Spatiotemporal wall pressure forecast of a rectangular cylinder with physics-aware DeepUFNet

2025-08-05

AI Summary

To address the spatiotemporal prediction challenge of wall pressure distributions around rectangular cylinders, this paper proposes DeepUFNet—a physics-informed deep learning model that synergistically integrates U-Net’s spatial encoding capability with Fourier Neural Network’s frequency-domain modeling strength. A novel physics-driven high-frequency loss control mechanism is introduced, employing a dynamic weight β to enhance fidelity in capturing high-order frequency fluctuations and pressure variance, while enabling robust spatiotemporal extrapolation from sparse inputs. Trained on a high-frequency wall-pressure database acquired from wind-tunnel experiments, DeepUFNet achieves excellent agreement with measurements across multiple metrics: statistical moments, instantaneous pressure evolution, power spectral density, spatial pressure distribution, and spatiotemporal correlation. It significantly outperforms existing single-time-step spatial prediction methods, demonstrating superior accuracy, generalizability, and physical consistency.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization

Abstract

The wall pressure is of great importance in understanding the forces and structural responses induced by fluid. Recent works have investigated the potential of deep learning techniques in predicting mean pressure coefficients and fluctuating pressure coefficients, but most of existing deep learning frameworks are limited to predicting a single snapshot using full spatial information. To forecast spatiotemporal wall pressure of flow past a rectangular cylinder, this study develops a physics-aware DeepU-Fourier neural Network (DeepUFNet) deep learning model. DeepUFNet comprises the UNet structure and the Fourier neural network, with physical high-frequency loss control embedded in the model training stage to optimize model performance, where the parameter $β$ varies with the development of the training epoch. Wind tunnel testing is performed to collect wall pressures of a two-dimensional rectangular cylinder with a side ratio of 1.5 at an angle of attack of zero using high-frequency pressure scanning, thereby constructing a database for DeepUFNet training and testing. The DeepUFNet model is found to forecast spatiotemporal wall pressure information with high accuracy. The comparison between forecast results and experimental data presents agreement in statistical information, temporal pressure variation, power spectrum density, spatial distribution, and spatiotemporal correlation. It is also found that embedding a physical high-frequency loss control coefficient $β$ in the DeepUFNet model can significantly improve model performance in forecasting spatiotemporal wall pressure information, in particular, in forecasting high-order frequency fluctuation and wall pressure variance. Furthermore, the DeepUFNet extrapolation capability is tested with sparse spatial information input, and the model presents a satisfactory extrapolation ability

Problem

• Forecast spatiotemporal wall pressure on rectangular cylinder
• Improve deep learning accuracy with physics-aware model
• Validate model using wind tunnel experimental data

Innovation

• Physics-aware DeepUFNet for spatiotemporal pressure forecast
• UNet and Fourier neural network combined architecture
• High-frequency loss control optimizes model performance