Two identical blank nozzles have been processed by EDM (Electro Discharge Machining), obtaining two different orifice geometries; for the sake of brevity, they are called “Round” and “Slot”.

The two nozzles were designed with the aim of making comparisons at the same conditions, namely the same fuel amount per injection shot and the same opening/closing time of the needle. This condition is verified and discussed in the results section by analyzing the mechanical hydraulic behavior of the nozzle.

Fig. (1) Cross-sections of the investigated nozzle orifices.

The two nozzles are multi-hole type (4-hole); concerning the former, it is based on a cylindrical standard hole and its simple layout is assumed as the reference case in the current investigation. In the latter layout, the slot cross section is implemented. Such a shape is realized in the attempt to achieve the same nominal geometrical flow area of the circular case. Dealing with non-circular holes a further concern is introduced, due to the freedom to orient the cross section in respect to the injector axis. This concern has been addressed in advance, on the basis of 3D-CFD nozzle flow simulations, whose results were collected and published in [19Chiatti, G.; Palmieri, F. Hole cross section shape influence on diesel nozzle flow. SAE Tech. Pap., 2013.]. The adopted configuration and the details of the tested nozzles are shown in Fig. (1), where the major axis of the section is normal to the injector axis.

According to Fig. (1), the chosen Slot cross section is formally related to the circular one by the following equations (1-4), which are directly derived from the adopted section geometries. Table 1 summarizes the relevant nominal dimensions of tested nozzle holes.

(1)

(2)

(3)

(4)

Table 1
Round and Slot cross-section nominal dimensions.

After drilling, the two nozzles have been hydro-ground. In order to avoid uncontrolled alterations of the Slot shape, the process has been limited admitting no more than 3% flow rate increase during steady hydro-grinding conditions. Hence, on the Round nozzle, the hydro-grinding was made up to reach the same flow rate of the Slot at the same steady flow conditions. In the experiments, the two prototype nozzles have been fitted, in turn, to the same injector body of a second generation common rail system.

The investigation was carried out on the injection system test bed at Engineering Department of Roma TRE University. The set-up includes a complete common rail injection system (high-pressure pump, rail and solenoid injector), an auxiliary system for handling and temperature control of the injection fluid (ISO 4113 compliant oil), a long-tube device for ROI measurement and an optical diagnostic system for the fuel spray.

The ROI measure is based on the well-known Bosch Tube system [20Bosch, W. The fuel rate indicator: A new measuring instrument for display of the characteristics of individual injection. SAE Tech. Pap, 1966.
[http://dx.doi.org/10.4271/660749] ], in which the pressure wave is measured by means of a piezo-resistive pressure sensor (4067-type by Kistler, Winterthur, Switzerland). The same sensor is used to measure the pressure signal at injector inlet. ROI trends are obtained by averaging 256 injection shots. Injected quantities are gravimetrically measured by considering 1280 injection shots.

The spray visualization technique here adopted is backlit-imaging, based on the use of a DBI (Diffused Back Illumination) system, synchronized with the energizing signal of the injector solenoid valve. With high accuracy in time, an electronic controller implemented in the TB-ECU (Test Bed-Electronic Control Unit) is used to position the strobe-light emission in respect to the solenoid energizing signal. The sketch of the system is shown in Fig. (3), while the technical characteristics of the backlighting and digital imaging system are reported in Table (2). The injection of fuel takes place in a closed chamber, under non-evaporative conditions. The chamber is quiescent, operated with air (7 bar) at room temperature (25°C).

Table 2
Backlighting and digital imaging system specification.

In the current investigation, the spray behavior is analyzed visualizing the shadowgraphs on two distinct planes, both containing the axis of the nozzle orifice (y-Axis) and orthogonal to each other, as shown in Fig. (2); for the sake of simplicity, the plane also containing the minor axis of the Slot (z-Axis) is called “minor plane”, while the other one, containing the major axis of the Slot (x-Axis), hole is called “major plane”.

Fig. (2) Reference axes on the slot section (left); planes of spray visualization (right).

Fig. (3) Sketch of the spray visualization system.

The visualization process operates as follows. The low pressure fuel handling system feeds the main pump at fixed pressure and controls the fluid temperature to the level set for the tests. The main drive is started, operating the high pressure pump at 900 rpm and maintaining the pressure in the rail under closed-loop control of pressure regulator. The electronic controller sends the excitation signal to the injector. After a certain delay time ΔT, the TB-ECU sends the driving signal of the strobe lamp, which emits light for a fixed time period (50 ns). Such a fast flash “freezes” the development of the spray at a well-defined instant after the start of the solenoid excitation signal, with high accuracy; the system is able to analyze the evolution in time of a periodic event by building a succession of images obtained from multiple events by increasing, step by step, the delay time ΔT. In the present investigation, the uncertainty introduced on the time axis is bounded within 0.06 μs. Such a value is here regarded as negligible in comparison to the time scale (tenths of millisecond) of fuel injection process. Due to the occurrence of cyclic dispersion, the visualizations are repeated for each delay time at which the evolution of the spray is frozen, until the ± 5% accuracy in penetration distance and spray angle is achieved. As highlighted by relevant contributions in the field [21Kostas, J.; Honnery, D.; Soria, J. Time resolved measurements of the initial stages of fuel spray penetration. Fuel, 2009, 88(11), 2225-2237.
[http://dx.doi.org/10.1016/j.fuel.2009.05.013] , 22Eagle, E.W.; Morris, S.B.; Wooldridge, M.S. High-speed imaging of transient diesel spray behavior during high pressure injection of a multi-hole fuel injector. Fuel, 2014, 116, 299-309.
[http://dx.doi.org/10.1016/j.fuel.2013.07.120] ], cyclic dispersion is always encountered when the time evolution of the main macroscopic spray parameters are investigated (penetration, velocity of the spray-tip and opening angle of the spray); the dispersion of these parameters has been highlighted by cameras with very high frame-rate (e.g. 100.000 fps), able to record the time evolution of each single injection event. Surely the opportunity to study the evolution of a single event remains a key-factor in the spray research field, but then it should be noted that, due to the cyclic dispersion, it is necessary to repeat the analysis on more events to determine the average effective values that represent the spray behavior. Thus, for the purposes of the current study, the use of the strobe-technique is considered appropriate. Moreover, in the current framework, the strobe technique has some advantages over the high speed video technique. In fact, as a large number of frames per second is not managed, the use of very high resolution Active Pixel Sensors (APS) is possible (e.g. full-frame 24X36mm CMOS sensors - 36.2 megapixels) that produces high quality images.

The sprays from the two orifices under investigation were characterized in terms of the spray-tip penetration distance (i.e. the maximum penetration distance of the spray) and the cone angle of the spray, as a function of injection pressure. All the tested cases took place at the same chamber density. The raw images, once acquired, are post-processed in a self-developed code. All images undergo a common process, aimed at extracting the contour of the sprays. The procedure provides for the selection of the field of interest of the image, the gray scale conversion, the background image subtraction and the operation of thresholding to isolate the spray region and its boundary. Once the spray contour is isolated, additional steps allow obtaining the jet penetration and the spray angle. The gray-threshold value here adopted is 10/255, where 0 is black and 255 is white. Fig. (4) shows a sample of the image post processing.

Fig. (4) Image post-processing; raw image (A); gray scale conversion (B); background subtraction (C); spray region extraction (D).

The spray shape is characterized by spray penetration and spray angle. The spray penetration is obtained by measuring the distance of the extreme point of the jet boundary from the hole exit section. The trend over time of the jet penetration is determined on the basis of the temporal spacing between one image acquisition and the other, which was set equal to 25 μs.

The spray angle is calculated according to what reported in the literature by [23Desantes, J.M.; Payri, R.; Salvador, F.J.; Gil, A. Development and validation of a theoretical model for diesel spray penetration. Fuel, 2006, 85(7-8), 910-917.
[http://dx.doi.org/10.1016/j.fuel.2005.10.023] , 24Shao, J.; Yan, Y. Digital imaging based measurement of diesel spray characteristics. IEEE Trans. Instrum. Meas., 2008, 57(9), 2067-2073.]. After the thresholding operation, in each frame, the spray region area is calculated. The spray angle is defined as the angle of the isosceles triangle having as the value of the height the jet penetration. The opening angle was also evaluated for fractions of the spray penetration, with the aim of obtaining information on the radial development of the jet along the direction of penetration. As previously mentioned the tested prototypes are multihole type, with four holes; thus, the visualizations were conducted isolating the attention on just one of the four sprays.

The analysis of the characteristics of the sprays has been performed considering three different injection conditions, as reported in Table 3.

Table 3
Test cases

The investigation on the spray development was preceded by the characterization of the mechanical-hydraulic behavior of the two nozzles, using the anechoic tube to reconstruct the Rate of Injection (ROI) trends over time.

The amount of injected fuel has been shown to be independent of the nozzle type, for all the pressure levels considered (Table 4); the maximum deviation is observed in the cases at 1000 bar, and it does not exceed 2.5%.

Table 4
Injected mass.

Once it has been found that the tested nozzles are able to inject the same fuel quantity with injections of the same duration, the ROI trends are analyzed. The adopted Slot does not alter the mechanical-hydraulic behavior of the nozzle; in fact the start, the end and the duration of the injection do not depend on the nozzle type.

At the lower pressure level, ROI trends are equivalent, as shown in Fig. (5). At higher pressures (Figs. 6 and 7), the ROI of the slot cases shows some deviations from the standard cases. Such deviations are located in the central injection phases. Indeed, the ROI values around to the maximum show some dependency on the type of nozzle; in the Slot case, the values reached before the maximum exceed those of the circular hole; this behavior is reversed after the maximum, in which the values relating to the Round case prevail. This behavior probably reflects that the three-dimensional flow pattern within the nozzle is affected by the hole shape. The viscous forces at the interface between fuel and the nozzle surface can also reflect the nozzle hole shape so that, among the fuel properties, viscosity could be the major factor influencing the flow development.

Fig. (5) Comparison between ROI (Rate of injection) curves – 700 bar injection pressure.

Fig. (6) Comparison between ROI (Rate Of Injection) curves – 1000 bar injection pressure.

Fig. (7) Comparison between ROI (Rate of Injection) curves – 1400 bar injection pressure.

The layout of the nozzle hole cross section influences the spray development, depending on the injection pressure. At relatively low pressure (700 bar), the penetration trends are quite similar, (Fig. 8).

Fig. (8) Spray penetration distance vs. time; 700 bar rail pressure.

As the injection pressure increases (1000 bar injection pressure, Fig. (9) the spray produced by the Slot hole is characterized by a more marked penetration.

Fig. (9) Spray penetration distance vs. time; 1000 bar rail pressure.

The higher is the injection pressure, the higher is the difference between the trends, Fig. (10). Compared to the 1000 bar case, the instant at which the trends detach from each other is moved backward in time and the difference in the penetration exceeds 20 mm.

Fig. (10) Spray penetration distance vs. time; 1400 bar rail pressure.

The results reported in the following show that the nozzles hole layout causes a significant influence also on the spray angle. The differences were sought both close to the hole outlet section and in the so-called “steady” spray portion, in agreement with what is reported in the literature [23Desantes, J.M.; Payri, R.; Salvador, F.J.; Gil, A. Development and validation of a theoretical model for diesel spray penetration. Fuel, 2006, 85(7-8), 910-917.
[http://dx.doi.org/10.1016/j.fuel.2005.10.023] , 24Shao, J.; Yan, Y. Digital imaging based measurement of diesel spray characteristics. IEEE Trans. Instrum. Meas., 2008, 57(9), 2067-2073.].

The near spray region is identified by the distance from the hole exit section, when it equals 60 times the diameter of the round hole, that is to say 16.2 mm. According to such a definition, the near spray angles are compared; in case of Slot layout a greater opening of the spray is found.

This effect is somewhat appreciable at 700 bar, (Fig. 11), but it becomes significant with increasing injection pressure. In fact, while in the Round case the jet opening grows slightly due to the rise of pressure, the Slot layout widens about 20 degree the spray angle at a maximum injection pressure, (Fig. 12).

Fig. (11) Near spray angle vs. time; 700 bar rail pressure, XY plane.

Fig. (12) Near spray angle vs. time; 1400 bar rail pressure, XY plane.

Also considering the overall spray region, differences in the opening angle are observable. The injection pressure level still exerts an interesting role on the spray development. At low pressure, in the first 100 μs of injection, the Slot hole produces wider sprays, reaching double values ​​in respect to the standard case, (Fig. 13). Such effect is reduced at higher pressure levels; the high values were seen in the first instants of injection disappear and the angle values are similar to the Round case; significant differences, around ten degrees, are instead observed in the second half of the injection process, (Fig. 14).

Fig. (13) Steady spray angle vs. time; 700 bar rail pressure, XY plane.

Fig. (14) Steady spray angle vs. time; 1400 bar rail pressure, XY plane.

The visualization of spray angle needs to be completed on the “minor” plane (ZY plane). With reference to Fig. (2), the results obtained on the minor plane are reported in the following, pointing out the remarkable points of interest; in the case of round hole, the near angle opening on the ZY plane tends to exceed the values on the XY plane, Fig. (15). Since the Round hole is perfectly circular, the causes of such a difference are tied to the flow conditions at the outlet section of the hole which depends, in turn, on the whole flow pattern within the nozzle.

Fig. (15) Near spray angle vs. time for the two different visualization planes; 1000 bar injection pressure.

Regarding the Slot hole, the development of the near angle on the ZY plane shows a first phase in which the trend is dominating. In this phase the greater spray angle occurs in the minor axis direction of the Slot section, with about twenty degrees maximum difference. This condition is reversed through the progress of injection, since the max spray angle is found along the minor axis, (Fig. 15).

Also the trends of steady spray angle depend on the geometry of the hole, as shown in Fig. (16). For both the nozzles, it was observed that the spray angle is greater on the ZY plane; in the case of the Slot geometry, this behavior is more marked.

Fig. (16) Steady spray angle vs. time for the two different visualization planes; 1000 bar injection pressure.

From a quantitative viewpoint, Fig. (17) shows the trends of the ratio between the steady spray angles on the two viewing planes, for the Round and Slot cases. It is evident that the Slot section alters the spray symmetry, doubling the same effect observed on the standard geometry.

Fig. (17) Steady spray angle XY/ZY ratio vs. time; 1000 bar injection pressure.