Testing the BASTARD

[Cory]

During Hurricane Karl last year, Josh noticed that the wind speeds that were reported by the Kestrel instrument seemed to be too low. My hypothesis was that heavy rain was being driven into the Kestrel's wind impeller, and also cluttering up the airstream and slowing down the rotation, which affected the wind speed. To test that idea, Josh came to Texas and we came up with two different tests.

For the first experiment, we aimed a high pressure air nozzle at the instrument and took measurements of the wind speed, then poured water into the air stream to see if there was an affect (by the way, we put the nozzle too close to the instrument and measured a 133 KT wind speed before the impeller disintegrated ). There was an effect, but after further consideration we questioned whether the sudden presence of the water in the air stream was causing the slowdown, and not the water hitting the impeller.

So, for a second experiment, we rigged a 2-liter soda bottle with a nozzle through its cap, filled it with water, and attached 12 feet of flexible plastic tubing to the nozzle. We attached a wooden support to the BASTARD pole and positioned the end of the tube just below and in front of the impeller, then attached the whole thing to the roof of the car. We drove at 60 mph to get a steady wind reading, then began shooting the water into the wind in front of the Kestrel. The Kestrel recorded data at 2-second intervals, and we later uploaded it to the computer. During analysis, it was clear that the water (and probably the heavy rain) does have an impact on the wind speeds.

Readings:

Test 1: 10 KT decrease from 55 KT baseline = 18% decrease.

Test 2: 20 KT down from 63 KT baseline = 29% decrease.

Test 3: 7 KT down from 60 KT = 12% decrease.

Test 4: 12 KT down from 65 KT = 18% decrease.

Average = 19% decrease in wind speed when water droplets are ingested into the impeller. Josh will use this correction during future chases to see if it's realistic.

The graphs from the four tests, showing the decreases:

#1:

#2:

#3

#4

Video of air nozzle and driving tests:

A few pics:

[phil882]

Nice work guys! Do you think you have any explanations for this rain induced wind reduction? Perhaps the closed nature of the Kestrel allows it to be more greatly affected by outside effects like rainfall in comparison to a three cup anemometer or open propeller anemometer.

[wxmeddler]

That's awesome data guys! Have you thought about sending the results to Kestrel?

Also, I didn't get the the chance to ask Josh at the the conference but I noticed that the temperature quickly drops when interacting with rain. Quicker than air can actually cool or how it feels. Have you guys experienced something of the same thing?

[Cory]

Nice work guys! Do you think you have any explanations for this rain induced wind reduction? Perhaps the closed nature of the Kestrel allows it to be more greatly affected by outside effects like rainfall in comparison to a three cup anemometer or open propeller anemometer.

My guess is that the plastic impeller is so small and sensitive that the rain is hitting the blades and flinging around inside the housing, which causes it to slow down. The Kestrels were designed to be very sensitive for testing low wind speeds for shooting ranges, wildfires, air conditioning, etc., and not for being blasted by a hurricane with heavy rain. I had an email exchange with one of the engineers at Kestrel, and he thinks that's what's happening as well. Hopefully the correction factor obtained from these tests will help with more realistic wind speeds.

We looked at other portable anemometers, but there's just nothing on the market that can compare to the ease of use, portability, and Bluetooth datalogging capabilities of the Kestrel. Hopefully we'll soon be testing it again in a real hurricane.

[phil882]

My guess is that the plastic impeller is so small and sensitive that the rain is hitting the blades and flinging around inside the housing, which causes it to slow down. The Kestrels were designed to be very sensitive for testing low wind speeds for shooting ranges, wildfires, air conditioning, etc., and not for being blasted by a hurricane with heavy rain. I had an email exchange with one of the engineers at Kestrel, and he thinks that's what's happening as well. Hopefully the correction factor obtained from these tests will help with more realistic wind speeds.

We looked at other portable anemometers, but there's just nothing on the market that can compare to the ease of use, portability, and Bluetooth datalogging capabilities of the Kestrel. Hopefully we'll soon be testing it again in a real hurricane.

Thanks for the reply, as that was my suspicion as well. Well I sure hope to be seeing the Kestrel in action very soon (Re: This week!)

[Isopycnic]

Josh, do you own any shirts with sleeves?

back on topic.

How does the kestral compare to other portable gauges with the same conditions?

[k***]

Josh, do oyu own any shirts with sleeves?

lol

[turtlehurricane]

I'm not sure if your second experiment actually takes out the effect of precipitation loading on the wind speed. The lower momentum of the water droplets (since they lose momentum quickly in the face of air resistance) should slow down the air stream. However, in a hurricane the water droplets are already in equilibrium with the air flow (well relative to the car experiment at least), so this effect isn't present. I haven't heard of anemometers accounting for precip like this so it might not be good to correct it, though I may just be ignorant on the subject.

[OKpowdah]

I'm not sure if your second experiment actually takes out the effect of precipitation loading on the wind speed. The lower momentum of the water droplets (since they lose momentum quickly in the face of air resistance) should slow down the air stream. However, in a hurricane the water droplets are already in equilibrium with the air flow (well relative to the car experiment at least), so this effect isn't present. I haven't heard of anemometers accounting for precip like this so it might not be good to correct it, though I may just be ignorant on the subject.

Agreed. I don't think the water droplets in the air affect the reading, but the anemometer impeller getting wet could make a difference since then the added mass on the impeller would cause it to slow.

[Cory]

I'm not sure if your second experiment actually takes out the effect of precipitation loading on the wind speed. The lower momentum of the water droplets (since they lose momentum quickly in the face of air resistance) should slow down the air stream. However, in a hurricane the water droplets are already in equilibrium with the air flow (well relative to the car experiment at least), so this effect isn't present. I haven't heard of anemometers accounting for precip like this so it might not be good to correct it, though I may just be ignorant on the subject.

That was considered and it's why we did two different types of tests. We really have no way to perfectly simulate rain in a hurricane, so this is the best we could do, although the main point was that water being present on the impeller and in the airstream flowing through the housing is messing up the free spin of the impeller and slowing the measured speeds--not that hurricane-force rain drops are causing the problem, although it may. Josh definitely noticed the Kestrel reporting wind speeds too low in Karl, the Kestrel engineer agreed that the rain could be causing the problem, and the tests at least showed that water getting into it does affect it, although some of it is likely due to interference in the airstream due to the water loading (which we can't avoid).

There's also a possibility that Josh's original Kestrel was defective. He had it replaced after Karl but it has never been tested against reality. If the readings seem realistic during the next hurricane, it'll help narrow down the problem to a bad instrument and would show flaws in our experiments.

[Ed Lizard]

Have you guys ever considered pitot tube wind measurement? As far as I know, rain does not affect aircraft speed measurements.

[Uncle Bobby]

Guys,

The water is causing the air speed to drop.

In the first test, you mixed air moving at one speed with water moving at zero. With the water having much greater density than the air, this effect is obviously quite significant.

Then, did you think that getting the water moving at 60mph (above the car) would compensate? No. Not at all, because the water's speed relative to the Kestrel is still zero. Sure, they are both moving at 60mph, but, again, the relative speed is zero, and the water was being released so close to the unit that it had very little opportunity to accelerate to a relative speed of 60.

Do it again, but this time position the Kestrel as far rearward on the vehicle as possible, and position the sprayer at the front... even ahead of the front bumper if possible! You may want to increase the amount of water being released somewhat.

[Cory]

Guys,

The water is causing the air speed to drop.

In the first test, you mixed air moving at one speed with water moving at zero. With the water having much greater density than the air, this effect is obviously quite significant.

Then, did you think that getting the water moving at 60mph (above the car) would compensate? No. Not at all, because the water's speed relative to the Kestrel is still zero. Sure, they are both moving at 60mph, but, again, the relative speed is zero, and the water was being released so close to the unit that it had very little opportunity to accelerate to a relative speed of 60.

Do it again, but this time position the Kestrel as far rearward on the vehicle as possible, and position the sprayer at the front... even ahead of the front bumper if possible! You may want to increase the amount of water being released somewhat.

The main idea wasn't necessarily that the rain drops themselves are hitting the impeller at hurricane speed and slowing it down--it's that the water is spending time being flung and splashed around against and inside the impeller blades and housing and causing problems (essentially adding "clutter" to the total airstream flowing through the impeller housing). All we really needed to do was get some water disrupting the smooth, free spin of the impeller, which is what we accomplished. By the way, the % drop in wind speed we observed in the experiments was close to what Josh observed in Hurricane Karl.

The test with the car was done to remove the condition of a strong static airstream from the compressor blasting directly into the impeller,which was greatly affected by the steady stream of water being poured directly into it, since it slowed and deflected the airstream. If you watch the car video closely, most of the water splattered against the Kestrel body and not directly into the impeller, because the hose was positioned below the impeller so that the stream would be shredded by the wind and blown into it, rather than shooting a thick stream of water into it.

[Uncle Bobby]

The main idea wasn't necessarily that the rain drops themselves are hitting the impeller at hurricane speed and slowing it down--it's that the water is spending time being flung and splashed around against and inside the impeller blades and housing and causing problems (essentially adding "clutter" to the total airstream flowing through the impeller housing). All we really needed to do was get some water disrupting the smooth, free spin of the impeller, which is what we accomplished. By the way, the % drop in wind speed we observed in the experiments was close to what Josh observed in Hurricane Karl.

The test with the car was done to remove the condition of a strong static airstream from the compressor blasting directly into the impeller,which was greatly affected by the steady stream of water being poured directly into it, since it slowed and deflected the airstream. If you watch the car video closely, most of the water splattered against the Kestrel body and not directly into the impeller, because the hose was positioned below the impeller so that the stream would be shredded by the wind and blown into it, rather than shooting a thick stream of water into it.

I understand all that, but the fact remains that your results couldn't be more corrupt.

Let's say you're standing outside in perfectly calm air holding your Kestrel which is, of course, reading 0. Then Josh grabs a garden hose with a sprayer nozzle and blasts the Kestrel with water from 20 feet away. The Kestrel now reads greater than 0. What useful data do you glean from that experiment? Of course, the answer is "Absolutely nothing". Same with what you have done.

The water slowed the air speed. It HAD to!

You need to test two different scenarios with only ONE variable, but you have multiple variables. What's that called? Experimental control?

[Cory]

If you want to have a civil disagreement, then fine. Until then I'm done with you.

[HurricaneJosh]

Doing the experiments with Cory was a blast. He's really handy, building all these contraptions and rigs.

The other benefit of this trip: Cory inspected the BASTARD and gave it a clean bill of health! It had been banged up a little by flying debris during Hurricane Karl, but he looked at the component parts and said it's fine. (I felt like I was bringing a dog to the vet.)

And Cory made a small enhancement to the instrument: he replaced the black taping around the four anchor magnets with fancy little rubber boots.

[HurricaneJosh]

Re: the Kestrel... As Cory mentioned, it remains the best option for our uses at this point, due to 1) portability, 2) datalogging capabilities, and 3) Bluetooth capability (meaning no wires are needed).

The device collects data, and then you download the data to the computer via Bluetooth, from which point you can generate graphs, etc., like the ones above. The device can sample as frequently as every 2 seconds, which is pretty high-res.

I am soooooo itching to take the BASTARD/Kestrel out in the field Friday. Even if it's a nighttime landfall, it'll be hawt just to collect data in the core.

[Mallow]

If you want to have a civil disagreement, then fine. Until then I'm done with you.

One pretty simple way to test the airstream disruption vs water-in-the-instrument issue is to see how long the Kestrel takes to "recover" back to the 60mph background speed after you stop spraying the water. If it takes longer than however long it's supposed to take to spin back up to 60mph in "normal" conditions, then you know the water inside the instrument is at least partly to blame (which I suspect is the case, as do you).

[Cory]

One pretty simple way to test the airstream disruption vs water-in-the-instrument issue is to see how long the Kestrel takes to "recover" back to the 60mph background speed after you stop spraying the water. If it takes longer than however long it's supposed to take to spin back up to 60mph in "normal" conditions, then you know the water inside the instrument is at least partly to blame (which I suspect is the case, as do you).

Good idea! I have my Kestrel here, so I'll mess around and try that. Thanks for the non-douchebag suggestion.

[Mallow]

Good idea! I have my Kestrel here, so I'll mess around and try that. Thanks for the non-douchebag suggestion.

FWIW, I suspect both factors are at play, but given your attempts to mitigate the airstream-disruption idea, I think that the water-in-the-instrument-induced-friction is the more significant factor.