Cavitation erosion refers to severe material damage caused by collapsing vapor structures near walls. During the collapse of vapor structures, high intensity pressure waves up to several GPa are emitted that can lead to damage of nearby surfaces. Compressible numerical flow simulations enable the numerical prediction of cavitation erosion by spatial and temporal resolution of such pressure impacts. However, these simulations usually only provide point-based pressure data. To obtain numerical substitutes for cavitation erosion pits, further post-processing steps are required, including clustering methods of the point-based pressure data. Using algorithms, spatially and temporally contiguous areas featuring high pressure are recognized and defined as clusters. In this contribution, we test the applicability of various clustering algorithms in the field of artificial intelligence (AI) and machine learning (ML) for this clustering process. We employ established algorithms as k-means and Density-Based or Density-Ratio Based Spatial Clustering of Applications with Noise (DBSCAN, DRSCAN). The clustered data and the performance of the clustering process are compared with a problem-specific, physically motivated algorithm. Additionally, the simulation results reveal new insights into the spatial and temporal distribution of potentially erosive events.     «

Cavitation erosion refers to severe material damage caused by collapsing vapor structures near walls. During the collapse of vapor structures, high intensity pressure waves up to several GPa are emitted that can lead to damage of nearby surfaces. Compressible numerical flow simulations enable the numerical prediction of cavitation erosion by spatial and temporal resolution of such pressure impacts. However, these simulations usually only provide point-based pressure data. To obtain numerical...     »