Development of test rig for flow measurements in common rail injector nozzles
CARL-JOHAN CEDERBERG JOHANNES HÖRNER BJÖRK
Master of Science Thesis Stockholm, Sweden 2010
Development of test rig for flow measurements
in common rail injector nozzles
Carl-Johan Cederberg Johannes Hörner Björk
Master of Science Thesis MMK 2010:21 IDE 030 KTH Industrial Engineering and Management
Machine Design
SE-100 44 STOCKHOLM
Master of Science Thesis MMK 2010:21 IDE030
Development of test rig for flow measurements in common rail injector nozzles
Johannes Hörner Björk Carl-Johan Cederberg
Approved
2010-02-12
Examiner
Priidu Pukk
Supervisor
Priidu Pukk
Commissioner
Scania CV AB
Contact persons
Fredrik Wåhlin & Per Österlund
Abstract
This master thesis was commissioned by the Injection Performance department (NMCX) at Scania CV in Södertälje, Sweden. The main purpose was to develop a new test rig for fuel injector performance measurements. A new test method exists that provide more accurate measurements and studies of multiple injections due to less oscillation. The method also gives the possibility to perform hole‐to‐hole comparisons of the flow through the injector nozzle holes.
Based on the test method referred to as the impingement method, a test rig for nozzle flow measurements was designed and built. The aim of the thesis was to describe the development and design work of the test rig. This method consists of measuring the sprays impact force with a pressure transducer. The collected data is then used for analysis of the injector’s performance.
Using contemporary methods in product development as well as skills and knowledge obtained during the education the design work was conducted. By establishing a requirements specification and producing a number of design concepts the product gradually took shape. Detail design and its iterations were made with the aid of 3DCAD software which led to the final design. Most of the components were manufactured at the Machining Workshop at Scania in Södertälje.
When the design was manufactured and assembled it was tested to ensure its function. A number of modifications were made and the test rig is operational and impingement
s an
measurements are now possible. The test result d gathered data are satisfactory.
The master thesis has resulted in a functional test rig that will help the understanding of ommon rail injectors.
c
Examensarbete MMK 2010:21 IDE030
Utveckling av provrigg för flödesmätningar genom munstycke hos common rail-injektorer
Johannes Hörner Björk Carl-Johan Cederberg
Godkänt
2010-02-12
Examinator
Priidu Pukk
Handledare
Priidu Pukk
Uppdragsgivare
Scania CV AB
Kontaktpersoner
Fredrik Wåhlin & Per Österlund
Sammanfattning
Detta examensarbete utfördes på uppdrag av avdelningen för insprutningsprestanda (NMCX) på Scania i Södertälje. Projektets huvudsyfte var att utveckla en provrigg för att studera bränsleinjektorers prestanda. En ny testmetod existerar som ger noggrannare mätningar av multipla injektioner tack vare mindre oscillationer i mätmetoden som sådan.
Mätmetoden ger möjligheten att utföra hål till hål‐jämförelser av flödet genom hålen i injektormunstycket.
Baserat på mätmetoden omnämnd som impulsmetoden, utvecklades och tillverkades en provrigg för flödesmätningar genom injektormunstycket. Målet med examensarbetet var med detta att beskriva utvecklings‐ och konstruktionsarbetet av provriggen. Mätmetoden baseras på att med en tryckgivare mäta sprayernas kraft när de träffar givarens spets. Den insamlade datan används sedan för att analysera injektorns prestanda.
Utvecklings‐ och konstruktionsarbetet bedrevs genom att använda vedertagna metoder för produktutveckling samt tillämpa färdigheter och kunskaper som erhållits under utbildningen. Produkten utvecklades stegvis genom att ta fram en kravspecifikation och ett antal koncept som utvecklades och utvärderades mot specifikationen. Detaljkonstruktionen och dess iterationer utfördes med hjälp av 3DCAD‐mjukvara vilket slutligen ledde till den färdiga utformningen. Det flesta av komponenterna tillverkades på Scanias mekaniska verkstad för prototypframställning.
När konstruktionen var tillverkad och monterad testades den för att säkerställa dess funktion. Ett antal modifikationer gjordes och riggen fungerar som den ska och mätningar kan nu utföras. Mätresultaten och insamlad data är tillfredställande.
Resultatet av examensarbetet är en fungerande provrigg som kommer att hjälpa förståelsen
och utvecklingen av common rail‐injektorer.
Foreword
This master thesis was conducted at the Injection Performance Department (NMCX) at Scania in Södertälje, Sweden, during the period September 2009 to February 2010.
The project was supervised by Fredrik Wåhlin and Per Österlund at Scania and Priidu Pukk at he Royal Institute of Technology, Stockholm.
t
Carl‐Johan Cederberg & Johannes Hörner Björk
Stockholm, February 2010
Nomenclature
Symbols Description
Re Reynolds Number (1)
u
We Weber Number (1)
m/s)
d Velocity of a fluid (
µ
Droplet diameter (m)
ρ Viscosity (kg/ms)
g/m
3)
σ Density (k
F
Surface tension (N/m
2) Force (N)
M N)
f
omentum Flux (
M&
m&
I a (m
4)
Mass Flux (kg/s)
W
bArea Moment of Inerti
)
σ
axSection Modulus (m
3m
(Pa)
z Maximum stress (1)
h
i
Number of teet
m Real module (1)
i Gear Ratio (1)
Abbr eviations Description
CAD Computer Aided Design
achining
CAM Computer Aided M
NO
xNitrous Oxides
tter
Injection
PM Particulate Ma
XPI (e)Xtra High Pressure
CR Common Rail
HPP High Pressure Pump
LPP Low Pressure Pump
r
HPL High Pressure Line
HPC High Pressure Connecto
IMV Inlet Metering Valve
it
ECU Electronic Control Un
MDV Mechanical Dump Valve
RPS Rail Pressure Sensor
VCO Valve Covered Orifice (Injector Type)
MW Machining Workshop at Scania R&D
&D
UTTC Instruments Supply Depot at Scania R&D
y Department at Scania R
T artment at Scania R&D
UTMR Materials Technolog
NMB Strength Testing Dep
HPH High Pressure Hose
DOF Degrees of Freedom
splay
S tal–Oxide–Semiconductor
LCD Liquid Crystal Di
DMO Double‐diffused Me
FSP Flat Strike Plate
PSP Pointy Strike Plate
IC Integrated Circuit
RECT Electronic Components and Testing at Scania R&D
Table of Contents
1. Introduction ... 11
1.1 Background ... 11
1.2 Purpose ... 11
1.3 Problem Formulation ... 11
1.4 Limitations ... 12
1.5 Aim ... 12
1.6 Method ... 12
2. Context ... 13
2.1 The XPI System ... 13
2.2 Emissions and Legislation ... 16
2.3 Fuel Sprays and Combustion Characteristics ... 17
2.4 The Measuring Method... 17
2.6 Preliminary Study ... 20
2.7 Inputs and Requirements ... 22
3. Implementation ... 25
3.1 Planning ... 25
3.2 Preliminary Study ... 26
3.3 Design Requirements and Evaluation ... 26
3.4 Ideation ... 27
3.5 Concept Development ... 28
3.6 Designed Parts ... 37
3.7 Standard Components... 54
3.8 Production Parts ... 58
3.9 Tolerances and Fits ... 58
3.10 Development of Support Systems ... 59
3.11 Electronic Parts ... 59
4. Results and Analysis ... 63
4.1 Testing ... 63
4.2 The Final Design ... 70
4.3 Time Plan Outcome ... 71
5. Conclusions and Recommendations ... 73
5.1 Conclusions ... 73
5.2 Recommendations ... 73
7. References ... 75
8. Thanks ... 77
Appendix A - Drawings ... ii
Appendix B - Standard Parts List and Fasteners ... xx
Appendix C - Time Plan ... xxi
Appendix D - Test Results Test 2 ... xxii
Appendix E - Circuit Board Schematics ... xxiii
Appendix F - C-code ... xxv
• Writing the thesis and other documentation
• Presentation
The assignment is estimated to be finalized within 20 weeks time.
1. Introduction
1.1 Background
Stricter legislation on emissions from diesel engines increases the demands on the engines and their injection systems. Decisions made by the EU have led to the Euro 5 and 6 regulations regarding including emissions of nitrous oxides (NO
x) and particulate matter (PM). Herein lies a problem, the reduction of emissions is strongly dependent on the combustion characteristics which in turn affect the engine’s performance. The injection pressure very much determines the combustion rate. Higher injection pressure increases the level of atomization of the fuel, decreases the amount of soot but increases the NO
xlevels. High injection pressure is however wanted to ensure high power output and overall performance.
To meet the Euro 5, and later Euro 6, standards Scania has with its associate Cummins Inc.
developed the XPI common‐rail fuel injector. The XPI injector is electronically controlled and can therefore operate independently of camshaft angle. High injection pressures are thus available at any time, irrespective of engine speed. Scania wants to study the behavior of the XPI injector in order to maintain high performance and still meet the emission standards.
A measuring method has been developed that helps to enhance measurements of fuel injectors’
sprays, in order to implement this method a test rig is to be developed.
1.2 Purpose
The purpose of this master thesis is to describe the development process of the test rig and the design of the finished apparatus and its functions. Secondly the purpose is to present the context and background data needed to construct the test rig.
1.3 Problem Formulation
The assignment consists of designing and building a test rig for common rail‐injectors based upon an existing measuring method. The test rig is to be used to measure and understand the properties of the flow through the nozzle of the injector. A prototype of a test rig already exists but needs to be improved to ensure accurate measuring results and simplify the measuring procedure. CAD models are to be produced in order to obtain the drawings needed for the manufacture of the test rig. Once the parts for the test rig are manufactured they are to be assemb led into the finishe d apparatus. The problem is divided into the following stages:
• Prelimina ry study
•
• Planning
• Ideation
•
Concept development
•
Calculations
• CAD‐modelling and drawings
• ems
Manufacture
•
Development of support syst
•
Assembly and Testing
ons
Adjustments and alterati
further adjustments are made to fix problems that might occur.
Continuously during the project notes are taken of gathered data, thoughts that arise and conclusions. Other documentation such as drawings and photos is also produced and collected.
This material is later compiled in the final paper which is finalized after the completion of the test rig. The paper is then handed in by which the project is presented and thereby ended.
1.4 Limitations
Since the design work will occupy most of the time, less time will be spent on literature studies and more focus is to be put on the design work itself, this to ensure that the design is tested and completed within the scope of the thesis and with that, guarantee that it’s operational.
1.5 Aim
The aim of the thesis project is to complete the design, manufacturing and assembly of a test rig for common rail injectors and also to test and ensure its function.
1.6 Method
The project starts with a preliminary study to provide the necessary knowledge required to solve the assignment. This comprises; reading of papers, literature and articles on the subject;
inspecting drawings of an existing test rig and the prototype; contacting and consulting specialists within the field; and getting familiar with the injector and other existing parts that have to be implemented in the test rig. The preliminary study also comprises acquiring knowledge and skills in the CAD‐software CATIA V5, the programme used by Scania today and consequently in the development of the test rig.
Early in the project a time plan is established to ensure that the thesis is completed within time.
Discussions within the group and with supervisors (at Scania and KTH) are conducted which aims to establish solutions to different problems associated with the design and what specifications and features have to be included in the test rig. A requirements specification is established based on these discussions.
Using standard procedures in product development a number of concepts are produced. These procedures, or methods, comprise brain‐storming and discussing different scenarios connected to the use, manufacture and assembly of the product. These scenarios are basically a breakdown of the required functions and features for the test rig. The methods are backed up by geometrical studies and dimension analyses to determine if a solution can be realized. Design solutions are continuously evaluated regarding their performance of the desired functions. Concepts are sketched for presentation and necessary calculations are made.
When a number of concepts are developed they are examined in detail to the requirement specification. The concept with best correspondence to the specification is selected for further
il d
development. Through iterations and with the aid of CAD‐software deta esign is performed.
A final concept is produced and scrutinized, possible alterations and modifications are made.
Drawings of the final design are made to enable manufacture. During the manufacturing of parts, support systems for the test rig are developed.
When the parts have been manufactured they are assembled and the rig is tested to ensure that
it’s operational. If the test turns out to be satisfactory the rig is considered completed and if not,
2. Context
This chapter describes the XPIsystem as such, how combustion affects the emissions, the test method that exists and the theoretical background to explain the results given from the test method. Further the input and results from the prestudy will be declared in order to describe the background for the decisions made in the design process.
2.1 The XPI System
XPI stands for (e)xtra high pressure injection. This relates to the high injection pressures in the engine, maximum average injection pressure is 1800 bar and the max peak pressure is 2400 bar.
The XPI is a common rail (CR) system, i.e. all injectors are fed by the same pressurised volume, the accumulator, also known as the rail. The rail is in turn fed by the high pressure pump (HPP), in contrast to ordinary diesel engine systems with unit injectors where the pressure is induced with one cam‐driven pump element per injector. The main benefits of a CR system, such as the XPI system, is that the injection timing and duration is independent of camshaft angle, the average injection pressure is higher compared to unit injectors, injection pressure can be chosen independently of engine speed and multiple injections are possible. All of these properties enable higher performance and lower emissions.
Injector w/ Electronically Controlled Pilot Valve (Only 1 Shown)
Low Pressure Pump
High Pressure Pump Accumulator
Mechanical Dump Valve
Rail Pressure Sensor
Inlet Metering Valve Fuel Filters
High Pressure Connectors
High Pressure Line
Fuel Manifold
Figure 1. Overview of the XPI system
2.1.1 Functional Description of the XPI System
Fuel is fed from the fuel tank into the pre‐filter where fuel is filtered and water is separated. The
fuel is then fed through a heater to prevent clogging of the fuel filter in cold climates. The fuel
continues to the low pressure pump (LPP in figure 1). The LPP sucks the fuel from the tank and
feeds the pressure filter and HPP, the LPP is mechanically driven by the engine. In the pressure
filter the fuel is once again filtered to remove smaller debris that could damage the HPP and the
injectors. Fuel from the filter is sucked into the HPP where the fuel pressure is increased to
approximately 2000 bar; the fuel is then fed to the rail. The amount of fuel that enters the HPP,
and thereby the rail pressure, is controlled by the inlet metering valve (IMV). The IMV is in turn
controlled by the electronic control unit (ECU). The rail distributes fuel via high pressure fuel lines (HPL) to the injectors, in other words all injectors and the rail are pressurized during operation. The rail is equipped with two key components; the mechanical dump valve (MDV) and the rail pressure sensor (RPS). The RPS gives information to the ECU about the actual rail pressure, and the ECU in turn will adjust the rail pressure when needed. The MDV is fitted on the rail to prevent over‐pressurization of the system. It is set to open at 3100 bar and then lower the pressure to 1000 bar. The injectors are fed with fuel from the rail via HPL and high pressure connectors (HPC), see figure 2.
High Pressure Connector = HPC
HPC pocket (in injector)
HPC nut
Injector clamp
HPL nut High Pressure Line (HPL)
Figure 2. Cross-section view of injector, HPC and HPL.
2.1.2 The XPI Injector
The HPC fills the cavity volume of the injector with pressurized fuel. The plunger is lifted via an electronically controlled pilot valve. On top of the pilot valve is a ball seat where a small ball is situated, see figure 3. When the electronic signal for injection is sent the armature becomes magnetic which lifts the ball. This results in a reduced pressure within the control chamber. The back pressure sucks the plunger upwards and the holes in the nozzle are opened, the injection is initiated. When the signal is cut off so is the magnetic field and a spring in the retainer forces the armature and ball back into position, the injection is completed.
Armature Ball
Pilot Valve
Control Chamber Plunger
Figure 3. Close up view of pilot valve plunger.
Retainer, control valve Spring adjusting screw
Stroke adjustment shim Armature plunger
Spring retainer
Nozzle Combustion seal
Injector body Floating sleeve
Stroke shim Ball Retainer, seat
Stator spring
Stator Spring disc
Armature
Over travel spring
Ball retainer
Valve seat
Plunger seal
Spring
Plunger
Nozzle retainer
Figure 4. Overview of the XPI injector in cross-section view.
The injected fuel spray is ignited by the heat generated in the compression by the pistons’
upward movement. Surplus fuel from the injector is led away from the injector through a channel in the cylinder head exiting in the fuel manifold. The fuel manifold outlet transports the return fuel back to the tank.
Figure 5. Cross-section- and bottom view of the nozzle.
The XPI injector has a multi‐hole nozzle of sac‐type. This means that the lower part of the plunger, also referred to as the needle, does not cover the nozzle holes and thereby leaves a small volume beneath its tip, a sac. This is to compare with VCO‐type nozzles (Valve Covered Orifice) where the needle covers the nozzle holes resulting in smaller volume underneath the needle. The benefit of the VCO‐type is that after the injection no sac‐volume remains. These small volumes of fuel in sac‐type injectors which is injected at low pressure at the end of the injection may cause increased smoke emission [1]. The downside of VCO‐type nozzles is their sensitivity to needle misalignment. Alignment errors causes fuel to pass the needle which give rise to variations in flow between the holes which also effects combustion and thereby emissions.
17°
The standard XPI injector nozzle has eight evenly spread holes with a hole‐to‐hole angle of 45°.
The spray angle is 17° relative to the cylinder head’s bottom (73° relative the injector’s centre axis).
2.2 Emissions and Legislati n
The modern direct injection (DI) diesel injection operates much more efficiently than its predecessors regarding exhaust gas emissions of nitrous oxides (NO
x) and particulate matter (PM). One of the technological advancements that have helped this development is new fuel injection systems. In the diesel engine the fuel injection process strongly determines emission formation. To achieve a good combustion the fuel injector has to atomize and vaporize the fuel during the injection. If the injector fails to do this the result is higher soot and particulates formation. This derives from the fact that the core of a fuel drop when ignited forms a soot particle due to the pressure induced from the burning shell. The shell compresses the core which in turn forms a hydro carbon particle. If the drop is small this phenomena is reduced since the
o
ignition and burn of the drop is initialized almost instantaneous.
The current Euro 5, and in 2013 Euro 6, standards aim to lower the allowed emissions from heavy duty vehicles, see table 1 [7].
NO
x[g/kWh] PM [g/kWh]
Euro V 2.0 0.02
Euro VI 0.5 0.01
Table 1. EU Emission Standards for HD Diesel Engines, g/kWh
The standards comprise other types of emissions besides NO
xand PM but they are however the hardest to reduce [1] why this thesis and other studies are focused on these two types of emission.
To achieve this high rate of atomization very high pressure gradients are found in modern diesel injection nozzles. The great difference in upstream and downstream pressure can cause cavitation to occur in the nozzle holes. The cavitation on one hand helps the spray breakup and also prevents buildup of coke and deposits within the nozzle holes. Coke build up effects the mass flow through the nozzle holes and it is sought to prevent this in order to maintain long term combustion performance. High levels of cavitation are however unwanted, too much cavitation can erode the nozzle holes which influences the spray characteristic negatively. When a hole has begun to erode the eroded surface intensifies the cavitation formation which leads to further erosion. The erosion of nozzle holes deeply effects fuel spray impulse, mass flow, penetration etc. Cavitation in the nozzle hole can also cause hydraulic flip, a phenomena where the fluid flows separated from the hole walls due to propagated cavitation. The results are non‐
cavitating flow and the formation of a hydro jet instead of a spray that is; no vaporization. This causes insufficient combustion, particle formation and can damage the cylinder walls.
The goal of diesel engine development is to find a way to minimize the soot formation whilst
maintaining moderate combustion temperatures. High temperatures promote the formation of
NO
xbut also provide higher power output.
2.3 Fuel Sprays and Combustion Characteristics
Since the fuel spray in general, and fuel spray momentum particularly, plays a key role in the combustion and emission process it is desirable to investigate its characteristics.
As diesel fuel exits the nozzle under high pressure, the fluid breaks up into a spray due to strong turbulence in the flow. This is the primary breakup mechanism of the spray. A secondary breakup of the spray can occur by shear force induced by the surrounding air. The breakup of the spray is affected by Reynolds and Weber numbers, which in turn are related to the velocity of the fluid, see equation 1 and 2. The Reynolds number is defined as,
u d
Re ρ μ
= ⋅ ⋅ (1)
and the Weber number as,
u d
2We ρ σ
= ⋅ ⋅ . (2)
W
u here:
– velocity
d iameter
– droplet d
µ
– viscosity
ρ
– density
σ – surface tension.
Once the velocity is determined the Reynolds and Weber number can be calculated. Typically a turbulent flow gives a Reynolds number >2300. A turbulent flow is necessary to ensure the formation of a spray. The Weber number on the other hand determines the rate of spray breakup, the higher Weber number the greater the break‐up. These two numbers govern the fuel drop diameter and is consequently a measure of atomization level and an important factor when estimating combustion characteristics. With data on spray velocity obtained, along with other parameters connected to fuel spray characteristics, penetration of the fuel spray can be calculated.
2.4 The Measuring Method
Presently Scania primarily performs tests with a measurement method using a so‐called rate tube in order to investigate fuel injection rate. Rate tube tests are carried out by injecting fuel into a liquid filled volume. The injection causes a rise in pressure which is measured. The pressure signal is sampled at high speed and the resulting curve correlates to the rate of fluid flow through the injector nozzle.
The work on momentum flux measurement of DI‐diesel sprays using impingement transducers by Mikael Lindström [1] is the foundation of this master thesis. Lindström has found that there is much potential in this measuring method to examine fuel spray characteristics.
In this application as well as in Lindstöms work the Kistler 4065A200 pressure transducer is
used, see figure 6. The transducer measures pressure from a fluid trough changes in voltage
caused by compression of a piezoelectric element.
Figure 6. Kistler 4065A200 Pressure Transducer.
In this application the spray momentum is measured when the spray hits the tip of the transducer. The impact force deflects the membrane on the tip which in turn compresses the piezoelectric element, see figure 7. The inertia of the transducer membrane is negligible and the test performed by Lindström and others show no signs of oscillation [1]. Given that the spray hits the tip perpendicularly the impulse produced by the impact can be calculated into a force.
The force is the momentum flux of the spray. The method is referred to as the impingement method [1]. Since the Kistler transducer measures pressure (50 mV/bar) rather than force, Lindström has calculated a factor to interpret the induced voltage into force. In his work he found that 1 volt is the equivalent to 17.84 N, i.e. the force‐volt factor is 17.84 N/V. The force is
Seat for conical ring
the product of mass flow and the fuels velocity.
When the transducer is struck repeatedly by the spray the metal of the transducer membrane expands due to increased temperature. The thermo elastic deflection gives rise to an unwanted behavior that affects the momentum flux curve. Instead of leveling out to zero the post flank value is negative [1]. To avoid such behavior the transducer tip is equipped with a strike plate, in accordance with M. Lindströms work.
Figure 7. Schematic of measuring method [6] and transducer positioning.
The measurement produces an impulse curve that describes how the force varies over time. The mpulse is given by:
i
I = ∫ F dt ⋅ = ∫ mu dt & ⋅ . (3)
Where;
u – is the fluid’s velocity and the mass flow.
W
m& –
hen the data of the fuel spray impulse is obtained further investigations may be initiated to
examine the behavior of the injector. If the mass of the injected fuel is measured, the sprays
velocity can be determined and furthermore Reynolds and Weber number can be calculated.
There are a number of benefits with this type of injection measurement method compared to other measuring methods such as rate‐tube tests. The impingement method provides the possibility to carry out tests with very short separation between multiple injections. The oscillations that occur after each injection event in rate‐tube tests make observation of multiple injections difficult, see figure 8. The oscillations have to dissipate before the next injection can be measured accurately. In the impingement method the surrounding media is air, why oscillations are minimal not to say non‐existent, see figure 8.
Figure 8. Rate-curve with one injection (left) and impulse curve with multiple injections (right).
Seen in figure 8 (left) is a rate‐tube measurement of a main injection, the impulse curve (right) displays a pilot, main and post injection measured with the impingement method. One can clearly see the difference in oscillations. The oscillation amplitudes of the rate‐tube measurement are significantly larger compared to the impingement method and the time for full
p c
dissi ation of os illations is much longer with rate‐tube measurements.
The measuring method can also provide knowledge of what injection pressures and other circumstances that cause the spray to collapse. Higher injection pressure provides greater atomization to the point where hydraulic flip occurs. This behavior can be observed by examining the impulse curve from the transducer. With increased pressure one would assume that also the momentum of the spray would increase. If the force doesn’t increase with risen injection pressure one can suspect that the flow has collapsed or that the hole is choked.
2.4.1 The Existing Prototype
During his licentiate paper M. Lindström developed a prototype to perform impingement
method tests. The prototype basically consists of a ring with eight evenly distributed holes with
a 45° angle separating them, see figure 9.
Figure 9. Lindströms prototype.
The ring is fastened with screws to the bottom side of a stock cylinder head. The transducer is then mounted in one of the holes.
2.6 Preliminary Study
2.6.1 Conversations and Discussions with Mikael Lindström
From discussions with both supervisors and Mikael Lindström it was given that the alignment of the transducers has to be as precise as possible. Since the axial position of the nozzle holes is constant, no adjustment would be needed in this degree of freedom (DOF) given that the mount for the transducer is designed and manufactured with accuracy. The angular position of the nozzle holes however, varies since there is no rotational lock on the injector nozzle which aligns the holes in a predestined way. Because of this, angular adjustment had to be a feature in the design. This led to the conclusion that the transducer/s had to be able to rotate relative to the injector, to ensure that the operator is able to adjust the angular alignment. Furthermore discussions with Lindström gave that in order to determine the angular position a series of test sprays are conducted. Once the impulse curve amplitude peaks the transducer’s tip is aligned
u
and perpendic lar to the spray.
Two possible solutions were discussed to solve the alignment procedure; the first, visual determination via a “peep hole”; the second, the ability to rotate the transducer and at the same time study the impulse curve. Option one was soon discarded since the holes are very small (Ø0.18 mm) and such an adjustment could imply great uncertainties. Left with option two there were some further problems to be dealt with. Since the test rig and the computer equipment are separated by estimably five meters and a wall, adjustment would require at least two persons, one person to rotate the transducer and the other one to study the impulse curve induced by the transducer. This seemed like an intricate and uncertain adjustment method. There is also a safety aspect to regard. The rig will be operated with high pressures and possible malfunctions could be hazardous.
2.6.2 Holetohole Variations
Desantes [4] et. al. describes that hole‐to‐hole variation regarding diameter and inlet radius
greatly affects the cavitation number and atomization rate. Contemporary manufacturing
methods induce variations in hole parameters. It is therefore interesting to examine these
variations. The XPI injectors have eight nozzle holes, and the use of eight transducers was
discussed. This however was discarded due to spatial limitations but mostly to economic factors
since the transducers are expensive. The outcome of the discussions led to a cassette that would support four transducers. Once manufactured the test rig will able to operate with up to four transducers. If equipped with four transducers only one transducer has to be aligned with a hole and the three others will also be aligned given that hole‐to‐hole angle tolerance is made with high precision. In order to measure the four other holes the cassette is rotated 45°.
Numerous discussions took place regarding how many transducers were to be used in the rig.
This also relates to how many holes there had to be in the fixture, later called the transducer cassette, for the transducers. The benefits of using one transducer is; cost reduction; the possibility to test different types of injectors (angular spacing of nozzle holes), without any alterations to the design; and also minimize possible variations in simultaneous hole measurements caused by inaccuracy in calibration of different transducers. On the other hand the use of only one transducer would mean longer test cycles and does not give the possibility to measure spray variations simultaneously.
2.6.3. The Effects of a Pressurized Injection Chamber
To simulate the environment of a real diesel engine the possibility to pressurize the injection chamber was discussed. Such a design requires fine tolerances and fits which would complicate the design. It is although desirable to mimic the engine environment to the greatest extent to obtain the best test results possible. Therefore the effects of a pressurized injection chamber were investigated.
A drawing of another injector test rig designed by Scania's business associate Cummins was studied, see section 2.6.6. This rig is designed to be pressurized during testing. To determine the difference between a pressurized and a non pressurized injection chamber Jing Li at Cummins was contacted. Cummins hadn’t, at the time, tested their rig and testing wasn’t to be conducted within a month or so. The scarcity of time didn’t allow waiting for these results because development and design work had to be initiated in order to complete the project in time.
The effect of pressurization is however, according to Jing Li, decreased spray penetration.
Decreased penetration derives from decreased momentum since the gas in the chamber is denser. The result is of course a decreased measured momentum if using a pressurised injection chamber compared to measurements performed in a non‐pressurised environment. The lack of back pressure in the injection chamber also leads to higher cavitation within the nozzle holes.
Exactly what the effect the increased cavitation has on test result is uncertain, but to design a rig with a pressurized test cylinder would severely increase the rigs complexity. A design with a pressurised injection chamber would aggravate the design and would conclusively imply difficulties to implement the necessary moving parts and sealing for these.
2.6.4 Distance between Transducer Tip and Nozzle Holes
The distance from nozzle hole to transducer tip was another factor that was an area of concern.
It was found in the literature studies that this distance varied from study to study. Ganippa et. al.
found that test results improve with decreased distance [5], but that the distance shouldn’t be
less than 1 mm. Distances shorter than 1 mm affects the sprays properties and the measuring
results. In the study of Desantes et. al. [4] a distance of 5 mm is used while Lindström in his
licentiate thesis measures impingement at approximately 4.5 mm. It was hard to make any cross
reference between these studies since they all used different input parameters such as injection
pressure and duration. No conclusion was made as to what distance that produces most accurate
measurement readings. M. Lindström was consulted for further input; his thoughts were that the distance is irrelevant as long as the entire spray strikes the transducers membrane.
2.6.5 Sealing the System
A brief conversation with one of the test cell technicians, Daniel Bohman, disclosed a problem with the prototype produced by Lindström. Since the prototype wasn’t hermetically closed the fuel spray filled up the test cell with fuel mist during a test, this is of course an unwanted behaviour.
2.6.6 The Cummins Rig
To gain more knowledge about the test method and get inspiration for the design drawings of a test rig developed by Scania’s associate Cummins was studied, see figure 10.
Transducer Adapter
Figure 10. Cross-section view of Cummins rig.
Here the transducers are mounted in an adapter which in turn is screwed into the injection chamber wall (indicated by arrow in figure 10). The injection chamber is pressurized with nitrogen gas to mimic the environment of a real cylinder in a diesel engine. The main cylinder can rotate and thus moving the adapters and the transducers mounted therein. This enables measurements of all eight injector nozzle holes despite only four mounting places for transducers.
2.7 Inputs and Requirements
It was finally decided that the test rig is to be designed without a pressurized test cylinder, although the effect of lacking injection‐chamber pressure leads to increased penetration. If needed the test results from Scania and Cummins can be compared to determine variations of the rig’s behaviors and physical properties.
By initiative of supervisor Fredrik Wåhlin the rig is to be equipped with an appendix volume.
The appendix volume serves as substitute for the excluded volumes in HPLs, HPCs and injectors
that exist in a real engine, the reason for this being to dampen and mimic the pressure waves
within the rail of a real motor.
To enable hole‐to‐hole measurements the transducer cassette is to be designed with four mounting holes. The transducer cassette, or simply the cassette, has to be able to rotate. The rotation serves two functions; the angular alignment of the transducer/s; and the ability to measure all holes of the nozzle. With the aid of a stepping motor the transducer cassette is remotely manoeuvred at the same time as the user inspects the impulse curve thus enabling single person operation and measurements. Centering of transducers is of importance. As to say radial distance has to be maintained regardless of rotation. This was specified from the supervisors, radial distance may vary within 0.25 mm hole‐to‐hole.
To ensure that the test rig can be used with other types of injectors it was decided that the transducer cassette is to be interchangeable. This is to say design it in a way so that it can be replaced with another hole‐to‐hole geometry without any other alterations to adjacent components. The reason for this being that further development of injectors and nozzles might imply a different set of holes.
The transducer cassette has to guarantee alignment of the transducer tip relative the nozzle hole in all DOF. The radial distance alignment is to be adjusted with the thread on the transducer. No axial adjustment functions are to be implemented since the axial position of the injector nozzle and thus the sprays has no variability. The transducer’s axial position relative to the injector is governed by the mounting holes of the cassette. If misalignment would occur the combustion seal (see figure 4) thickness can be alternated using additional washers or shims.
Instead of altering a stock cylinder head into a desirable design a new one is to be made. The custom designed cylinder head, later called injector fixture, has to possess all the necessary functions that a non‐combusting cylinder head would have besides interfaces for parts and components intended for this application. The concerns with using a stock cylinder head in the design are that it’s heavy, bulky and contains holes and cavities for air inlet and exhaust outlet.
These holes are situated where the intended transducer cassette was to be placed. This could imply interface problems regarding the fixation of the transducer cassette onto the cylinder head and possible leakage of fuel mist through the holes. While these were indeed problems, the time it would take to design a custom cylinder head was taken into account. Since the adaptation of a cylinder head for the test rig met with several problems it was decided to manufacture a specific injector fixture for the test rig instead. This design was considered to cause fewer problems in the development of the other adjacent components.
A standard high pressure line (HPL) is to be used to connect the accumulator (rail) with the high pressure connector (HPC).
The injection chamber had to be bottle tight even though not pressurised. This to ensure that
fuel fog doesn’t fill up the test cell during operation and prevent leakage.
Easy to transport.
Remote operation of transducer positioning.
• Possibility to test injectors with different number of holes.
3. Implementation
Chapter three describes how the project was planned and how the design work was conducted.
Further it describes how and why design decisions were made and the result thereof, the design of parts and selection of standard parts.
3.1 Planning
During the preliminary study the planning of the project was also conducted. This included the establishment of a time plan and dividing the project into different stages. The purpose was to facilitate the solution of the problem and design of the test rig and to ensure that the project was completed in time. The full time plan chart can be found in appendix C and the problem stages in section 1.3.
3.1.1 Important Events and Dates
A time plan was devised to help and secure the realisation of the project. The time plan was augmented with important dates during the course of the project, listed below. These dates serve to guide the wor k resources into the right dire ction.
• Project starts 2009‐08‐31
pt
•
• Choice of conce 2009‐09‐17
• Start of manufacturing 2009‐10‐05
• Final concept 2009‐10‐15
leted
• mpleted
All CAD models and drawings comp 2009‐10‐22 support systems co
• n for presentation
Initiate testing, 2009‐11‐30
tio
• d
Start prepara 2010‐01‐18
Thesis finishe 2010‐01‐25
• Presentation 2010‐02‐12
3.1.2 Specification of Requirements
The planning also comprised the establishment of a specification of requirements. The specification’s purpose is to list the functions and features required of the product based on the information gathered from the preliminary study. The specification’s requirement is the basis of the decision‐making process.
Required functions and features
• Ability to measure all sprays, either by having a pressure transducer for eac h nozzle hole
•
or by using a fewer amount of transducers that are moveable.
Ability to adjust the transducer’s angular position relative the nozzle holes.
• Adjustability of the transducer’s radial distance between the tip/s of the transducer/s and the nozzle holes.
• Reinforced striking surface of the transducer to avoid therm o‐elastic deflection, pitting
• d fuel.
of the tip and limiting spray bounce.
• Collection, return feed and removal of air of the injecte
•
Avoid leakage of test fuel and fuel fog.
sure line between rail and injector.
• test cell.
Standard high pres
• Ability to operate rig with the use of available
•
requirements. The common conception is that the simpler the solution the greater the benefit.
The requirements and evaluation method was applied since the scope of the master thesis is short. The supervisors wanted a functional product, preferably within the project time. All unnecessary features and styling were discarded. The intention was to maintain an “as simple as it gets” design rule in order to speed up manufacture and the rig’s completion. The reason for this being that the design had to be tested in order to evaluate its function and in the case of malfunction and still have time left for further development and adjustments within the scope of the thesis.
• The distance is allowed to vary within 0.25 mm after repositioning.
• Guarantee that no personal injuries can occur provided that the equipment is properly operated.
Desired functions
• transducer.
• Pressurisation of test cylinder.
±180° rotation to allow use of only one
• Real time monitoring of injection tests.
3.2 Preliminary Study
The preliminary study was conducted by reading three papers and a licentiate thesis. All of these in the field of flow through diesel injector nozzles and measurements methods of spray momentum. The author of the licentiate thesis, Mikael Lindström, was also contacted at different stages in person, via phone calls and mail contact. His experiences and knowledge were of great use to the development of the test rig. The preliminary study also comprised the study of drawings of a test rig developed by Cummins Inc. A number of questions arose why Jing Li at Cummings was contacted. Furthermore the study included getting acquainted with the XPI system as a whole, its functions and parts, by studying in‐house education material, visiting test cells and examining the parts.
The users of the test rig are the supervisors of the project, thereby the user research and product development intertwined. A big part of the pre‐study was spent on talking to the supervisors regarding the test rigs functions and requirements.
3.3 Design Requirements and Evaluation
In order to optimize the design work and evaluate detail and part designs the project was focu sed on three main areas to ensure the success and com pletion of the test rig:
Use ‐ does the solution ensure the function and usability?
Manufacture ‐ is the solution supported by available manufacturing methods?
Assembly ‐ does the solution allow assembly with regard to neighbouring parts?
These three requirements have to be fulfilled in order to certify that a design can leave the drawing‐table.
Besides these three a solution is weighed regarding to its complexity versus the possible benefit
it provides. This is to say, does the solution benefit from increased complexity or can the same
function be obtained with lesser means. If so use the simpler solution that fulfils the design
All design decisions were made to ensure that the test rig will imitate the real motor environment.
3.4 Ideation
The two key components in the design were considered to be the injector fixture and transducer cassette. This was due to the fact that the main function of the rig, measuring spray momentum, is governed by these two components. The transducer’s position relative the injector nozzle relies solely on how the interface between the two is designed. Most of the ideation and concept work regarded the design of these two parts and their mutual interfaces. But these two parts do not alone serve a functional test rig.
Early in the ideation process the conception was that the transducer had to be fixated relative the injector nozzle. The nozzle in turn is a fix part of the injector which in turn has to be fixed in the cylinder head and aligned with the HPC for fuel supply.
To help the ideation process an overview sketch was made, or rather a list, of necessary parts.
All the parts serve a function that was thought of as crucial for the complete operation, see figure 11. The sketch was meant to serve as a basis for ideas of the test rig design.
Figure 11. Overview of required parts.
The rig is fed with fuel using an HPP in a motor test cell via an accumulator (rail). The accumulator in turn feeds the injector via a standard HPL and a HPC (not depicted in the figure).
The injected fuel spray hits the transducers tip and the piezoelectric element compresses
producing a current. The signal is amplified and recorded by a computer. The excess fuel is collected in a fuel container which also is adapted for a suitable hose connection. The fuel is evacuated via a hose to a drain tank where air and other particles are removed.
To adjust transducer position and alter measured hole the transducers cassette is rotated with a stepper motor via a transmission. The motor in turn is controlled by a control system, consisting of logic and motor driver circuits located in the operating device. It was decided to use a gear wheel transmission to transfer the movement to the cassette, which in turn required some sort of bearing.
3.5 Concept Development
3.5.1 Product Architecture
Although the key components for the rigs main functions were given most attention, they and their neighboring parts had to be designed with concurring interfaces. The concept development consisted in finding different solutions of how to position the necessary parts and how to design their interfaces.
The different concepts mostly regarded the part‐to‐part layout, also known as the product architecture. The solutions were produced to meet the demands on high precision as well as simplifying the measurement procedure. The product architecture concepts aimed to solve the problems of ensuring main‐ and secondary functions. There were three main ideas to meet this demand, standard bearing housing design, car hub bearing design or single bearing design all with different cassette‐ and support structure design, see figures 12‐14.
Figure 12. Standard Bearing Housing Design, orange indicates bearing.
This design consisted of two bearings fixating the rotating parts, one located in the injector
fixture and the other below the transmission on the cassette. The benefits of this design were the
capacity to withstand large forces.
Figure 13. Car Hub Bearing Design, orange indicates bearing.
The car hub design was based on using an existing bearing from a car wheel hub. This type of bearing can withstand great bending torque although only one supports point. This is also the strength of this design. Only one bearing reduces the number of components and thus the complexity of support structures.
Figure 14. Single Bearing Design, orange indicates bearing.
The single bearing design is similar to the car hub design but utilizes a standard roller bearing instead. Much like the car hub design the single bearing design’s strength is low complexity due to the decreased number of components and simpler support structures.
The car hub bearing concept was discarded since the inside diameter of a car hub generally is too small for the injector with surrounding support structure to fit through. To obtain a bigger diameter a bigger hub bearing could be used but the then the remaining dimensions would cause fixation and weight problems. The conception was that these types of hub bearings are greatly over‐dimensioned for this application.
The classic axle bearing design was also discarded. This solution requires two supports since
there are two bearings. To solve this, an extra mounting plate was added, wherein one of these
bearings were to be fitted. The solution was however not believed to ensure that the radial
alignment was guaranteed after rotation due to possible misalignment of the two housing points.
This design though had one key benefit; it gave a good place for fitting of the motor, indicated with arrow in figure 15.
Suggested motor fitting
Figure 15. Mounting fixture concept for classic bearing housing.
The single bearing design was finally chosen for further development and detail design. The contact forces from the transmission are too small to jeopardize the supporting and rotating function of the design since the motor torque is only 1.25 Nm.
The placing of the motor and transmission was finally decided to be above the transducers. To further investigate other possibilities more concepts were produced. One idea was to place the motor and transmission beneath the transducer mounting holes of the cassette, see figure 16.
This layout could however result in errors if the transducer cable gets tangled in the transmission. This concept also show an idea were the injector fixture is placed in a tilted position, the reason being to simplify ocular inspection of the transducer’s position. The tilted design was however discarded after discussions with engineer T. Flink, at NMCX, see section 3.6.1.
Figure 16. Motor and transmission placed beneath transducers.
To ensure that transducer cable wouldn’t inflict with transmission movement the motor and
transmission were placed above the transducer and with that above the transducer mounting
holes in the cassette, see figure 17.
Figure 17. Final layout concept of transmission and motor.
It was decided early in the project to use a modular product design due to uncertainties regarding the specifications for the rig. The aim was to maintain interfaces as flexible as possible since the complete detail design of each part wasn’t expected to be completed simultaneously.
For instance the injector fixture, which had the longest manufacturing time, had to have interfaces towards adjacent parts that guaranteed function and assembly. Therefore most emphasis was focused on the cassette and injector fixture interface. This interface was detached from other interfaces to enable the completion of these two parts while other parts were designed.
3.5.2 MockUp
A mock‐up of the rig was produced to determine the approximate height of the test rig, due to the fact that simply estimating the height was difficult. The mock up was made from A3 printer paper rolled into tubes, a box made of folded paper and a paper bag to represent the supporting plate.
Fig re 18 Mock-Up.
The mock‐up gave an appropriate height that enabled inspection of the transducer’s position and overall assembly. The height was set to 950 mm.
u .
3.5.3 Geometrical Studies and Dimension Analysis
To get approximate dimensions a geometrical study was conducted. The most critical component in the test rig regarding its position is the transducer. Its position is governed by its placement in the transducer cassette. The closest distance between nozzle and transducer tip was studied to estimate the cassette’s dimension.
Figure 19. Geometrical model of the relation between the nozzle radius and the distance a.
The inclination of the nozzle orifice is 17° and therefore the transducer need to be positioned in the same angle to be able to accurately measure the force of the spray. The tip of the transducer has a diameter of 5 mm and the distance at which the nozzle and the transducer collide was calculated. This was done by simple trigonometric studies of the components relative positions.
he distance, a, was calculated according to equation 4 based n the geometry in figure 19.
T o
4.68 cos 17 ⋅ ( ) ° ≈ 4.48 mm . (4)
This is the closest possible theoretical position of the transducer but since the calculations were made using a simplified model (figure 19) of the injector as a cylinder without the chamfer that is present in reality the distance will be shorter, see figure 20.
Figure 20. The real design of the nozzle and the simplified shape used for the calculations.