NYUAD-KAIST collaboration

NYUAD - KAIST collaborative research

Dr. Kim is actively engaged in a collaborative research program with the Korea Advanced Institute of Science and Technology (KAIST), focusing on spray flame dynamics through high-fidelity simulation methods. He and Prof. Hyung Sub Sim of KAIST have been working closely in this area since their respective postdoctoral appointments at Argonne National Laboratory and Sandia National Laboratories. Prof. Sim specializes in probing transient ignition phenomena using advanced high-speed optical diagnostics developed at Sandia. Within the current program, Dr. Kim plays a key role in mentoring KAIST Ph.D. students and offering technical guidance, helping them deepen their understanding of the underlying physics by leveraging high-fidelity simulation approaches such as large-eddy simulation (LES).

This research initiative combines high-speed optical diagnostics and numerical simulation to investigate unsteady ignition process, especially focusing on the low-temperature kinetics and its impact on the flame dynamics. This study benchmarks the Engine Combustion Network (ECN) Spray-A3 flame under engine relevant conditions. Therefore, the insights gained through this research initiative can provide guidance in the pursuit of developing future-generation advanced compression ignition (ACI) engines.

Pre-burn Constant-Volume Combustion Vessel (CVCV)

Figure shown on the left presents the constant-volume combustion vessel (CVCV) used in this study. Sprays were injected into the optically accessible chamber with its volume of 1.15 liters. The vessel is initially filled with a premixed combustible gas mixture. Following spark ignition, the mixture completely, causing both the pressure and temperature inside the vessel to rise. The pre-burn mixture reaches a peak temperature of nearly 1800 K in less than 200 ms. After the pre-burn event, the temperature gradually decreases. This cooldown period allows sufficient time for the pre-burn gases to reach chemical equilibrium. Fuel injection occurs when the desired temperature and pressure conditions are achieved.

High-Speed Optical Diagnostics (PLIF, Schlieren, and OH* chemiluminescence)

Planar laser-induced fluorescence (PLIF) is an optical diagnostic technique widely used for flow visualization and quantitative measurements. In this study, the PLIF technique allows to visualize quantitative excitation of formaldehyde (CH2O) and Polycyclic Aromatic Hydrocarbons (PAH) as indicators of low-temperature ignition and soot precursor. The PLIF configuration used in this study includes a quasi-continuous burst-mode Nd:YAG laser (QuasiModo TM , Spectral Energies, LLC). The laser system provides 355-nm radiation at a high-repetition rate of 50 kHz over a 6.0 ms period, corresponding to 300 sequential 70-mJ laser pulses formed into a 45 mm × 0.5 mm laser sheet. Unlike the typical HCHO PLIF setup, a non-intensified high-speed CMOS camera was used in this study to capture the visible bands of the CH2O spectrum.

Diffused Back-Illuminated Extinction Imaging (DBI-EI) setup

The DBI-EI setup was used to identify the liquid phase boundary of the spray; thereby capturing the liquid penetration length. This setup includes an ultra-fast LED (406 nm, ~15 nm FWHM), a field lens, a large-area (100 mm diameter) engineered diffuser, and a high-speed camera (Vision Research Phantom v2512) fit with a 50-mm f/1.2 Nikkor lens and a 2-diopter close-up lens. The light collected by the camera was spectrally filtered by a 405 ± 5-nm band-pass filter combined with an absorptive neutral density filter of OD = 1.0. The choice of a short incident light wavelength was motivated by the higher diagnostic sensitivity to early incipient soot particles as a result of a higher extinction coefficient.

Test Condition: ECN Spray-A3 flames

This study employs a ECN Spray A-3 test condition using the A-3 injector which has a 93 µm of orifice diameter and k-factor of 3.2. Details of the benchmarking Spray A-3 operating condition is described in the table (left).

A commercially available CFD software package, CONVERGE was used to facilitate the high-fidelity large-eddy simulations (LESs). The computational domain was configured to imitate the Sandia CVCV setup with the injector mounted on the upper surface of the domain facing downward. Based in the base grid scale of 8 mm, fine grid points were set to be clustered along the spray plume trajectory by employing fixed embedding and adaptive mesh refinement (AMR). The minimum grid scale yields to 62.5 µm, which is relevant scale to capture important mixing scale observed in the Spray A-3 injector mixing process.

Non-Reacting Spray: Evaluation on the liquid/vapor penetration length (LPL/VPL)

Movie_nonreacting_DBI_LES_SCH_3x1_935K.mp4

The injected fuel initially enters in the liquid phase and subsequently undergoes evaporation, forming a fuel-vapor boundary as it evolves. Simultaneously identifying the extents of both liquid and vapor phases in the experiment is crucial, as such measurements provide validation for the predictive accuracy of the LES simulation in terms of fuel-air mixing behavior.

The animation and figure on the left illustrate the spatiotemporal evolution of the spray development observed in both the experiment and the simulation, demonstrating strong agreement and confirming the reliability of the numerical setup.

Reacting Spray: Characteristics of ignition process

During the ignition process, low-temperature chemical kinetics promote the formation of formaldehyde (CH₂O), which serves as an indicator of the ignition onset. The animation below visualizes the spatiotemporal evolution of CH₂O formation observed in the PLIF diagnostics (left) and predicted by the LES simulation (right) under the ambient air temperature of 935 K.

Movie_PLIF_LESCH2O_AJKLDK935K.mp4

Additional tests were conducted by varying the ambient temperature from 800 K to 1019 K. Both the experimental measurements and LES simulations successfully captured the CH₂O distribution and its inception behavior. The key findings are as follows:

In the experiments, a PAH signal was detected under high ambient temperature conditions. (Dynamics of PAH signal was not modeled in the LES)
Both experimental and LES results reveal that the CH₂O inception region shifts upstream as the ambient temperature increases.

Following the low-temperature kinetic stage, high-temperature reactions become dominant, leading to the formation of OH* within the flame. The spatiotemporal behavior of this process is examined here. The animation below presents the evolution of the high-temperature reaction zone, visualized through OH* chemiluminescence in the experiment and predicted by the LES simulation.

Movie_OHstar_LESOH_AJKLDK935K.mp4

Additional tests were conducted by varying the ambient temperature from 800 K to 1019 K. Both the experimental measurements and LES simulations successfully captured the quasi-steady state OH (OH*) structures. Key observations are summarized below:

The experiment shows a more upstream-localized OH distribution.
At lower ambient temperatures, a noticeable discrepancy appears between the LES-predicted and experimentally measured flame lift-off lengths (FLOL), warranting further discussion, which is a scope of upcoming study.

Page updated

Google Sites

Report abuse