Wind speeds

[Ralph1]

I have noticed an annoying problem with the version of CumulusMX(build3053) I am using. We have had a lot of wind of late and frequently I notice on the dashboard under wind speeds the gust high shows a much higher wind speed than the value which appears using #wlatest. For example the highest value of #wlatest yesterday at 14:23 was 41.2knots but the value which showed on the dashboard under gust high at the same time was 46.8knots which seems a very large difference. So I have to go in and modify the maximum reading in my system to correspond to the higher value shown. This happens very frequently and it is not a "one off" occurrence. Surely if the dashboard value shows a certain value, it should be visible
in the data stream?

Regards,
Ralph1

[sfws]

the highest value of #wlatest

Surely you should be looking at webtag <#wgust> as that is the web tag that reports the gust (irrespective of which build you might use).

[Ralph1]

Thanks for your reply.
According to Mark #wgust is something to do with Davis stations. He said that #wlatest should get the right answer. OK I'll change to #wgust and see if that solves it.

Ralph1

[Ralph1]

Thanks for your reply. I've tried substituting #wgust instead of #wlatest. Unfortunately #gust remains at the last high reading and so does not change. So the current wind speed as shown by #wgust remains constant at whatever was the last high gust. I don't know for how long this remains at the last gust level. So I have had to put back #wlatest as the current wind speed as it was before. But my problem remains; why does #wlatest apparently sometimes miss the #wgust figure? If it does, how can I get the #wlatest figure to momentarily set to #wgust, whenever it changes?

Ralph1

[water01]

I think the clue is in the Wiki.

<#wlatest> Current wind speed reading from console. Corresponds to 'latest' on the Cumulus main screen.

<#wgust> The highest wind reading in the last 10 minutes. Corresponds to 'gust' on the Cumulus main screen.

<#wspeed> The 10-minute average, if you have Cumulus set to calculate a 10-minute average. Otherwise, it's the latest 'wind' value from the console (i.e. the current speed as determined by the station). Corresponds to 'average' on the Cumulus main screen.

So the last two are 10 minute averages wheres #wlatest is the direct reading from the console.

The gust high for the day is held in

<#wgustTM> Today's maximum wind gust

So the highest gust is held there and #wlatest shows you the currect reading which unless the reading of the wind gust and wind latest occured at the same time they would always be different.

So I think you are imagining a problem that doesn't exist.

[Ralph1]

Thank you David for the information. I had not noticed (until this morning the parameter for maximum wind gust of the day #wgustTM). That could be a great help. However I output my wind speeds (and all other significant weather parameters) to a database and it is necessary that when a gust occurs it should appear in the data stream which is supposedly #wlatest. My complaint is that it doesn't and so my database which is fed directly with #wlatest misses out on maximum gusts occasionally, possibly more often than not. One solution would be to keep watching #wgust and when it goes higher than the previous value, (which would have to be stored) it momentarily replaces #wlatest with the #wgust value. But the logic to achieve that is a little complicated. Perhaps that is the way to go. Otherwise may database misses out on many of the maximum gust events which is very annoying as I have to go in and modify the values to correspond.

John

[water01]

Could this have to do with your data logger value as outlined in these two parts of the FAQ in the Wiki. https://cumuluswiki.org/a/FAQ#The_.E2.8 ... m_my_graph

[mcrossley]

Cumulus attempts to read the data from the Instromet every two seconds or better - but that does not guarantee it will read every value, occasionally it misses readings as it has to do other things than read the data stream - like update the loggers pointers. Also are you really logging the <#wlatest> wind speed every 1.5 seconds to a database, the realtime interval can only be whole seconds, so you'd have to be logging it every second!

[Ralph1]

Hello Mark,
Yes, I do believe that my system is logging the wind data once per second. That is what surprises me - I would have thought that would guarantee that gusts were not missed, since as you say according to the WMO definition it's the maximum average speed over any 3 seconds.
If I could work out the logic to do it, I would do something like the following:
first of all create a value #wgustold = #wgust. Then if #wgust > #wlatest and #wgust > #wgustold then set #wlatest = #wgust and set #wgustold=#wgust
So long as #wlatest < #wgust then #wlatest would pass through as normal. The high value of #wgust only lasts 10 mins so after that I gather it resets.
I can see that there is more to do before such a system will work correctly. It needs a proper logic diagram to find an optimum solution.
I will keep thinking about it.

I know the daily maximum daily gust is available from the system(which I didn't know before).

John

[mcrossley]

The Instromet documentation is shall we say a bit light. But the live value it supplies is just a single wind speed - do you know if that is a 3 second average, or an instantaneous reading?

But anyway, if Cumulus is showing the correct gust value then it must be reading it from the live stream - it has no other source. So your problem looks like the live values you are logging do not include all possible values, I think the realtime update of <#wlatest> probably cannot keep up with rate of data input - which I've looked up the code and has a 0.5 second delay in the loop, how that translates to a real world data read rate I don't know. Which also implies that the value Instromet supply is a instantaneous reading not a 3 second average.

[Ralph1]

That's a good point, Mark. I'll get onto Instromet and see how often the samples are generated. The only other evidence I have is the analogue read-out from the wind on the dial - this reads
the data directly and it is that which makes me doubt that we are seeing the correct values. (It often tends to go higher than the guat readings). It moves fairly instantaneously and although it's an analogue display, it used to agree fairly closely with the digital figure I used to have before I used the Instromet Metplus box.

Regards,
John

[sfws]

The high value of #wgust only lasts 10 mins so after that I gather it resets.
...
I would do something like the following:
first of all create a value #wgustold = #wgust. Then if #wgust > #wlatest and #wgust > #wgustold then set #wlatest = #wgust and set #wgustold=#wgust
So long as #wlatest < #wgust then #wlatest would pass through as normal.

I know nothing about Instromet, and I'm not yet using Cumulus MX, but reading what you say in the first part of that quote, it occurs to me that you might want to make an adjustment using the Cumulus configuration file.

I don't know if this would help you, but why not try setting the period between changes to the gust value to either 1 minute so it becomes the most frequently allowed by Cumulus or to (24 x 60) minutes if you want one reset per day?

Quoting https://cumuluswiki.org/a/Cumulus.ini#Section:_Station

Code: Select all

PeakGustMinutes=N

This parameter (attribute=value) is added directly to the "Cumulus.ini" file, it does not appear on a configuration screen. This line can be added in the [Station] section of the Cumulus.ini.

With this parameter, N is the number of minutes over which Cumulus calculates its highest wind gust.
Without this parameter, the default of 10 minutes, the international standard duration, is used.
If you select 60, then Cumulus would calculate rolling 60-minute highest wind gust speeds.

I realise this does not solve the second part of what you said in that quote, but it might affect how often you would need to do it?
Hopefully you will solve all your issues in consultation with Instromet?

[Ralph1]

Thank you for that additional information. Not sure that I want to change the period between peak gust measurements, all I really want to do is to get the peak gust to appear in #wlatest. Weather Display does this quite nicely if it is being used to decode the signals from scratch but CumulusMX seems to be aimed at applications where the peak gusts are required to be extracted from the data stream.

I will continue in my efforts to combine #wlatest and #wgust but it is not as easy as it sounds - it would have been easier if CumulusMX had kept #wgust in the data stream of #wlatest.

John

[sfws]

However I output my wind speeds (and all other significant weather parameters) to a database
...
I will continue in my efforts to combine #wlatest and #wgust but it is not as easy as it sounds

I still do not grasp what you are trying to do, but prsumably you have written a script to get the stream of statistics into the database?

If all you want is somehow to get the largest of two numbers output within that script, then actually that is simple.

If your script language is PHP, then get Cumulus to process this snippet of code and use the output variable for your display (the realtime interval is the fastest frequency that Cumulus can do this):

Code: Select all

<?php
$latest	= max( <#wlatest>,<#wgust>);

If your choice of script language is JavaScript, it can do calculation with slightly different syntax:

Code: Select all

latest = Math.max( <#wlatest>,<#wgust>);

If you want more help, I suggest you provide everybody with a copy of exactly what you are using to update the database, is it the database loading facility in Cumulus MX?

[mcrossley]

I hope that you are just creating a data file locally for Cumulus to process, then accessing it from the local file system? FTP is slow.

Another thought. Don't use realtime processing at all! Just subscribe to the WebSockets data feed Cumulus provides - ws://<ip_address>:8998/ws - that pushes out a JSON data block that includes the latest windspeed, and it is updated every time Cumulus gets a data block from your Instromet.