Pre-processing unsteady flows for clutter reduction in vortex visualization

Visualization of flow features, such as vortices, aids in analyzing complex unsteady (turbulent) flows and thus facilitate human cognition of flow phenomena. Typical data of unsteady flows are vector fields of velocity in four spatiotemporal dimensions (4D). Often empirical data of unsteady flows suffer from unknown measurement errors (i.e., undesired vectors). The undesired vectors lead to clutter in visualization. The clutter degrades information content and obstructs user interaction for analyses. As a feasibility study, this paper presents adaptive nonparametric approach (ANPA) to pre-process three-dimensional (3D) flow data at individual time frames. The pre-processing aimed to eliminate undesired vectors in the data. Tested on 4 empirical datasets with different intensities of vortex shedding, ANPA reduced undesired vectors, and thereby visual clutter, from the flow data without distorting vortices. The reduction was quantifiable by the decreased number of false positives in vortex identification. These observations indicate a potential to extend ANPA to pre-process 4D flow data for vortex visualization.