Search Results

Accurate modeling of the internal flow and spray characteristics in fuel injectors is a critical aspect of direct injection engine design. However, such high-fidelity computational fluid dynamics (CFD) models are often computationally expensive due to the requirement of resolving fine temporal and spatial scales. This paper addresses the computational bottleneck issue by proposing a machine learning-based emulator framework, which learns efficient surrogate models for spatiotemporal flow distributions relevant for static coupling injection maps, namely total void fraction, velocity, and mass, within a design space of interest. Different design points involving variations of needle lift, fuel viscosity, and level of non-condensable gas in the fuel were explored in this study. An interpretable Bayesian learning strategy was employed to understand the effect of the design parameters on the void fraction fields at the exit of the injector orifice.

In order to improve understanding of the primary atomization process for diesel-like sprays, a collaborative experimental and computational study was focused on the near-nozzle spray structure for the Engine Combustion Network (ECN) Spray D single-hole injector. These results were presented at the 5th Workshop of the ECN in Detroit, Michigan. Application of x-ray diagnostics to the Spray D standard cold condition enabled quantification of distributions of mass, phase interfacial area, and droplet size in the near-nozzle region from 0.1 to 14 mm from the nozzle exit. Using these data, several modeling frameworks, from Lagrangian-Eulerian to Eulerian-Eulerian and from Reynolds-Averaged Navier-Stokes (RANS) to Direct Numerical Simulation (DNS), were assessed in their ability to capture and explain experimentally observed spray details. Due to its computational efficiency, the Lagrangian-Eulerian approach was able to provide spray predictions across a broad range of conditions.