Wind speed measurement.

[Jenki]

I'm posting as a non Boffin, in search of ideas.

I want to set up a weather station so in the future I can use this for weather compensation, but also want this information to have around the house. 

Additionally I have a project in mind that having wind speeds available, and being able to use windspeed as a signal may help.

 

So I suppose the question is can I hack an of the shelf system and interrogate the data from say an RPI? Or should I go down the route of building one based in an RPI?

 

At some point I will want to use an Rpi to monitor the house, and create a PV divert, but this will stretch my abilities as it will require some remote / wireless monitoring as my Electricity meter will be 50M from the house.

But all this said I don't want to be tinkering with it all the time and it needs to work without me here.

[SteamyTea]

A Davis Vantage Pro2 is what you need, can interface directly with an RPi.

 

Have you looked at WeatherUnderground to see what weather stations are near you, and what sort they are.

You can sometimes get the info on the hardware they are using.

 

[RobLe]

I made a crude windspeed monitor, using a couple of small temperature sensors and a resistor.  The resistor is glued to one of the temp sensors, and left on, I think consuming 100mW continuously so that it warms up that temperature sensor a bit.  How much it warms it up by, depends on the windspeed.  I calibrated the response into m/s by comparing it to a nearby weatherstation.  I wanted it so I could try and correlate home heating versus windspeed.... I haven't found a correlation at all.  I did find one versus light level from a solar panel 🙂  Nb: we have an air tight house with MVHR, likely why windspeed didn't seem to matter.  Maybe if I had infinite time I would plot these things out. 

 

 

[SteamyTea]

1 hour ago, RobLe said:

I made a crude windspeed monitor, using a couple of small temperature sensors and a resistor

I have often thought something along those lines should be easy to do. Cars use a similar idea in the Mass Airflow Meter.

I tried to make an ultrasound one. Never got it to work.

https://hackaday.com/2013/08/21/ultrasonic-anemometer-for-an-absurdly-accurate-weather-station/

May have to revisit it sometime.

 

[Jenki]

3 hours ago, SteamyTea said:

A Davis Vantage Pro2 is what you need, can interface directly with an RPi.

way too expensive for me..

[SteamyTea]

1 hour ago, Jenki said:

way too expensive for me..

You will end up getting one in the end.

I bought a cheap one, soon packed up.

[Radian]

On 07/11/2022 at 15:08, RobLe said:

I made a crude windspeed monitor, using a couple of small temperature sensors and a resistor.  The resistor is glued to one of the temp sensors, and left on, I think consuming 100mW continuously so that it warms up that temperature sensor a bit.  How much it warms it up by, depends on the windspeed.  I calibrated the response into m/s by comparing it to a nearby weatherstation.  I wanted it so I could try and correlate home heating versus windspeed.... I haven't found a correlation at all.  I did find one versus light level from a solar panel 🙂  Nb: we have an air tight house with MVHR, likely why windspeed didn't seem to matter.  Maybe if I had infinite time I would plot these things out. 

 

 

You can eliminate the temperature sensor if you power the resistor with a constant current source. The voltage measured across the resistor is then proportional to the cooling effect of the airflow. You need to have quite a high resistor temperature in order to minimise the effects of ambient temperature (which of course is another parameter you'll be measuring anyway so can be compensated for). But for this reason I've smashed 6V bulbs to expose their filaments which can be held just on the verge of incandescence to maximise the accuracy.

 

I did once try getting wind direction by having three such sensors in a triangle and attempting to correlate changes in velocity between them but I did it so long ago that I was using a BBC Micro to do the processing and never got sensible results. Quite fancy revisiting that idea with some ARM power.