A series of optical diagnostics was used in a light-duty optically accessible Diesel engine to investigate the in-cylinder thermofluidic conditions, the spray and combustion characteristics, and compression-ignition (CI) knock. Most of the experiments were performed with Spray B from the Engine Combustion Network’s (ECN) series of standardized injectors and experimental conditions.

To determine the in-cylinder flow field, particle image velocimetry (PIV) was performed for both low-swirl and high-swirl conditions. From the resulting flow fields, the time evolutions of the turbulent kinetic energy and the swirl ratio were calculated. The in-cylinder temperature field at compression top-dead center was obtained from laser-induced fluorescence (LIF) of toluene evaporated into the intake charge. At low swirl, the ensemble-average temperature field was found to be almost uniform throughout the piston bowl. At high swirl, temperatures in the bowl are lower and radially inhomogeneous with a thicker boundary layer.

The ECN Spray B (three jets, injector 211201) was investigated by “fire-deck retro-reflection” with illumination by a pulsed light-emitting diode (LED). The liquid and the gas phase of the fuel spray were captured simultaneously with high temporal resolution (about 50 000 frames/second) and identified in the images in post-processing. The effects of ambient temperature and density, injection pressure, swirl level, and hole-to-hole variation on liquid length, vapor penetration, spray dispersion angle, ignition delay, ignition location, flame luminosity, and heat release rate were investigated.
The effects of swirl and hole-to-hole variation were found to be complex and are likely to be related to thermofluidic conditions and spray/flow interactions that are beyond the scope of this work. The quasi-steady liquid length and the ignition delay measured in this work were compared to data from the literature on Spray B in high-pressure vessels and engines.
The effect of facilities (in the sense of chamber size and flow field) was found to be less significant than the effect of the different diagnostics used in the various investigations.

To clarify the effect of diagnostics on the detection of ignition delay, multi-spectral high-speed imaging was performed. Chemiluminescence (CL) in the UV (from OH* and other species) and in the visible spectral range was recorded simultaneously by an intensified camera and a color camera, respectively.
The area-averaged flame luminosity in the flame contours, obtained with by segmentation via dynamic thresholding, was compared with pressure-based metrics close to ignition.
The results show that the OH*-CL and visible-CL signals roughly match but have different sensitivities to the low- and high-temperature reaction. The low-temperature reactions also make traditional heat-release analysis unreliable for detecting the ignition delay. The net pressure rise (the fired pressure minus the motored one) was a better metric.

Finally, knocking combustion after compression ignition was investigated. High-speed image sequences of combustion luminosity were processed based on the concept of optical-flow. The resulting flow fields show a “sloshing” movement of the flame that is due to acoustic resonance.
The corresponding oscillation frequencies were obtained from the velocity-vector time-series. They match those from pressure-trace analysis and theoretical calculations. Here, knock originated from auto-ignition of the end gas, resembling knock in spark-ignition engines. In one very severe cycle, a shock wave was seen.