Air density estimates

Post Reply
BillOsler
Posts: 116
Joined: Mon Apr 04, 2016 10:10 am

Air density estimates

Post by BillOsler »

I may have this wrong ... my fluid dynamics training was a long time ago and mostly revolved around liquids, not gases, but IIRC the resistance from air depends on air density. Air density depends on a combination of barometric pressure and the mix of gases. Since the oxygen, carbon dioxide and nitrogen components have fairly stable ratios the main source of change in the gas mixture is the water vapor content, ie: humidity.
Aeropod measures air pressure directly and presumably uses that data to calculate altitude change and wind speed. I assume that the details are proprietary.
When Isaac displays information in the program it reports air pressure but not air density unless I have missed that data somewhere. When Isaac exports CSV files it does report air density.
That raises a question: Does the Aeropod also measure humidity? Or does Isaac impute a typical value for purposes of the air density calculation? It seems like using an imputed value would be a potential source of problems comparing ride data from one day with ride data from another day but measuring humidity would present some technical challenges.
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Air density estimates

Post by Velocomp »

We don't have a humidity sensor.

I don't recall anywhere in our calculations where we make an assumption about humidity. Remember that, because we measure air pressure, not wind speed, we don't have to know the air density for power/CdA calculations. Wind speed is derived from air pressure, so there is an assumption there about density. But since we don't use wind speed for power calculations, we don't worry about potential errors in wind speed.
John Hamann
BillOsler
Posts: 116
Joined: Mon Apr 04, 2016 10:10 am

Re: Air density estimates

Post by BillOsler »

OK. I'd have to work through the equations but intuitively at least the resistance from air would be related to the pressure differential * CdA and that might get around needing humidity for the CdA calculation ... but the Connect IQ app does report wind speed and the CSV files exported by Isaac report air density as well as headwind relative to the bike. Don't those calculations require making an assumption about humidity?
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Air density estimates

Post by Velocomp »

The equation we use to convert air pressure to wind speed does not make an assumption about humidity.
John Hamann
ted.kelly
Posts: 27
Joined: Fri Apr 19, 2019 12:40 pm

Re: Air density estimates

Post by ted.kelly »

Does the equation AeroPod uses to covert air pressure to wind speed have air density in it?
If it does, how is the air density value determined?
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Air density estimates

Post by Velocomp »

ted.kelly wrote: Tue Apr 23, 2019 10:13 am Does the equation AeroPod uses to covert air pressure to wind speed have air density in it?
If it does, how is the air density value determined?
Yes. We get air density from barometric pressure sensor measurements.
John Hamann
BillOsler
Posts: 116
Joined: Mon Apr 04, 2016 10:10 am

Re: Air density estimates

Post by BillOsler »

Velocomp wrote: Tue Apr 23, 2019 12:46 pm
ted.kelly wrote: Tue Apr 23, 2019 10:13 am Does the equation AeroPod uses to covert air pressure to wind speed have air density in it?
If it does, how is the air density value determined?
Yes. We get air density from barometric pressure sensor measurements.
That's the problem. In order to get air density from pressure you have to know or impute a water vapor pressure or dew point or some equivalent information. Without the water vapor pressure you could calculate the standard pressure of dry air ... but the cyclist isn't moving through dry air.
In the process of trying to make sense of the numbers i encountered today (temp reported as 83 F, pressure 1012 mbar from Isaac and air density (rho) 1.1451 kg/m^3 in the CSV file) I ran some comparisons using calculators at:
https://www.omnicalculator.com/physics/ ... ir-density
and https://www.gribble.org/cycling/air_density.html (computational cyclist)
and https://barani.biz/apps/air-density/
If I use the numbers as stated the humidity necessary to make the numbers line up is about 120%.
If I assume that the reported values are rounded to the nearest digit then using a temperature of 83.49 F and a pressure of 1011.5 mbar the corresponding humidity to obtain a density of 1.1451 kg/m^3 is still about 115% according to the calculators.
I'm having a hard time making sense out of the various estimates.
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Air density estimates

Post by Velocomp »

BillOsler wrote: Mon Apr 22, 2019 10:46 pm I may have this wrong ... my fluid dynamics training was a long time ago and mostly revolved around liquids, not gases, but IIRC the resistance from air depends on air density. Air density depends on a combination of barometric pressure and the mix of gases. Since the oxygen, carbon dioxide and nitrogen components have fairly stable ratios the main source of change in the gas mixture is the water vapor content, ie: humidity.
Aeropod measures air pressure directly and presumably uses that data to calculate altitude change and wind speed. I assume that the details are proprietary.
When Isaac displays information in the program it reports air pressure but not air density unless I have missed that data somewhere. When Isaac exports CSV files it does report air density.
That raises a question: Does the Aeropod also measure humidity? Or does Isaac impute a typical value for purposes of the air density calculation? It seems like using an imputed value would be a potential source of problems comparing ride data from one day with ride data from another day but measuring humidity would present some technical challenges.
I am returning to the original posting, because I want to make sure I provide the correct information.

We measure air pressure, then use the usual conversion formula to convert the air pressure readings into wind speed.

One of the important factors used in the conversion formula is "rho". The value rho varies with temperature, elevation (barometric pressure), and humidity.

In our devices we adjust the value of rho in accordance with changes in barometric pressure and temperature. We don't adjust rho for relative humidity (RH).

In our view, humidity correction is minor. The graph below, from here https://www.gribble.org/cycling/air_density.html illustrates this.

The maximum variation, due to RH, at sea level (where the RH effect is greatest), going from 0% RH to 100%RH is 0.8%.

For typical day to day rides here in Florida, the RH variation might be from 50% to 80% (at 22C, roughly dew point of 7.5C to 10C). The effect on rho of this change in RH is 0.07%.

The effect of RH on rho is less at higher altitudes.
Attachments
Screen Shot 2019-04-27 at 10.13.49 AM.png
Screen Shot 2019-04-27 at 10.13.49 AM.png (177.06 KiB) Viewed 4255 times
John Hamann
BillOsler
Posts: 116
Joined: Mon Apr 04, 2016 10:10 am

Re: Air density estimates

Post by BillOsler »

Thank you for the detailed response. That makes sense.
I'm not sure that the air density (rho) reported by Isaac in the exported CSV files matches well with the temperature and pressure values reported for the ride file but the differences are small.
Post Reply