Next-Generation Turbulence Modeling: Integrating Classical Methods, Machine Learning, and Quantum Computing

سال انتشار: 1404
نوع سند: مقاله کنفرانسی
زبان: انگلیسی
مشاهده: 41

فایل این مقاله در 14 صفحه با فرمت PDF قابل دریافت می باشد

استخراج به نرم افزارهای پژوهشی:

لینک ثابت به این مقاله:

شناسه ملی سند علمی:

TETSC05_013

تاریخ نمایه سازی: 17 دی 1404

چکیده مقاله:

Turbulence, a nonlinear, chaotic, and multiscale phenomenon, governs fluid motion in both natural and engineered systems. This paper provides a comprehensive analysis of classical and advanced turbulence modeling techniques, emphasizing their evolution, applications, and integration with emerging computational technologies. Traditional methods such as Direct Numerical Simulation (DNS), Reynolds-Averaged Navier–Stokes (RANS), and Large Eddy Simulation (LES) have laid the foundation for modern hybrid approaches like Detached Eddy Simulation (DES) and Scale-Adaptive Simulation (SAS), which balance accuracy and computational cost. Recent developments highlight the growing influence of machine learning (ML), physics-informed neural networks (PINNs), and quantum algorithms in overcoming the limitations of classical models. These tools enhance predictive capability and efficiency by uncovering hidden patterns within turbulent flow data and accelerating solution convergence. Applications span aerospace, energy, environmental, and biomedical fields—from optimizing aircraft aerodynamics and wind turbine efficiency to predicting cardiovascular flow instabilities and pollutant dispersion. The paper also explores the role of digital twins, which integrate real-time sensor data with adaptive LES models, enabling dynamic updates for industrial systems. Despite significant advances, challenges persist in model interpretability, validation, and scalability, particularly in data-driven and hybrid frameworks. The future of turbulence modeling lies in the synergistic convergence of artificial intelligence, high-performance computing, and quantum simulation, promising breakthroughs toward a universal turbulence solver capable of unifying physics-based and data-driven paradigms for complex, real-world fluid dynamic problems.

کلیدواژه ها:

Turbulent flow modeling ، Computational fluid dynamics (CFD) ، Machine learning in fluid mechanics ، Multiscale turbulence

نویسندگان

Aliasghar Azma

School of Hydraulic Engineering, Faculty of Infrastructure Engineering, Dalian University of Technology, Dalian ۱۱۶۰۲۴, China

Yakun Liu

School of Hydraulic Engineering, Faculty of Infrastructure Engineering, Dalian University of Technology, Dalian ۱۱۶۰۲۴, China