Understanding Turbulence Kinetic Energy ProfilesUp to Velocimeters - profiling
Understanding Turbulence Kinetic Energy ProfilesPosted by Kaan Koca at April 22. 2015
We carried out simultaneous measurements of the flow velocity with Vectrino II and Particle Image Velocimetry in order to identify where reliable velocity measurements start above the bed. The details of the experiments are as below:
Sampling rate: 75 Hz
Water depth: 30 cm
Profiler range: 40 - 74 mm
Profiling resolution: 1 mm
Bed material: Timber frame
Data collection: .xyz coordinates
Velocity algorithm: Adaptive (once)
I am currently experiencing some problems with the turbulence statistics. Turbulence seems to be very sensitive away from the transducer head. Urms and Vrms decrease sharply at approximately 5 cm away from the transducer and then increase on the other side (herein referred as a "bump"). This behavior in the signal is present for all the tests we conducted over smooth and rough bed. I also came across similar results in a few posts here and in a paper published by MacVicar et al. (2014): A quality analysis of the Vectrino II instrument using a new open-source MATLAB toolbox and 2D ARMA models to detect and replace spikes.
After reading some comments here, I have used the method by Hurther & Lemmin (2001) to estimate the variances of the noise as a recommended method. However, there still exists the so-called bump in the turbulence statistics. Please see the attached figures for an example test, showing the mean longitudinal velocity, turbulence kinetic energy, together with SNR, correlation, amplitude, and estimated noise (variance) profiles.
It appears to me that the shape of the TKE profile is not mainly caused by the noise. I have the following questions and would be glad if anyone could kindly assist me:
1- Has anyone seen similar turbulence statistics before? Does anyone happen to know how to deal with that?
2- Might the shape of the profile be a result of an unwanted probe calibration?
Any assistance is greatly appreciated. Thank you!
Re: Understanding Turbulence Kinetic Energy ProfilesPosted by Magnus Grøtterud at April 29. 2015
This is an ongoing support issue that is being discussed over email. It would be great if others had any experience with the matter that they would like to contribute with Kaan and other users.
Re: Understanding Turbulence Kinetic Energy ProfilesPosted by Siddhi Joshi at April 30. 2015
Hello- I am studying TKE profiles obtained using an ADV and have found Soulsby 1983 has described the theoretical profiles of the TKE (let me know your email if you would like the attachment as it is too big). The TKE shear stress correspondingly also increases with proximity to the bed. Whilst I am not sure of how to solve your calibration issues, I think you should also calculate the Turbulence Intensity- TI as 100*RMS u'/Umean, as described in http://www.cfd-online.com/Wiki/Turbulence_intensity and then plot this to see the local level variation in turbulence. The percentages give you an idea of the "turbulence cells" present in your profile. Then you will be able to identify the cause of the turbulence in the profiles and whether they are due to the ADV configuration... If you have any further references you feel are relevant I would be interested in reading them too!
Re: Understanding Turbulence Kinetic Energy ProfilesPosted by Amirreza Ghasemi at July 16. 2015
Did you figure out how to solve your problem? I have the same problem for TKE. I will be appreciate if anyone can help me about this problem.
Re: Understanding Turbulence Kinetic Energy ProfilesPosted by Kaan Koca at February 07. 2017
It has been some time, and we have been working with Vectrino Profiler, from which we (hopefully) gained some more experience. The reason for the local minimum of turbulent kinetic energy around the sweet-spot of the profiler can be due to the decorrelation of the signals (see Figure 1 in http://www.tandfonline.com/eprint/IYesmh57qFARYhnYuunI/full). The beams observe the same particles at the sweet-spot, correlation is higher and noise is lower (see supplement), but further away the beams observe different regions of the sampling volume. As a result, turbulence characterization is expected to be most reliable within 3 mm around the sweet-spot. Furthermore, I think that one should understand how the probe potentially affects (if it affects) the flow field within the sampling volume to better understand the instrument. Please note that we have not yet found the opportunity to use the upgraded Vectrino probe.