Deep Learning and Artificial Intelligence
In many practical applications empirical (0-D and 1-D) models are implemented to use efficient and robust computational algorithms. Traditional methods for developing and tuning of these simplified models usually combine engineering sense (and physical intuition) with basic statistical optimization techniques on limited data sets.
The rise of the state-of-the-art high-performance computing facilities as well as advanced simulation tools can provide us with high fidelity data sets (big data) to analyze high-dimensional, complex dynamical systems. Big data allow us to develop new machine learning algorithms like deep neural networks (DNNs). In the last decade, DNNs have become a dominant data mining tool for big data applications. Here, we develop new DNNs architecture and the data-driven adaptive algorithms to increase the accuracy of the turbulence models and different approaches in multi-phase flows based on the generalized framework of Computational fluid dynamic (CFD).