Page 1 of 1

Questions about wind readings

Posted: Thu 09 Nov 2017 5:50 am
by lloydh
I have my Davis 6410 temporarily installed at the front of the house (not a good position) just while I am testing and I think I am very close, it will soon be installed on the peak of a gable on the roof which should be in reasonably clear air.

On the Arduino I am reading the wind speed and wind direction from the Davis every 10 seconds. I save the readings and keep the last 5 readings, I then take the average of the 5 readings and that goes to the station as wlatest and bearing, I also send the maximum speed reading and its bearing of the 5 readings as windgust and windgustbearing. For the average I use the formula from http://www.webmet.com/met_monitoring/622.html with a lot of help from http://www.scadacore.com/2014/12/19/ave ... ind-speed/. If there is no wind for the 5 readings I set both wlatest and bearing to zero.

On the station I do the 10 minute averages for wind and wind gust using the same formula as above and also save the maximum wind gust.

For the dominant wind direction I started by saving avgbearing (and wspeed) every 10 minutes but only if there was wind. I then did a circular average on the number of saved values only and then convert it to a cardinal reading.

The wind rose, I guestimated that the 16 values were 16 cardinal points going clockwise from N so every 10 minutes I cleared the wind rose data and then added the saved wspeed values to their correct cardinal point and this seems to work.

For wind run I add all the saved wspeed readings and divide by the number of 10 minute periods since midnight to give me the true average km/h and then I divide by the number of hours since midnight (to 2 decimal places) to give me the wind run in klm.

Since finishing that I have changed the saving of avgbearing and wspeed from every 10 minutes to every 5 minutes giving me a maximum of 288 readings each day but I didn't notice any real change of what I saw on the gauges after I did this.

All I am asking is am I going in the right direction or have I made mistakes, the attached graphic is what the gauges look like. I've done all of this on a second Raspberry and I will transfer it to the main station when I am happy with it.

Re: Questions about wind readings

Posted: Thu 09 Nov 2017 12:11 pm
by mcrossley
It sounds reasonable to me. The only significant differences to how I think Cumulus handles things are:

Cumulus just passes the last reading for 'wlatest' and 'bearing' rather than a rolling 50 second average.

Cumulus uses the 10 minute high gust value for 'wgust'.

Does 'windgustbearing' exist? I don't think the gauges use it, though it would be a nice addition to CMX. The problem is that it may not be available if it is read from either the logger or Davis records values (ignoring other station types).

Re: Questions about wind readings

Posted: Sat 11 Nov 2017 7:27 am
by lloydh
Mark, thanks for your comments and it looks like I am there.

I have a lot of extra variables and lists in the background for calculations and sometimes for graphs. I try to keep to the names on your Wiki but there are probably some that are not being used anymore and I have been doing some housekeeping in the past few days to try and tidy things up.

I made a decision in the early days to do a small amount of averaging on the sensors because I am only taking readings every minute.

I may look at things like BearingRangeFrom etc at a later date maybe but I have already had one comment from a friend saying it looks too complicated already.

Re: Questions about wind readings

Posted: Sat 11 Nov 2017 11:07 am
by mcrossley
lloydh wrote:I may look at things like BearingRangeFrom etc at a later date maybe but I have already had one comment from a friend saying it looks too complicated already.
I always take the view that my web page is mainly for me - given the numbers of visitors it isn't a competitor for BBC/weather! - so I put up what I want to see and if others find it useful or conversely over technical so be it :lol: