Multi-level ensemble Kalman Filter algorithms enhanced by Convolutional Neural Networks

The present research work proposes advancement for Data Assimilation strategies using Convolutional Neural Networks (CNN). More precisely, multi-level algorithms for the Ensemble Kalman Filter are enhanced by CNN tools, with the objective to reduce the discrepancy in the prediction between ensemble realizations performed with different models. The proposed methodology is assessed via the analysis of the flow through a cascade of NACA 0012 profiles for Reynolds R e = 1 000 and Mach M a = 0 . 5 . Depending on the angle of attack α , unsteady features of the flow can be observed. The results indicate that the usage of the CNN tools, which are trained using data from the DA procedure, significantly augments the accuracy of the low-fidelity models with little augmentation in computational costs. It is shown that the usage of the CNN tools provides a faster convergence of the Data Assimilation algorithms, which leads to a significant gain in terms of computational resources required. • Extended comparison of ensemble Data Assimilation techniques available in the literature using an online CFD-DA architecture. • Development of a Machine Learning, CNN-based, predictive tool to enhance the accuracy of the low-fidelity models included in the multi-fidelity algorithms. • Integrating the Machine Learning tools into the Data Assimilation online platform to perform on-the-fly DA predictions.