Caracterización de flujos turbulentos en ingeniería hidráulica

Size: px

Start display at page:

Download "Caracterización de flujos turbulentos en ingeniería hidráulica"

Samson Waters
5 years ago
Views:

1 acústicos : Caracterización de flujos turbulentos en ingeniería hidráulica Departamento de Hidráulica Facultad de Ciencias Exactas, Ingeniería y Agrimensura Universidad Nacional de Rosario Junio de 2011

2 Outline 1 2 of turbulent 3 (ADV) 4, signal and turbulence parameters computation

3 The nature of turbulent The flow is unsteady, irregular, random and chaotic. (

4 The nature of turbulent motions of many scales. (

compressed turbulent gaseous cross-stream. Courtesy of M.

5 More detailed observations in laboratory experiments Breakup of an aircraft engine turbulent liquid fuel jet injected into a compressed turbulent gaseous cross-stream. Courtesy of M. Herrmann Hydraulic jump formed downstream a sluice gate. Courtesy of E. Trierweiler

6 The study of turbulent (Pope 2000) Navier-Stokes equations (1845) (laminar or turbulent fluid ) Modelling: theoretical studies, aimed at developing tractable mathematical models that can accurately predict properties of turbulent. Discovery (the process of finding information): experimental (or simulation) studies aimed at providing qualitative or quantitative information about particular. of turbulent

7 The study of turbulent Flow visualizations of turbulent Flow measurements

8 Flow measurements Defining the optimun... Measurement technique., signal and turbulence parameters computation (including confidence intervals)....in order to accurately characterize turbulent flow.

9 Measurement technique PIV Hot-wire LDA ADV UVP PTV ADCP Laboratory (scale of experimental facilities) Field

10 , signal and data analysis techniques. Requirements for making the water velocity signal representative of the turbulent process in the water. strategy (sampling frequency and sampling time). Error analysis (adequate signal post-). Computation of turbulence parameters (including confidence intervals).

11 Flow measurements Defining the optimun... Measurement technique., signal and turbulence parameters computation (including confidence intervals)....in order to accurately characterize turbulent flow.

12 (ADV) Courtesy of Sontek R

13 (ADV) ADV measures 3-D water velocity components in a remote sampling volume using the shift principle. Courtesy of Sontek R

14 The effect Courtesy of Sontek R

15 ADV acoustic operation Single-Pulse (ADCPs) Pulseto-Pulse Coherent (ADVs) Measures shift in frequency between emitted and reflected pulses. Spatial resolution is a function of the frequency of the emitted pulse (3m-3MHz and 100m-500KHz). Measures the phase shift between the reflected pulses at time t and t + τ. Higher spatial resolution (1cm-10MHz). Introduces ambiguity errors (aliasing).

16 Determination of bistatic velocity v i = C dφ 4πf ADV dt, i = 1, 2, 3. Courtesy of C. Kraus

17 Dual pulse repetition rate Unequal pulse repetition rates, τ 1 and τ 2, separated by τ D. Courtesy of S. McLellands T s = 3(τ 1 + τ D + τ 2 + τ D )

18 Cartesian coordinate system transformation Bistatic velocity (v i ) Cartesian coordinate system (u i ). u 1 u 2 u 3 = a 11 a 12 a 13 a 21 a 22 a 23 a 31 a 32 a 33 v 1 v 2 v 3 the transformation matrix a ii is provided by the manufacturer. ADV computes velocity u i at f s = 1/T s

19 Sampled and recorded velocity Filtering effects due to the sampling strategy used by ADVs. ADV samples velocity at f s. User records velocity at f R. Courtesy of S. McLellands User records velocity u i at f R

20 ADV data validation Signal-to-Noise Ratio (SNR) > 15dB. Courtesy of S. McLellands

21 ADV data validation Correlation (COR) > 70%. Low COR values indicates loss of coherence during pulse propagation through the fluid ( noise). G ii [cm 2 /s] Turbulence spectrum G uu G vv G ww 5/3 G ii [cm 2 /s] White noise f[hz] f[hz]

22 to keep in mind... Pulse-to-pulse coherent method and ambiguity errors (spikes). technique and loss of coherence during pulse propagation through the fluid ( noise). strategy and f s and f R (filtering effects).

23 Flow measurements Defining the optimun... Measurement technique., signal and turbulence parameters computation (including confidence intervals)....in order to accurately characterize turbulent flow.

24 T = L/U c Definition of the optimum sampling Determination of the sampling time T serie. T is the turbulence time scale. L is the energy containing eddy length-scale. U c is the convective velocity. T serie 20T for mean velocity (ε < 10%). T serie 400T for variance and power spectrum (ε < 5%).

25 Dimensionless frequency F = f R L/U c > 20. f R is ADV recording frequency. L is the energy containing eddy length-scale. U c is the convective velocity. Definition of the optimum sampling Determination of the sampling frequency f R. Courtesy of C. M. García

26 Example - Open Channel flow time T serie and sampling frequency f R?.

27 Spikes. Signal and turbulence parameters computation including uncertainty analysis noise. Sources of error in ADV s turbulence measurements. Filtering effects. Uncertainty analysis for defining confidence intervals. Moving Block Bootstrap method (MBB).

28 Sources of error - Spikes 200 Raw turbulence water time signal (open-channel flow laboratory). 200 u[cm/s] u[cm/s]

29 Sources of error - Spikes Identification and replacement of spikes using PSTM algorithm (Goring and Nikora 2002). u[cm/s] u[cm/s]

30 Sources of error - noise White noise characteristics (Voulgaris and Trowbridge 1998). noise do not affect mean velocities (its mean is equal to zero). Reynolds Stresses are not affected by noise. noise produces decorrelation of the signals. Thus, temporal scales are low biased. kinetic energy is biased high.

31 Sources of error - noise G ii [cm 2 /s] Turbulence spectrum 10 0 G uu 10 1 G vv G ww 5/ f[hz] Identification and quantification (plateau at high frequencies). G ii [cm 2 /s] Turbulence spectrum f[hz] G uu G vv G ww 5/3

32 Sources of error - Filtering effects Due to ADVs digital averaging (García et al. 2005). f s = 1/ t x ADV sampling frequency. f R = 1/ t y = f s /N recording frequency. Conceptual model Courtesy of C. M. García

33 Sources of error - Filtering effects F = f R L/U c 20, filtering is about 10% (variance at f R is 90% of the variance at f s ). On variance. Courtesy of C. M. García

34 Uncertainty analysis - confidence intervals Statistical errors due to sampling a random signal Moving Block Bootstrap method (García et al. 2006). Block length and turbulence integral time scale. Courtesy of C. M. García

35 Sources of error - Variance computation, García and Lopardo (2011b) Signal post- technique and uncertainty analysis of ADV s turbulence measurements on free hydraulic jumps Journal of Hydraulic Engineering, ASCE. RECV [%] spikes (stage 4) 60 noise (stage 7) sampling strategy (stage 10) x/(z 2 z 1 )

36 Confidence intervals, García and Lopardo (2011a) Discussion of Energy dissipation and turbulence production in weak hydraulic jumps by E. Mignot and R. Cienfuegos. Journal of Hydraulic Engineering, ASCE. U is lower than 3% of measured values. < uw > between 15% and 60% of measured values. < uw > du/dz between 40% and 200% of measured values.

37 Collaboration Universidad Nacional de Córdoba, Facultad de Ciencias Exactas, Físicas y Naturales. Universidad Nacional del Litoral, Facultad de Ingeniería y Ciencias Hídricas. Instituto Nacional del Agua.

38 Research group UNR

39 Thanks!

Acoustic Doppler Velocimeters (ADV) Performance Curves (APCs) sampling the flow turbulence

Acoustic Doppler Velocimeters (ADV) Performance Curves (APCs) sampling the flow turbulence Carlos Marcelo García 1, Mariano I. Cantero 1, Yarko Niño 2 and Marcelo H. García 1 1 Ven Te Chow Hydrosystems