In the dual-hole nozzle-hole arrangement studied here the flow pattern established in the sac volume led to differences
in flow entry conditions for the two nozzle holes. This results in differences in the cavitation pattern in the two holes. At
inception of cavitation the first structures observed were in the free stream, close to the nozzle hole inlets and towards
the nozzle hole walls facing the bottom of the injector. It appears that they occur at the centre of vortex structures
in the flow. As the cavitation number increased, the flow regions filled with cavitation stretched further down streaminto the holes. The cavitating regions were located towards the bottom nozzle hole walls. In the top hole, cavitation was
then seen around the entire rim of the nozzle hole entrance, whereas in the bottom hole it only cavitated around half of
the hole entrance. At high cavitation numbers, the cavitation filled the entire nozzle hole exits. As expected the cavitation leads to flow asymmetries across the nozzle holes, with decreased velocities towards the walls where cavitation occurs as well as downstream of those cavitating regions. The transient injections display a similar pattern, with the cavitation starting at the entrance of the holes, spreading downwards. They also display the same asymmetry
with cavitation along the nozzle hole side towards the bottom of the injector. Ballistic imaging of the spray formation
implied the existence of a thick cone of large contiguous liquid structures, which would indicated aggressive, catastrophic primary break-up similar to that found in effervescent atomizers. In the near-nozzle jets observed in the
engine, a larger spray cone angle was observed for the top hole, whereas no large deviations from the geometrical
hole directions were observed. In order to further our understanding of in-nozzle flows and their effect on fuel jet development, it would be interesting to carry out those types of studies also inside pressurized and heated optical test chambers in the future, in order to more closely mimic actual engine conditions. It would also be of interest to study nozzle hole arrangements closer to actual production geometries.