Parallel Social Group Optimization for physics-informed calibration of RANS turbulence models: Accuracy, robustness, and generalization

Puntos clave

• Improved accuracy of turbulence models is achieved through parallel social group optimization techniques, demonstrating notable advancements.
• The analysis reveals enhanced robustness across various turbulence scenarios, suggesting these methods can withstand different conditions effectively.
• Robustness is quantified through statistical validation, showcasing significant reductions in error rates under specific turbulence conditions.
• These findings may enable better modeling of fluid dynamics in engineering applications, but further testing in real-world settings is necessary.