Society of Robots - Robot Forum
General Misc => Misc => Topic started by: atherophage on February 21, 2009, 02:39:03 PM
-
I'm exploring options for a datalogger to record the performance of a wind turbine's RPM. At the same time, to capture the windspeed also would be big plus. If data were recorded once per hour that would work.
I've seen some commercial units with a lot more built-in capacity (and cost) than I need. I'd guess the core of these expensive units is a small micro controller of sorts all dressed up in a fancy box, for $1200.00
Is it possible to take the output from a simple anemometer and route it through a micro controller?
I've seen magnetic sensors for the RPM data, I think, on the Parallax site; has anyone seen a simple wind speed device?
At this point, I do not even know what I need to know to ask the proper questions.
Thanks for your time,
Ath
-
Being a PIC micro guy, if I were going to roll my own data logger, I would use a dev board like this:
http://www.sparkfun.com/commerce/product_info.php?products_id=19 (http://www.sparkfun.com/commerce/product_info.php?products_id=19)
Add an EEPROM chip to that empty socket you see:
http://www.sparkfun.com/commerce/product_info.php?products_id=525 (http://www.sparkfun.com/commerce/product_info.php?products_id=525)
Most mid-range microcontrollers (like the one I referenced) are capable of accepting analog or digital inputs. The micro can then scale or format the data any way you like and store it in the EEPROM chip. The data can then be retrieved via the serial or USB port. Of course, software will need to be written for the micro.
I don't know what outputs are available from typical anenometers, so I don't know how easy it would be to interface to a micro.
-
Many thanks!
-
At the other hand, being an AVR guy... I can simply say that this is a easy to do job....
But you have to think about coding....
Building a data logger myself, I know the pain of limited storage space.... So coding is critical...
I just though of a possible way to do the data logging, but I'm really only testing this out...
Say you have a 128 byte array... that means only 128 8bit values...
Say that you take measurement every 2 seconds...
That means that at 256 seconds no memory left...
What you can simply do is to reduce the values to half...
but also decrease the resolution...
So you can take measurements every 4 seconds...
but to reduce the already full memory you simply do this:
You get the average of the n and n+1 bytes and store them at the array....
You simply need some maths, but I'm not even done with the code... so
no pseudo code so far.... sorry....
Best regards,
Lefteris
Greece
-
Is it possible to pipe the captured information directly to a running computer or attached flash drive?
Thanks,
Ath
-
You can output the data from the microcontroller into a hyperterminal on the computer, and the hyperterminal will save everything in a text file.
-
I was hoping this would be a possibility.
Thanks,
Ath
-
My Axon comes with free DAQ and oscilloscope software . . . takes minutes to get it working.
-
It would have to b quite elementary for me to get it working in mere minutes. I'm just starting out here. I'll get one; hope to prove you correct.
I bought some sort of sensor (a tiny black piece of plastic with three leads) a hall sensor maybe? It was cheap; I also picked up a few of the recommended magnets. Unfortunately for me, no instructions.
Thanks for the info.
Ath
-
Out of curiosity, what is your predicted max RPM? Chances are you can just run your program in a simple while loop and not even bother with interrupts.
-
The turbine is not a standard horizontal axis unit. It does not have typical blades. Expected RPM should be around 700 and up to 3k in a good wind. It's fast. The generator is the limiting factor.
Many unknowns @ this point.
Thanks for asking,
Ath
-
Up to 50 Hz? For a turbine that's fast but painfully slow for a mcu. How many blades?
Can this be a quick hack for testing, or does it need to be long term datalogging?
(trying to think of the fastest/simplest implementation)
-
The turbine is patent pending, finally. For this initial test prototype (still maybe 3 years away from production) a short term datalogger would do. We just need to tweak the rotor speed, initially. After that, rotor speed is irrelevant; then the information needed is wind speed to power production. Power output is displayed on most inverters.
Generators can only spin so fast before they go up in smoke.
To give a picture of something not at all like the turbine in question: http://www.mariahpower.com/ (http://www.mariahpower.com/) but it comes close. I spoke with these people but they either could not or would not tell me what they used to collect their data.
A magnet could be attached to the base and a sensor could pick that up: a bicycle speedometer.
Obviously more data the better.
2 days ago I ordered the Axon. I already have some materials from Parallax. The kid plays with the left overs.
Thanks for your interest.
Ath
-
http://www.inspeed.com/wind_data_logging/Windworks.asp (http://www.inspeed.com/wind_data_logging/Windworks.asp)
Wonder if inside is an Axon?
-
hmmmm looks like an Arduino actually . . .
-
Just a side note on the annemometer. Although you're probably looking for a more professional one than my suggestion as its for testing a final production unit.
I've built an annemometer using a simple I.R. sender/reciever pair placed side by side behind a 10x4 simprop (other proprietry props are available ;D). The prop is attached to a shaft on low friction bearings.
Every 1000 rpm = 1mph wind speed (approximately). Paint a small fraction of the rearside of 1 blade black and set the threshold of the I.R. signal so the mcu doesn't pick up the second blade.