Improved RANS Predictions Using Diffusion and Dissipation Sensitivity for Supersonic Jets Guided by WALE LES

This study improves RANS model predictions for a non-axisymmetric supersonic jet by incorporating turbulence anisotropy and by analyzing sensitivity of diffusion and dissipation using validated LES data. It is demonstrated that uncalibrated (baseline) RANS models significantly underpredict turbulence production in jet shear layers, leading to prolonged potential cores and reduced diffusion towards the jet centerline. To address this, the diffusion term is modified in the ω-equation of the k-ω SST model and the dissipation term in the k-equation for the quadratic constitutive relation (QCR) RANS, guided by WALE LES. This approach is extended towards a six-equation Reynolds Stress Model (RSM). The destruction term coefficient in the dissipation rate equation for six-equation RSM is fine-tuned guided by WALE LES data. During the fine-tuning process, various error metrics are tracked and are presented in this paper. The modifications enhance turbulence production, diffusion, and jet mixing, in turn reducing the potential core lengths while capturing turbulence anisotropy in the jet nearfield. The simulations utilize a widely available commercial solver, with validation provided through comparison with Particle Image Velocimetry (PIV) data for various flow conditions. Improved turbulence production and diffusion along with improved validation cases after conducting sensitivity analysis are presented.