Contour Analysis ............................................................................ 26

5.2 Infrared Emitter/Detector Analysis .......................................................................... 28

5.2.1 IR Emitter/ Detector Location and Positioning ............................................... 28

5.2.2 IR Emitter/ Detector Accuracy.......................................................................... 29

5.3 Data Acquisition.......................................................................................................... 29

5.4 Test Enclosure Design................................................................................................. 31

5.4.1 Structural Analysis ............................................................................................. 31

5.4.2 Pressure Relief System Design........................................................................... 32

5.4.3 Fuel Injector Adapter Design ............................................................................ 33

6.0 Sensor System Conclusions .............................................................................34

7.0 Detailed Design.................................................................................................34

7.1 Accelerometer Sensor System .................................................................................... 34

7.1.1 Design ................................................................................................................... 35

7.2 Data Acquisition System............................................................................................. 36

7.2.1 Fuel Injector Control.......................................................................................... 36

7.2.2 Sensor Input ........................................................................................................ 37

7.2.3 Data Reporting .................................................................................................... 38

7.2.4 DAQ Hardware ................................................................................................... 38

7.3 Test Enclosure ............................................................................................................. 39

7.3.1 Test Enclosure ..................................................................................................... 39

7.3.2 Injector Positioning System ............................................................................... 40

8.0 Results ...............................................................................................................41

9.0 Future Considerations .....................................................................................43

10.0 References .........................................................................................................44

Appendix A. Concept Development Documents..................................................................45

Appendix B. Accelerometer Sensor System Proof of Concept...........................................46

Appendix C. Infrared Emitter/Detector System Proof of Concept ..................................48

Appendix D. Infrared Emitter/Detector Experimentation.................................................51

Appendix E. Strain Gage Sensor System Proof of Concept ...............................................55

Appendix F. Feasibility Assessment .....................................................................................60

Appendix G. Accel. Sensor System Diaphragm Material Exp. ..........................................62

Appendix H. Test Enclosure Structural Analysis................................................................69

Appendix I. Fuel Injector Adapter Stress Analysis ...........................................................78

Appendix J. Final Design Return on Investment Analysis ................................................84

Appendix K. Bill of Materials................................................................................................85

Table of Illustrations

Table 1: Brainstorming Pro's/Con's ............................................................................. 10

Figure 1: Test Enclosure Concept ................................................................................. 14

Figure 2: Fuel Injector Positioning System Concept 1 ................................................ 15

Figure 3: Fuel Injector Positioning System Concept 2 ................................................ 15

Figure 4: Fuel Injector Adapter Concept ..................................................................... 16

Table 2: Estimated Return on Investment.................................................................... 18

Table 3: Scheduling Assessment .................................................................................... 19

Table 4: Design Objective Checklist ............................................................................. 21

Table 5: Materials Tested for Diaphragm.................................................................... 25

Figure 5: Decay Envelope - 3000 Series Aluminum..................................................... 26

Figure 6: Square Diaphragm Modal Analysis.............................................................. 26

Figure 7: Circular Diaphragm Modal Analysis ........................................................... 27

Table 6: Modal Analysis Results-Non-Contoured Diaphragm................................... 27

Figure 8: Contoured Diaphragm................................................................................... 27

Figure 9:Contoured Diaphragm Modal Analysis ........................................................ 28

Table 7: Excel Test Report............................................................................................. 30

Figure 10: Test Enclosure Structural Analysis - Design 1 .......................................... 31

Figure 11: Test Enclosure Structural Analysis with Center Support – Design 2 ..... 32

Figure 12: Test Enclosure Loading ............................................................................... 33

Figure 13: Test Enclosure and Injector Positioning System....................................... 33

Figure 14: Sensor Assembly Cross Section................................................................... 35

Figure 15: Accelerometer Fixture Assembly................................................................ 35

Figure 16: PCB 12605 Ceramic Shear Accelerometer ................................................ 36

Figure 17: LabVIEW Operator Interface .................................................................... 37

Figure 18: DAQ Hardware Flow Diagram................................................................... 39

Figure 19: Test Enclosure and Control Cabinet .......................................................... 40

Figure 20: Injector Positiong System ............................................................................ 41

Table 8: Completed Design Objectives Checklist ........................................................ 42

Delphi Fuel Injector Drip Test Stand 04014 8

1.0 Recognize and Quantify the Need

1.1 Project Mission Statement The mission of the design project is to research, design, and implement an automated fuel injector drip sensor system and test enclosure to replace the current manual system. The sensor system must be accurate, repeatable, allow unattended operation, and be low cost. The test enclosure must be a quiescent, explosion proof chamber.

1.2 Product Description The project involves the investigation and determination of a sensor that can detect drips of fuel from multiple automotive fuel injectors simultaneously. The sensor will feed a signal to a LabVIEW based program to record the quantity and time of the drips. The control of the injector will be accomplished using LabVIEW software and the sensor/ drip-timing algorithm is required to be merged into the existing injector control system. From a design perspective, requirements include a multiple injector positioning system as well as the individual injector sensor systems, and the associated fuel and electrical services. The test enclosure should be a quiescent, nitrogen purged explosion proof chamber capable of withstanding full vacuum load. Key project requirements are safety, automated testing, test repeatability, easy set-up, and low cost.

1.3 Scope/ Limitations

The Fuel Injector Drip Test Stand should be a stand-alone system with the ability to test multiple injectors simultaneously. The system must adhere to all relevant safety codes and regulations regarding gasoline vapor control. The test stand must also accommodate multiple injector types. Programming must use LabVIEW software to ensure compatibility with the current injector control system. All data acquisition should also employ LabVIEW software. The accuracy of the sensor system must meet or exceed that of the current manual system. Total project funding for both the sensor system and test enclosure should not exceed $7000.

1.4 Stakeholders

The stakeholders of the Fuel Injector Drip Test Stand design project include Delphi Corporation Technical Center Rochester. This includes the Engineering Services, Spray Laboratory, Fuel Injection Systems, Product Development, and Application Groups departments. Other entities of interest include Kate Gleason College of Engineering and automotive original equipment manufacturers.


1.5 Key Business Goals The Fuel Injector Drip Test Stand should reduce manpower required for testing. The current test system places emphasis on manual labor and allows for single fuel injector testing. The test stand will increase sample size capability and laboratory throughput, as well as decrease man-hours required for testing.

1.6 Financial Analysis

The Fuel Injector Drip Test Stand should repay project investment within six months. Manpower cost savings should be accomplished through the creation of an automated test system that allows for unattended operation. Individual injector testing costs should be reduced through simultaneous testing of multiple injectors.

1.7 Preliminary Market

The Fuel Injector Drip Test Stand is an engineering tool with limited used beyond port fuel injection testing at Delphi Technical Center Rochester.

1.8 Order Qualifiers

Order Qualifiers include labor and manpower savings, unattended operation, and automated analysis. Set-up time per injector shall be less than ten minutes. Also, the system will allow for testing of multiple injector types.

1.9 Order Winners

Order Winners include quick and simple set-up of the injector sample set. The test stand will allow for the test of multiple injectors simultaneously. Also, the test stand allows for the drip testing of fuel injectors in an array of orientations.

1.10 Innovation Opportunities The Fuel Injector Drip Test Stand allows for the development of a new sensing system and test enclosure, or the application of an existing system to a unique problem of detecting fuel drips from automotive fuel injectors. 2.0 Concept Development The Concept Development stages of the Fuel Injector Drip Test Stand included several steps to help create and define ideas for possible solutions to the sensing system and test enclosure design problem. Included is a description of several brainstorming sessions, empathy methods, and consensus building activities used to facilitate this process.

An initial brainstorming session was utilized to create a long list of 25 concepts [Appendix A]. Further scrutiny of these concepts narrowed the list to four concepts, of which employed vibration, strain gauge, electro-mechanical methods, and optics as possible solutions to the design project.

Of these four ideas, three were selected for a group drawing session. Each member of the design team drew his or her interpretation of one of the three concepts on paper for five minutes, after


which the concept drawing was passed to the next person. This continued until all members had contributed to each concept drawing. The group drawing session allowed team members to place their ideas on paper, as well as spur thought in others minds at the same time.

Empathy, or role-playing, methods were also used to resolve problems and misunderstandings associated with the design concepts. In class, each team member was assigned a component of the project, including the sensor, fuel drip, data acquisition unit, detection signal, and wires to connect the electric components. The team member was to role play their component and interact with the other components, simulating the actual interactions in the final product. The process was acted out to determine any problems that would arise, and how the system might be designed to react in such a case. This method emphasizes the need for careful attention to interfaces throughout the system. This process identified such problems as software programming issues, input arguments, output display features, and sensor location ambiguity.

Another group drawing session was used in further concept development stages. Ideas were written on the board and team members were allowed to freely contribute fresh ideas, or add to previous contributions, while continuing to build on the short list. The result was several fairly well defined and workable concepts. The session also sparked a number of questions and concerns from the group, which led to a pros/cons session. Each of the short list items was itemized to determine the positive features and possible drawbacks. At this point in the concept development, the electro-mechanical concept was eliminated due to unavoidable concerns regarding isolation of the electrical components.

Concept Pros Cons

Vibration Accuracy

Price Mounting apparatus for accelerometer

Resolution Electric contact shielding

Optical Low system interaction Detection field Signal output

Part availability Electro-Mech. Low Cost Electric contact isolation

Strain Gauge Low Cost

Easy mounting – flexible setup Off-the-shelf accessibility

Material used to show strain -strength vs. flexibility -fuel resistance

Table 1: Brainstorming Pro's/Con's

2.1 Accelerometer Sensor System A proof of concept experiment for the accelerometer sensing system was completed to determine the ability of the system to detect low mass drips [Appendix B]. The experiment was set up to show that an accelerometer would be able to detect an impulse traveling through a diaphragm with sensitivity required to give an adequate voltage output reading. The conclusion of this experiment was that the accelerometer concept is indeed a valid concept that has a high probability of success in a drip sensing application.


The next step in the concept development process was a team meeting with field representatives from PCB Piezoelectronics Inc. This meeting confirmed that the use of an accelerometer as an acceptable sensing tool in this application. Additional equipment required for the operation of this system includes a signal conditioner and power supply. Concerns about budget arose throughout the course of the session. It was determined that the cost of the sensors would depend on the sensitivity needed to detect a drip of fuel. Approximate cost estimates for the discussed systems ranged from $3,900 to $4,500. Safety issues were also discussed. This included certified intrinsically safe accelerometers, hermetically sealed integral cables, and a Zener barrier to protect against power surges.

Relevant information resulting from this meeting included PCB’s service/return policy, which guarantees 100% satisfaction, lifetime service or replacement, and full money back during warranty period. The approximate life cycle of an accelerometer is equal to approximately one hundred million cycles. The lead-time for an accelerometer is equivalent to standard delivery time, providing that the system does not require any specialty parts. Mounting requirements eliminated the use of magnetic adhesion, but strongly suggested the use of stud mounting.

Initial testing provided an approximate voltage and frequency range to determine the appropriate accelerometer. Detailed bills of materials and technical drawings have been created based on the data gathered in the concept development phase of the accelerometer sensing system.

2.2 Infrared Emitter/ Detector Sensor System A proof of concept experiment was conducted for the Infrared Red sensing system [Appendix C]. The conclusion of this experiment was that an IR sensing system would detect drips as they fall through the infrared beam, resulting in an easily traceable change in flux between the emitter and detector. This experiment proves the validity of the IR sensing concept. The low cost of such a sensing system made the concept an attractive fit to the design. All test system components are also readily available at local electronics stores.

Questions were raised about the detection field and the range of the IR beam, as problems arose regarding undetected fuel drips. Other issues included concerns with set-up time and calibration of the emitter detector, interference from the fuel film that develops on the test chamber walls and sight windows in the saturated environment, as well as material property restrictions inherent to acquiring emitters of the same wavelength as that of the sight windows.

A meeting with Dr. Robert J. Bowman, Electrical Engineering Department Head, Rochester Institute of Technology, provided answers to questions that arose during IR sensor testing. Dr. Bowman had completed similar work regarding drip detection in medical applications. His insight provided a basis for understanding the logic behind the system, as well as a thorough understanding of solutions to possible problems associated with the IR sensor system. Dr. Bowman proposed the use of automatic gain circuits to address the issue of fuel vapor build up on sight windows and the degradation of output signal. He also expressed thoughts of


matching emitter wavelength with that of Lexan windows to eliminate interference between the emitter and test enclosure. Infrared field width and intensity were also addressed in the meeting. Through the use of a refracting lens, the sensor beam could be spread to encompass a diameter of 1/8” to 1/4”. A small current draw of 25mA per set of infrared emitters/ detectors would be required, and a circuit could be designed to minimize the effect of signal drop over time. This signal drop could be due to the system’s lens becoming blocked by dirt or fuel mist over time. Through the use of an infrared system with a higher intensity or lens system, the detection ability of the system could be increased. Infrared lasers can be purchased in a variety of sizes, increasing the detection field capability. Data transfer from a receiver can be accomplished using series or parallel ports, which leads to reduction in signal conditioning costs. An IR sensing system would also have expansion capability with the ability to record volume and drip size through pulsing the infrared beam.

As a result of this meeting, it has been determined that the IR sensing system concept is both valid and feasible. Further experimentation and research was done to determine the best setup for the sensing system [Appendix D]. Results from experimentation showed that an emitter/ detector pair is capable of accurately detecting drips. Problems arose due to a lack of a positioning system to consistently position the set in the same location.

2.3 Strain Gauge Sensor System The Strain Gauge Sensor System employed the use of a cantilever-mounted catch basin that would deflect in the presence of a fuel drip. This deflection would induce a strain in the basin that would be sensed by a strain gauge. The strain changes the resistive properties of the strain gauge that induce a change in voltage as recorded by the LabVIEW software. Preliminary experimentation was completed to demonstrate the validity of the strain gauge concept as a design solution. Conclusion of this experiment showed that the concept is valid [Appendix E]. A strain gauge attached to a 0.25” thick aluminum bar displayed a 3 mV change. However, this experiment raised a number of concerns. A primary concern of the strain gauge system was inadequate adhesives and protective coatings. This is not only a durability issue, but also leads to a safety concern. The breakdown of strain gauge adhesive could lead to a potentially hazardous situation. If live wires were to be exposed in the test enclosure, arcing could occur, setting off a combustion process in the fuel saturated environment. Also, a situation resulting from exposed wires and contacts could result in the failure to detect fuel drips. Another, and perhaps less substantial concern, is possible deformation of the material to which the strain gauge would be mounted.

As a result of this experimentation, follow up research was done to investigate the types of adhesives and protective coatings that could be used to mount and protect a strain gauge in this


application. Adhesive manufacturers would not give estimates about the life cycle of their products used in this particular application. Similarly, strain gauge manufacturers will not guarantee that recommended adhesives would withstand a fuel-saturated environment for any length of time. Possible base materials were also researched to use with the strain gauge sensor system so that the correct amount of deflection would occur with each consecutive drip. One possible coating material was identified but the vendor specified that this material has a potentially short life cycle and may also be too stiff. Given the flexibility requirements of the application, any damping of the output level from the strain gauge would result in unacceptable output levels.

2.4 Test Enclosure The design of the test enclosure encompasses many different functions and needs. The enclosure must be able withstand any explosion which may occur due to combustion of gasoline vapors inside of it. Release of pressure from explosion is accomplished using a pressure relief panel on top of the chamber, venting excessive internal pressures. The chamber design must also incorporate a fuel injector mounting system that is capable of testing the fuel injectors at different angles. Most injectors will be tested at a 45-degree angle. The test enclosure is designed to test injectors thru angles of 0 to 90 degrees by increments of 5 degrees. Perhaps most importantly, the test enclosure must withstand a maximum vacuum pressure of 14.7 psi (1 atm). The chamber must be able to withstand full vacuum load without collapse. The enclosure must have a door that will allow the operator quick access to setup fuel injectors in under 10 minutes. The door must be also able to withstand the force exerted by vacuum. In addition, the design of the chamber must also accommodate nitrogen purge lines and flow meter probes. The enclosure must be sealed for all wiring that passes into the chamber. Gasket material used as a sealant around the blowout plate on top of the chamber and around the Lexan doorframe is a concern. Possible fuel resistant materials include Buna-N and Viton. The chamber must include a vacuum trough that will evacuate the test chamber of fuel vapor. The trough will also serve as a place to connect the vacuum line to the bottom. The trough should be equipped with a 2.5”diameter quick disconnect cam and groove fitting.


Figure 1: Test Enclosure Concept

The use of steel for the test enclosure chamber walls was ruled out due to the corrosive nature of the material. It is possible to coat or plate the metal, but the price of this additional operation would push the total cost of the enclosure beyond the allotted project budget. Design of the chamber walls started with ½” alloy 6061 aluminum.

2.5 Fuel Injector Positioning System The fixture must accommodate at least 80% of Delphi fuel injectors that are currently being produced. A number of different Delphi fuel injectors were provided to the team for the purpose of standoff design. The operator must also have the ability to rotate the fuel injectors through a range of 0o-90o. Fuel injector mounting cannot affix the fuel injector to the test stand at the spraying end. The pre-test set-up cannot exceed 10 minutes per fuel injector. The fuel injector standoff fixture uses interchangeable parts to hold the fuel injectors in place. The standoff adapters vary in length to place the each fuel injector tip at the same location. The standoff design concept simulates connection to an automotive style fuel rail. This provides quick and easy clip-on performance, with the ability to accommodate the 80% of Delphi fuel injectors as required by the customer. This concept also has expansion capabilities to encompass a broader array of fuel injector styles. The proposed design will allow the operator to rotate six injectors through the 90o range. The header plate of the injector fixture will travel a 6-½” radius. The complete fixture will be located inside of a chamber whose walls provide a surface to tap for ½ -20 bolts, which will provide the axis of rotation for the fixture. To lock the fixture at a specific angle a pin style locking mechanism is provided.


Figure 2: Fuel Injector Positioning System Concept 1

Figure 3: Fuel Injector Positioning System Concept 2

Injectors will attach to the varying length adapters using a traditional clip style mount that is used in fuel rail applications. This allows for use of readily available in-house fuel injector clips that provide low changeover time between injector test sets. The first positioning system concept allows for reposition of all six injectors in one motion. The second concept allows for individual injector orientation changes, but also increases set-up time if all injectors are to be tested in the same orientation.


Figure 4: Fuel Injector Adapter Concept

3.0 Feasibility Assessment Sensing a drip from the tip of a fuel injector is the primary function of the design project, as most of the feasibility assessment was spent looking at three different sensing systems. Although the project is to include a test enclosure and fuel injector positioning system, the sponsor felt that if a sensing system couldn’t be developed that the entire project would be unfeasible. The first sensing system proposed was an infrared emitter / detector system which senses the drip as it passed through an infrared beam. The second proposed system is an accelerometer adhered to the underside of a diaphragm. The diaphragm will transmit the impact of a drip to the accelerometer. The third option was to mount a strain gauge on a cantilever beam, which would deflect when a fuel drip falls from the fuel injector. Ultimately, after performing a feasibility assessment of the three sensing systems using five feasibility subheadings, the group was able to decide on which options to pursue and design. Through the use of Pugh’s Method, the weighted method, and a radar plot it became apparent that further development of the accelerometer sensing system and the infrared emitter / receiver system would be needed for the fuel drip sensing portion of this project [Appendix F]. Furthermore, development of both systems was justified because sensing performance, economic viability, and safety are major facets of the guidelines of this project that were set forth by the project sponsor.

3.1 Technical Feasibility Using a simple setup in the Vibration’s Lab at RIT, the group was able to verify that an accelerometer will react to the impact of a drip. The accelerometer sensor system will require a base and a diaphragm. A design has been generated for the base and will have to be fabricated. The skills of the group make it possible to fabricate such a base in the RIT machine shop. Experiments are currently underway to determine a suitable material for a diaphragm. Based on preliminary testing, many commonly available materials can function effectively as a diaphragm. A resilient adhesive or a mechanical fastener must be used to attach the accelerometer to the


diaphragm in order to ensure long life in a fuel environment. The base and diaphragm will easily fit into the space required while providing substantial area for a drip to land. LabVIEW will acquire an input signal of voltage or current versus time, however a signal conditioner is required to amplify and filter the output from the accelerometer for the Data Acquisition System. The strain gauge system will require a bonding agent that will permanently and rigidly attach the strain gauge to a cantilevered beam. The cantilevered beam can be made from a thin piece of aluminum. A coating agent will also be required to protect against spark and corrosion in the fuel environment. The bonding agent and coating agent selected will have to endure a harsh environment and repeated cycles of loading and cleaning. The system is very compact and can be easily mounted within the test enclosure. An amplifier and a filter will be required to prepare the signal for LabVIEW. A simple test was preformed using a common strain gauge mounted to a small piece of aluminum and driven through a Wheatstone bridge. The magnitude of the output was far too low for our intended use. The experimental results led the group to the conclusion that a very thin beam would be needed that could pose bonding agent durability problems. The technology being employed for the infrared emitter / receiver has been demonstrated to work in similar situations. Our group has also demonstrated this technology with simple parts bought for under $10 at Radio Shack. Special parts that will be needed are a printed circuit board that can be made at RIT for a low cost. Size is not an issue with infrared detection because the entire system can be contained outside of the harsh environment. However, windows will need to be included in the design for the infrared system to operate. Inputting the signal into LabVIEW can be done via an RS-232 cable or parallel port. Signal conditioning and amplification can be easily integrated into the circuit design. To detect every drip, alignment of the system will need to be very precise that could result in an intensive design.

3.2 Performance Feasibility Since preliminary testing of the strain gauge was unable detect a drip, it was the weakest of the three sensing options in the performance feasibility category. Preliminary testing has shown that the accelerometer is capable of detecting multiple drips from the fuel injector. Most manufactures of accelerometers guarantee repeatability and life cycle for approximately five years. On the other hand, the team is currently developing the infrared emitter / receiver sensing system, with aiming of the emitter being the main problem. Whether or not the infrared system is going to accurately detect multiple drips is still unclear.

3.3 Economic Feasibility The chart below shows the return on investment time for each of the three different drip-sensing systems. The chart was made assuming the three sensing systems are both reliable and capable of testing the same number of fuel injectors per week. The accelerometer sensing option is the most expensive, followed by the strain gauge sensing system. Conversely, the infrared emitter / receiver sensing system is the most inexpensive.


Sensor System Estimated Return on Investment

Accelerometer Sensing System

7 months

Strain Gauge Sensing System

8 months

Infrared Emitter / Detector Sensing System 6 months

Table 2: Estimated Return on Investment

3.4 Safety Concerns Evaluation of concept safety has been broken down into three categories: heat concerns, electrical concerns, and compliance with federal, state, and Delphi safety codes. The main reason behind the safety concerns being, of course, that the drip tester will be used in a high fuel concentration environment. The three concepts were each rated on a relative scale of low, medium or high based on the level of safety concern for that concept. The strain gauge and the accelerometer each have electrical components that will need to be inside the vacuum chamber. This could result in electrical and heat concerns, making it necessary to minimize this risk by sealing or containing these components. The vision system makes it possible to place all electronic components outside of the chamber, eliminating the need for certain precautions.

3.5 Scheduling Matters Due to the tight time restrictions for project completion, it is necessary to analyze the scheduling feasibility of each concept. Scheduling feasibility has been broken down into part lead-time, lab time required vs. lab time available, LabVIEW programming, purchasing lead-time, and research and development. The results are displayed in the table below. Each item has been broken down into the estimated number of weeks needed for the completion of each project segment. It should be noted that lab time is estimated as the number of weeks that will be required to satisfy any access requirements. It should also be noted that purchasing lead-time is based on Delphi policy that any component order greater than $2000 needs to be approved by the purchasing department.


Scheduling Concern Strain gauge Accelerometer Optical/IR System Order lead-time 1 week 1 week 1 week Lab Time 2 2 3 Programming 3 3 3 Purchasing 0 1 0 Research and Devel. 2 1 2 Total: 8 weeks 8 weeks 9 weeks

Table 3: Scheduling Assessment While total estimated time is important, the primary feasibility concern is the research and development time that must be completed to satisfy the project deliverables. This includes the lab time required, which can and should be completed simultaneously with research and development. Using these criteria for analysis, the strain gauge and the vision system concepts both have an R&D completion time of two weeks. The accelerometer system leads under this heading with a completion time of 1 week, due to a greater knowledge base in this area. Based on the scheduling analysis, all three of these systems are similar in feasibility and will not be written off due to scheduling issues. 4.0 Design Objectives and Specifications A number of design objectives and performance specifications have been determined to provide a ruler for the team to measure the fuel injector drip test stand performance.

4.1 Design Objectives The team reviewed the project goals to devise a set of design objectives. The purpose of this is to have a set of guidelines and small goals, organized in a rudimentary schedule, through which the product is designed, built, tested and finally delivered. The list of these objectives is as follows. The primary objective of the project is to create an automated injector drip-monitoring system. This will keep the operator from having to visually inspect injector tips during quality control tests. This goal lies at the core of the project and the design elements must include this objective at every phase.

1) The sensor must be able to distinguish between pulsating fuel vapor spray and drips. The accuracy of the drip-monitor is important for the accuracy and effectiveness of the system.

2) The signal generated by the drip sensor system shall be transferred to a time referenced

graphical representation in LabVIEW. The Delphi team requests that the error detection records be kept in the same system (LabVIEW) that is currently being used, for ease of interpretation and viewing.

3) The total setup time for the drip sensor system shall be less than 10 minutes (per fuel

injector tested).


4) The secondary objective of this project will be to create a containment unit in which the injectors will be installed, powered, and monitored. The test unit must be designed in such a way as to, at a minimum, match the performance of the existing containment environment.

5) The test enclosure shall be designed for future accommodation of multiple injectors. The

unit, in its entirety (including the sensory units and containment facility), will be able to duplicate the efficiency and accuracy of the primary and secondary objectives in a multiple injector environment.

6) Finance for the finalization of the project objectives must be kept below the Delphi-

approved budget of $7000.00

4.2 Performance Specifications The team set a list of requirements that must be met by the final product in order for the design-project to be considered a success. The deliverable product must, at a minimum, meet all of these requirements in order to function per customer requests. These specifications have been kept in mind while designing the automated fuel injector drip-monitoring system. The list of objectives is shown below.

A) Systems (both the Sensing system and the enclosure system) must accommodate at least 80% of the fuel injectors currently being tested by Delphi. Delphi has supplied a sample of different injector types that will be mounted in the test enclosure. The drip-monitoring system must adapt to differences between injectors as well as monitor each with equal accuracy and reliability.

B) Test enclosure must be able to accommodate the multiple injector types and change the corresponding orientation. Per customer request, the test enclosure must accommodate multiple fuel injector orientations ranging from 0o-90o.

C) Software used to create the time referenced graphical representation of the drips must be compatible with Delphi’s current version of LabVIEW. Therefore LabVIEW version 7 is required.

D) The formats for the technical drawings must be compatible with Delphi’s standards. By implementing ANSI standards and (.dwg) format for all drawings, compatibility is insured.

E) The design must be in compliance with Delphi’s internal safety codes as well as industry safety standards. The following codes and standards must be followed: • NFPA • NEC and local and state building codes • All OSHA safety requirements • NFPA 79


• NEMA 12 • SLI sound level specs

Delphi currently uses English standard sizes for all its hardware applications. Therefore, all hardware, such as nuts and bolts, used for the drip-monitor system and containment facility shall be in US customary sizes (English standard sizes).

1. Fit into assigned workspace

2. Withstand full vacuum

3. Capture entire range of drips @ 45o injector angle

4. 4 to 6 fuel injectors to be tested at once

5. Setup time less than 10 minutes per test per injector

6. Programming in LabVIEW

7. Adhere to standards and codes pertaining to Delphi considerations for drip test stand (section 3.9)

8. IPS to accommodate orientation range (0o–90o)

9. Automated drip detection

10. Explosion proof quiescent test chamber

11. All drawings in .dwg format

12. Under Delphi budget of $7000.00 / Return on Investment within 6 months

13. Fuel resilient exposed components

14. Resolvable error detection (distinguish fuel drip from vapor)

15. Sensor repeatability

Table 4: Design Objective Checklist

4.3 Design Practices During design stages and experimentation, certain engineering principles were enlisted to aid the evolution of the project. These include computer aided finite element analysis, vibration analysis, statistical methods, the Plan-Do-Check-Act philosophy, and discipline specific knowledge. Together with these tools, the organization and dedication of the team created a successful constant flow of new ideas and feedback. Examples of engineering tools utilized to answer questions and solve problems are in the diaphragm analysis, enclosure structure, and fuel injector standoff design. The data and results of these processes are shown later in this report. The diaphragm analysis, discussed in sections 5.1.2 and 5.1.3, utilized the methodical test practice of “Plan-Do-Check-Act” to gather uniform data. Mechanical vibration equations were applied to the data and complete analysis is pending


final design decision. Concurrently, FEA analyses of different diaphragm designs were evaluated through the use of IDEAS CAD software. Through the use of COSMOS FEA software, the loads and dimensions of the test enclosure were easily modified for optimal design and safety factor. Likewise, the fuel injector standoff was also modeled in COSMOS FEA software. To estimate the loads acting on both CAD models, mechanics equations of force and pressure were used [Appendix I].

Other examples of utilizing engineering principles, practices, and intuition: Design for Manufacture – All components, assemblies, and sub-assemblies can be built utilizing machining facilities located at Rochester Institute of Technology Assembly - Complete system consists of four sub-assemblies that allow for concurrent building of each. Safety - Design process implemented safety concerns and measures were taken to ensure a safe work environment.

4.4 Code Compliance Design of the sensing system and test containment unit should adhere to all safety codes and regulations that are currently adhered to by Delphi as well as other relevant and applicable regulations. It has been determined that Delphi processes and equipment should adhere to the following codes:

• NFPA – 79 • NEC • Local/State Codes • OSHA Codes • NEMA • SL – 1

Upon investigation of these codes, it has been determined that there are no mandatory codes or regulations that apply to the sensor system design. Delphi currently observes all state and local codes that apply to the handling of flammable materials as well as codes observed that are applicable to the storage of the Delphi gasoline tanks. These should all include requirements for fire extinguishers, protective clothing, and exposure limits. These codes do not fall under the scope of this project and will not be addressed by this document. OSHA does have documentation of non-mandatory standards that it recommends for the safe processing of flammable and explosive liquids that will be incorporated into the design documentation. A summary of the steps required in order to comply with this standard are included below. Documentation required for the actual sensing system and test containment unit: [1910.119(d)]

Processing documentation must include: A block flow diagram or simplified process flow diagram of the system, maximum values for all intended materials


entering and leaving the system (including fuel, air, and electricity), safe upper and lower limits for such items as temperatures, pressures, flows or compositions, and an evaluation of the consequences of deviations, including those affecting the safety and health of employees.

Equipment documentation must include: All bills of materials, piping and instrument diagrams, electrical classifications, ventilation and relief system designs and design basis or other safety systems used, any relevant design codes and standards used, and material and energy balances.

The employer should provide written documentation that all equipment “complies with recognized and generally accepted good engineering practices”. A hazard analysis should be performed on the completed system that is “appropriate to the complexity of the process and shall identify, evaluate, and control the hazards involved in the process” [1910.119(e)]. This analysis should evaluate the potential hazards, consequences of those hazards and documentation of any alarms or detection systems that will be included in the design. This hazard analysis should include:

• “What-if” analysis • Hazard and operability study (HAZOP) • Failure modes and effects analysis (FMEA) • Fault tree analysis

This standard also includes information about how the analysis should be completed, including such recommendations as the inclusion of an experienced employee in the evaluation methods, previous experience in process operations, and documented recommendations that should include actions to be taken, a schedule of when actions should be completed, and who should complete these actions. Also included is a schedule of how often a new hazard analysis should be completed and what documentation employers should maintain [refer to 1910.119(e), sections 6, and 7] Documentation should be provided regarding proper and safe operating procedure [1910.119(f)] including:

Steps for each operating phase, initial startup, normal operations, normal shutdown, emergency operation and emergency shutdown (including restart procedures, and the conditions under which an emergency shutdown should be implemented). Documentation should also include normal operating limits and any consequences that could result in deviations from these limits, any special or hazards, and safety systems and their functions.

All documentation included under this provision should be made readily available to employees and reviewed annually. Training specifications are included and will be the responsibility of Delphi personnel to ensure that effected employees will be able to work in a safe and compliant manner [1910.119(g)].


The next set of standards applies to the mechanical integrity of the system design [1910.119(j)]. It is specified that these standards need be applied only to:

• Piping systems (including piping components such as valves) • Relief and vent systems and devices • Emergency shutdown systems • Controls (including monitoring devices and sensors, alarms, and interlocks) • Pumps

For these items, documentation must be provided for:

Maintenance, inspection and testing of equipment (including schedules and procedures), and any relevant expected replacement dates.

Important employer information is also included in section 1910.119(j) and should be observed by the appropriate staff at Delphi. The following 5 sections [1910.119(k), (l), (m), (n), and (o)] include important employer information regarding hot work permits, management of change, accident investigation, emergency planning and response plans pertaining to the entire plant in the event of an emergency situation, and compliance audits and should be observed appropriate Delphi staff. The next section of standards is information regarding trade secrets [1910.119(p)]. The OSHA standards state that all safety considerations should be regarded as more important than trade secret information, and relevant information should be made available to those developing the documentation required by document 1910.119 including processing instructions, hazards analysis, and safety features and procedures. There is also a provision stating that trade secrets should not prevent employees from receiving or having access to any relevant safety information or documentation provided under document 1910.119. The employer may request confidentiality agreements to be signed before such information is disclosed. 5.0 Problem Analysis and Design Synthesis

5.1 Accelerometer Sensor System Design The Accelerometer Sensor System is composed of a sensor fixture, diaphragm and accelerometer. Problems that arose during accelerometer system considerations included safety, electrical isolation, diaphragm material selection and contour.

5.1.1 Safety Concerns Safety concerns are addressed using a Certified Intrinsically Safe accelerometer that includes a sealed integral cable. Electrical isolation will be achieved using an insulated cable connected to a power supply mounted outside the test enclosure.

5.1.2 Diaphragm Material Selection To choose an appropriate material for the diaphragm, characteristics of elasticity, damping, and durability will need to be compared versus the price and fabrication of each material. Initially, a plastic material was considered optimal due to its low modulus of elasticity. Several samples of


common plastics were tested in the Vibration lab under uniform conditions [Appendix G]. Later, two samples of aluminum with different thicknesses were also tested in a similar fashion.

Material Thickness, in (mm) Material 1 Low Density Plastic 1 .037 (.939) Material 2 Low Density Plastic 2 .069 (1.75) Material 3 Plexiglass .116 (2.95) Material 4 High Density Plastic 1 .078 (1.98) Material 5 High Density Plastic 2 .061 (1.55) Material 6 6061 Aluminum .024 (.610) Material 7 3000 Aluminum .005 (.127)

Table 5: Materials Tested for Diaphragm

An initial proof-of-concept experiment was carried out with assistance from Dr. Kochersberger. A 100mV/g accelerometer with a mass of 2 grams and frequency range to ten thousand hertz was used to record the impact from a water drip. The accelerometer was mounted to the inside bottom of a clear plastic cup. The cup was then inverted, and water drips were released at a height of two inches above the surface. Using the OROS software, waves were captured at frequencies of two hundred hertz and ten thousand hertz. At two hundred hertz, a shockwave with a maximum peak of approximately .300 g was recorded. The wave damped out in approximately fifty milliseconds. At ten thousand hertz, a shockwave with a maximum peak of approximately .750 g was recorded. This wave also damped out in approximately fifty milliseconds. Tests were also done with previous drips forming a pool on the surface. The responses had less amplitude and in some cases failed to trigger the software recorder. The experiment succeeded in proving that the concept was valid. It also became apparent through this experiment that the choice of material, size, and shape for the diaphragm would be crucial in producing reliable results during repeated use of the sensor system. Diaphragm material selection will be based on the criteria of responsiveness and quick decay of transients. For the accelerometer to detect the drip impulse clearly, the diaphragm must be composed of a low mass, lightweight material featuring high stiffness and resilience to fuel. This will allow a clean detection of the impulse with damping characteristics allowing any ensuing transients to decay in a short period of time. With the accelerometer attached to the underside of the diaphragm, it is essential that the mounting adhesive used must be resilient in the fuel environment and adhere to the selected material. Experimentation was conducted that created a mock up of the current fuel injector test system. An array of materials were tested to determine which combination of stiffness and damping ratio would provide the best accelerometer response and detection resolution. Results from experimentation showed that 3000 series aluminum with a .006” thickness would be the best diaphragm choice. Calculated natural frequencies of 196 Hz (Analytical), 125 Hz (Simulation), and 134 Hz (Experimental) and damping ratio of .0459 were determined for the material. With this information, a decay envelope plot was created to determine material response to a unit impulse. It was determined that the sensor package could accurately detect up to 10 drips per second, as the system was able to damp a unit impulse to less than 3% within 100 milliseconds.


0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

time, sec

g

Decay Envelope

XaXIdeasXexp

Figure 5: Decay Envelope - 3000 Series Aluminum

5.1.3 Diaphragm Contour Analysis

Preliminary testing of various diaphragm materials has been conducted to determine contour properties. Characteristics of interest were response and transient decay to in impulse input. The first sets of tests were conducted to simulate a contoured diaphragm shape. This was accomplished by cleaning the test surface before each successive drip. The following test sets allowed the drips to puddle on the diaphragm surface in a progressive manner. This was done to simulate a non-contoured diaphragm and determine the need for a spherical shape or base fixture offset. Conceptualized designs of three different shapes for the accelerometer diaphragm were modeled to evaluate the modal characteristics of each. Using properties of 3400 aluminum alloy, the first six modal shapes were evaluated for two of the designs. The third design was more complicated and only the primary mode was modeled using this same material. Through the accelerometer testing and the modal analysis, graphs revealed that some of the second or later modes may have excited due to the location of the drip impact on the diaphragm.

Figure 6: Square Diaphragm Modal Analysis


Figure 7: Circular Diaphragm Modal Analysis

The first diaphragm design was a 2-3/10” square and 0.005” thick. The second diaphragm was a disc design with a diameter of 3 ½” and .006” thick. At a mesh element length of 0.1”, the following natural frequencies were found for these two designs.

Mode Diaphragm 1 (Hz) Diaphragm 2 (Hz) 1 163.03 125.22 2 322.86 247.00 3 346.65 247.85 4 565.35 464.9 5 641.54 469.03 6 681.25 561.92

Table 6: Modal Analysis Results-Non-Contoured Diaphragm The third diaphragm design was a dome construction surrounded by a flat border. The dome had a height of ½” and diameter of 3 ½”. The border width was ½”, and a thickness of 0.005”. The model was meshed with element lengths 0.1” and 0.08”. Both of these simulations resulted in a first mode frequency greater than 9000 Hz. Since the expected frequency range of the accelerometer to be used in the final design would have a maximum value of 10000Hz, it did not seem logical to evaluate this design at any higher modes.

Figure 8: Contoured Diaphragm

Based on the preliminary brainstorming, it was hypothesized that the diaphragm would need to be dome shaped to provide a clean surface for successive drips in order to eliminate possible damping by drip accumulation. Following more organized and rigorous testing, the data revealed that pooling on the surface of the diaphragm would prevent later drips from being detected. Also, the increased cost of manufacturing of a dome shaped diaphragm leads to the conclusion that a flat diaphragm design will be sufficient for this project. Experimentation was conducted regarding an offset position that places the diaphragm at an angle. This would allow successive drips to disperse away from the impact concentration and eliminate the build up of fuel on the surface of the diaphragm. Results from experimentation showed that a 7.5o offset would allow the drips to run to the low side of the diaphragm and not impede the impact of successive drips.


Figure 9:Contoured Diaphragm Modal Analysis

The distance between the injectors regulates the size of the diaphragm. The diaphragm could have a diameter as large as five inches, but the design must also include a proper base. Therefore, the expected outer diameter of the base will be four inches with a fastening ring in place, sandwiching the diaphragm between the base fixture and fastening ring. The effective size of the diaphragm design is 3.00” diaphragm catch basin, a 3.5” outside diameter and 7/32” thickness of the base and ring.

5.2 Infrared Emitter/Detector Analysis Testing carried out has shown this concept able to detect a drip of water at several distances separating the emitter and detector. With the use of a simple circuit, a voltage spike of 0.5V was repeatedly observed when the emitter and detector were placed at a distance of approximately one inch. A more complex circuit was created to amplify the output voltage. The second test had an approximate distance of 12” between the emitter and detector, and averages of 1.8-volt increases were recorded. The basis of this design relies on the ability of an infrared beam to remain undisturbed as it passes through the test enclosure, is reflected off of a precisely positioned mirror, and passes outside the box where it would be sensed by a detector. As previously stated, the primary advantage of this concept is that all electrical components of the system could remain outside of the enclosure. Concerns due to the infrared beam’s alignment include the diffractive properties of the window and signal degradation are discussed in the following. The second section covers other concerns such as the distinction of a drip versus the fuel spray and the width of the infrared beam to detect all drips from an injector. While these sections may attempt to justify or alleviate those concerns, further testing will be required to refine this sensing system.

5.2.1 IR Emitter/ Detector Location and Positioning By use of an optical detection system, the fuel saturated test enclosure will remain free of any electrical instrumentation minus the fuel injector command input. While this sensor concept eliminates the need for holes through which wires would need to pass, it will require the addition of windows to the design of the test enclosure. As long as the natural frequency of the Lexan window and infrared frequency are similar, there should be no interference in the beam passing through the exterior of the test enclosure. To limit circuitry to one side of the stand, the final sensor design will contain an emitter and detector adjacent to one another. This will create the need for a reflective surface inside the test enclosure. The alignment of the emitter, detector, and


reflective surface will be accomplished through careful measurements and an integrated light emitting diode (LED) that will increase in brightness as alignment is achieved. Throughout the course of testing, a layer of fuel may build up on the surfaces of the mirrors and windows. The result of this additional fluid layer through which the infrared beam will have to pass can cause degradation of signal intensity. To overcome the loss in signal strength, it will be necessary to include an automatic gain control circuit into the design. The circuit will monitor and normalize input voltage to the reference voltage during the test.

5.2.2 IR Emitter/ Detector Accuracy The sensor must also distinguish between the drip and the fuel vapor pulsing from the injector tip. While the difference in diameter between a drip and a fuel spray particle is extremely large, on the order of 106, the high concentration of fuel spray particles creates a situation where the infrared system may be erroneously triggered. To counter this concern, the selection of an infrared intensity will be crucial. It must be high enough to pass through smaller fuel spray particles but not so high that a drip will go undetected. If testing reveals that this goal intensity cannot be obtained, it may be necessary to redesign the test enclosure so that the infrared beam passes perpendicular to the spray path or passes well below the injector tip where the spray will be more diffused. When the height of the infrared beam from the injector tip is found, there is still the challenge of aligning the beam with the drip location. The final selection of the infrared system will need to be such that an emission angle from the emitter is wide enough to cover the width of the injector tip but narrow enough to sense the interruption caused by the drip. This will also be dependent on the intensity and beam diameter.

5.3 Data Acquisition Initially included as part of the project scope, the sensor system shall communicate the time of a drip through the use of LabVIEW software. The team is therefore required to integrate the signal output from the sensing system into LabVIEW. The signal consisting of either analog voltage or a digital pulse will be read by LabVIEW and compared with a preset limit. Before the signal reaches the computer, it first must be amplified, filtered, and sent through circuitry and cables. The actual details about these will be dependent on the sensing system. The output from the accelerometer sensing system will be an extremely low voltage signal. Low noise cables will be needed to keep the signal free from noise before it enters the conditioner module. In the conditioner module, the low voltage signal will be amplified and filtered with onboard circuitry. The module may require a chassis box depending on the supplier chosen for the signal conditioner. From the chassis, the signal will pass through a cable that will be fastened to an input card in the computer. Through use of National Instruments software, the signal from each accelerometer will be designated a channel number. The infrared sensing system will use a different signal conditioning components to transmit the output signal to the data acquisition system. At this stage of the design, it is proposed that a printed circuit board will be attached to the test box in order to combine the separate signals coming from each sensor system. A coupled pair of an emitter and a detector for each fuel


injector will send their signal through the circuit to a microprocessor. The processor will combine the signals from each detector and transmit the data to the computer via an RS-232 cable. This cable will mate up with a serial port on the rear of the computer. From there, the National Instruments software will interpret the data and divide it into corresponding channels for each injector. The LabVIEW software will begin the fuel injector testing by first allowing the operator to initiate a purge cycle for all of the injectors. This cycle will carry out a process to run the injectors for a predetermined time and then terminate. Upon input from the operator, the test subprogram will begin its process. The output from the program will be the existing frequency command to the fuel injectors that is currently used with the current test stand. The command cycles the injector through a pulsing sequence at a set frequency for ten minutes. This command will have to be duplicated for the multiple injectors in the new test box. The input from either sensor system will communicate with LabVIEW through analog or digital channels. Setting initial time to zero at the onset of the test, the program will create an individual waveform for each input signal. A positive limit level will be compared to the input signal and will initiate a timestamp when the limit is first crossed in elapsed time from the beginning of the test. The program will then delay reading of the sensor for approximately a half second to allow the accelerometer or infrared detector to reestablish unexcited equilibrium. At later drip detection, the program will repeat the process of taking a timestamp and then waiting for the signal to stabilize before analyzing the signal again. The LabVIEW program will also have front panel inputs for type of and information for each injector. Upon completion of the ten-minute test, the operator will be prompted for location and filename to save the data from the test. Through the use of an Excel macro, it is proposed that the file will be opened, the data charted, and the file resaved in Excel spreadsheet format. An example of the excel file content is shown below.

Table 7: Excel Test Report As shown, the first cell in the worksheet will contain the date and time of the test. For each injector, the model and information about the test will be shown followed by the timestamps in seconds for each drip.

Injector Tip Drip SummaryTest Description:

EWO# Part No.: Fuel Type:Date: Driver Type: Fuel Pressure: 380 kPa (rel. to chamber)

Requester: Pulse Width: 2.5 ms Injector Angle: 45 degreesOperator: Pulse Period: 100 ms Connector Angle: vertical, up

Fuel Press. Cham. Vac. 1st Drip 2nd Drip 3rd Drip 4th Drip 5th Drip6 380 0 01:15 01:32 01:57 02:21 02:44 2527 380 0 01:17 01:29 01:57 02:21 02:46 2320 380 0 01:22 01:33 02:04 02:30 02:55 2223 380 0 01:19 01:33 02:02 02:27 02:51 243 380 0 01:33 01:50 02:20 02:48 03:15 21

M3.5 Development Tip Drip Phenomena Investigation: CARFG-2 drip compare to stoddard testing in injector lab, 4HDSFD821FG1 CaRFG-2 (aka CARB Phase II)27-Oct-03 LIF / PFI JustinOtto Muller-GirardLee Markle

>40 mg fuel collected from tip after 10 minutes

P/N S/NSpray Conditions Injector Tip Drip Performance

Time to Drip Total Drips in 10 min.

Comment

>40 mg fuel collected from tip after 10 minutes25 mg fuel collected from tip after 10 minutes20 mg fuel collected from tip after 10 minutes31 mg fuel collected from tip after 10 minutes


5.4 Test Enclosure Design The design team is required to design an explosion proof, quiescent test chamber that would house both the sensor system and fuel injector positioning system. The enclosure should allow operator access to test samples, positioning system adjustment, electrical supplies, and fuel connections. The test enclosure is 34” long x 22” wide x 17.5” tall, with a 30” x 14” x ½” thick Lexan viewing window. This size would accommodate the required range of motion of the injector positioning system, as well as keep fuel spray away from the Lexan window. The complete front panel is removable, and allows the operator full access to the inside of the enclosure. A large O-ring type gasket is adhered to a recess in the back of the panel to ensure sealing. This design also ensures sealing of the Lexan window, as the viewing panel is undisturbed during fuel injector change over.

5.4.1 Structural Analysis Structural analysis needed to be performed on test enclosure proposals in order to determine wall material, wall thickness, and vision window size and location. A finite element analysis was performed using Solid Works Cosmos Express software. The analysis took into consideration different configurations of the test enclosure to design a safe, reliable, and efficient operating enclosure that took into account customer requirements and material efficiency (Appendix H).

Figure 10: Test Enclosure Structural Analysis - Design 1

The structural analysis focused attention to structural stress associated with testing under vacuum. In this situation, a pressure of one atmosphere acts on the outside walls of the test enclosure. A material thickness too low would allow the walls to buckle. If a material with large thickness is selected, unnecessary expenses are incurred.


Figure 11: Test Enclosure Structural Analysis with Center Support – Design 2

Structural Analysis of the first test enclosure design, without a center support, yielded unfavorable stresses and deflection of enclosure walls. The maximum von Mises stress was 8461 psi. A second analysis was performed after adding a ½” thick aluminum plate that divided the enclosure at the center. The analysis showed a dramatic decrease in stresses at the edges of the window opening, as well as a decrease in the deflection of the back wall. The maximum von Mises stress was 3909 psi. The design provided a factor of safety above 2.0. With the center panel added for strength, it will require two injector positioning systems to be built. The more complicated design will increase the price of the enclosure, but the increased amount of safety now present outweighs the economic downside.

5.4.2 Pressure Relief System Design Per customer request, a pressure relief system needed to be designed in order to vent a pressure buildup due to enclosure content ignition. Through customer suggestion, it was required that the relief system be a removable panel that was at least 60% of the surface area of the top panel. This design would allow the unfixed top panel to allow any ensuing pressure build up to be relieved through blow off, similar to the design of an explosion proof room, in which roof panels are designed to blow out.


Figure 12: Test Enclosure Loading

Figure 13: Test Enclosure and Injector Positioning System

5.4.3 Fuel Injector Adapter Design To keep sensor position consistent and eliminate the need to reposition sensor assemblies when changing fuel injector types, a fuel injector adapter system was hypothesized. A set of adapters that accommodates 80% of Delphi fuel injectors was needed to position the fuel injector tip at the same location regardless of length. To accomplish this, a set of adapters for each type of fuel injector was developed. Three different length adapters were conceptualized to position the fuel injector tip at the positioning system axis of rotation. The adapters would quickly thread into the positioning system at one end, with quick disconnect fuel fittings. At the opposite end, injectors would attach utilizing the same clip-style technology that affixes the injectors to vehicle fuel rails. To ensure that the adapters would withstand the high fuel pressures associated with injector testing, finite element analysis was conducted to verify material thickness would withstand internal pressures [Appendix I]. Results verified that a minimum wall thickness of 0.25” would


withstand internal fuel pressures up to the specified maximum 87 psi fuel pressure. A maximum stress of 430 psi was determined from the finite element analysis. A safety factor of over 22 was determined for each of the adapters. This would allow a significant decrease in material thickness, but concerns with machining and material availability justified the added wall thickness. 6.0 Sensor System Conclusions The design team was forced to make a decision between the accelerometer sensor system and the infrared detection system. The team weighed a number of factors into the decision that included cost, availability, support, and experimental results. The accelerometer system provided a robust solution to the problem. The cut and dry theory behind the system made it the top candidate for the project. The proposed system was ready to be built with components previously quoted and for the most part, readily available. The suppliers that had been contacted provided continuous support and feedback, as well as future support regarding calibration and maintenance. The experimental results from testing provided a sound foundation to build from. The system also included a return on investment of 7 months. The infrared system was discarded due to a number of outstanding questions. This included mounting procedures, calibration, and commercial availability of sensor pairs. Troubles arose when possible mounting procedures were questioned. If the sensor package were mounted outside the test enclosure, calibration would have to be completed before each test run to ensure that the emitter and detector aligned. Also, IR emitter and detector pairs capable of spanning the depth of the enclosure were incapable of detecting particles as small as the fuel drips. If the sensor package was mounted inside the test enclosure, fuel vapor would trigger a false read. Build up on lenses and mirrors would also have posed detection problems. With these considerations in mind, the accelerometer sensor system was chosen after a meeting with the Spray Lab team at Delphi. The meeting laid out each sensor system with a detailed Bill of Materials. The Delphi team felt that additional costs of the accelerometer sensor system were justified based on system performance during testing.

7.0 Detailed Design

7.1 Accelerometer Sensor System The sensor system is composed of (6) accelerometers, (6) diaphragms, (6) annular clamp-rings, (6) base fixtures, (6) accelerometer cables, and (1) signal conditioner. A 100 mV/g accelerometer is securely adhered to the center of a 3.50” (88.9mm) diameter, 0.006” (0.15mm) thick aluminum diaphragm. The diaphragm is clamped at the outer boundary between a 0.25” (6.35mm) tall x 0.216” (5.49mm) thick x 3.50” (88.9mm) outer diameter annular clamp ring and 2.00” (50.8mm) tall hollow cylinder base fixture.


Figure 14: Sensor Assembly Cross Section

The accelerometer cable is sealed to the top of the accelerometer and routed through the backside of the base fixture. One assembly is positioned 4.00” (101.6mm) below each of the 6 fuel injector tips.

The signal conditioner supplies the accelerometer with a constant 5.0 volts. This provides the reference voltage for drip detection. Upon impact, the falling drip will excite the diaphragm with an impulse force. The acceleration of the diaphragm will be directly transferred to the accelerometer. This will cause a change in the piezo-resistive properties of the accelerometer and return a voltage spike to the signal conditioner.

Figure 15: Accelerometer Fixture Assembly

7.1.1 Design The sensor system is required to detect drips from 6 separate fuel injectors throughout the course of a 10 minute test cycle. To accomplish this, the package needed to be responsive to the fuel drip and capable of detecting multiple drips throughout the test cycle. A lightweight accelerometer was used to keep system mass as low as possible. A PCB ceramic shear, internal conditioning, 100 mV/g model 12605 with mass 2 grams was selected.


Figure 16: PCB 12605 Ceramic Shear Accelerometer

The diaphragm is a key component of the sensor outfit. A diaphragm with a large mass and high stiffness would not transfer energy to the accelerometer. If the diaphragm failed to quickly damp the oscillation, a drip may be missed. With this in mind, diaphragm material selection was based on characteristics that included response magnitude and decay envelope. A natural frequency of 1232 rad/s and damping ratio of 0.0459 were determined for the circular aluminum plate.

The aluminum plate was able to provide the required responsiveness as well as damp a unit impulse to less than 3% in 100 milliseconds. This provides the sensor system with the capability to accurately detect up to 10 drips per second.

To address pooling of successive fuel drips on the top of the diaphragm, a 7.5o offset was implemented into the bottom of the base fixture. This allows the fuel to run to the low side of the diaphragm without impeding the impulse force of the subsequent drips. A 0.25” (6.35mm) wide x 0.100” (2.54mm) tall channel milled into the bottom face of the annular clamp-rings allows fuel to run off the top of the diaphragm and down the side of the base fixtures.

7.2 Data Acquisition System

7.2.1 Fuel Injector Control To cycle the fuel injectors, a digital signal is created in LabVIEW with selectable cycle period, injector on-time, and total test time. These values are controlled through the front panel display of the LabVIEW program before the operator begins the test. Due to two types of solenoids being tested at Delphi, a switch is also present to invert the signal sent to the injectors. To prepare the injectors for testing, a purge cycle is initialized. The separate panel on the program’s display is present for the shorter purge cycle. The operator can enter a different period width, injector on-time, and total cycle duration for the purge cycle. A button is also present to allow the operator to create a constant injector spray. During the purge cycle, the controls for starting the test cycle are disabled to prevent the operator from trying to start the test cycle until the injector purge is complete. The test cycle has independent controls for the settings of period, injector on time, and total test duration. To generate a steady and constant output, the LabVIEW program generates a square wave array from zero to one. The frequency and duty cycle are controlled by the values that the operator inputs for period and injector on time. This data array is sent to the digital output control that generates a five-volt pulse via the internal PCI card and external BNC board. Due to


the lack of a driver box that will divide the digital signal into six separate 13 volt lines, the program currently uses the existing test program and ISA card to drive the fuel injectors. This card is attached to Delphi’s “Justin” box that has outputs for the fuel injectors. To address safety concerns, an emergency stop button will be placed on the outside of the system control tower. In the case of a need to immediately turn the system off, the emergency stop will cut power to the complete sensor system. The fuel injectors will return to the closed position at 0 volts, and fuel will no longer be supplied to the test chamber.

7.2.2 Sensor Input

The LabVIEW program detects drips using the sensor system input. During system testing, a maximum sensor input of 100 millivolts occurred. This voltage is based on the output sensitivity of the accelerometer. The program reads the accelerometer voltage at a frequency of 1000Hz with an internal memory buffer of 2000 samples. After a fuel drip excites the sensor system, the program executes a time stamp by changing the case structure loop to a true case as the threshold voltage is exceeded. The current elapsed test time is added to an array of previous time stamps separated by a tab constant. This time stamp array remains unchanged through the utilization of a shift register on the outer while-loop of the program. The array passes directly through the inner case structure loop when the limit has not been exceeded. The program continues to monitor the input signal from the accelerometers until the elapsed test time matches the total test time selected by the operator.

Figure 17: LabVIEW Operator Interface

The data acquisition system also has the capability to begin monitoring drips after a predetermined test cycle time has elapsed. The system allows the operator to control the delay until drip detection begins. The fuel injectors will cycle during the prescribed delay time and


any detected drips will be ignored. During this, the elapsed test time is compared to the delay time and passes a true statement to the stop terminal of the DAQ sub-VI when the delay time has not passed. When delay time is surpassed by elapsed test time, the statement changes to false, informing the DAQ sub-VI to begin monitoring the input signals. At the bottom of the LabVIEW front panel, a graph of the scaled sensor input versus time is created. This provides a check for the operator to compare the displayed array times to the actual sensor input. The impact of a drip can be seen on screen as a spike in output voltage. The graph can be switched to display each of the six drip sensors.

7.2.3 Data Reporting

Information boxes on the front panel of the LabVIEW interface allow the operator to enter data for the test being performed. Items included are filename, test description, EWO designation, test requestor and operator, fuel type, fuel pressure (kPa), chamber vacuum (kPa), fuel injector orientation angle, and electrical connector position. For each fuel injector, individual part number and serial number fields are available. When the test is completed, the data from these boxes along with the drip time array for each injector are exported to a preformatted Excel sheet. If the test description field or the EWO numbers are left blank, the operator will be prompted for this information before the data is saved under the filename path. The file will be saved in the Excel format. The process of the program is carried out through a sequence structure in LabVIEW and several sub-VI’s supplied by Delphi.

7.2.4 DAQ Hardware The data acquisition system, or DAQ, is compromised of several different components for the output and input from the LabVIEW program. The output from LabVIEW is currently sent to an internal ISA card in the computer. This ISA card communicates with a driver box known as a “Justin” box. The driver box divides the command signal into separate channels for each fuel injector. These channels are connected to the fuel injectors through wires fed into the test enclosure through rubber stoppers. They are then attached to the fuel injector through a common Delphi supplied plug. This completes the output portion of the DAQ. In the future, the signal from LabVIEW will be sent to the PCI card currently used for input of the accelerometer signals. The digital pulse signal will be relayed to the external BNC board that will produce a zero to five volt TTL signal. This pulse signal will be fed into a driver box that has yet to be manufactured at Delphi. It is assumed that the driver box will have individual switches to enable or disable the output to each fuel injector. From the driver box, the signal will continue to the fuel injector as in the previously described setup. For the input signal, the voltage source originates in the signal conditioner box. Each accelerometer is attached to a port on the signal conditioner using a coaxial cable. In addition, the signal conditioner also outputs a voltage to the BNC connector board. Each of the six channels occupied by the BNC board relay the voltage from the designated accelerometer. The BNC board is attached to the computer’s internal PCI card through a 68pin cable. LabVIEW amplifies and monitors the voltage from these channels by means of its DAQ Assistant sub-VI.


Figure 18: DAQ Hardware Flow Diagram

7.3 Test Enclosure

The design of the test enclosure had to ensure compliance with a number of key requirements. First, the enclosure was designed to be explosion-proof. This is in a sense that, if ignition of fuel were to occur in the test chamber, the top of the enclosure will release internal pressures towards the back of the enclosure to protect the operator. The test enclosure is also capable of withstanding a full vacuum load. This allows for fuel injector testing in an environment similar to an intake manifold setting, as well as evacuation of hazardous fuel vapors. The third requirement of the test enclosure is to incorporate an injector positioning system that would allow drip testing throughout an orientation range of 0o-90o. The enclosure is also required to accommodate all pertinent injector fuel and electrical supplies, data acquisition apparatus, vacuum, and nitrogen purge services.

7.3.1 Test Enclosure The test enclosure is 34” wide x 21” deep x 17.5” tall. It is constructed from 0.5” (12.7mm) thick 6061 aluminum walls that are welded at the seams. The walls are reinforced with 1.00” (25.4mm) square 6061 aluminum tubing with 0.125” (3.175mm) wall thickness. The size of the enclosure was determined based on the size of the injector positioning system, full operator access, and to ensure that fuel spray cones do not intersect. The center span is reinforced with a 0.5” (12.7mm) thick 6061 aluminum load bearing plate positioned at the center of the enclosure. The reinforcements were made to ensure that the enclosure is capable of withstanding the high external pressures associated with vacuum loading. A finite element analysis was performed on the test enclosure to determine structural stresses. A full vacuum load was simulated using COSMOS FEA software. The analysis was performed to ensure that the enclosure was capable of withstanding vacuum load and the appropriate material and thickness was selected in the most cost efficient means. The enclosure design provided a factor of safety of 2, with negligible displacements at the top panel and back wall. A full vacuum load is capable based on results from the FE analysis. The small displacements seen ensure that sealing of the top plate and front panel is accomplished. The removable front panel provides the operator with full access to the test chamber. This design also allows for quick and simple set-up. During testing, the front panel is affixed to the test enclosure using Steco clamps. Vacuum sealing is achieved using a large O-ring that is


adhered to a recess in the panel. A polycarbonate viewing window allows the operator to monitor drip testing.

Figure 19: Test Enclosure and Control Cabinet

Threaded fuel supply connectors were machined to connect services inside and outside the enclosure. The outer diameter of the connectors was threaded into the ¾”-10 tapped holes in the back of the enclosure. 1/4”-NPT male ends of the connectors allowed for a barbed fitting on the inside end and quick disconnect fitting on the outside end of the connector. Fuel injector and accelerometer electrical services were routed to the inside of the test enclosure through holes in the back panel. Rubber vacuum plugs were hollowed to allow the two fuel injector wires and one accelerometer wire to pass through each. The combination of interference fit and vacuum forces causes the rubber plug to compress around the outside diameter of the wires and form an air tight seal.

7.3.2 Injector Positioning System The injector positioning system utilizes (2) fuel injector fixtures. Each fixture, mounted on separate sides of the center wall, holds (3) fuel injectors, (3) fuel injector adapters, (3) sets of fuel supplies/ electrical services, and a locking system to hold each injector fixture at the desired orientation. Orientation changes are supported by a spring loaded locking system that allows changes to be made in 5o increments. The orientation change is made by pulling the pin from the locking plate and rotating the injector positioning fixture to the desired angle. The pin releases into the correct dial plate hole and locks the system at the desired orientation. To keep set-up time at a minimum, the injector positioning system was designed to place the fuel injector tip at the same point for all injectors. This requires no repositioning of the sensor assemblies for testing of different injector models. To accommodate 80% of Delphi fuel


injectors, an adapter system was developed. The adapters are threaded into the top plate of the positioning system. Fuel services are connected at the top of the adapter. The injectors are mounted at the opposite end of each adapter using the same clip-style technology that is used to affix the injector to the fuel rail in vehicle applications.

Figure 20: Injector Positiong System

8.0 Results System testing included procedures to ensure that design goals and objectives were met. This included sensor repeatability testing, LabVIEW program testing, and vacuum testing of the enclosure. The accelerometer sensor system was capable of detecting fuel drips repeatedly. Testing at Delphi included positioning the injector at 45o and running a number of 10-minute tests. Throughout each test run, the sensor assembly successfully reported every drip occurrence to the data acquisition system. The base offset also forced fuel to disperse from the center of the diaphragm, providing a clean surface for the subsequent drips. The data acquisition system testing required some troubleshooting. After performing test runs, it became apparent that input threshold values needed to be adjusted. A value of .04 mV was determined based on inherent system noise and the magnitude of accelerometer input. This change ensured that false drip detection did not occur. Problems also arose when drips were detected and then reported twice by the data acquisition system. Accelerometer input crossed the predetermined threshold twice during impulse decay. This triggered the system to report 2 drips. To address this, a DAQ sampling delay was implemented for 0.5 seconds after the initial threshold crossing. This gave the diaphragm enough time to return to steady state before sampling resumed, thus resolving the problem of false drip reporting.


The test enclosure was positioned on a moveable cart in the spray lab. Appropriate fuel, electrical, and vacuum services were attached to the system. The system was then tested for vacuum and fuel leaks. After making the first vacuum test, it became apparent that the deflections of the Lexan viewing window allowed air to leak past the seal. To address this, a clay fillet was positioned at the seam where the window met the aluminum front panel. This resulted in an air tight seal around the viewing window capable of withstanding the largest vacuum load the Delphi spray lab is capable of. A vacuum gauge attached to a fitting on the enclosure reported an absolute internal pressure of 10 kPa.

1. Fit into assigned workspace

2. Withstand full vacuum

3. Capture entire range of drips @ 45o injector angle

4. 4 to 6 fuel injectors to be tested at once

5. Setup time less than 10 minutes per test per injector

6. Programming in LabVIEW

7. Adhere to standards and codes pertaining to Delphi considerations for drip test stand (section 3.9)

8. IPS to accommodate orientation range (0o–90o)

9. Automated drip sensing

10. Explosion proof quiescent test chamber

11. All drawings in .dwg format

12. Under Delphi budget of $7000.00/ Return on Investment under 6 months

13. Fuel resilient exposed components

14. Resolvable error detection (distinguish fuel drip from vapor)

15. Sensor repeatability

Table 8: Completed Design Objectives Checklist A look into the Completed Design Objectives checklist shows that the team was able to provide an automated solution to the unique problem of detecting fuel drips from injector tips. Major design accomplishments include unmanned drip sensing, decreased set-up time, ability to test multiple fuel injectors simultaneously through a 90o range, and test enclosure capable of withstanding full vacuum load.


The team failed to provide a design solution that complied with the approved $7000 budget. Total system cost was $8613, with a return on investment of 7 months based on a cost breakdown analysis [Appendix J]. This is one month over the required 6-month payback period. A detailed Bill of Materials is provided [Appendix K]. Included are all drip test stand assemblies and the corresponding components. The Bill of Materials provides component descriptions, quantity, part number, supplier name, supplier representative, supplier contact information, price per component, estimated cost and actual cost. System totals are also provided in a manner that divides the drip test stand into the three major assemblies. 9.0 Future Considerations Throughout the build process, a number of considerations for drip test stand improvement were hypothesized. This includes revisions to the test enclosure, fuel delivery system, injector adapters, and injector input sequence. The revisions would allow for more flexibility in the drip test stand system, as more operator freedom would be allowed. Upgrades to the test enclosure include changes to the fuel delivery system and the addition of a quartz viewing window. A fuel delivery manifold capable of individual fuel pressure adjustments would allow the operator to simultaneously test a set of fuel injectors at different individual fuel pressure. By replacing the Lexan viewing window with a quartz panel, the use of clay to seal the mating surfaces would not be needed. An appropriately sized quartz panel would not deflect as much as the Lexan window, and allow the mating surfaces to form an air tight seal. By increasing the quantity of fuel injector adapters to accommodate 80% of Delphi injectors, the spray lab would have the ability to test the remaining Delphi injectors and those of other manufacturers. The addition of a digital driver box would take advantage of the LabVIEW drip timing algorithm provided with the data acquisition system. Delphi is provided with the capability of tuning individual fuel injector pulse sequences through the LabVIEW user interface. On screen adjustments allow the operator to test a set of injectors at more than one injector on-time and period.


10.0 References Brian K. Thorn. Personal Interview. 6 February 2004. Bowman, Robert. Electrical Engineering Sophomore Practicum Reference Text. Dept. of

Electrical Engineering @ Rochester Institute of Technology. 2004 Dr. Kevin Kochersberger. Personal Interview. 27 January 2004. Dr. Robert Bowman. Personal Inteview. 2 February 2004. Markle, Lee E. Delphi Drip Test Stand Considerations. Delphi Technical Center Rochester:

2004. Omar Anbari. Personal Interview. 22 Jan. 2004

OSHA website, OSHA.gov. Feb. 14, 2004.Occupational Safety & Health Administration, 200 Constitution Avenue, NW, Washington, DC 20210. <http://www.osha.gov/pls/oshaweb/owadisp.show_document?p_table=STANDARDS&p_id=9760>

OSHA website, OSHA.gov. Feb. 14, 2004.Occupational Safety & Health Administration, 200

Constitution Avenue, NW, Washington, DC 20210. http://www.osha.gov/pls/oshaweb/owadisp.show_document?p_table=STANDARDS&p_id=9763

Texas Instruments Real World Signal ProcessingTM. Interface Selection Guide. 9 Jan. 2004. 9

Jan. 2004 <http://focus.ti.com/lit/ml/sszt009/sszt009.pdf> Travis, Jeffrey. LabVIEW for Everyone Second Edition. New Jersey: Prentice

Hall PTR, 2002


Appendix A. Concept Development Documents

1. Weight of one drip 14. Elec. Properties of tip 2. Collecting total drips 15. Motion detector 3. Photograph 16. Change in spray pattern 4. Vision system 17. Strain gauge on F.I. 5. Laser refraction 18. Paper roll/ drop counter 6. Reflective properties of tip 19. Pressure sensor 7. Photon Sensor 20. Temperature change 8. Accelerometer 21. Paddle wheel sensor 9. Surface characteristics 22. See-Saw sensor 10. Unskilled labor 23. Combustion approach 11. Measure weight of injector 24. Wave sensing system 12. Vibrating drum 25. Battery acid technique 13. Length of injector

Concept Development Long List of Concepts


Appendix B. Accelerometer Sensor System Proof of Concept

Purpose: Prove through experimentation the validity of the accelerometer concept. Through faculty assistance, the experiment will prove that a noticeable output from an accelerometer will be significant enough to indicate when a drip occurs. Equipment: DAQ: OROS Software System Accelerometer: PCB C66 100mV/g Fluid: Tap water Material: Inverted plastic Silo cup Other: Eye dripper Experimental Setup:

Procedure:

• The accelerometer was attached to the inside bottom of a clear plastic cup. The bottom of the cup was approximately 2” in diameter and 5” tall. The cup was then inverted.

• A special low noise cable was attached to the accelerometer. The other end was connected to the OROS BNC black power strip.

• The OROS software package was used to read the accelerometer output. Dr. Kochersberger changed settings in the software to have the accelerometer read at 200 Hz and 10 kHz.

Accelerometer


• Using the eye-dripper, water was released at an approximate height of 2” above the cup’s base.

• Results:

• At 200 Hz, the software recorded an acceleration shockwave with a maximum peak of approximately .300 g. At 10 kHz, the software recorded an acceleration shockwave with a peak of approximately .750 g. At a higher frequency, an initial small wave could be seen as the drip began its impact on the diaphragm followed by a much larger wave sensing the entire impact of the drip.

• As water began to pool up on the diaphragm, the impact acceleration as seen by the accelerometer was much less and in some cases failed to trigger the freeze frame limit. The waves as seen in the software had an initial high spike which damped out to nothing in an unknown time period.

• This experiment did prove that the concept was valid and that with the correct accelerometer, diaphragm and software, the impact from a drip could be sensed.

• Some concerns that arose from this meeting were: o Background noise and vibration in the lab o Use of proper cables to eliminate as much noise as possible o The use of a fuel resistant adhesive to attach the accelerometer to the diaphragm o Accumulation of fuel on the diaphragm to the extent where subsequent drips

would not be detected


Appendix C. Infrared Emitter/Detector System Proof of Concept

Objective: The primary objective of this experiment was to validate the use of an Infrared phototransistor receiver to create an optical sensing unit that will detect the motion of a 3.75mm diameter free falling drip. The properties of the circuit (emitter, detector configuration) were studied to determine the necessary resistor values and predicted output voltage levels. Theory: The infrared emitting diode will be directed toward the detecting transistor and will act like an input into the base. Current will be allowed to flow through the transistor while the base voltage is higher than the threshold voltage. Thus under normal transmitter to receiver operation, the output voltage will be close to the input voltage. When there is interference in the beam path, the base shuts off the current flow and a voltage drop is created. Equipment: 1 Prototyping Board 1 1k� resistor

2 150� resistor 1 Infrared Emitter and Detector (#276-0142 RadioShack -- $2.99) 1 5v power supply 1 Digital Multimeter

Parameters: The emitter and detector infrared pair were pointed toward each other with a distance of two inches between them. The drip was produced from a handheld eyedropper and positioned to fall between the two diodes onto a paper towel. Procedure: The circuit was constructed as shown in Figure 1 and 2. Figure 3 has been included to show the active diode beam in the visible light spectrum. After the circuit was powered with a 6VDC source (note ratings shown in figure 4), the output voltage was monitored with a digital multi-meter. As each drip fell through the beam path, a constant change of at least 0.2 volts was displayed.

Figure 1: Infrared Emitter/ Detector Circuit Schematic


Figure 2: Infrared Emitter/ Detector Board Layout

Figure 3: Infrared Emitter/ Detector Outfitted with LED

Results: The three liquid mediums used in this experiment were tap water, Coca-Cola, and vegetable oil. The most significant voltage change was found using Coca-Cola (ranging from 0.5V to 0.7V). It was observed that the beam gave the strongest signal when the two diodes were aligned, and the voltage change was related to how precisely the drip broke the beam path. The ambient light was decreased to eliminate noise and produced larger voltage changes for each drip. Also, an operational amplifier could be added to increase the output reading that could then be transmitted into a data acquisition unit.

Type of Fluid Output Voltage Range from series of drips Water (tap) 0.2V – 0.4V Coca-Cola 0.5V – 0.7V Vegetable Oil 0.3V – 0.5V

For this experiment, the response time of the sensor was not analyzed. This will be determined in later experimentation by relating the output to the measured drip height. However, for this experiment the drips were released at an approximate height of two inches above the LED’s.


Figure 4: Infrared Emitter/ Detector Specifications Conclusion: Proof of concept was achieved by this set of experimentation. The infrared optical sensing system displays real world feasibility for detection of a fluid drip. An infrared pair of emitter and detector can successfully detect a freefalling drip. The parameters of the infrared system will be determined through further testing and calculations.


Appendix D. Infrared Emitter/Detector Experimentation

Objective: Evaluation of the infrared optical sensor circuit used in the Electrical Engineering Sophomore Practicum course text for detecting fluid drips. A secondary objective was to determine if this particular circuit would be able to transmit the drip detection through an RS-232 serial communication link. The circuit’s configuration and LabVIEW compatibility will also be determined. Equipment:

1 “Electrical Engineering Sophomore Practicum” text, Dr. Robert J. Bowman 1 Figure 7.4 Complete Schematic of the Optical Communication Link p. 60 (as

reference) 1 Sophomore Practicum supply kit (all necessary components, minus (1x) 2pF

capacitor) 2 Prototyping board 2 3V power supply 1 oscilloscope 1 lab created RS-232 null modem cable (pins 3 and 5 on female 9pin connector)

Experimental Setup:

Figure 1: Schematic from page 60 of Electrical Engineering Sophomore Practicum text


Figure 2: Graph paper with alignment of optical beam and placement of drops. X represents a no – read and

circle represents a positive reading.

Parameters: H1 = 7”: height of dropper tip above the IR path

D1 = 12”: distance between Emitter tip and Receiver tip Output taken from pin post R23 (1k�) on output of Op Amp 2A (TP7 on Schematic)

Procedure: The oscilloscope was adjusted to read the output of operational amplifier two after the 1k� resistor (pin 1). This is representative of the signal that would be passed to the computer via the RS-232 serial cable. The two prototyping boards were assembled with an infrared emitter circuit and an infrared detector circuit. The LED’s were separated by approximately 12”. A piece of graph paper was positioned below the beam path as a reference for the drop pattern. Constant 3VDC power was supplied to each circuit. Measurements were made on the oscilloscope at a magnification of ten. Adjustments set the oscilloscope to measure five volts per division versus ten milliseconds per division. The dropper was held at the constant height of seven inches above the infrared path and each drop was aligned through sight to fall through the beam pattern. The position for the drip’s impact was marked on the graph paper for reference following each test. The following screen captures were obtained for five of the drips during the experiment. Results: The voltage output from the five drips was taken from the oscilloscope.

Figure 1: Vmin = -156.2mV; Vmax = 30v; Trigger Normal; Distance: 11” tip to tip; 10x magnification; 5V increments; 10mSec. period;


Figure 2: Measurement recorded with out use of graph paper as placement reference.

Figure3: Indirect drop pattern fall (without graphical placement reference).

Figure 4: Drop # 3 on graph.


Figure 5: Drop #5 on graph

Figure 6: Drop #7

Figure 7: Drop # 9 Conclusion: It was determined from this experiment that the proper values for an accurate RS-232 level transmission could be attained by passing the drip through the beam. It was also determined that the speed of the drip (or measurement period) and location of the drip are very important in accurately determining drips. The proper beam pattern and drop alignment will be a necessary research area. However this experiment proves that it is possible to detect drips of liquid using infrared optics at distances greater than one inch. At this point, it is still only theorized that the values will be easily passed to the computer via RS-232 level serial or parallel communications. However, with the current predictable output values, it is proven that whether there is a parallel, serial, or microprocessor data link, the values will be able to transmit at a predictable and accurate rate.


Appendix E. Strain Gage Sensor System Proof of Concept

Purpose:

To prove the strain gauge concept by determining whether a strain gauge attached to a material will pick up a voltage change when a small force is applied. Also, to determine the time and effort involved in preparing the material and the mounting of the strain gauge to that material. Equipment: Strain gauge: EA-06-240LZ-120 Test material: Aluminum: Length: 3 ¼ inches (8.2 cm) Non-strain gauge end width: 5/8 inch (1.5 cm) Strain gauge end width: 9/16 inch (1.4 cm) Thickness: 1/8 inch (0.3 cm) Fluid: Water

Power supply: Hewlett Packard E3631A Triple Output DC power supply Multimeter: Tektronix CDM 250 Digital Multimeter Strain gauge application kit: GAK-2-200 Other: Eyedropper, Prototyping board, 3 (100 ohm) resisters, connectors (wire), C-clamp Experimental Set-up:


Procedure: Strain Gauge, Part #EA-06-240LZ-120, was attached to the top of an aluminum test specimen and wired to a three-wire Wheatstone quarter bridge circuit as shown in Figure 1. Steps for preparation and mounting are included in the “Student Manual for Strain Gauge Technology” from the Vishay Measurements Group.

Figure 1 – Quarter Bridge (Electrical Engineering Sophomore Practicum)

Figure 2 – Aluminum Springboard

Surface Preparation • Solvent degreasing

Sprayed CSM-2 solvent degreaser on specimen and cleaned entire surface with a clean gauze pad.

• Surface Abrading While keeping specimen surface wet with M-Prep Conditioner A, 320-grit silicon-carbide paper was used to wet-lap the surface until the surface was clean and bright. The surface was then wiped dry with a clean gauze pad using a clean surface of the pad with each wipe.

• Layout Lines Crossed perpendicular lines were place on the specimen surface to mark the desired location of the strain gauge with a medium hard lead pencil. Conditioner A was applied to the surface and a clean swab was used to scrub the surface until the tip was no longer discolored. The


surface was then dried using a clean gauze pad each time wiping from inside the cleaned surface outwards. The layout lines were very faint after this last cleaning.

• Neutralizing The surface was kept wet with M-Prep Neutralizer 5A and then scrubbed with a clean cotton swab. This is done to provide optimum alkalinity for the adhesives. The surface was then dried using a clean gauze pad each time wiping from inside the cleaned surface outwards.

Strain Gauge Application

The strain gauge was removed from the acetate envelope with tweezers and placed on cleaned empty gauge box with the bonding side down. A four to six inch piece of M-Line PCT 2A cellophane tape placed over the gauge and wiped firmly across. The tape was then lifted at a 30-45 degree angle to avoid over-stressing the strain gauge until completely removed. The strain gauge and tape were placed over the specimen surface so the triangle marks on the gauge were over the layout lines. The tape with the gauge was wiped at a shallow angle onto the surface. The end of the tape opposite the solder tabs was lifted until the gauge was free and placed back over itself. M-Bond 200 Catalyst was applied on the exposed back of the strain gauge in a very thin coat. It was then allowed to dry for one minute. One drop of M-Bond 200 Adhesive was placed at the junction of the tape and specimen. Very quickly, the tape was brought over at a 30-45 degree angle and slowly and firmly pressed and wiped over the specimen surface with a clean gauze pad. Pressure was then applied over the strain gauge area for at least one minute and then left to sit for an additional two minutes. The tape was then slowly pulled back over itself leaving the strain gauge attached to the surface.

To ensure maximum cleanliness, the following should be avoided: o touching surface with fingers o reusing swabs or gauze pads or wiping back and forth o contaminating area by dragging swabs or gauze pads from unclean surface onto cleaned

surface o cleaning solutions evaporating on surface o allowing prepared surface to sit between steps and before bonding

With the gauge in place, single conductor copper wires were soldered onto the solder tabs. These were then soldered to the three-conductor lead-in wire shown in Figure 3. The lead-in wire was then attached to the quarter bridge circuit shown in Figure 4.


Figure 3 - Lead-in wires attached to strain gauge

Figure 4 – Quarter Bridge setup (Wheatstone bridge)

The white wire (second lead) was sent to the Tektronix CDM 250 Digital Multimeter for readout. The other end of the digital multimeter was attached between the two resistors shown in Figure 4. The two wires (red and yellow in Figure 4) were attached to the Hewlett Packard E3631A Triple Output DC power supply and 5 V was sent to the circuit. The sample was held by a C-clamp to a tabletop (Figure 5). A small force, by pressing on the end of the springboard, provided a springboard deflection of about a ¼”.

Black wire (third lead) from gauge into positive side

Red wire (first lead) from gauge into negative side

To the power supply (positive and negative)

Digital Multimeter positive or negative side placement


Figure 5 – Springboard test specimen attached to tabletop

Results: From a deflection of ¼”, the total change in voltage was 3 mV. This experiment proved that the strain gauge concept was valid. The total time to prepare the simple test specimen was 45 minutes. Conclusion:

o Sample that was used for the test was too thick for the strain gauge to sense a fuel drip. o Possible safety issue due to the exposed strain gauge wires. Will need a protective

coating. o Research has shown that the adhesives and protective coatings have a very small pre-

application and post-application life. Loss of fuel injector test time will be a result of this. Replacing the adhesives and protective coatings constantly will cost money that should not have to be spent.

o Deformation of the sample over an unknown time will result in frequent replacement. Preparation time needed to apply strain gauge to the testing springboard unacceptably time consuming.


Appendix F. Feasibility Assessment

Weighted Evaluation Worksheet Baseline Concept: Accelerometer

Attribute Relative Weight

Concept 1 - Baseline

Concept 2 - IR Sensor

Concept 3 - Strain Gauge

Cost 0.19 3 5 4 Skill Requirements 0.14 3 2 3

Setup Time for Sensor 0.24 3 2 2 Performance 0.29 3 3 2

Safety 0.10 3 4 2 Life Cycle 0.05 3 3 1

Repair 0.00 3 2 1

Raw Score - 21 21 15 Normalized Score - 3.03 3.13 2.49


Radar Plot Data

Sensing Option Setup Time Repair Cost R & D Progress Performance

Life Cycle Safety

Accelerometer 1 1 0 1 2 1 2 2 Infrared 2 2 2 2 2 1 2 1 Strain Gauge 1 0 1 2 0 0 1 0

*Note: 0 = low, 1 = medium, 2 = high

Feasibility Assessment Radar Plot

Sensor System Feasibility Radar Plot

0

1

2Setup Time

Repair

Cost

R & D

Progress

Performance

Life Cycle

Safety

Accelerometer Infrared Strain Gauge


Appendix G. Accel. Sensor System Diaphragm Material Exp.

Introduction: An important part of the accelerometer sensing system of project 04014 is the diaphragm to which the accelerometer is mounted. The purpose of this experiment is to determine which conditions will help to optimize the output signal of the accelerometer. To complete this experiment, an accelerometer will be affixed to the underside of a standard sample, supported by a base. Drops of alcohol (used to simulate gasoline) will be released onto the top of the sample and the response will be measured in units of gravitational force (gs). Several factors will be tested, including sample material, sample thickness, height from which the drop is released, and whether or not the sample is cleaned off in between trials. This experiment has been set up as a factorial design and analysis will be completed using Minitab. Regression techniques will also be used to analyze the effects of sample thickness on the response. In addition, the data will be used to determine the decay envelopes of the different materials. This data will be used to confirm the choice of diaphragm material. Purpose of Decay Analysis: To determine the magnitude(s) and damping ratios from a drip impulse based on varying materials, material thickness, pooling effect and drip input height. The data should be recorded in a manner in which it can be analyzed to find the optimal material that will give the desired output magnitude(s) and decay envelope. Discussion of factors: Sample material: Seven samples, including six different materials are each cut into rectangles of approximately the same size. Five of the materials are plastics of varying stiffness and two of the samples are aluminum*. Sample thickness: Each of the plastic samples is of a different thickness and will be analyzed for a trend. The two aluminum samples are of significant difference in thickness and will be compared. This factor will not be part of the factorial design in the data analysis section. Height: Data will be collected for drops released from 2 different fixed heights above the sample. “Clean vs. Dirty”: These two designations represent different experimental conditions. For the “clean” condition, the alcohol was wiped off of the sample before each trial. For the “dirty” condition, the alcohol was left to build up between trials. The basis for including this factor is to determine whether a build-up of gasoline will significantly limit the sensing capabilities of the accelerometer during implementation. If it is determined that this is the case, considerations will need to be included in the design of the diaphragm to ensure that the overall accuracy of the system is not affected.


Discussion of response variable: The response variable for this experiment is reported into a program called Oros as a graph of gs over time. It has been determined that the most important indication of a good material will not simply be the highest resulting output but instead a clear response, or a high signal to noise ratio. This will be indicated in this experiment as the difference between the first two peaks in the graph of the response. A high value will indicate a large difference between these first two peaks and therefore a clearer response than a low difference. Therefore the implementation conditions of choice will be those that result in the highest value of peak difference.

Material Thickness, in (mm) Material 1 Low Density Plastic 1 .037 (.939) Material 2 Low Density Plastic 2 .069 (1.75) Material 3 Plexiglass .116 (2.95) Material 4 High Density Plastic 1 .078 (1.98) Material 5 High Density Plastic 2 .061 (1.55) Material 6 6061 Aluminum .024 (.610) Material 7 3000 Aluminum .005 (.127)

Table 1 *Note: Material 3 has been excluded from analysis due zero response to drip Equipment: DAQ: Oros Software System Accelerometer: PCB C66 12605 Fluid: S-L-X Denatured Alcohol; ρ= .789 g/cm3 Material: 2-3/10” square samples, varying thickness, stiffness Other: Hollow Cyl. Standoff (Dia 50 cm), Eye dropper, Eye dropper stand Experimental Set-Up:


Parameters:

h1= 177.8mm h2= 228.6 mm

Procedure: Five different 58 mm square samples were suspended and adhered to the top of a hollow cylindrical standoff. Data was recorded for each of the 4 different tests performed on the samples. The first test involved the eyedropper fixed at h = h1 and the preceding drips were cleaned from the diaphragm surface between successive drips. The second test was conducted with the dropper fixed at h = h1 again and the drips were allowed to pool on the diaphragm surface. The third test was conducted in the same manner as the first, with the exception being h = h2. The fourth test followed the directions of the second test, with the exception being h = h2.

The data was recorded by the Oros Software package and exported as text files to disk. Using logarithmic decrement procedures, damping ratios of each material were calculated for each drip. The damping ratio values were then averaged by grouping to accurately capture material properties. Equations for calculating damping ratio:

Minitab Results:

The experimental data was analyzed using Minitab to determine which, if any, of the factors tested had an effect on the vibration response caused by the drip. By analyzing this general factorial experiment, an ANOVA table can be constructed to determine the statistical significance of any of the three factors, diaphragm material, drip height, or a “clean” or “dirty” surface. We can see from looking at table 1 that each of the three factors appears to have a statistically significant difference on vibration response to better than .05% accuracy. The most significant factor appears to be material, followed by drip height, followed by a clean vs. a dirty surface.


Analysis of Variance for Difference in response, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Material 5 10.63529 10.67544 2.13509 313.55 0.000 Height 1 0.08487 0.08231 0.08231 12.09 0.001 Clean/Di 1 0.02902 0.03118 0.03118 4.58 0.034 Material*Height 5 0.27090 0.26885 0.05377 7.90 0.000 Material*Clean/Di 5 0.16771 0.16176 0.03235 4.75 0.000 Height*Clean/Di 1 0.01690 0.01666 0.01666 2.45 0.120 Material*Height*Clean/Di 5 0.13913 0.13913 0.02783 4.09 0.002 Error 144 0.98055 0.98055 0.00681 Total 167 12.32438

Table 2 Once we know which factors are significant, we can look at response plots to determine which characteristics of the setup yield the most ideal vibration response. Looking at table 2 we can see the effect that the different variables have on the response. According to the first graph on table 2, the best material to use for the diaphragm is material 7 (the .127mm thick aluminum). The preferred height would be 7 inches and the surface of the diaphragm should be clean during use. Given the conditions and the materials tested, this is the best possible scenario. Knowing the effect of several of these factors will aid in the design process, especially with the new information about diaphragm material and a clean or dirty surface.

Clean/DirtyHeightMaterial0.80

0.65

0.50

0.35

0.20

Diff

eren

ce i

Vibration Response vs. Material, Height, and Clean/Dirty Surface

Table 3

Error, or residual, analysis for this experiment shows that there is little bias in the experimentation. The residuals appear to be both random and normally distributed.


Decay Results:

Once it was determined which factors affect the output caused by the diaphragm vibrations, the decay envelopes of the different materials were examined based on consistent ability to damp the impulse response. This should identify the material with the best dampening ratio, which would be the most ideal for obtaining a clear signal. The .127mm (.005”) thick aluminum sheet provided the best results as seen in the consistency between the clean and “pooled” results. The .127mm aluminum plate possessed damping coefficient values in the range from .08 - .108, which proved to be the most consistent of all materials. A qualitative look at the corresponding response plots shows that the .127mm aluminum plate also provides the quickest decay of the impulse. This will aid in LabVIEW programming, as a DAQ software programming requires the diaphragm to return to steady-state within a prescribed ½ second. This result is consistent with the Minitab analysis, which identified the .127mm thick aluminum sheet as the material with the best response.

A key project requirement is that the diaphragm/accelerometer setup is able to detect multiple drips throughout the course of a test. Pooling of previous drips was a concern as damping increases as drips puddle on the diaphragm. The results of the experiment show that the .127mm aluminum plate is capable of detecting successive drips with fuel on the diaphragm. This concludes that the diaphragm can be designed without intricate contours complicating the build process.

Conclusion: Using two different types of analysis, and ideal material for the sensing system diaphragm has been identified. Of the materials tested, the .127mm thick sheet of aluminum is the most ideal material for this application. The responsive nature of the aluminum combined with its high damping ratio provide a medium that will give a strong signal that will damp out quickly when impacted with a drip of fuel. This will provide a clear, easily detectable signal in LabVIEW that can be easily distinguished from any noise created by the accelerometer. Ultimately, this new information will aid in the creation of a system capable of accurately sensing drips

Unusual Observations for Difference in Response Amplitude

Obs Differen Fit SE Fit Residual St Resid 16 0.46747 0.63432 0.03119 -0.16685 -2.18R 26 0.40813 0.59587 0.03119 -0.18773 -2.46R 27 0.79021 0.59587 0.03119 0.19434 2.54R 83 0.42261 0.24206 0.02917 0.18055 2.34R 84 0.48484 0.24206 0.02917 0.24278 3.15R 93 0.14328 0.30765 0.03119 -0.16437 -2.15R 108 0.40813 0.15114 0.03119 0.25700 3.36R 144 0.58011 0.73344 0.03119 -0.15333 -2.01R

R denotes an observation with a large standardized residual.


Residual Analysis

16014012010080604020

0.3

0.2

0.1

0.0

-0.1

-0.2

Observation Order

Res

idua

l

Residuals Versus the Order of the Data(response is Differen)

0.30.20.10.0-0.1-0.2

20

10

0

Residual

Freq

uenc

y

Histogram of the Residuals(response is Differen)


Decay Analysis

Diaphragm Material Analysis Mat 1-7

ZipLock Container ZipLock Container Plexiglass Floppy Disk Case Paper Clip Case .037" thick .069" thick .116" thick .078" thick .061" thick mat1c h1 mat2c h1 mat3c h1 mat4c h1 mat5c h1

δδδδ 0.592562 δδδδ 0.37144 δδδδ no δδδδ 0.19651 δδδδ 0.76678 ζζζζ 0.093886 ζζζζ 0.05901 ζζζζ response ζζζζ 0.03126 ζζζζ 0.12114

ΤΤΤΤd 0.010427 ΤΤΤΤd 0.00429 ΤΤΤΤd ΤΤΤΤd 0.00370 ΤΤΤΤd 0.00388 ωωωωd 0.065513 ωωωωd 0.02696 ωωωωd ωωωωd 0.02322 ωωωωd 0.02435 ωωωωn 0.066098 ωωωωn 0.02705 ωωωωn ωωωωn 0.02325 ωωωωn 0.02471

ZipLock Container ZipLock Container Plexiglass Floppy Disk Case Paper Clip Case .037" thick .069" thick .116" thick .078" thick .061" thick mat1c h2 mat2c h2 mat3c h2 mat4c h2 mat5c h2

δδδδ 0.273398 δδδδ 0.37172 δδδδ no δδδδ 0.15342 δδδδ 0.75538 ζζζζ 0.043456 ζζζζ 0.05906 ζζζζ response ζζζζ 0.02441 ζζζζ 0.11936

ΤΤΤΤd 0.005483 ΤΤΤΤd 0.00410 ΤΤΤΤd ΤΤΤΤd 0.00376 ΤΤΤΤd 0.00393 ωωωωd 0.034453 ωωωωd 0.02578 ωωωωd ωωωωd 0.02365 ωωωωd 0.02472 ωωωωn 0.034525 ωωωωn 0.02588 ωωωωn ωωωωn 0.02366 ωωωωn 0.02508

Aluminum Plate Aluminum Plate Aluminum Plate Aluminum Plate .024" thick .024" thick .005" thick .005" thick mat6c h1 mat6d h1 mat7c h1 mat7d h1

δδδδ 0.56568 δδδδ 0.694736 δδδδ 0.52976 δδδδ 0.09698 ζζζζ 0.08967 ζζζζ 0.10988 ζζζζ 0.08401 ζζζζ 0.10866

ΤΤΤΤd 0.00498 ΤΤΤΤd 0.00378 ΤΤΤΤd 0.01703 ΤΤΤΤd 0.01628 ωωωωd 0.03127 ωωωωd 0.02372 ωωωωd 0.10701 ωωωωd 0.10231 ωωωωn 0.03153 ωωωωn 0.02403 ωωωωn 0.10778 ωωωωn 0.10387

Aluminum Plate Aluminum Plate Aluminum Plate Aluminum Plate

.024" thick .024" thick .005" thick .005" thick mat6c h2 mat6d h2 mat7c h2 mat7d h2

δδδδ 0.19827 δδδδ 0.14421 δδδδ 0.59468 δδδδ 0.65925 ζζζζ 0.03154 ζζζζ 0.02295 ζζζζ 0.09422 ζζζζ 0.10434

ΤΤΤΤd 0.00381 ΤΤΤΤd 0.00384 ΤΤΤΤd 0.01524 ΤΤΤΤd 0.01476 ωωωωd 0.02396 ωωωωd 0.02410 ωωωωd 0.09575 ωωωωd 0.09276 ωωωωn 0.02398 ωωωωn 0.02411 ωωωωn 0.09661 ωωωωn 0.09379

Table 4


Appendix H. Test Enclosure Structural Analysis

Objective: A test enclosure box will be designed and modeled in SolidWords CAD software. The material selected for the box will be Aluminum Alloy 6061 that has a yield strength of 9000psi. The final test enclosure design will have a pressure relief system, a Lexan window, structural strengthening members, and attachment points for the fuel injector fixture. Theory: The use of steel for the test enclosure chamber walls was ruled out due to the corrosive nature of the substance. It is possible to coat or plate the metal, but the price of this additional operation would push the total cost of the enclosure beyond the allotted project budget. Design of the chamber walls started with ½” alloy 6061 aluminum. The internal dimension of the test box was designed to fit the fuel injector fixture. Note all meshes were run with a mesh size of 0.5 inches. Option #1

• ½” thick alloy 6061 aluminum walls, welded together • Subjected to an internal vacuum of 14.7 psi (1 atmosphere) • Yield strength of material: 9000 psi • Max von Mieses: 8461 psi • Factor of safety: 0.71


Option #1.1

• ½” thick alloy 6061 aluminum walls, welded together with aluminum angle used for extra support to side joints only

• Subjected to an internal vacuum of 14.7 psi (1 atmosphere) • Yield strength of material: 9000 psi • Max von Mieses: 7121 • Factor of safety: .84

Option #1.2

• ½” thick alloy 6061 aluminum walls, welded together with aluminum angle used for extra support to all internal wall joints

• Subjected to an internal vacuum of 14.7 psi (1 atmosphere) • Yield strength of material: 9000 psi • Max von Mieses: 5024 • Factor of safety: 1.19


Option #1.3 • ½” thick alloy 6061 aluminum walls, welded together with 1” aluminum tubing across

top span length wise and across bottom depth wise • Subjected to an internal vacuum of 14.7 psi (1 atmosphere) • Yield strength of material: 9000 psi • Max von Mieses: 862 psi • Factor of safety: 6.96

Option #1.4 • 1/4” thick alloy 6061 aluminum walls, welded together with 1” aluminum channels

(1/8” thick) across top span length wise and across bottom depth wise • Subjected to an internal vacuum of 14.7 psi (1 atmosphere) • Yield strength of material: 9000 psi • Max von Mieses: 1226 psi • Factor of safety: 4.74


Project sponsor requests that the box have a blowout panel on top that is roughly 60% of the total surface area of the top of the structure

Option #2

• 1/4” thick alloy 6061 aluminum walls, welded together with 1” aluminum square tubing (1/8” thick) across top span length wise and across bottom depth wise

• Subjected to an internal vacuum of 14.7 psi (1 atmosphere) • Yield strength of material: 9000 psi • Max von Mieses: 36570 psi • Factor of safety: 0.25

Option #2.1

• 1/4” thick alloy 6061 aluminum walls, welded together with 1” aluminum square tubing (1/8” thick) across top span length wise, across bottom depth wise, and braced between along the back wall

• Subjected to an internal vacuum of 14.7 psi (1 atmosphere) • Yield strength of material: 9000 psi • Max von Mieses: 21340 psi • Factor of safety: 0.42


Option #3 • 1/4” thick alloy 6061 aluminum walls, welded together with 1” aluminum square

tubing (1/8” thick) across top span length wise and across bottom depth wise • Subjected to an internal vacuum of 14.7 psi (1 atmosphere) • Yield strength of material: 9000 psi • Max von Mieses: 10800 psi • Factor of Safety: .83

Option #3.1 • 1/2” thick alloy 6061 aluminum walls, welded together with 1” aluminum square

tubing (1/8” thick) across top span length wise and across bottom depth wise • Subjected to an internal vacuum of 14.7 psi (1 atmosphere) • Yield strength of material: 9000 psi • Max von Mieses: 7805 psi • Factor of Safety: 1.15


Option #4 • 1/4” thick alloy 6061 aluminum walls, welded together with 1” aluminum square

tubing (1/8” thick) across top span length wise and across bottom depth wise • Division panel added to strengthen box • Subjected to an internal vacuum of 14.7 psi (1 atmosphere) • Yield Strength = 9000 psi • Max Von Mises = 3909 psi • Factor of Safety = 2.05

Separate Top Panel

• ½” thick aluminum alloy 6061 • Subjected to an internal vacuum of 14.7 psi (1 atmosphere) • Yield strength of material: 9000psi • Factor of safety on top: 2.35


Lexan Window • ½” thick Lexan window on test enclosure door • Subjected to an internal vacuum of 14.7 psi (1 atmosphere) • Yield strength of material: 9500 psi • Max von Mieses: 1772 psi • Factor of safety: 4.01

Conclusion: The fourth design of the test enclosure has a factor of safety greater than 2. The mid-panel added for strength will require two injector positioning systems to be built. While this and the more complicated design will increase the price of the enclosure, the increased amount of safety existing in this design outweighs the economic downside.


Red arrows show internal forces on the enclosure due to vacuum pressure. Magenta arrows show distributed force on the enclosure from the top blowout plate and the force from the door. The force provided by the top plate and the doors are a result of pressure from the vacuum applied to their surface area. Force exerted by pressure relief plate on top of box Force due to weight of top plate Volume of top plate = 11” x 27” x 0.25” = 94.5 in^3 = 0.00155 m^3 M = volume of plate x density of aluminum M = (0.00155m^3) x (0.00273kg/ m^3) = 4.23 e-06 kg F = (4.23 e-06 kg) x (9.81 m/s^2) F = 4.148e-05 N � (negligible) Force due to vacuum pressure on top plate Area of top plate effected by vacuum = 14” x 27” = 297 in^2 Force exerted by top plate on structure = Pressure x Area F = (14.7 psi) x (297 in^2) = 556.6 lbf = 24,716 N


Force exerted by door on front of structure Force due to vacuum pressure on top plate Area that force from door is exerted on = 23” x 9.5” = 218.5 in^2 F = P x A F = (14.7 psi) x (218.5 in^2) = 3211.95 lbf = 14,287 N


Appendix I. Fuel Injector Adapter Stress Analysis

Considerations: The fixture must accommodate at least 80% of Delphi fuel injectors that are currently being produced. A number of different Delphi fuel injectors were provided to the team for the purpose of standoff design. The operator must also have the ability to rotate the fuel injectors through a range of 0o-900. Fuel injector mounting cannot affix the fuel injector to the test stand at the spraying end. The pre-test set-up cannot exceed 10 minutes per fuel injector. Design Attributes: The fuel injector standoff fixture uses interchangeable parts to hold the fuel injectors in place. These standoff parts vary in length to place the each fuel injector tip at the same location. The standoff design concept simulates connection to an automotive style fuel rail. This provides quick and easy clip-on performance, with the ability to accommodate the 80% of Delphi fuel injectors as required by the customer. This concept also has expansion capabilities to encompass a broader array of fuel injector styles. The proposed design will allow the operator to rotate six injectors through the 90o range. The header plate of the injector fixture will travel a 6-½” radius. The complete fixture will be located inside of a chamber whose walls provide a surface to tap into for ½ -20 bolts which will provide the axis of rotation for the fixture. To lock the fixture at a specific angle a pin style locking mechanism is provided. Injectors will attach to the varying length standoff using a traditional clip style mount. This allows for use of readily available in-house fuel injector clips that provide low changeover time between injector test sets.

Changing “L” in the above model by the formula (6.5” – L = Length from Injector clip to Tip) keeps each injector tip at the axis of rotation. To keep set-up time to a minimum, the same clip style attachment used to attach the fuel injector to a fuel rail is utilized here. Using standard attachment clips ensures that each injector is held as far as possible from the spraying tip. This also eliminates the need for more special parts. Not drawn in this model are threads (1” – 14) that will locate and hold the injector and adapter assembly in the rest of the fixture. Also not drawn are threads (9/16” – 18) that will attach a fuel


line to the top of the injector fixture. These things will provide a very quick set-up time with a minimum of tools. Material Selection Considerations: Material must withstand internal fuel pressure Resistant to fuel Relatively easy to machine Lightweight for operator convenience Low cost Aluminum alloy 2024, which is a wrought alloy, has been chosen for a variety of reasons. The strength of the material is more than adequate to constrain the fuel pressure. Aluminum is naturally corrosion resistant and can be anodized to enhance these properties. Aluminum is soft enough to be machined easily. Aluminum is very low cost compared to 316 stainless steel which would give similar corrosion resistance. A thickness of 0.25” has been chosen for the top and side plates for a few reasons. One is the 1” –14 thread specified on the top of the adapter needs enough depth to hold. Assembly is a second reason. A simple weld will hold the material together and 0.25” is necessary to hold a 90 degree corner without any other support. Models: FullAssembly.SLDASM (Assembly using General Adapter) FullAssemblyEXP.SLDASM (Exploded version of assembly) TopplateThreadless.SLDPRT Righthandside.SLDPRT Lefthandside.SLDPRT GeneralAdapter.SLDPRT Analysis: The purpose of this section is to provide results from stress analyses done on a proposed fuel injector adapter intended for use in an autonomous fuel injector drip sensing system. The COSMOSXpress package contained within Solidworks 2003 was used to determine stresses and possible failure points within the design. A common material, 2024 Aluminum, with a yield strength of 11,000psi was specified for all parts. The internal pressure that can be expected is 601.3 kPa (superimposed 500kPa gage fuel pressure and 101.3kPa because of vacuum load; equal to 87.02psi).


COSMOSXpress Analysis of General model of Fuel Injector Adapter. Model Constrained at 1-14 Threads. Model made from 2024 Aluminum. Pressure load on all inside surfaces (including chamfers) of 87.02 psi. Minimum Factor of Safety found through model was 22.6347.


COSMOSXpress Analysis of Fuel Injector Adapter Designed for FI 4222. Model Constrained at 1-14 Threads. Model made from 2024 Aluminum. Pressure load on all inside surfaces (including chamfers) of 87.02 psi. Minimum Factor of Safety found through model was 25.2407.


COSMOSXpress Analysis of Fuel Injector Adapter Designed for Multec Long (9111A). Model Constrained at 1-14 Threads. Model made from 2024 Aluminum. Pressure load on all inside surfaces (including chamfers) of 87.02 psi. Minimum Factor of Safety found through model was 25.2407.


COSMOSXpress Analysis of Fuel Injector Adapter Designed for Multec Short (D117A). Model Constrained at 1-14 Threads. Model made from 2024 Aluminum. Pressure load on all inside surfaces (including chamfers) of 87.02 psi. Minimum Factor of Safety found through model was 24.5677.

Conclusion: The highest load concentration encountered in any of the designs is under 500psi, which is far less that the yield strength of 11000psi. The highest concentration of stress that occurs in each model is around the injector securing end of the fixture. In order to alleviate the stress concentration a round or fillet will be added to dissipate the stress concentration at the interface. The lowest factor of safety calculated by COSMOSXpress is 22.63 which are more than adequate. Wall thickness could be decreased to save on material costs, however ease of machining and handling considerations provide justification.


Appendix J. Final Design Return on Investment Analysis Return on InvestmentCurrent Method Yellow Boxes indicate parameters that can be changed

Assumptions:Cost of manpower to run test: 50 $/hour

Number of hours spent running tests: 4 hours / week(assuming setup and run times)

Number of injectors able to be simultaneously tested 1 injector (s)

Time to setup and run test 24 minutes

Number of injectors tested per week 10 injectors / week

Avg time to test one injector 24 minutes

Calculations: Cost of testing over six month period 5,200$ Cost of testing over one year 10,400$ Cost of testing per fuel injector 20$ per fuel injector

Using New Sensing System Method

Assumptions:Cost of manpower to run test: 50 $/hour

Number of hours spent running tests: 4 hours / week(assuming setup and run times)

Number of injectors able to be simultaneously tested 6 injector (s)

Time to setup and run test 40 minutes(with total # of injectors)

Number of injectors tested per week 20 injectors / week

Avg time to test one injector 6.67 minutesCalculations: Cost of testing per fuel injector 6$ per fuel injector

Payback Time ( Initial Investment / Annual Savings )

Cost of new system 8,613$ Injectors tested per year: 1040 injectors

Time to Payback (assuming same # of fuel injectors are tested per week by the engineers)Payback Time: 0.57 years


Appendix K. Bill of Materials Sensor System

Components Qty. Part # Supplier Contact Price Estimated Cost Actual CostQuick Bond Gel (mounts acceleromters) 1 080A90 $9.00 $9.00 $9.00Line Powered, ICP sensor signal conditioner, 8 channel 1 482A18 PCB Terry McCarville $1,552.50 $1,552.50 $1,552.50Accelerometers - PCB Piezoelectronics 6 U352C66 Piezoelectronics 1-800-828-8840 $197.50 $1,185.00 $1,185.00Coaxial Teflon Cable - PCB Piezoelectronics 1 002P10 $199.80 $199.80 $199.80Shim Stock Roll 0.006" thick, Aluminum 1 3L933 Grainger, Inc. - $24.30 $24.30 $24.30Aluminum 3" diameter, 0.216" thickness, 3 feet 2 - Metal Supermarket - $15.00 $30.00 $30.008-32 x 1/2", hex head screws 24 - - - $0.03 $0.72 $0.00

Total $3,001.32 $3,000.60

Test EnclosureParts for Plumbing Test Box Qty. Part # Supplier Contact Price Estimated Cost Actual Cost

Stainless Steel Female 1/4 NPT 12 51205K152 McMaster & Carr - $3.16 $37.92 $0.00 to Female 1/4 NPT 90 degree Elbow -Stainless Steel Male 1/4 NPT to barbed hose fittings 12 4885K13 McMaster & Carr - $14.36 $172.32 $0.00Brass Quick Disconnect to Female 1/4 NPT 6 4860K153 McMaster & Carr - $5.95 $35.70 $0.00Stainless Steel Straight Coupling 1/4 NPT both ends 5 4464K12 McMaster & Carr - $2.07 $10.35 $0.00Stainless Steel Square Head Plug 5 5481K15 McMaster & Carr - $1.39 $6.95 $0.00Fuel Line 3/8 ID, 12 FEET 9 5645K23 McMaster & Carr - $0.96 $8.64 $0.00Hose Clamps 7/16" - 25/32" 10 pack 2 55411K12 McMaster & Carr - $5.05 $10.10 $0.003/4-10 Nut 9 - T & k - $0.33 $2.97 $2.973/4-10 Washer 9 - Lumber - $0.33 $2.97 $2.973/4-10 Threaded Rod 2 Foot Length 2 - Lowes - $3.58 $7.16 $7.16#2 Stoppers 50 pack 1 9545K14 McMaster & Carr - $11.72 $11.72 $0.00Clay 3/4 lb 1 1334T24 McMaster & Carr - $2.23 $2.23 $0.00Locktite Gasket Sealer 30515 1 35125A66 McMaster & Carr - $13.11 $13.11 $0.00Tube Locktite Thread Sealer 927 1 355445A67 McMaster & Carr - $2.98 $2.98 $0.006061 Aluminum for Fixturing 1.5" Round stock 18 - Metal Supermarket - $9.88 $177.84 $9.88

Total $502.96 $22.98

Test Box Fabrication Qty. Part# Supplier Contact Price Estimated Cost Actual CostSteco Clamps, Series 331 3 5071A53 McMaster & Carr - $55.60 $166.80 $166.80Buna A, 0.4" thick gasket 1 9477K43 McMaster & Carr - $56.70 $56.70 $56.702.5" diameter male can and groove fitting 1 51415K56 McMaster & Carr - $22.50 $22.50 $22.5012" x 24" lexan viewing window 1 8184K19 McMaster & Carr - $240.00 $120.00 $120.004' X 12' sheet of 1/2" thick Aluminum Alloy 6061 1 - Accufab $892.00 $950.00 $950.004' X 8' sheet of 1/4" thick Aluminum Alloy 6061 1 - Accufab Mike Masters $355.00 $355.00 $355.003' X 3' sheet of 1/8" thick expanded aluminum 1 - Accufab (607)-273-3706 $80.00 $80.00 $80.00Fabrication Cost - - Accufab $1,588.00 $1,588.00 $1,588.00

Total $3,339.00 $3,339.00

Data Acquisition SystemComponent Qty. Part# Supplier Contact Price Estimated Cost Actual Cost

NI Multifunction I/O & NI-DAQ 1 PCI-6071E National Steve Schiffauer $1,695.00 $1,695.00 $1,695.00Noise Rejecting, Shielded BNC Connector Block 1 BNC-2110 Instruments (585)-392-2772 $295.00 $295.00 $295.00Shielded Cable, 2 meters 1 SH1006868 $195.00 $195.00 $195.00

Total $2,185.00 $2,185.00

DESIGN AND IMPLEMENTATION OF A MULTIPLE...

Documents

Transcript of DESIGN AND IMPLEMENTATION OF A MULTIPLE...