A physics-informed diffusion model for high-fidelity flow field reconstruction

• Published in: Journal of computational physics 2023-04, Vol.478, p.111972, Article 111972
• Main Authors: Shu, Dule, Li, Zijie, Barati Farimani, Amir
• Format: Article
• Language: English
• Subjects: Computational fluid dynamics; Denoising diffusion probabilistic models; Super-resolution
• ISSN: 0021-9991; 1090-2716
• DOI: 10.1016/j.jcp.2023.111972

Machine learning models are gaining increasing popularity in the domain of fluid dynamics for their potential to accelerate the production of high-fidelity computational fluid dynamics data. However, many recently proposed machine learning models for high-fidelity data reconstruction require low-fidelity data for model training. Such requirement restrains the application performance of these models, since their data reconstruction accuracy would drop significantly if the low-fidelity input data used in model test has a large deviation from the training data. To overcome this restraint, we propose a diffusion model which only uses high-fidelity data at training. With different configurations, our model is able to reconstruct high-fidelity data from either a regular low-fidelity sample or a randomly measured sample, and is also able to gain an accuracy increase by using physics-informed conditioning information from a known partial differential equation when that is available. Experimental results demonstrate that our model can produce accurate reconstruction results for 2d turbulent flows based on different input sources without retraining. •Deep learning helps to reduce the computational cost of fluid dynamics simulation.•Deep learning models can reconstruct high-fidelity data from low-fidelity input.•Many existing deep learning models require low-fidelity data for model training.•Our diffusion-based model removes the requirement for low-fidelity training data.•Physics informed condition is utilized by our model to improve data reconstruction.