smoothing of apparent wind data

Hi everybody, is there an easy way or plugin available to smooth apparent wind data? Actually it’s mostly jumping around as the wind vane raw data. Thanks !
23 Replies
gregy
gregy4mo ago
there is a plugin, however i have had difficulty with it working reliably to produce averages etc. i have seen others make similar comments.. but you could try it.
PaddyB
PaddyB4mo ago
I much prefer looking at influx data in grafana to see trends for most data, max 5 minutes to see the gusts. Much more useful imho
Teppo Kurki
Teppo Kurki4mo ago
but if you only want / can display a number smoothing / averaging can be very useful
Hannes
HannesOP4mo ago
I saw that already. I’m not yet using node red on my setup, so i would prefer a way without nodered.
Teppo Kurki
Teppo Kurki4mo ago
I am not aware of anything else with reported success
Charlie
Charlie4mo ago
I have had good results with using node-red flows to average apparent wind speed and direction.
David Godin
David Godin4mo ago
If you use KIP you can change the sample path rate but it’s not an average. I could add this.
Kalymnos
Kalymnos4mo ago
I use Dejitter plugin, working good for me
Hannes
HannesOP4mo ago
I will try Yes, using KIP. Tried it today morning, but dejitter works only with the first data. Second and more data are not generated.
Phil
Phil4mo ago
Not sure if this will work for you, but if you can store the last displayed value, you can do a simple weighted average with the new value. I've used this in my instrument code: https://github.com/philseeley/boatinstrument/blob/964c128049901ecabe7d1079f2bd8789fddacaf2/lib/data.dart#L18 Also if you're dealing with wind angles, I do a similar thing but using vectors: https://github.com/philseeley/boatinstrument/blob/964c128049901ecabe7d1079f2bd8789fddacaf2/lib/data.dart#L7
Hannes
HannesOP4mo ago
I tested today dejitter during sailing. After a few reboots it is running now properly for True Wind Angle, Speed through Water and Windspeed. I tried several data frames. I decided „7“ is a good value for my sensors (Raymarine Wind & DST810 Triducer)
David Godin
David Godin4mo ago
What’s the use case? Have you tried out Data Chart widget in combination with Datasets? This will give you a visual graph of Actual and Simple Moving Average Sample Rate should be enough over 1 sec or less, no?
Hannes
HannesOP4mo ago
Use case is still the data displayed on the Mastdisplays. Actually especially TWA/AWA is jumping around like the raw data. With dejitter it’s much better. B&G/Raymarine & Co. displays have individual smoothing/averages implemented.
David Godin
David Godin4mo ago
Over what elapse?
ray_sabado
ray_sabado4mo ago
The ability to dampen data would be very helpful for improving the readability. I believe B&G do up to 10 seconds in 1 second intervals. This uses a moving average to display the data. Helps a lot in a pitching sea state.
David Godin
David Godin4mo ago
You can do this with the Data chart widget. But it's a chart view. Have you tried to increase the path's Sample rate to 1000 ms?
Teppo Kurki
Teppo Kurki4mo ago
checked out dejitter plugin's code, seems like it is not suitable for smoothing wind data as radialAverage seems to be buggy
GitHub
Angles are radians in Signal K · Issue #1 · amirlanesman/signalk-da...
Angles are radians in Signal K so the conversions from degrees and back to degrees should not be there.
Hannes
HannesOP4mo ago
Yes i did, but i did not recognize an effect at all.
David Godin
David Godin4mo ago
500ms more is hard to see. Try 2000ms
Pantera
Pantera2mo ago
I have been using avg-paths pluggin and seems to work fine. And im testing another pluggin for eliminating outliers. My instruments keep from time to time sending data corrupt the display and seems very usefull
petsi
petsi4w ago
some time ago.. I was trying the effect and somebody giveme this node red and it works fine
petsi
petsi4w ago
the problem is the apparent wind angle Teppo Kurki help's me a lot but here in discord somebody give's this node red flow and it works fine
Want results from more Discord servers?
Add your server