When running CFD simulations with turbulence on, I often need to specify inlet parameters such as the turbulent intensity, turbulent viscosity ratio and/or the turbulent length scale. If these parameters are already known for a given experiment, I have no problems validating experimental data.

But I'm curious to know how these parameters are measured experimentally. How can I directly measure the turbulent kinetic energy and turbulent viscosity from a single experiment? Due to the inherent randomness of turbulence, would I need a lot of repeated experiments to measure it or can it be done in a single experiment?

How is this typically done in practice? Knowing that there is a trade-off between accuracy and cost, what would be a low cost measurement technique that is minimally intrusive to the overall flow while providing accuracy to within an order of magnitude with the fewest number of repeated experiments?


1 Answer 1


One method is to use a hot wire anemometer. This (about 50 pages, ignoring the product-specific stuff at the end) is a good read.

In principle, a hot wire anemometer is simply a metal wire (often platinum-coated tungsten wire with a diameter of a few micrometers) with an electrical current passed through it. The rate of heat loss from the wire depends on the air flow velocity over it, and the resistance of the wire changes with temperature.

A simple way to use the probe is to supply either a constant voltage or constant current, and measure the change in electrical resistance. A more common method in practice is to use an electrical circuit with a feedback loop that attempts to keep the resistance of the wire (and hence its temperature) constant. The time-dependent electrical power supplied to the wire is then equal to the amount of cooling from the air flow, which can be converted into the time-dependent flow velocity.

Hot wire probes are physically small (the length of the "hot wire" is of the order of 1 or 2 mm), and because of their low thermal inertia they can measure at frequencies up to 10kHz or higher to resolve short timescale features of turbulent flow.

  • 1
    $\begingroup$ It would be helpful to have an explanation or summary of the technique rather than a link as an answer. For one, links are rarely permanent. But more importantly, it would be helpful to understand the physics behind how the measurement is taken. $\endgroup$
    – Paul
    Commented Jul 25, 2016 at 3:11
  • $\begingroup$ @Paul If you want to summarize the document yourself and answer your own question, that's OK by me. Personally, I don't have the either time or inclination to do that amount of work for free. $\endgroup$
    – alephzero
    Commented Jul 25, 2016 at 20:02
  • 1
    $\begingroup$ Considering the complexity of the turbulence, how efficient is this method ? by the way the link to the article does not work anymore $\endgroup$ Commented May 6, 2020 at 7:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.