An Image and State Information-Based PINN with Attention Mechanisms for the Rapid Prediction of Aircraft Aerodynamic Characteristics

Prediction of aircraft aerodynamic parameters is crucial for aircraft design, yet traditional computational fluid dynamics methods remain time-consuming and labor-intensive. This work presents a novel model, the image and state information-based attention-enhanced physics-informed neural network (ISA-PINN), which significantly improves prediction accuracy. Our model incorporates the following innovations: the designed attention module dynamically extracts hidden features from pattern data while selectively focusing on relevant dimensions of target information. Meanwhile, the image-information fusion module combines multi-scale geometric features derived from aircraft images to enhance the overall prediction accuracy. By embedding aerodynamic equations, the model maintains physical consistency while enhancing interpretability. Extensive experiments validate the effectiveness of our model for rapid aircraft aerodynamic parameter prediction, achieving a significant reduction in prediction error that improves performance by 29.25% in RMSE and 37.99% in MRE compared to existing methods. A 6.12% error increase on the test set confirms the model’s robust generalization ability.