Is there a recognised way to calculate current and peak wind speed?
I am taking pulses from an anemometer, storing them every second then using the last 5 seconds of stored values to calculate the wind speed, so averaged over 5 seconds, updated every second.
Then I take the current speed and if it is higher than the last stored peak wind speed update the peak wind speed with the current wind speed. I hold the peak speed for 15 seconds and if it has not been updated in those 15 seconds I reset it to the current speed and start over with the new lower peak speed.
This was based on what seemed right to me, but I have nothing much to go on. Is averaging over 5 seconds to get the current speed reasonable? What is standard practice? What do professional meteorologists do?
I want to end up with a graph of wind speed, but this will have 1 measurement per 15 minutes, which is even more problematic as I will be averaging over 15 minutes. I guess peak isn't such an issue, it is just the highest speed measured over the 15 minutes.
PerryBebbington:
Is there a recognised way to calculate current and peak wind speed?
A little bit of googling says no - it seems to be arbitrary. Peak wind speed measurements apparently tend to be for a 24 hour period and count if they were sustained for a few seconds - three was one example I saw.
Current wind speed is usually averaged over a number of minutes. I saw 1, 2, 3, 10 mentioned.
But it looks like it varies enough that you may as well adopt the Bebbington standard with what you already coded.
I've only recently ordered the wind speed/direction/rain fall kit to add to my weather station.
@PerryBebbington, your postings about questions relating to weather measurements have been very helpful.
Define wind peak. Is it the fastest wind measured for the day, hour, minute?
I, recently made a IR based heart beat detector. The trick to getting a steady heartbeat is steady pressure on the IR by the finger. Before I put the LED to a clamp, I tried various ways of damping the effects of physical finger pressure on the IR.
I tried different methods and found that this PeakDetection Library worked good enough to not require a clamping device on the finger. It took me quite a number of hours to get how to set the 3 parameters but well worth the effort.
I'm with @wildbill, use the PerryBebbington method of peak wind detection.
I have been applying LinearRegression (LR) to my Air Pressure readings. While seeing predictive future air pressures to within .92 to .98. I have been surprised as to how the LR calculations show the storm 'intensity'.
PerryBebbington:
Is there a recognised way to calculate current and peak wind speed?
I am taking pulses from an anemometer, storing them every second then using the last 5 seconds of stored values to calculate the wind speed, so averaged over 5 seconds, updated every second.
Then I take the current speed and if it is higher than the last stored peak wind speed update the peak wind speed with the current wind speed. I hold the peak speed for 15 seconds and if it has not been updated in those 15 seconds I reset it to the current speed and start over with the new lower peak speed.
This was based on what seemed right to me, but I have nothing much to go on. Is averaging over 5 seconds to get the current speed reasonable? What is standard practice? What do professional meteorologists do?
I want to end up with a graph of wind speed, but this will have 1 measurement per 15 minutes, which is even more problematic as I will be averaging over 15 minutes. I guess peak isn't such an issue, it is just the highest speed measured over the 15 minutes.
Thoughts and comments please.
I thank you
My Arduino anemometer computes the wind speed once a second(Arduino second) and stores it in the next available entry in a 15 entry array. Then the array is used to compute the average for those 15 seconds.
If the 1-second speed is greater than the last maximum, it replaces the old maximum.
My wife resets the Arduino every morning after recording the maximum.
The anemometer head is about 4 ft. above the patio roof. So is subject to the wind speeding to move around the house. It also is sensitive to the wind turbulence from the poll barn structure.
To be correct for wind speed, the anemometer needs to be 10 meters high and away from any turbulence from buildings or any other obstruction.
I have several whirley gigs set up about 6 ft.above the ground. The turbulence even away from a building is tremendous. All the distorts the TRUE wind speed.
Paul
Paul,
You are doing similar to me then, but over 15 seconds instead of 5. Is your 15 as arbitrary as my 5?
At best my anemometer is 1m above my roof line, just looking at it it is obvious that the wind going up the roof will affect it. I'm not sure I want a pole 10m high and I don't have a big, open space clear of trees and buildings! I like the view of trees outside my window!
I could probably get it 3 to 4m above the roof line without too much difficulty, probably attract funny looks from the neighbours
I wonder what the objective of the measurement is?
If you are studying (for example) the forces on the gear box of an windmill then fast variations are relevant, in other applications they are not. What is the objective of your experiments? I think the targeted use of the measurements are key to know how to filter best what you measure.
Best Regards,
Johi.
What is standard practice? What do professional meteorologists do?
I once given a guided tour the meteorological department at Reading University when I was down there giving an other group a talk.
To my surprise they were more interested in the wind vector, that is both speed and direction. To do this they used an ultrasonic device to measure the wind speed in three dimensions.
It looked like three fingers placed orthogonally, with a matching set a very short distance away. From the size of it the frequency was more in the MHz range than the standers 40KHz.
From reading the posted info, I can see that a burst is defined as a sustained wind speed of .2 seconds. My thinking is that to get a good sample of a .2 second burst wind speed should be sampled at 2X the .2 seconds in the least.
I see the definition of highest wind speed is the highest wind speed measured in a 24 hour period.
I took from the reading that a rolling average of 3 seconds of wind speed is the minimum data set and that the sample of 3 seconds can be increased.
It would seem that a 3 seconds rolling average of sample time of .1 second would equal the minimum number of data points needed to get wind speed data set and to be able to measure gusts.
I'd go with 5 seconds rolling average with .025 second sample time as the least number of data points for wind measurement.
PerryBebbington:
Paul,
You are doing similar to me then, but over 15 seconds instead of 5. Is your 15 as arbitrary as my 5?
At best my anemometer is 1m above my roof line, just looking at it it is obvious that the wind going up the roof will affect it. I'm not sure I want a pole 10m high and I don't have a big, open space clear of trees and buildings! I like the view of trees outside my window!
I could probably get it 3 to 4m above the roof line without too much difficulty, probably attract funny looks from the neighbours
Perry
The bottom line is" wind speed ALWAYS measures turbulence. That is what causes wind in the first place. I guess the point is to minimize the amount of turbulence.
My 3-cup anemometer head is commercially made, but has plastic cups rather than the aluminum cups of the real model. Must have been a prototype as the the real one is about $200 and I got mine for $10.
I used the speed calculation formula for the real one.
The highest reading I got was 270MPH when a down burst with a huge load of hail hit the cups and broke one. Of course the lowest reading is zero.
I also had to replace the internal reed switch in the head. It had a SINGLE blade and apparently it weakened or work-hardened after years of use. I had a normal two-blade switch of the same size and that has worked properly ever since.
Good luck!
Paul
