Gasoline Fuel Injector Spray Measurement and Characterization A New SAE J2715 Recommended Practice

Transcription

1 SAE TECHNICAL PAPER SERIES Gasoline Fuel Injector Spray Measurement and Characterization A New SAE J2715 Recommended Practice David L.S. Hung Visteon Corporation David L. Harrington General Motors Corporation (Retired) Anand H. Gandhi Ford Motor Company Lee E. Markle Delphi Corporation Scott E. Parrish General Motors Corporation Joseph S. Shakal TSI Incorporated Hamid Sayar Siemens Corporation Steven D. Cummings Chrysler LLC Jason L. Kramer Robert Bosch LLC Reprinted From: Combustion & Flow Diagnostics & Fundamental Advances in Thermal & Fluid Sciences, 2008 (SP-2178) 2008 World Congress Detroit, Michigan April 14-17, Commonwealth Drive, Warrendale, PA U.S.A. Tel: (724) Fax: (724) Web:

2 By mandate of the Engineering Meetings Board, this paper has been approved for SAE publication upon completion of a peer review process by a minimum of three (3) industry experts under the supervision of the session organizer. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE. For permission and licensing requests contact: SAE Permissions 400 Commonwealth Drive Warrendale, PA USA permissions@sae.org Tel: Fax: For multiple print copies contact: SAE Customer Service Tel: (inside USA and Canada) Tel: (outside USA) Fax: CustomerService@sae.org ISSN Copyright 2008 SAE International Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE. The author is solely responsible for the content of the paper. A process is available by which discussions will be printed with the paper if it is published in SAE Transactions. Persons wishing to submit papers to be considered for presentation or publication by SAE should send the manuscript or a 300 word abstract of a proposed manuscript to: Secretary, Engineering Meetings Board, SAE. Printed in USA

3 Gasoline Fuel Injector Spray Measurement and Characterization A New SAE J2715 Recommended Practice Copyright 2008 SAE International David L.S. Hung Visteon Corporation David L. Harrington General Motors Corporation (Retired) Anand H. Gandhi Ford Motor Company Lee E. Markle Delphi Corporation Scott E. Parrish General Motors Corporation Joseph S. Shakal TSI Incorporated Hamid Sayar Siemens Corporation Steven D. Cummings Chrysler LLC Jason L. Kramer Robert Bosch LLC ABSTRACT With increasingly stringent emissions regulations and concurrent requirements for enhanced engine thermal efficiency, a comprehensive characterization of the automotive gasoline fuel spray has become essential. The acquisition of accurate and repeatable spray data is even more critical when a combustion strategy such as gasoline direct injection is to be utilized. Without industry-wide standardization of testing procedures, large variablilities have been experienced in attempts to verify the claimed spray performance values for the Sauter mean diameter, D v90, tip penetration and cone angle of many types of fuel sprays. A new SAE Recommended Practice document, J2715, has been developed by the SAE Gasoline Fuel Injection Standards Committee (GFISC) and is now available for the measurement and characterization of the fuel sprays from both gasoline direct injection and port fuel injection injectors. A primary motivation for the development of the standardized procedures for test configuration, data acquisition, data reduction and reporting was to achieve significant reductions in the test-to-test and laboratory-to-laboratory variabilities of such reported spray data. All of the major areas of fuel injector spray testing and characterization are addressed in detail in the document, including spray imaging, high-resolution patternation and drop sizing by both phase-doppler interferometry and laser diffraction. Valuable lessons regarding the definitions and interpretations of commonly-used spray parameters were learned during the development of the J2715 document, and these are presented and discussed. Based upon the five years of committee discussions and consensus decisions, five key recommendations on fuel spray measurement and characterization are made to the worldwide automotive industry. The first, and most important, recommendation is that the Recommended Practices in SAE J2715 be utilized by the spray laboratories of all automotive companies and injector

4 manufacturers in place of the myriad of in-house test protocols that are currently being used. To evaluate and quantify the efficacy of the new Recommended Practices in J2715, a comprehensive program of round-robin spray characterization tests was designed, and is currently being conducted in the spray testing laboratories of six injector manufacturers and end-users worldwide. This two-year testing protocol will be completed in mid-2008 and will compare the results of in-house testing procedures to those obtained using the J2715 procedures. This round-robin test program is described in detail, and an informative example of the initial test results for the direct-injection spray angle and spray-tip penetration is provided and discussed. INTRODUCTION The new SAE J2715 document [1] has been developed by the SAE Gasoline Fuel Injection Standards Committee (GFISC), and addresses the subject of characterizing automotive gasoline fuel sprays. The document introduces new recommended practices and definitions to achieve this. The objective of the development of J2715 is to promote uniformity in this field throughout the worldwide automotive industry. In this SAE Recommended Practice document a large number of detailed test protocols for fuel spray characterization are developed and promulgated. To accompany these testing procedures, a specific set of standardized test conditions are established and uniform data reduction and reporting procedures are derived. The SAE GFISC Committee is comprised of representatives from fuel injector suppliers, vehicle manufacturers, automotive end users and spray testing laboratories, many of whom are fuel system specialists who are concerned with fuel spray measurement and characterization. With an accurate fuel spray characterization becoming increasingly important for combustion system development for both gasoline directinjection and port fuel injection applications, the development of the SAE J2715 standards document was considered a critical priority for the automotive community. SCOPE OF SAE J2715 These new recommended practices apply to the sprays from automotive fuel injectors that are used in both port fuel injection (PFI) and gasoline-direct injection (G-DI) engine applications. The SAE J2715 recommended practice document contains the detailed background, test procedures and data reduction protocols for nearly all fuel spray characterization metrics that are applicable to automotive applications. It is intended to be utilized in conjunction with other SAE J documents that address other (non-spray) injector performance metrics such as flow curve measurement and leakage testing. These are SAE J1832 [2] for Port Fuel Injectors and the forthcoming SAE J2713 for Gasoline Direct Injectors. PURPOSE OF SAE J2715 The development and publication of the new SAE J2715 Recommended Practice document has six main purposes. These are, in the chronological order of development: 1. To standardize the use of nomenclature and definitions specifically related to automotive gasoline fuel injector spray measurements. 2. To identify and define the key metrics which constitute the characterization of automotive gasoline fuel sprays. 3. To establish detailed test procedures and recommend test equipment and methods to measure and quantify these key metrics. 4. To establish recommended procedures for the data reduction and computation of spray characterization parameters for automotive gasoline fuel sprays. 5. To establish the detailed protocol and format for spray data reporting in order to facilitate both the interpretation and the third-party verification of the test results. 6. To facilitate the adoption and usage of the gasoline fuel spray definitions, standard conditions, testing protocols, data reduction procedures and data reporting practices in the new SAE J2715 document by the worldwide automotive industry. FUEL SPRAY VARIABLES UTILIZED FOR SPRAY CHARACTERIZATION The quantification of the characteristics of sprays from automotive gasoline fuel injectors is a very important, but specialized, area of measurement. The injection of fuel in an engine is a very rapid, transient event of a few milliseconds duration, and the resultant spray of atomized liquid fuel is not easily characterized. Parameters such as the mean drop diameter, the spraytip penetration, the fuel mass distribution and the cone angle (or spray angle for G-DI) are critical to the selection of an injector for a given application. Ideally, the values would be determined with minimal measurement error by standardized procedures if the performance metrics of one injector are to be meaningfully compared to the performance of another. Standardized testing protocols are also necessary if the spray performance parameters being claimed by an injector manufacturer are to be independently verified by an end user or third-party test laboratory. Some measurements, such as those of spray-tip penetration, involve macroscopic spray parameters, while others, such as the mean drop size, involve the determination of microscopic characteristics. The older and more common type of fuel spray in the automotive industry is that obtained from the port fuel injector. This type of injector is discussed in detail in

5 Reference [2]. This relatively low-pressure unit produces a fairly coarse spray over a 2 to 15 millisecond time of injection, and exhibits a mean drop diameter of 50 to 150 microns. In contrast, the spray from the newer gasoline direct injector produces a much finer spray over a 0.5 to 5.0 ms time of injection, providing a mean drop diameter of 10 to 25 microns. The fuel pressure utilized in G-DI injectors normally exceeds 8 MPa, and, for the latest designs, may exceed 20 MPa. Several important regions of a developing G-DI fuel spray are schematically illustrated for a swirl-type injector in Figure 1. There may or may not be after-injections, ligaments, spray fingers, or even a prominent sac spray, depending on the basic design of the injector. These are not the characterization metrics themselves, but are general descriptors of areas that may affect the values of the measured spray parameters. Swirl G-DI Injector Inwardly-Opening Slit-Type G-DI Injector Casting-Net G-DI Injector Multi-Hole G-DI Injector G-DI Injector Outwardly-Opening Air-Assist G-DI Injector Figure 2 G-DI Injector Sub-Classification by Nozzle Design [3] Figure 1 Description of Observed Regions of a Transient G-DI Fuel Spray [3] Figure 2 depicts a representative set of the G-DI injector sub-classifications by nozzle design. In addition, there are newer piezoelectric-actuated G-DI injectors for gasoline applications. There are corresponding subclassifications for PFI injectors that include the common single spray, dual spray, hollow-cone and vacuum airassist types. The details of G-DI injector and spray subclassifications are provided in reference [3], which also documents numerous G-DI spray measurement techniques and variations in the measured values of spray characterization parameters for both ambient and on-engine conditions. In this reference the effects of wide ranges in the values of non-standard test conditions on the resultant G-DI fuel spray are illustrated and discussed, including the effect of variations in fuel temperature, fuel pressure and downstream ambient pressure. Reference [4] provides a general overview of the spray characteristics of atomizing nozzles, not necessarily just those of automotive injectors. The emphasis in this reference is upon the physics of how sprays are formed, and on the parameters that influence atomization, rather than upon the details of spray measurement procedures. The key metrics of an automotive gasoline fuel spray have been well established in the automotive industry by means of both engine testing and computational fluid dynamics modeling. This is due to the influences of these parameters on engine operation, combustion and emissions that have been verified over many years in a multitude of engine configurations. These key metrics and the associated parameters for an automotive gasoline fuel spray are identified and defined in detail in SAE J2715, and may be summarized as follows: 1. The angular extent of the spray envelope. This is the Cone Angle for a PFI injector spray, and is the Spray Angle for a G-DI injector spray. 2. The fuel mass distribution within the fuel spray for a PFI injector, including the location of the centroid of that fuel mass relative to the injector axis. 3. The spray-tip penetration characteristics and a key penetration point for the fuel spray from either a G-DI injector or a PFI injector. 4. The drop-size distribution, the Sauter mean diameter (SMD) and the D v90, or 90% cumulative volume diameter. The latter two are generally accepted for automotive fuel spray applications as the most meaningful statistical parameters derived from the drop-size distribution. The meanings of the terms are as follows: the SMD represents the diameter of a drop having the same ratio of volume to surface area as the collection of all drops in the measured dropsize distribution curve; the D v90 represents the diameter of a drop for which the volume of all smaller drops in the measured drop-size distribution curve is 90% of the total volume of all drops in the curve. It is

6 a shorthand indicator of the largest drops in a spray from either a PFI injector or a G-DI injector. 5. Additional parameters that numerically characterize the resultant spray geometry from injectors having special features or options, such as a bent-spray injector (G-DI or PFI) or a PFI dual spray injector. These parameters are the Cone Bend Angle for a bent-spray (offset) PFI injector, and the Spray Bend Angle for a bent-spray G-DI injector. For a PFI dual spray injector the additional geometric parameters are the individual Cone Angles of the two plumes and the Separation Angle between them. KEY CONSIDERATIONS IN G-DI AND PFI SPRAY MEASUREMENTS Automotive gasoline fuel sprays can be characterized by defining the important spray metrics and then specifying a series of proven measurement techniques to numerically quantify those metrics. It is important to note, however, that certain measurement techniques are much more suitable for PFI sprays than for G-DI sprays, and vice versa, due to the nature of the sprays. This is the result of significant differences between PFI sprays and G-DI sprays regarding the mean drop size, light scattering, vaporization rate and penetration distance. For example, the angular extent of the direct injection spray (the spray angle) can be best quantified via imaging, whereas the angular extent of the PFI spray is best quantified by high-resolution (H-R) mechanical patternation. Although a single large drop scatters more light than a single small drop, it may easily be observed experimentally using a simple strobe light that the collection of relatively large drops in a PFI spray scatters significantly less light than the collection of smaller, but much more numerous, drops in a G-DI spray. As a result, the imaging of a PFI spray generally yields a poorly defined spray boundary. This leads to possible errors in the quantification of PFI spray geometry by imaging. Hence, with the exception of PFI spray-tip penetration, imaging of PFI sprays is not recommended for the determination of spray characterization parameters. The patternation restrictions regarding G-DI fuel sprays are discussed in the sub-section on Patternation Considerations. The importance of quantifying the spray delivered by an automotive gasoline fuel injector has resulted in a significant number of research projects and technical developments over the past decade, and a number of excellent representative papers are listed as references [5, 6, 7, 8, 9]. In almost all cases the work is an application of a new measurement tool or technique to quantify one or more fuel spray metrics that are defined. This includes the measurement techniques that are recommended in J2715, as well as planar optical imaging and the application of optical patternation using laser fluorescence. Taken as a whole, these research projects on spray characterization do indeed provide insight into the problem, and all such work is invaluable to a committee that is attempting to introduce standardization into that field. Each measurement tool and procedure that was reported in the technical literature was evaluated as to whether it was feasible for definitive measurement, and as to whether it should be considered as an accessible and reliable technique. A number of emerging optical techniques, including optical patternation, were considered to be still under rapid development, and were deemed to be not yet suitable for use in an industry-wide standard. PULSE WIDTHS AND TIMING CONSIDERATIONS IN IMAGING It was apparent in the development of SAE J2715 that there was little uniformity throughout the automotive industry in the use of the term pulse width, or in the definition of time in presenting spray data such as tip penetration or spray images. It was ascertained that this is a contributor to the scatter in the reported spray data from imaging. This lack of uniformity becomes quite important for G-DI injection systems, due to the fact that the pulse-control strategies can be significantly more complex than those of PFI systems. Many, but not all, G-DI drivers incorporate a designed driver charge delay (DCD) time between the initiation of the logic pulse from the engine control unit (ECU) and the command pulse of the driver to open the injector. This is normally done to ensure a maximum voltage level of the driver capacitor just prior to actuation. The interrelationships of four key timing traces are illustrated schematically in Figure 3. In the progression from the ECU logic pulse to the fuel delivery event there are alterations due to a possible DCD, as well as the effects of the individual times of mechanical opening and mechanical closing. The actual fuel delivery time between the first appearance of fuel at the injector tip (the SOF; start of fuel) and the end of fuel (EOF) is not generally equal to either the logic pulse width or the injection pulse width. Figure 3 Key Definitions in the Correct Usage of Pulse Width Terminology

7 70There are further important considerations in selecting a representative time for spray imaging. The primary electronic logic pulse from the engine controller may be delayed by the DCD time, following which the actual injector opening time for the lifting of the injector pintle from the seat will be delayed by the mechanical opening time of the injector. The appearance of the first fuel at the injector tip, which is the initiation of the spray event, will not occur until these two time periods have elapsed. The values obtained from spray imaging are significantly affected by the image time selected, due to the fact that the spray is rapidly developing (and penetrating) with time. Hence, the flash time relative to other time markers in the system electronics must be chosen carefully, and should be clearly indicated when the acquired data is reported As the SAE J2715 document was being developed it became evident that there were numerous in-house procedures for selecting the time to use in imaging the spray (the flash time), which included many combinations of whether or not to account for the DCD and/or the injector opening time. Therefore, it has been quite commonplace to not know the precise time basis that was used in procuring the data when spray imaging data are being presented. Without such information the data cannot be verified by retesting. Thus, there are two basic questions that must be addressed prior to the imaging of sprays for the determination of the spray angle, the spray-tip penetration and the spray bend angle of a G-DI fuel spray; what precise imaging time is to be used, and when does that time start? As is evident in the timing diagrams in Figure 3, the DCD time (if it is present) should be ascertained and taken into account in establishing and reporting the time to be employed for spray imaging, particularly if the initiation of the electronic trigger phasing is the start of the logic pulse from the ECU. In the SAE J2715 document these considerations are fully discussed, and standardized procedures are recommended. For example, the fundamental imaging time for a G-DI spray is established as 1.50 ms after the time of the first appearance of fuel at the injector tip (SOF), which is independent of the duration of the mechanical opening time and the presence or absence of a DCD. This does, however, require a determination of the time of first fuel relative to the start of the ECU logic pulse. For PFI injectors the fundamental imaging time (used if required for penetration only) is specified as 5.0 ms after SOF. PATTERNATION CONSIDERATIONS The application of mechanical patternation (mass collection) to direct injection fuel sprays is not recommended due to the significant percentage loss of fuel that is associated with the reduced penetration and rapid evaporation of the very small drops. This results in a significant percentage of the fuel in each spray event that is never collected and measured in the patternator collection cells. Thus, it is specified in SAE J2715 that mechanical patternation should not be applied to G-DI sprays. Mechanical patternation measures the 2-D fuel mass distribution, and the cone angle is derived from the radius corresponding to 90% cumulative mass as measured from the centroid of that distribution. The lack of a measured mass distribution for a G-DI fuel spray means that the usual cone angle calculation based upon such a distribution cannot be conducted, thus, strictly speaking, there is no known cone angle for the spray from a G-DI injector. The confusion regarding the angular extent of a G-DI fuel spray is traceable to the common practice within the industry to use the term cone angle for both PFI and G-DI sprays. For G-DI sprays, a common industry practice is to obtain a photo of the spray, measure the angle of the spray boundary at some distance down from the tip, and then denote this as the cone angle. It is here emphasized that the value acquired in this manner does not at all correspond to the mass-based PFI cone angle, and is not a measurement of the same physical characteristics. The solution to this practice that is recommended in SAE J2715 is to create a new term for the imaged angular extent of the G-DI spray, as imaged and analyzed according to a strictly defined protocol. This new recommended term is the G- DI Spray Angle, which is addressed in detail in the section on G-DI Spray Angle Measurement. EXPERIMENTAL TEST PROTOCOLS IN SAE J2715 Detailed test protocols for nearly every fuel spray characterization variable have been developed, and are sequentially listed in SAE J2715 according to the class of injector that is being tested. The test configurations and procedures for the measurement, data reduction and reporting of fuel spray characteristics are only briefly outlined here. The specific details on the test procedure and data reduction may be found in the SAE J2715 document. In general, undertaking the task of measuring the key fuel spray variables such as the mean drop size or the cone angle is not to be taken lightly. Highly specialized laboratory facilities, safety equipment and test instrumentation are required, and, even with such facilities, significant time may be required to configure and align the test optics, and to acquire the many data points. Nearly all fuel spray characterization data is obtained in dedicated spray laboratories that are also laser optical laboratories. This implies that the technical staff conducting the tests must also be highly skilled in operating the instruments. The level of sophistication for the required instrumentation is higher for the measurement of the spray microscopic properties such as drop sizing than for the overall macroscopic properties such as spray-tip penetration. For these reasons, it is recommended in SAE J2715 that spray testing be done in an order starting with the geometric, macroscopic characteristics, which are obtained by imaging or patternation. Once those spray parameters are determined, the measurements can progress towards the detailed microscopic measurements, such as drop sizing. The macroscopic measurements are generally less difficult to perform and, in addition, can provide information that significantly facilitates the

8 microscopic measurements. The order of testing for the series of complete spray metrics is recommended in SAE J2715 as follows: For G-DI sprays, imaging should be performed first, followed by drop sizing. For PFI fuel sprays, patternation should be the first test, followed by imaging for penetration (if required), then drop sizing. SPRAY GEOMETRIC PARAMETERS FOR G-DI USING IMAGING Spray imaging is a technique employed to produce a two-dimensional, static, digital image representing a three-dimensional, time-varying spray plume. The primary requirement in determining the spray geometric parameters is to image the entire spray, which precludes the use of light-sheet techniques. A uniform light source, a diffuser, the injector and the camera should be configured to generate backlit images with the entire spray being illuminated (Figure 4). To freeze the motion of the drops in the spray, a pulsed, triggered light source of sub-microsecond duration such as a pulsed laser or a flash lamp is required for use with standard digital cameras. Figure 4 Schematic of the Backlit Imaging Test Configuration Figure 5 Spray Imaging - Backlit Image of a G-DI Fuel Spray SPRAY GEOMETRIC VARIABLES FOR PFI USING HIGH-RESOLUTION PATTERNATION The two-dimensional fuel mass distribution of the spray in a plane that is orthogonal to the injector axis can be ascertained by collecting the liquid drops from a large number of consecutive fuel spray pulses in the cells of a mechanical patternator. The term "fuel spray mechanical patternation" implies an intrusive measurement of the spray to obtain the distribution of the fuel mass in two spatial dimensions. Specifically, the cone angle and other spray geometric parameters of a PFI injector spray can be computed from the mass distribution that is obtained. Figure 6 shows a schematic representation of the test configuration that is used to measure the mass distribution within the spray of a PFI injector. A continuous light source with a camera capable of very short exposures (such as a gated-intensified camera) may also be employed. The light (or camera) trigger is to be activated with adequate time-delay controls from the injector driver circuit. There are additional requirements for the electronic system that controls, synchronizes and properly phases the injection event, light source and camera shutter. As illustrated in Figure 5, backlit illumination is recommended for more uniform imaging of the spray, and it is very suitable for the later determination of the spray boundaries. The data reduction procedure requires the use of specified image processing techniques such as subtraction of the nonspray background image and image thresholding to locate and define the spray boundaries. Key spray parameters such as the spray angle, spray bend angle and the axial spray-tip penetration distance are determined by spray imaging and image analysis. This is discussed in more detail in the section on G-DI Spray Angle Measurement. Figure 6 Test Configuration for the Measurement of the Mass Distribution in a PFI Fuel Spray using a High Resolution Patternator

9 The measurement of fuel mass distribution may be obtained by sensing the volume of liquid fuel in each collection cell. This is generally done by using another sensing variable or, in a method that has been generally superseded, by weighing the fuel that is collected in a series of precision cells. In automated systems, the fuel mass in each cell may be computed by means of a calibration algorithm for capacitance, laser-beam attenuation or liquid-column-height sensing. The mass of fuel in each of the 250+ collection cells provides the two-dimensional distribution of fuel in the collection plane. This is graphically represented in Figure 7, with the three shades of gray indicating three ranges of fuel mass. The key spray parameters that are computed from the fuel-mass distribution of a H-R patternation test are the cone angle, the bend angle and the angle of the fuel-mass centroid relative to the centerline of the injector electrical connector. An additional (second) cone angle and a separation angle are also acquired and reported if the spray is from a dual spray PFI injector. This is illustrated in Figure 8. Any of the H-R patternator designs represent a significant enhancement in resolution over the original SAE eight-ring mechanical patternator that was first introduced to measure the cone angle of a throttle-body injector or port-fuel injector more than 20 years ago [2]. In this low-resolution, annular-ring fixture the fuel collected in each of the eight rings is separately drained and weighed. This simple patternator is still in limited use, although it has largely been supplanted by H-R patternators that have automated computer control. Application of the eight-ring annular patternator is limited, as only the PFI cone angle can be obtained, and dual spray units cannot be evaluated. As compared to the older eight-ring patternator, H-R patternator designs having 250 to 625 cells, with spacing between adjacent cells of 3 to 6 millimeters, exhibit a five-to tenfold enhancement in the potential measurement resolution of a fuel mass distribution. DROP SIZING LASER DIFFRACTION AND PHASE- DOPPLER INTERFEROMETRY The specialized laboratory instruments that have been developed and applied to the measurement of drops in automotive fuel sprays may be divided into two major types: laser diffraction instruments and phase-doppler instruments. With the proper selection of optics, either of these two laser-based instruments may be alternatively utilized to measure and record the sizes of drops within any automotive fuel spray, whether PFI or G-DI. However, there is one important caveat that must be considered for these alternate measurement techniques; there is no known conversion or relationship between the two results. Strictly speaking, the results for the characteristic drop sizes as determined by laser diffraction should not be compared directly to the results obtained by means of phase-doppler interferometry. Figure 7 Representation of a Single-Plume Spray Pattern in a High- Resolution Mechanical Patternator Figure 8 Representation of the Separation Angle and the two Cone Angles in a PFI Dual-Spray Patternation Measurement Laser diffraction systems determine the particle distributions from collected light that has been scattered by all of the spray drops that are present in a cylindrical beam of laser light, as illustrated in Figure 9. The scattered and collected light is used to infer a spatiallyintegrated drop-size distribution. The diffracted light from all of the drops present in the laser beam that are within the working distance and size range of the receiving optics is collected in an annular array of photodetectors. All of this collected light contributes to the determination of the overall drop-size distribution. Laser diffraction systems yield an integrated result for the spray in one measurement, hence the measurement time is relatively short. It is important to note that variations of key spray metrics along the laser beam are not obtained, nor is any information obtained on the drop velocity distribution. The key spray drop-size parameters obtained using laser diffraction are the SMD (Sauter mean diameter), D v50, and D v90. The D v50 and D v90 are drop diameters that are defined statistical moments of the drop-size distribution that indicate the median and largest drops in the spray, respectively.

10 Figure 10 A phase-doppler Interferometry Measurement System for Drop Sizing using a Forward-Scattering Test Configuration Figure 9 A Laser Diffraction Measurement Configuration for Drop Size Measurement in Automotive Fuel Sprays Phase-Doppler interferometry (PDI) is another type of instrument that can be used to measure the drop-size distribution of the fuel sprays. Phase-Doppler systems perform measurements at a point in a spray. This is accomplished by monitoring the light that is refracted by the individual drops of a spray that pass through a small probe volume that is illustrated schematically in Figure 10. This is a very small volume that is defined by the intersection of two, or more, focused laser beams and a precision optical slit, or spatial filter. Wherever the probe volume is positioned within a spray, the complex refracted light signal from an individual drop that passes through that volume is interpreted, classified and, if it passes a series of logic tests, is converted to an accepted drop diameter and placed into a cumulative database. This database yields the drop-size distribution for that location at the end of the test. Unlike the laserdiffraction method of measurement, other locations in the spray are not interrogated while data at that location is being acquired. Typically, the probe volume is sequentially positioned at numerous sequential locations in the spray, yielding a measured spatial variation in spray characterization parameters. The individual measurement locations are specified to be along a specific straight line, which is designated in SAE J2715 as the radial scan line, and which normally scans the distance from the injector centerline to the outer spray boundary. As a result of having to index the probe volume through numerous positions in the spray, phase- Doppler testing is significantly more time consuming than laser diffraction testing. The key drop sizing parameters obtained using PDI are the drop-size distribution curves at key points within the spray, the SMD and the drop diameter that corresponds to the 90% cumulative volume point on the drop-size distribution curve, D v90. In interpreting the results of drop-size measurements, it is important to remember that automotive gasoline fuel sprays are pulsed sprays. The highly transient nature of a spray pulse is obvious, even with simple visualization using a strobe light. Although the drop-sizing statistics that are detailed in the SAE J2715 document are all based upon time-averaged measurements made over the entire injection event (in fact, over a large number of consecutive injection events), there are some PDI and laser diffraction instrument packages that can perform time-resolved drop-size measurements. This timewindowing has the additional capability of measuring only during brief time intervals within an injection event. This specialized equipment can differentiate between leadingedge drops and trailing-edge drops, and can separately measure the sizes of the main spray drops and the sac drops. However, such specialized equipment options are quite complex in application and interpretation. An additional consideration is that many PDI and laser diffraction instruments remain in service that do not have the time-windowing option. One of the goals of the new J2715 procedures was to have an equipment requirement that all spray laboratories could meet, and a test procedure that was not overly complex in configuration and interpretation. Hence, after much discussion by the GFISC committee, there was a consensus that all of the drop-sizing tests in the SAE J2715 document would be time-averaged, with the timewindowing option turned off if the drop-sizing equipment has such an option. With the exception of spray-tip penetration, this also applies to all of the other tests within the document. Controls on the air-purging rate of the spray enclosure are specified to prevent the measured, time-averaged results from being excessively influenced by the residual spray droplets. This is discussed in more detail in the section on Lessons Learned. INJECTOR TIP DRIPPAGE DURING REPETITIVE SPRAY EVENTS It is noted in SAE J2715 that the test that is developed for the injector tip drippage rate quantifies the propensity

11 of portions of the liquid fuel to collect and accumulate on the injector spray tip during repetitive spray events. This metric is only applicable to PFI injectors, not G-DI injectors, and is particularly meaningful for those PFI injectors having recessed tips or those providing two spray plumes having wide separation angles. Any fuel from a spray event that accumulates on the PFI injector tip will result in imprecise metering. It should be noted that tip drippage is not the same phenomenon as tip leakage. The leakage test is not a spray-related test, and is covered in SAE J1832 [2]. For an injector for which a small portion of the injected spray adheres to the tip surface, the liquid fuel accumulation continues until a critical mass is attained and the forces of adhesion to the tip surface are overcome by the weight of the fuel. At this point in time a large drop of liquid fuel falls from the tip. This drop size is normally hundreds of times larger than the largest drop in the spray, but is not measured in the test. For a specific mounting orientation of 45 from the horizontal, and for the standard operating conditions that are noted in Table 1, each large drop that falls during the test period is counted. The characterization value that is reported is the average periodic drip rate. STANDARD TEST CONDITIONS The establishment of standardized test conditions for a wide variety of test protocols is quite important for reducing variations in the final test values that are acquired and reported. The standard values of test conditions that are recommended in SAE J2715 are summarized in the second line of Table 1 for all eight test categories. The specified test fluid, n-heptane, is included as one of the standard test conditions. These standard test conditions for the eight basic categories of automotive fuel spray characterization were determined by a consensus of the SAE GFISC Committee representatives. This was achieved via lengthy committee discussions that followed a detailed analysis of each of the eight test categories. Each analysis considered the values that had historically been used by each injector OEM and automotive end user. OVERVIEW AND IMPORTANCE OF ROUND- ROBIN TESTING In a general round-robin test program a specific hardware set is sequentially provided to a series of testing laboratories. A test procedure is developed and codified, and each independent test laboratory conducts the required tests upon the full set of injectors. In a blind test, such as this particular program, the test results are not shared among the test laboratories, but are supplied to a central, non-laboratory administrator who compiles the results and statistics. The test injectors are then shipped to the next laboratory. This is continued until all of the laboratories have completed the testing. The data are then provided to a central, independent authority where data reduction, compilation, statistical analyses and comparisons are conducted according to strict preestablished procedures. Conclusions are drawn and the results are shared with all of the participating laboratories. Round-robin testing is particularly meaningful for determining laboratory-to-laboratory variability, and to evaluate the effectiveness of alternate procedures for test protocols and data reduction. In attempting to quantify the laboratory-to-laboratory error in measuring the various fuel spray characterization parameters, the advantage of having the exact same set of injectors being evaluated by multiple independent laboratories is evident. Each laboratory is to first measure and report the results using their own test procedures, test fluid and data reduction techniques, for the unique set of injectors that is supplied sequentially to each round-robin participant. DESCRIPTION OF THE ROUND-ROBIN TEST MATRIX In this specific round-robin test program, the test protocols are sets of detailed test procedures for all of the spray characterization variables that are listed in Table 2. For each parameter, such as the Sauter mean diameter, the test is to be conducted using two alternate protocols. The first is the established, in-house procedure for that parameter at that laboratory. The second is the new test protocol as specified for that parameter in SAE J2715. The hardware set consists of the specific injectors (and associated drivers) that are listed in Table 3. This set contains a current representative range of injectors for the worldwide automotive industry, including G-DI and PFI injectors of various designs and applications. Six of the automotive fuel spray laboratories from around the world are currently participating in this two-year experimental study. These are the laboratories that are associated with the co-authors of the paper. Each of the six laboratories is measuring values for the spray characterization parameters according to the corporate in-house test procedures at that lab, using the in-house designations for the test fluid, test conditions and datareduction procedures. Following the completion of testing according to the corporate in-house procedures, each lab is to then repeat all of the tests on the same injectors using the SAE J2715 procedures. The testing using in-house procedures is to be conducted first, with the testing using the J2715 protocols being completed by mid The J2715 procedures include not only the specified test protocols, but the standard test conditions and recommended data-reduction procedures as well. Each test laboratory is to report the results for both the in-house corporate procedures and the SAE J2715 procedures to the central data analysis authority. The matrix of results is then to be processed for each spray characterization variable in order to obtain both laboratory-to-laboratory variations and corporate-versus- J2715 variations. The full results from this two-year round-robin study will be presented in a separate SAE paper. Although not the main focus of this particular paper, some initial round-robin results for G-DI spray imaging are presented in the following sub-section.

12 Table 1 Summary of Standard Test Conditions for the Characterization of Automotive Gasoline Fuel Sprays Port Fuel Injection Gasoline-Direct Injection Qualitative Imaging High- Resolution Patternation Laser Diffraction Phase- Doppler Interferometry Drippage Test Imaging Laser Diffraction Phase- Doppler Interferometry Test Fluid Ambient Temperature ( C) Ambient Pressure (kpa) Fluid Temperature ( C) Initial Injector Temperature ( C) Fluid Pressure (kpa or MPa) Injection Pulse Width (ms) Injection Period (ms) Flow Measurement Injector Axis Orientation Injector Electrical Connector Orientation Measurement Height (mm) Axial Field of View (mm) Preconditioning or purging n-heptane n-heptane 21 ± 2 21 ± ± ± 5 21 ± 2 21 ± 2 21 ± 2 21 ± 2 Specified by application to within ± 1% 1000 ± ± 0.01 Specified by application to within ± 1% 5 ± ± 0.01 To be reported in mass flow units (g/s or mg/pulse) Vertical unless specified by application Specified by application and to be reported 45 ± 1 degrees from vertical 1000 ± ± 0.01 To be reported in mass flow units (g/s or mg/pulse) Vertical unless specified by application Specified by application and to be reported N/A 100 N/A N/A N/A 100 N/A 2000 pulses at 50 ms period 6000 pulses at 50 ms period 2000 pulses at 50 ms period

13 Table 2 Overview of Primary Spray Characterization Variables TEST LAB A TEST LAB B TEST LAB C Figure 11 Determination of the Spray Angle of G-DI Unit 7 Photos from the In-House Techniques of Three Spray Laboratories EXAMPLE OF INITIAL RESULTS FOR SPRAY ANGLE AND PENETRATION BY IMAGING Of necessity, each spray test laboratory has an established, in-house procedure for the measurement and data reduction of nearly every spray characterization variable. One of the early programs to be conducted in the round-robin test project was that of measuring and reporting the spray angle and spray-tip penetration of a specified G-DI injector. This injector is listed in Table 3 as Test Unit 7, and is a pressure-swirl G-DI unit. As of this time, three spray measurement laboratories have obtained characterization data on this particular unit. Table 3 Description of the Round-Robin Test Injectors TEST UNIT INJECTOR CLASS SPRAY TYPE DESIGN-INTENT SPRAY 1 PFI Injector Single-Spray Narrow ( 7.0 Degrees) 2 PFI Injector Single-Spray Narrow ( 7.0 Degrees) 3 PFI Injector Single-Spray Wide (25.0 Degrees) 4 PFI Injector Single-Spray Wide (25.0 Degrees) 5 PFI Injector Dual-Spray Cone (14.0 Degrees) Separation (19.5 Degrees) 6 PFI Injector Dual-Spray Cone (14.0 Degrees) Separation (19.5 Degrees) 7 G-DI Injector Pressure-Swirl 55.0 Degrees 8 G-DI Injector Pressure-Swirl 55.0 Degrees With the exception of a fixed fuel pressure of 8.5 MPa, each laboratory utilized the set of in-house procedures

14 specific to that corporation, including the configuration of the test, the specification of the test fluid and the choice of imaging method and image processing techniques, as would normally be conducted at that laboratory to obtain the angular extent of the G-DI spray (spray angle) and the spray-tip penetration. It may be noted in Figure 11 that three different spray illumination methods were utilized, with each being the established imaging method within that laboratory. Backlighting by means of a Nd-YAG laser was utilized at Lab A, with copper-vapor-laser backlighting used at Lab B. For Lab C, volume illumination of the entire spray by means of a strobe light was the method employed. For purposes of comparison, the three imaged fields of view have been adjusted so that each is 65 mm horizontal by 65 mm vertical. In Lab A the image was taken for a 1.50 millisecond injection pulse width at a time of 1.50 milliseconds after the start of the injection pulse, whereas the conditions of a 1.50 ms logic pulse width and 1.50 ms after the appearance of first fuel were used by Lab B. In Lab C a pre-specified injection pulse width was not utilized, but, instead, a floating pulse width was selected that provided a fixed fuel mass delivery of 10.0 milligrams per injection. This corresponded to an injection pulse width of ms, and the flash time for spray illumination was 1.00 ms after the first appearance of fuel. Two different test fluids were utilized by the three laboratories; n-heptane for Lab A and Lab C, and Indolene for Lab C. The post-processing of the image to obtain the associated spray angle and spray-tip penetration was performed according to existing protocols within each laboratory. In Lab A the spray boundaries were obtained along two lines orthogonal to the injector axis at 5 mm and 15 mm from the tip, and the spray-boundary points on these two lines were used to define two lines that each formed a half-angle relative to the injector axis. In Lab C the two selected lines that were orthogonal to the injector axis were at 1 mm and 10 mm from the tip, and the four points on the spray boundary defined two lines in space. The absolute angle between these two lines was measured, and was designated as the Spray Angle. In Lab B an algorithm from a commercial image processing program was utilized. This program effectively uses a line orthogonal to the injector axis for every horizontal row of pixels in the image to determine the left and right spray boundaries at that location. The program then fits a least-squares straight line through the points defining both the left and right boundaries. The angle between the two lines is the reported spray angle. This illustrates the range of in-house procedures that are currently being employed. The reported values for spray angle and spray-tip penetration are presented in Table 4 for discussion. Note that these data are the result of testing with inhouse procedures, not with the SAE J2715 procedures. The results of tests using the J2715 protocols are to be obtained by mid-2008, and will then be compared to the earlier results using the in-house procedures. The measured spray angle for the same exact injector, driver and fuel pressure is seen to range from 37 degrees to 61 degrees, and the reported spray-tip penetration varies from 44 mm to 50 mm. The wide range of reported parameter values that may be encountered with automotive fuel sprays, even for the same injector, is made evident by this example. Table 4 Reported Data from Three Test Laboratories for the Spray Angle and Tip Penetration of Test Unit 7 Using In-House Testing Methods G-DI SPRAY IMAGING; In-House Test Protocols Injection Pulse Width (ms) Image Distance(s) for which Spray Edges were Determined Reported Spray Angle (degrees) Main Tip Penetration at 1.50 ms (mm) Round-Robin Spray Test Laboratory Designation Lab "A" Lab "B" Lab "C" ms after Start of Injection Pulse 5.0 and 15.0 mm from Tip 1.5 ms after Start of Fuel All Horizontal Pixels - Commercial Image Processing Software 1.0 ms after Start of Fuel 1.0 and 10.0 mm from Tip None of these three values are inherently "erroneous"; they merely reflect the conditions and procedures that were used to obtain them. In fact, initial indications show that a large portion of the variation among laboratories for the G-DI spray angle is attributable to the lack of standardized procedures for setting up the test configuration, selecting the test fluid, specifying the imaging time relative to the ECU logic pulses, and in choosing the manner in which the images are to be interrogated to obtain the angle. A number of these critical choices that are made at all spray laboratories are answers to the basic questions of "when do you image the spray?" and "where do you measure the angle?" The time of the flash and the distance(s) from the injector tip must be specified, and each spray laboratory, of necessity, has an individual set of established guidelines for this. Even before the flash time is established, an operational pulse width must be selected when the injector is being configured for the test. It should be evident from the example that no spray laboratory can currently verify the spray measurement claims of another without obtaining very detailed information on all of the internal procedures that were used to obtain the reported performance data. The use of the standardized techniques provided in SAE J2715 should significantly improve this situation. G-DI SPRAY ANGLE MEASUREMENT AS AN EXAMPLE OF SAE J2715 STANDARDIZATION The values of the G-DI spray angle and the mean drop size are arguably the first two spray parameters that are requested when a particular G-DI injector is being considered. In fact, G-DI injectors are often subclassified and marketed based upon the incremental