Signal Postprocessing Technique and Uncertainty Analysis of ADV Turbulence Measurements on Free Hydraulic Jumps

• Published in: Journal of hydraulic engineering (New York, N.Y.) N.Y.), 2012-04, Vol.138 (4), p.353-357
• Main Authors: Romagnoli, Martín, García, Carlos M, Lopardo, Raúl A
• Format: Article
• Language: English
• Subjects: Applied fluid mechanics; Applied sciences; Buildings. Public works; Doppler; Doppler effect; Errors; Exact sciences and technology; Fluid dynamics; Fundamental areas of phenomenology (including applications); Hydraulic constructions; Hydrodynamics, hydraulics, hydrostatics; Instrumentation for fluid dynamics; Measurements. Technique of testing; Physics; Technical Note; Technical Notes; Time; Turbulence; Turbulent flow; Uncertainty; Variance
• ISSN: 0733-9429; 1943-7900
• DOI: 10.1061/(ASCE)HY.1943-7900.0000507

AbstractThis note presents experimental results on an error analysis of acoustic Doppler velocimeter (ADV) turbulence measurements in free hydraulic jumps incorporating a recent ADV signal processing technique. The work aims to quantify the influence of each error source in acoustic Doppler measurements (i.e., presence of spikes, Doppler noise, and filtering effects due to the ADV sampling strategy) on variance and integral time scale computations as well as to define the corresponding confidence interval for each reported value. Experiments were conducted for an upstream Froude number of 3.0. The relative error contribution of each source to both raw longitudinal variance and raw longitudinal integral time scale decreases downstream. Inside the zone where aeration effects and high turbulence levels are relevant, Doppler noise constitutes the most important error source. Further downstream, where aeration effects and turbulence levels diminish, the error produced by the presence of spikes can be neglected, whereas the effects of both Doppler noise and ADVs’ sampling strategy are balanced. The relative confidence interval for corrected longitudinal variance ranges between 30 and 10%, whereas the relative uncertainty level for the corrected longitudinal integral time scale varies from 80 to 50%. The results of this study point out the requirement for an adequate signal postprocessing technique and uncertainty analysis of ADV turbulence measurements.