Computational Fluid Dynamics (CFD) generates high-dimensional spatio-temporal data. Data-driven method approach to extracting physical information from CFD has attracted widespread concern in fluid mechanics. Good results have been obtained for some benchmark problems. However, the performance on complex flow field problems has not been extensively studied. This paper uses a dimensionality reduction approach to preserve the main features of the flow field, on the basis of which unsupervised identification of the flow field states is performed using a clustering approach, which applies a data-driven to analyse the spatio-temporal structure to complex three-dimensional unsteady cavitation flows. The result show that the data-driven method is able to represents the changes in the spatial structure of the unsteady flow field over time and to visualise changes in the quasi-periodic state of the flow. In addition, we demonstrate that the combination of principal component analysis (PCA) and Toeplitz Inverse Covariance-Based Clustering (TICC) can identify different states of the cavitated flow field with high accuracy, indicating that the method has great potential for application in complex flow phenomena.

No abstract

No abstract

No abstract

No abstract