New data-driven models predict pressure losses in turbulent flows

Researchers have developed two new data-driven models, one using kriging and the other a neural network (NN), to predict pressure losses in turbulent flows across perforated plates. These models were trained on experimental data and consistently outperformed existing empirical formulas. The NN and kriging models showed good agreement with experimental measurements and are suitable for practical computational fluid dynamics applications when implemented as a source term in RANS equations. AI

IMPACT Introduces novel data-driven approaches for fluid dynamics modeling, potentially improving simulation accuracy.

RANK_REASON Academic paper detailing novel modeling approaches. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.LG →

• neural network
• kriging

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →