CALIBRATION OF THE CONSTANTS IN THE KELVIN-HELMHOLTZ RAYLEIGH-TAYLOR (KH-RT) BREAKUP MODEL FOR DIESEL SPRAY UNDER WIDE CONDITIONS BASED ON ADVANCED DATA ANALYSIS TECHNIQUES
Ming Jia, Haiyang Pan, Ye Bian, Zonghan Zhang, Yachao Chang, Hong Liu
Abstract
The Kelvin-Helmholtz (KH) Rayleigh-Taylor (RT) breakup model has been widely utilized in the simulations of diesel spray, whereas the calibrations of the five constants in the KH-RT model highly depend on the operator's experience. To overcome the shortcoming, advanced data analysis techniques were introduced into this study by employing genetic algorithm (GA) and global sensitivity analysis for the simulations using the Reynolds-averaged Navier-Stokes (RANS) turbulence model. First, a genetic algorithm was used to optimize the five constants for the representative cases of diesel spray under different ambient temperatures and pressures. Based on the optimal solutions obtained from the GA optimizations and global sensitivity analysis, the dominant parameters affecting the predicted liquid penetrations of diesel spray, including CRT and Cb, are identified. By fitting these optimal solutions, two correlations for CRT and Cb are derived as: CRT = 3.2 × ρamb-0.32Тamb-0.2 and Cb = 4.24 × ln(ρamb) + 4.74. The change of the breakup length constant (Cb) for the introduction of the RT mechanism with the ambient density (ρamb) and the variation of the optimal child droplet size constant of the RT mechanism (CRT) with the ambient temperature and pressure (Tamb and ρamb) can be understood from the instability of the RT mechanism and the derivation of the RT breakup model, respectively. Extensive validations indicate that the derived correlations are suitable for diesel spray under wide ranges of ambient pressure, ambient temperature, injection pressure, and nozzle diameter for various experimental data sources in the literature.