SLIDE 1

Introduction:
The in-nozzle cavitation flow during the fuel injection process affects the
combustion behaviour in large 2-stroke marine engines. Transparent nozzle
geometries have been used to link the cavitation phenomena with spray and
combustion characteristics.
SLIDE 2
Initially, it is necessary to know the phenomena of cavitation. It occurs
when the local pressure drops below the vapour pressure of the liquid at local
temperature. This leads to the formation of vapour-filled cavities in the liquid.
This contributes to the atomisation of the fuel particles.
SLIDE 3
The figure shows a schematic top view illustration of a standard fuel injector
with its mounted nozzle tip together and an enlarged detail on the five orifices.
Three single-orifice nozzle types have been selected based on the design
characteristics of the five-orifice nozzles from large two-stroke marine Diesel
engine fuel injectors. A set of sharp-edged and hydro-erosive ground single-
orifice nozzles are investigated in a constant-volume chamber. The nozzle
geometries have been made out of transparent plastic and metal to get the
maximum valuable information. Quartz glass and polymethyl methacrylate are
the most used materials.
SLIDE 4
Hydro-erosive grinding (HEG) is a process for radiusing orifice edges by flowing
an abrasive fluid through.
The images show sharp-edged and hydro-erosive ground orifice of the three
nozzle types N101, N104 and N105. Dark areas within the walls of the orifice
represent gaseous flow, i.e. cavitation while bright areas represent liquid flow.
The fuel enters the nozzle main bore from the top left side, enters the orifice,
and flows downwards towards the exit of the orifice at the bottom of the
images.
SLIDE 5
The nozzle geometries along with the operating conditions of the spray and
combustion chamber and optical setup constitute the experimental setup.
---------------------------------------------------------------------------------------------------------
SLIDE 6
Three different nozzle types, whose detailed side view and top view are shown
in the figure, are investigated without and with hydro-erosive grinding,
resulting in a total of six data sets.
The nozzle N101 has a centrically arranged, orthogonal orifice, N104 an
eccentrically arranged, orthogonal orifice and N105 a centrically arranged but
angled orifice. Using one instead of five-orifice nozzles enables a better
assessment since the in-nozzle flow and sprays are not interacting. Also, the
computational time for CFD validation is reduced and the optical acquisition is
simplified. But this complicates the approach of extending the experimental
results to real nozzle designs. Moreover, no flow interactions between the
orifices occur that could also influence the cavitation behaviour.
SLIDE 7
The experiments were performed in a constant volume spray and combustion
chamber. Optical access is provided by multiple sapphire windows with
maximum visible diameter of 150mm. A regenerator, two high-pressure air
bottles and valves define the pressure and temperature of the air in the spray
chamber. This process to create the combustion atmosphere, also allows to
create strong swirl motion of the charge. The pressure and temperature in the
SCC are measured using piezo-capacitive pressure transducers.
SLIDE 8
Spray and combustion chamber operating conditions:
9 MPa back pressure at a temperature of 900 K and a swirl velocity of
approximately 20 m/s. The SCC cover was heated to approximately 450 K to
simulate the cylinder wall temperatures. A standard Diesel fuel with a density
of 815.9 kg/m3, a kinematic viscosity of 2.112 mm 2/s and a net heat of
combustion of 43.16 MJ/kg was used. Fifteen fuel injections have been
recorded per nozzle type, resulting in a total of 90 experimental cases.
SLIDE 9
OPTICAL SETUP:

To evaluate the start of ignition with a high temporal resolution of 1 MHz, an

optical light probe prototype with a 305–315 nm band pass filter and a
photomultiplier was used. Two high speed cameras were used in an alternating
setup together with a lens-coupled image intensifier and a diode laser as a light
source.
---------------------------------------------------------------------------------------------------------
SLIDE 10
The given figure depicts the schematic top view of the Spray and Combustion
Chamber with the optical setup. The top view shows the SCC with part of the
regenerator and the mounted injector. Diode laser (f) is provided on one side
as the light source. The light fibre (e) transmits the light to the collimator (d),
which aligns the light into a parallel beam. The diffuser plate (c) scatters the
light onto the optical windows. On the other side are the high-speed cameras
(h, n, q) with their individual mirrors (s, t, u), lenses (r, o, l), and filters (p, m).
Using differently coated mirrors allows the use of a single optical axis. Mirror
(u) reflects below 370 nm, the mirror (t) reflects below 700 nm and mirror (s) is
a broadband mirror.
SLIDE 11
The different acquisition rates vary in different Fields of View. So, the optical
setup II is zoomed in towards the exit of the nozzle orifice while the optical
setup I shows the maximal field of view provided by the optical window. The
optical setup I used the camera (q) and camera (n) while the optical setup II
used the camera (h) and camera (n);
The Spray and combustion chamber (a) with the mounted injector (d) and
window holder (c) as well as the regenerator (b) are illustrated in the figure.
The three different fields of view (FOVI, FOVII & FOVIII) are enlarged using
dotted rectangles.
SLIDE 12
IMAGE ANALYSIS:

The figure shows the evaluation sequence of a typical spray image. Image (i)
shows a raw diffuse back illumination image during injection with an ignited
spray, obtained from the optical setup I with FOV 1. The spray is isolated by
background homogenisation and mapping and cropping as illustrated in (ii).
Using a threshold, the spray contour is then evaluated using a tracing
algorithm modified by a stopping criteria in image (iii). Image (iv) shows the
evaluated contour with the cropped, raw image to demonstrate the accuracy
of the method for the acquired images. The evaluated contour data of the
spray allows us to analyse the data regarding their spray morphology like spray
penetration, spray angle, spray-lift-off length, spray contour, spray area, etc.
---------------------------------------------------------------------------------------------------------
SLIDE 13
Under results, we’ll be discussing the general characteristics, the spray
morphology, and combustion properties.
SLIDE 14
In general results, the influence of the hydro-erosive grinding and reduced flow
resistance can be visualized by plotting the in-nozzle pressure curves during
the fuel injection. The solid lines represent the non-hydro-erosive ground, and
the dashed lines show the Hydro-ground nozzles. The plot shows how the
pressure rises with HG, especially for the nozzle N104. This effect occurs due to
the reduced flow resistance by creating smooth inlet radii using HG. For both
versions of the nozzles, N104 stands out. The in-nozzle pressure development
remains very similar for nozzles N101 and N105, while the eccentric N104 has a
lower pressure when sharp-edged, and an even further reduced pressure rise
time when HG.
SLIDE 15
Spray morphology:
Using the image analysis, the different spray contours have been averaged and
plotted individually for the three nozzle types used. The spray angle is
proportional to the nozzle diameter. The hydro-erosive grinding process
enlarges the orifice diameters significantly. The red curves represent HG nozzle
types and blue no HG. HG nozzle types show widened spray angles, as seen
clearly. The eccentric N104 has the widest angle while the standard N101
generates the narrowest.
SLIDE 16
The given figure shows the unedited, flame luminosity and diffuse back
illumination images of the sharp-edged and the hydro-erosive ground nozzle
N104 during the fuel injection.
(b) and (e) show typical flame luminosity images acquired with 20 kHz during
the quasi-steady-state fuel injection. There is a slightly larger flame area and
shorter flame lift-off length with the hydro-erosive ground nozzle in (e). The
two images match the exact timing of the diffuse back illumination images (a)
and (d), and although the field of view is slightly different, a comparison
reveals identical flame contours. Using both optical measurement techniques
helps to distinguish the differences between flame and spray.
---------------------------------------------------------------------------------------------------------
SLIDE 17
Using the flame luminosity optical setup, the flame lift-off length during the
quasi-steady-state fuel injection has been evaluated.
The start of ignition massively delays with Hydro erosive grinding no matter the
nozzle type. The chart shows that the sharp-edged nozzle N101 has the most
repeatable and earliest START OF IGNITION while the HG nozzle N104 has the
latest. Hydro erosive grinding roughly triples the START OF IGNITION time,
which can be explained with the reduced level of atomization due to improved
flow resistance and due to the enlarged orifice diameter resulting in roughly
double the injected fuel mass.
Regarding injected mass, the nozzle N101 has the lowest amount per injection
whereas the other two sharp-edged nozzle types have a very similar value. The
HG nozzle N104 achieves the highest and most stable mass flow.
SLIDE 18
The flame lift-off length of the sharp-edged, standard nozzle N101 is the
shortest. While it elongates with Hydro erosive ground for the standard N101
and angled N105 nozzle types, it slightly shortens for the eccentric nozzle
N104. This diverse behaviour again shows that the eccentrically arranged
orifice design significantly stands out among the three nozzle types.
The ignition location data is depicted in the figure. The earlier START OF
IGNITION for the sharp edged nozzles is clearly visible with the smaller and
shorter sprays, compared to the much larger and further penetrated sprays of
the HG nozzles. The Hydro erosive ground version of the eccentric nozzle N104
has the smallest standard deviations indicating a highly repeatable and very
firm fuel atomization as well as mixture formation. The swirl motion in the
combustion chamber leads to a one-side dominance of the ignition spot for
most cases, however, the sharp-edged (no HG) eccentric N104 and angled
(N105) nozzle seem to be the exceptions.

CONCLUSION
Different nozzle geometries show extensive differences in spray morphology
and combustion behaviour with regard to the effect of hydro-erosive grinding.
The hydro-erosive ground nozzle types have enlarged spray angles and higher
mass flow rates due to the influence on in-nozzle flow pattern and the
increased orifice diameters. Linking the in-nozzle cavitation flow images with
the spray morphology and combustion results gives a detailed analysis of the
fuel injection behaviour.