Hydrodynamic Classification of Natural Flows using an Artificial Lateral Line and Frequency Domain Features

Jeffrey Tuhtan, Nataliya Strokina, Gert Toming, Naveed Muhammad, Maarja Kruusmaa, Joni-Kristian Kämäräinen

Friday 3 july 2015

14:15 - 14:30h at North America (level 0)

Themes: (T) Hydro-environment, (ST) Ecohydraulics and ecohydrology

Parallel session: 16F. Environment - Ecohydraulic

The classification of natural flow signatures using conventional measurements in field applications is a challenging task. In this work, we present a new flow sensing device consisting of a fish-shaped sensor body equipped with an artificial lateral line system. Hydrodynamic classification under turbulent flow conditions is demonstrated both in an outdoor laboratory flume with obstacles as well as in a natural environment near and behind a large boulder. Lateral line pressure sensor data acquired in the time domain are converted into a set of frequency domain features. The spectrogram feature classifiers proposed in this work are shown to provide robust and accurate estimates of complex flow conditions which were previously not possible under field conditions. Furthermore, we show that classification using stereo sensor pairs consistently outperforms single sensors.