Submitted to: International Association of Hydraulic Research International Symposium
Publication Type: Proceedings
Publication Acceptance Date: 3/17/2016
Publication Date: 7/13/2016
Citation: Ursic, M.E., Wren, D.G., Langendoen, E.J. 2016. A method for identifying boundary interference in PADV data. International Association of Hydraulic Research International Symposium. 8th International Conference on Fluvial Hydraulics. River Flow 2016. July 12-15, 2016 St. Louis, MO. pp. 137-138.
Interpretive Summary: Recent commercialization of profiling acoustic Doppler velocimeters (PADVs) has enabled researchers to measure velocities at high frequencies simultaneously at specified increments over the instrument measurement range. Data resulting from acquisition include information about the signal strength (amplitude) and correlation between measurements which are typically used to discern the quality of the data. Unfortunately, the use of these quality indicators are affected by the boundary, where the emitted acoustic signal is returned with high signal strength and an increase in correlation between velocity measurements. Increases in these quality indicators would be ideal in free-stream flows but are false positives when due to boundary interference. Data that are affected by the boundary exhibit an arbitrary reduction in turbulence and velocity quantities which could lead researchers to derive false conclusions about the flow field. The purpose of this paper is to describe a method developed to identify and remove data that are affected by boundary interference. Results were shown indicating potential problems from boundary interference as well as the effectiveness of the method.
Technical Abstract: Recent commercialization of profiling acoustic Doppler velocimeters (PADVs) has enabled researchers to measure velocities at high frequencies simultaneously at specified increments over the instrument measurement range. The quantity of data output by PADVs can be large, hence robust quality control algorithms are essential to remove anomalous data. While a variety of methods exist for removing or replacing anomalous data in a velocity time series, few methods exist for identification of bins within a profile that are contaminated by boundary interference. Measurements near the boundary exhibit a velocity reduction, caused by interaction of the boundary with the insonifying acoustic field, which is beyond what would be expected from boundary effects on the flow. This may lead to false conclusions about the flow field. Removal of data that is subject to boundary interference is difficult due to a spurious increase in instrument signal quality parameters, such as correlation or signal-to-noise-ratio, which typically indicate good signal quality. This paper introduces a new filter method based on the statistical analysis of beam amplitude profiles away from and near a solid boundary. Results show that downstream velocity and Reynolds shear stress are indeed significantly reduced in the near bed region for the unfiltered data with boundary interference.