Delphi Automotive Systems
For over half a century the fuel injection engineering community has been pursuing knowledge-based nozzle design optimization for high-pressure diesel injection. One sticking point has been the lack of understanding of the primary breakup process, which involves highly complex multiphase and multiscale fluid dynamics phenomenon. Experimental investigation of this problem for real applications is challenging because of the very high injection pressures (> 2500 bar), and very small nozzle hole sizes (<90 μm). So far there is no effective measurement technique available for the characterization of field turbulence inside the injector nozzle. In addition, optical measurements using the transparent nozzle technique can only access the cavitation phenomenon in scale-up nozzles at laboratory injection pressures, and cannot penetrate the very dense near-nozzle spray. High resolution x-ray phase contrast imaging (PCI) is suitable for the primary breakup process characterization, but is still not feasible for the simultaneous visualization of the in-nozzle flow and the near-nozzle spray, so it does not permit direct investigation of the cause and effect. Due to these limitations in knowledge and tools, fuel injector nozzle development largely relies on an empirical and iterative approach.
The Advanced Injection and Combustion Center at Delphi Automotive Systems Luxembourg has carried out detailed investigation of the primary breakup process using up-to-date theoretical and experimental tools. High-resolution level-set large-eddy simulation (LES) using ANSYS Fluent resolved the scales and dynamics of the nozzle flow and tracked the liquid—gas interface of the near-nozzle spray during the atomization process. Simultaneously, high-resolution x-ray PCI was applied to visualize the near-nozzle spray. The Fluent simulation successfully predicted the spray patterns recorded by the x-ray PCI images for an innovative nozzle design. LES was also applied to a number of different nozzle designs with contrasting geometric features to focus only on the in-nozzle flow dynamics. For each case, good correlations were found between the predicted vortex flow patterns and the measured spray patterns. It was discovered that the process of the fluid flow entering the nozzle hole triggers vortex shedding, which further initiates liquid surface deformation and ligament formation in the primary spray breakup. This finding explains why the injector nozzle design parameters (seat-sac, hole inlet rounding, taper, needle shape and needle lift) have a significant influence on the spray formation and provides a new understanding of the primary breakup mechanisms in high pressure fuel injection.
The correlations between the vortex flow pattern and the primary spray breakup pattern established in this work provide convincing evidence for the vortex-driven atomization mechanism. With this understanding, fuel injector nozzle designs can be optimized by control and optimization of the vortex. Furthermore, the LES workflow provides a diagnostic function for the nozzle flow dynamics, which overcomes the limitations of the available measurement techniques.