Greetings again,

This time I am wanting to put a wind sensor on my solar panel array. The idea is that if the wind picks up then I will rotate panels to the horizontal position to limit stress. The solar panel array is about 10' x 10, and it's at the top of a 6"x10" I-beam 15' in the air. It can get up to 5000 lbs of force with a good storm, hence the desire to go horizontal.

So back to wind sensors. I'd like something that is reliable. It seems every rotating wind sensor I see is both a bit expensive ($30+ each, and if I put 1 per tower then I'll need a dozen!). One sensor I've seen that looks interesting is this one:
http://shop.moderndevice.com/products/wind-sensor

It's called a 'hot wire'. Basically you run power through a wire and wind cools it. Measure temperature and you can translate that into wind speed. Interesting approach.

My question is if anyone has used this method? How reliable is this design? Damage from leaves in the wind/etc? Unreliable in rain?
Thanks
Tony

 

For accuracy you need to measure the relative temperature of a wire in the wind compared to a wire shielded form the wind to cancel the effect of ambient temperature.

Rain could be a problem as any moisture on the wire will cool it significantly and affect its accuracy.

 

Almost every automobile with modern emission controls use a hot wire anemometer as the air mass sensor. Which brings us to the point of being careful to use a type of metal that will not evaporate or oxidize or otherwise change its thermal, electrical, and electrothermal properties after being heated in the air for long periods or time. From memory, platinum wire is used in automobile air mass sensors. Don't know off-hand what other metals if any have the high melting point, low tendency to sublimate, and resistance to chemical reactions that platinum has and that should bring some interesting reading.

Its also sensitive to wind direction. It seems intuitively that if you set up two wires at right angles and calculated their individual wind velocities,you could then find the total velocity by taking the square root of the sum of the squared detected velocities. (It would be good to check those assumptions with somebody with a good grounding in physics.)

 

crutschow's comments are completely on point and the concerns he raises need to be addressed.

If you look closely at the image of that device there are two series of three holes separated by a black line. This is to cut the sensor elements from the electronics so you can enclose most of the circuit (especially that pot!) from the elements.

The sensor itself is just a resistor and a thermistor. I would expect them to last outdoors, and if they do not you could make your own sensor and replace at (annual?) intervals.

 

Why would you need more than one sensor? Is the wind going to be that much different from panel to panel? One sensor, feeding all of the different lowering mechanisms at the same time wouldn't work?

Unless the wind in Texas is different than in Ohio, will horizontal make that much difference? If the panels were mounted at ground level, maybe it would. But at the top of a beam 15' in the air, not so much.

 

Since you're already looking at angle sensors, you might just want to put one on a pendulum. With a little calibration, it'd work fine as a wind speed meter. Immune to rain, temperature, etc.

 

Many people have sucessfully used tungsten lamps with the glass carefully broken away as hot wire anemometers. They are usually used as a bridge with a second one to compensate for ambient temperature.

A good method (for any hot wire design) is to include it in a feedback loop that maintains the bridge in balance with constant temperature on the hot wire and use the current needed to maintain that temperature as your output parameter.

Protection from blown debris is probably just a mesh cage, but rain/moisture will severely affect your result. You can protect from vertically falling rain but fine mist or rain in strong wind may cause problems.

One final note. Hot wire anemometers by their very nature use a few Watts of power continuously, 5-10W maybe for something robust enough to use outside. Presumably if you have solar panels then you also care about wasted energy and therefore perhaps hot wire isn't the best choice for you. Have you considered a simple spinning cup design?

 

Many people have sucessfully used tungsten

I was thinking about tungsten as, "the other platinum". Thanks for confirming that.

I could also suggest a tiny thermistor. They are epoxy coated and can be heated with less than a watt. I'll bet I saw an old datasheet somewhere that demonstrates a hot thermistor anemometer. Probably from National Semiconductor, which is now Texas Instruments...I think.

 

Why not just build your own spinning type. They are accurate and, with the right materials, they can last for years. When my kids were young, I helped a teacher build one with some plastic sheet rolled into cones to catch the wind. He attached them to a rim from a junked bicycle. His was low-tech - just counted the number of times the white cone passed (the rest were orange). You could use a rotary encoder to enable automation.

 

Guys,

I appreciate all the comments. This is what I wanted, a really good discussion about the various options /methods (and also continuing to research it myself).

Here's my current thoughts...

First for the Hot Wire method, while it can work with no moving parts, and yes it is often the mass air sensor in cars, there is a significant chance of reliability issues outdoors over time.

Let me list my main objectives: RELIABLE, AFFORDABLE, NOT BE UGLY, ACCURACY NOT REQ'D (+/- 10 mph). My goal is to look for when winds are high, 40 mph or better, and use that to rotate horizontal. Someone asked how much difference that will make. I'm maybe 100 miles from the gulf coast, so each year we have at least some 60 mph winds. When it gets fast, the panels at dawn/dusk will be nearly vertical which could see 5000 lbs of force. The 15 feet in height is what makes this so critical, it's a giant leverage bar. Rotating it I can drop the force to 1/10th. I will have 5-10,000 lbs of concrete in each hole, and it will have lots of steel. But the force will all start at my 4x 1" diameter x 18" long anchor bolts. So this is a meaningful effort.

I want one per tower due to redundancy. If one wind sensor fails, I could lose 1 tower (and I may not know it has failed). If I only have 1 wind sensor, all my towers can tumble in a storm (expensive!!!). This is basically a risk appeasement, and it simplifies communication requirements (in fact my towers don't need to talk to each other at all). Easy scaling as well. Note that the towers are 40' from one another.


OK, back to design options. I don't want to have the spinning sensors because they all have bearings, and those fail sooner or later or both. I wouldn't mind spending $10-25 each if I could find a quality unit that mounts easily, but I'm not find them (quality mostly meaning high-end bearings).

I am considering something I saw mentioned. Simple description: take a ball hanging on a wire, put a tilt-sensor on it and translate the tilt into wind speed. I don't mean a ping pong ball, I mean more like a softball/volleyball. It needs to be substantial enough that I can ignore breezes (<20 mph) and weigh enough that strong wind doesn't make it 'dance' (bounce around making a measurement inaccurate). I don't want an inflated ball (maintenance), but a bowling ball might be too much (ok, it is way overkill & might not even move much in a tornado!). Maybe a baseball/softball is just the right size/weight?

I'm using a tilt sensor on the array already to sense what angle it is at. I could use the same circuit here (X-Y tilt sensor driving an analog signal to a simple microcontroller which digitizes it & communicates back to my main control board; or maybe a tilt sensor that drives a digital signal in the first place?).

I'm think on paper here. Thoughts, ideas, comments?
Thanks
Tony

 

Thermal conductivity sensors use a similar principle.

I would try a Wheatstone Bridge arrangement with two sensors in the bridge. One sensor is the active sensor exposed to the wind. The other sensor is encapsulated in the same housing but protected from the wind. This is going to provide temperature compensation.

Protect the entire sensor assembly from the rain.

I think this is going to work.

 

Simple description: take a ball hanging on a wire, put a tilt-sensor on it and translate the tilt into wind speed.

FWIW, I was thinking of a rigid "wire" that wouldn't bend appreciably in the wind, suspended from a ball-joint to allow tilt in any direction. I suppose if a vane was built into it to orient the joint in line with the wind, you could use a simple joint that flexes on just one axis.

 

Mechanical idea: use a tail to orient the swinging detector in the direction of the wind to simplify the switch mechanism.

 

Protection from blown debris is probably just a mesh cage, but rain/moisture will severely affect your result. You can protect from vertically falling rain but fine mist or rain in strong wind may cause problems.

Rain affecting the sensor would be sensed as stong winds, yes? If so, then that may be a good time anyway to rotate to safety, especially if the rain is due to a thunderstorm.

Rain also means clouds, so you ain't gathering many amp hours anyway.

On the other hand, you don't want to be rotated to safe long after a passing shower.