Ladybug / Default windspeed in UTCI calculator component

grasshopper ladybug-legacy
1

Hi,

I’m using the Ladybug OutdoorComfortComponent to calculate UTCI values.

This component takes as input the windspeed mesured at 10m. It says that “if no value is plugged [in the windspeed input] the component will assume a very low wind speed of 0.05m/s, characteristic of most indoor conditions”.

Do you know why this default value of 0.05m/s has been chosen? It seems strange since UTCI originally accepts windspeeds only above 0.5m/s. Actually, the component returns the same UTCI value when windspeed ranges from 0 to 0.5m/s: are these results valid?

Thank you very much.
Julien

2

Given typical human walking speed of about 1.5m/s, you can imagine how slow 0.05m/s to 0.5m/s would be - at low speeds like these, heat transfer is more in the realm of free convection rather than forced convection so it’s not too surprising that the resulting UTCI are more or less the same.

If you look at the outdoor wind speed profile for any location, wind speed of less than 0.5 m/s corresponds to Beaufort scale of 0, where smoke rises vertically.

1 Like
3

Joseph’s answer is a good one. I’ll just add that the only reason why I put a default of 0.05 m/s instead of 0 is that I know it’s virtually impossible to get air speeds below 0.05 m/s around a human (natural convection of body heat at least gives you some air movement). But it’s virtually equivalent to 0 in this case, especially when you realize that UTCI assumes that the person’s metabolic rate is that of a walking adult and 0.05 m/s is clearly less than walking speed.

I’d be amenable to changing the default to something around typical walking speeds of 0.75-1.5 m/s if everyone thinks it makes more sense.

4

Thank you @chris and @josephyang for these great answers.

It makes sense indeed that under a certain threshold the effect of wind on the human body becomes negligible. And that explains why UTCI does not change for windspeeds ranging from 0 to 0.5m/s.

However, if UTCI is defined for a default metabolic rate corresponding to “walking”, then that might explain why UTCI is originally defined for windspeeds larger than 0.5m/s only. If we calculate UTCI for “walking” metabolic rate but remain able to compute UTCI for null windspeeds, then the results should perhaps be labeled are “invalid” or “questionable” (for 0-0.5m/s).

Another possibility would be to let the user define the metabolic activity or the walking speed. I remember seeing UTCI calculators that let the user define walking speed, age, gender, height and weight of the person - the metabolism is then calculated for these values…

@chris If that can’t be done, it would make sense to me that the default wind speed is around 0.75-1.5m/s. Also I think most users use UTCI in outdoor conditions, so maybe the 0.05m/s default is questionable.

Thanks again.

5

Hey @JulienFBC ,

The metabolic rate of a walking adult is something that is written into the definition of the UTCI metric and is hard-coded in the math that computes it. To prove it, here is the online documentation of the UTCI method used by Ladybug Tools and, if you click on the [source] button for the universal_thermal_climate_index method, you will see that UTCI is computed using a long ~200-term polynomial approximation, which you can think of as a “best-fit line” between a few thousand human energy balance simulations simulations that were run to create the UTCI metric. While there was clearly an input for metabolic rate and clothing in these initial human energy balance simulations used to make the UTCI metric, there is no way to change this metabolic rate when using this polynomial approximation method.

So, long story short is that, if some software is giving you a UTCI value that allows you to change the metabolic rate, then the metric they are giving you is not UTCI but some other “feels like” temperature. From your description, it sounds like you were probably using Physiological Equivalent Temperature (PET), which you can actually compute with Ladybug using the Thermal Comfort Indices component

Anyway, it seems like there’s agreement among us that we should change the default wind speed for UTCI to be something close to typical walking speed rather than the current assumption, which is more applicable for indoor conditions. I can also see that there’s some documentation that “UTCI isn’t supposed to be used at speeds below 0.5 m/s”, which confirms our findings here. So now the question is what exact value of wind speed we should use?

If I may put forward a proposal, according to the UTCI documentation, the metabolic rates assumed in UTCI correspond to a person walking at 4 km/h or 1.111 m/s. So this would be the air speed at the height of the human. However, UTCI needs to have the meteorological wind speed (at 10 meters above the person), which is always assumed to be 1.5 times the speed at the height of the person. So multiplying 1.111 m/s times 1.5 gives us 1.666 m/s. If we wanted to obey the UTCI model to the letter, we could use 1.666 m/s as the default. But I am inclined to round down to 1.5 m/s just to make the number easier to remember and because there’s a bunch of approximating happening in these assumptions anyway so I don’t think a difference of 0.16 m/s is that meaningful. It’s also the number that I see @josephyang originally put forward.

What do we all think?

6

Hi @chris,

thanks for your enlightening reply!

You are right : the software that allowed me to set met and clo used these values to compute PET.

Regarding your suggestion to set the default windspeed at 1.5m/s: I did not know the assumption that WS(10m)=WS(person height)*1.5 - but I guess it’s derived from a log wind profile for a specific roughness length? So yeah the 1.5m/s makes sense to me!

I’d just add this: it is unclear to me weather the “air speed caused by walking at 4 km/h” is taken as default in the UTCI calculation. In the UTCI documentation that you shared, it is stated that:

“the additional wind speed induced by the movement of the person requires the calculation of a relative wind velocity. We assume that the observed meteorological wind blows at 90° of the direction of a walking person.”

So here is the question : if meteorological wind speed is set at 0, does the UTCI calculation considers anyway a 4 km/h air speed resulting from the default walking activity?
How is the relative “wind speed” computed? Since the wind airflow and the walking air flow are considered perpendicular, it’s unclear if the two airflows are to be added to substracted or else…

Many questions! I’ll take a look at the UTCI docs but I guess you got my point.

Thanks again for your help and commitment

7

Hi @JulienFBC ,

Glad the response was helpful and I had wondered about the same ambiguity myself. That is, does the polynomial already account for the walking speed of the person in it’s conversion from meteorological wind speed to “relative wind speed” at person height? But seeing a number of sources that say UTCI should not be used at wind speeds below 0.5 m/s makes me think that we should be taking care of the walking speed in our input to the model. Moreover, it seems this 0.5 m/s minimum is a recommendation that comes from the authors of UTCI.

So I think we should probably be at least changing the default to 0.5 m/s if not 1.5 m/s. If you manage to do some more research and let me know what you recommend, I’ll implement that. Otherwise, I’ll just change it to 1.5 m/s.

And I can’t find the source for the 1.5 conversion rate between meteorological wind speed and the wind speed at the height of the person but, yes, it comes form a log or power law wind profile curve.

8

I’m not terribly familiar with UTCI but it does seem like there is quite a bit of nuance in this metric.

Perhaps 0.5m/s may be the lower limit of air velocity where everything else being equal, there is no discernible difference in the rate of convective heat transfer from human body (i.e. https://www.engineeringtoolbox.com/convective-heat-transfer-d_430.html).

Let me try downloading and looking through the UTCI dataset from CDS also (https://cds.climate.copernicus.eu/cdsapp#!/dataset/derived-utci-historical?tab=overview). Quite curious to see what they’ve calculated.

9

Hi @chris chris and @josephyang josephyang,
thank you for the provided elements!

I’m a colleague of JulienFBC and I’m also interested in Ladybug UTCI calculator.

After some research I’m here to share with you some findings, even if a check on the source code will probably solve all the calculation questions.

According to http://www.utci.org/isb/documents/develop_utci.pdf “wind should be measured at 10 m height, but we need the wind velocity at a height representative of the human being, which has been defined as 1.1 m. So measured or computed wind data must be adjusted to that height. The additional wind speed induced by the movement of the person requires the calculation of a “relative” wind velocity. We assume that the observed “meteorological” wind blows at 90 degrees of the direction of a walking person”.

So the input air velocity should be the vector addition between u = walking speed = 4 km/h = 1,111 m/s and v = wind velocity at 1.1 m .

It is possible to obtain wind velocity at 1.1 m starting from wind velocity at 10m, which according to UTCI scale varies from a minimum of 0,5 m/s to a maximum of 17 m/s.

The formula, derived from http://www.envi-met.info/doku.php?id=apps:biomet_utci, should be:

v = wind_1.1m = wind_10m x [(ln(1,1/z0)) / ln((10/z0))]

where:
0,5 m/s <= wind_10m <= 17 m/s
z0 = roughness length, which depends on terrain category and which could change depending on the country’s standards (default value is generally z0 = 1m for urban areas).

Since the vectors of wind speed at 1.1m and the walking speed are perpendicular, the relative velocity to be used should be the addition of the two vectors, following the Pythagoras theorem:

R = relative velocity = sqrt(u²+v²)

If this procedure is correct, it’s a quite complex one to obtain the input for the Ladybug UTCI calculation. It could be simpler to add two inputs (wind velocity at 10m, roughness length) and the relative velocity calculation process should be added directly in the source code of the Grasshopper component.

Let me know what you think about that and if you find other references about the UTCI calculation process.

Best regards,
Matteo

10

Thanks for the research, @MatteoFBC . I still don’t understand from all of this what your recommendation is for the default wind speed when no meteorological wind speed is input (presumably indicating low wind or wind-protected conditions). Is it 0.5 m/s because this is the lower limit recommended for UTCI?

11

Hi @chris ,

Exactly! In synthesis, I recommend two inputs for UTCI airspeed:

  • The metereological wind speed at 10m (default value = 0.5 m/s since it’s the lower limit, which will be also the best possible approximation for the calculation in the absence of wind);
  • The roughness length z0 (default value = 1m since it represents urban areas).
12

Thanks, @MatteoFBC and @JulienFBC .

I just pushed a few changes that should make 0.5 m/s the new default UTCI wind speed across all of Ladybug Tools.

This includes the Legacy Ladybug Plugin:


… the new Ladybug[+] plugin:

… and the ladybug-comfort Python package:

So the change of default wind speed is official now.

As for converting wind speeds between different wind profiles, there is already a component for that in Ladybug Legacy. All that you have to do is pass your meteorological wind speed through that component along with your terrain type (which determines the roughness length when using the log profile) before plugging the wind data into the UTCI component. Then, you’ll be accounting for whatever surrounding terrain you’re studying. We plan to add a similar component into Ladybug[+] and add the profile conversion into the ladybug-core Python package soon.

Thanks again, all!

2 Likes
13

Thank you very much @chris for the new update and for the information about the Wind Speed Calculator component! I’m looking forward to trying it!

1 Like
14

Hi @chris. That was indeed a good update to the component. Chris the description of the component when one hovers over wind speed still describes it as 0.05m/sec (legacy version). I think it should be changed too to 0.5m/sec else it might create some confusion for users.

1 Like
15

Good catch @Asisnath . The component input name was spelled wrong in the docstring. I just fixed it:

1 Like