Enhancing prediction performance of experimental pressure on square cylinders with CFD data and deep transfer learning

Key Points

• Prediction performance significantly increased by integrating CFD data with deep transfer learning methodologies.
• This analysis identifies a 30% improvement in accuracy compared to traditional models.
• Assessment involved combining numerical simulation data of square cylinders with deep learning approaches.
• Highlights the need for novel techniques to better model pressure distributions in engineering applications.