Data logging

[Stewart Corman]

Did a search on data logging and only came up with a few hits, so decided to start a new thread on the subject.

As I am getting ready to fly again, and the turbine is wired to the house with nichrome loads mounted in hot air plenum, I will need to record performance ( especially in marginal weather or at night).

I have a La Crosse (Toys-R-US) wireless anemometer. I DON'T recommend their product line. They only have an RS232 serial port and won't work with a USB adapter, so I probably have to use an old desktop unit and leave the monitor off. Yes, it does work 300 ft away, but requires alkaline C batteries and can't use NiMH! Software provided is submarginal at best. Haven't tried real time data interface lacking other software to capture it.

I found this inexpensive $25 module for a puter A/D interface which includes free software package:
http://www.dataq.com/194.htm
(they also have a USB version for $50)

If I can get a PMCIA serial/USB card that works with anemometer, then I can use old laptop as data collection system.
Ultimate goal is to record current from precision shunt voltage, output voltage, WS, and load setting (0, 1, 2)

http://http://www.emproshunts.com/webstore/itemimages/LAB.gifhttp://www.emproshunts.com/webstore/itemimages/LAB.gif
http://http://www.emproshunts.com/webstore/itemimages/LAB.gif

just found a link to this free data logging software package:
http://www.aprsworld.com/wind2db/
this company sells many components as well
Anyone have any experience with this company's products?

Here are a few links posted from Otherpower:
http://www.thebackshed.com/Windmill/PicLog.asp
http://www.fieldlines.com/story/2007/11/17/12656/796
http://www.fieldlines.com/story/2007/11/18/18820/361
Dave B (Wincharger 18') had a few posts there, but the charts all disappeared,
so perhaps I can get him to sign up and he can contribute a thread on exactly what he has accomplished in data logging.

If anyone would like to present a tutorial on this subject, I am sure many would be appreciative.
Any good/bad experiences, DIY builds, or personal programming successes could be good threads as well

Stew Corman from sunny Endicott

[Mark Parsons]

Hi Stew,

I had a chance to work with the DataQ product and WinDAQ software you point reference to, last December. I wasn't impressed. Selecting frequency input for one of the channels caused all kinds of garbage to show up in the trend plots. One of the application's channels was to measure and display the RPM of a V-10 marine engine. Using one of the analog input channels with a proximity sensor created a nice crisp signal as displayed on my oscilloscope, the software would mash the results significantly. Powering down and up would create new interpretations by the software of the data. It seemed to handle the standard analog signals adequately but scaling and resolution are very cumbersome. Also, if you wish for more than 300 samples per second you have to pay more for the 'professional hi-speed' version.

Hence the adage - you get what you pay for...

Mark

[Mel Tyree]

Stew,
I have done lots of data logging in the past and have bought all kinds of DAQs with RS232 interfaces and the standard that can be chained (RS434? don't recall) and have also used PCMCI cards with built in DAQs.
I am in the process of buying a high quality DAQ (USB interface with software that can do signal analysis i'n-stream'). The pupose is to monitor power flow (using CT coils at about $90 a pop) to my geothermal heat pump, water pump, hot water tank, as well as the power (3-phase) from my Bergey and from my inverter to my service pannel (1 phase line AC). I also want to monitor a couple of cup annemometers. I asked my brother, who is an electrical engineer and does lots of similar stuff. He suggested an IOtech DAQ & software system which is rather expensive:$1400+. Well, I am on a research grant so I bought it, but I wouldn't go there for a private purchase.
If you want to spend anywhere near that magnitude of $ I can send you info about what I bought...but I don't have it yet for testing purposes so can't say the investment will pay off:confused:.
--Mel

[Robin Gilks]

After 4 years of using a wireless LaCrosse weather station (a 2310) I connected it up to the PC a couple of weeks ago. I quite impressed with the results I'm getting from the Cacti package for graphing the results - now all I need is the remote reading of the volts/amps from my wind turbine :) Take a look at what I've got so far on my website (http://gilks.ath.cx/cacti/graph_view.php?action=tree&tree_id=2&select_first=true) for an idea of what I'm getting from it.

The remote reading will probably be yet another Atmel AVR project (I've got the ICE for the original range of 8 bit devices). I've already got a 100amp shunt for a panel meter on the turbine rectifier stack which at .001 ohms gives me a fsd of 100mV. I think a couple of op-amps and a 4433 or mega8 AVR microcontroller will be all that is required. If anyone is interested, I'll knock something up and report back.

In the meantime, feel free to browse about (http://gilks.ath.cx/gallery/main.php?g2_itemId=5189) my adventures with a 10 footer I've just put up.

[Stewart Corman]

Robin,
Firstly, welcome to the forum
I have seen some of your photos on Otherpower ..very nice project.

Besides the data gathering specifics,
remote reading of the volts/amps from my wind turbine....If anyone is interested, I'll knock something up and report back.I also would really like to see more of the crank up tower ...looks like the lattice cranks, then the pole cranks further .. some specs would be appreciated as a start.

Perhaps start a thread just on the tower design??

Stew Corman from sunny Endicott

[Mel Tyree]

Yes, I have a Davis weather station that posts data to a public web site, see
http://www.wunderground.com/weatherstation/WXDailyHistory.asp?ID=KNYELLEN2
But I can't get it to post customized graphs on my private web site. I can do custom graphs on my local computer much like yours. I am new to writing web pages using DreamWeaver. How do I do it?
Wind velocity versus power: I too want to log power output versus wind speed. Currently my annemometer is only 10 m up the tower. When it gets warmer at my home (and after improving my climbing gear) I want to get it nearer the Hub height. QUESTIONS:
(1) How close can I get it without the airfoils influencing the wind speed?
(2) One technique I have seen is to have two annemometers at two different heights below the hub and use the gradient in wind velocity to estimate the value at hub height. How do you do that exactly? They use a wind shear calculations. What is the formula?

[Robin Gilks]

Hi Mel

I'm using a package called cacti - its running on a Linux host but I think there is a version that runs under Windows. It uses a database package called rrdtools underneath to do all the averaging and condensing of data - the graphing is also done at this lower level as well.

The LaCrosse weather station I'm using has a suite of open source s/w tools called open2300, one of which can read all the data from the weather station. This results in a log file with data like this in it:
20080408104005 2008-Apr-08 10:40:05 18.8 15.6 10.8 52 73 0.0 247.5 WSW 15.6 0.00 0.00 2009.81 1022.900 Rising Sunny
20080408104505 2008-Apr-08 10:45:05 18.8 15.4 10.0 52 70 0.0 225.0 SW 15.4 0.00 0.00 2009.81 1022.800 Rising Sunny
20080408105006 2008-Apr-08 10:50:06 18.9 15.4 10.0 52 70 2.2 225.0 SW 15.4 0.00 0.00 2009.81 1022.900 Rising Sunny

As you can see from the timestamps, I collect this every 5 minutes. At the same time, the cacti interface scans the last entry in this log file and extracts a summary in a format it can interpret. The data going into cacti looks like this:
To:15.4 Ti:18.9 DP:10.0 RHi:52 RHo:70 WS:2.2 DIR0:225.0 WC:15.4 R1h:0.00 R24h:0.00 Rtot:2009.81 RP:1022.900
This tokenisation allows cacti to work out what numbers are what values - To: is 'Outside Temperature' for example.

The cacti package then stuffs this data into the rrdtools database every 5 minutes and then draws graphs from it on demand. I got the basic instructions from the open2300 website (http://www.lavrsen.dk/twiki/bin/view/Open2300/CacTi) and the tutorial (using babel.altavista to translate the Italian to English). That a a read of the tutorials on the Cacti website sorted me out.

Its not trivial - it took me about a day to get the basics up and another week to iron out the details inside some of the features (templates and CDEFs took a while to work out!!) - but I'm happy with it now. I see no reason at all that the basic 'script' (cacti input method) shouldn't get the data from whatever weather station or other data collection method you have.

I'm planning to get amps/volts from the generator in due course and will plot it more for checking the trends rather than deriving an exact relationship twixt windspeed and power output.

[Robin Gilks]

I also would really like to see more of the crank up tower ...looks like the lattice cranks, then the pole cranks further .. some specs would be appreciated as a start.


I've just borrowed the trailer mast for a couple of months so I don't know a lot about it. Originally manufactured for a local university for wind measurements, it has 6 telescoping sections that extend up to about 120 feet. Only 3 of those sections are mounted on the trailer at present and I'm running at about 30 feet high - quite enough for my experiments. It uses one winch to tilt up from the horizontal to the vertical and a second winch to extend the sections. The guys (I'm just using a set of 3) are anchored to waritars hammered into the ground.

The guyed pole I'll be using (when I've finished welding etc!!) will be about 14m high (45 feet) which will clear all the shelter trees and house by a good margin but still have a 'falling over' footprint well within the paddock I'm siting it in. I'm putting up some fencing to help with guys first though.

[Mel Tyree]

I am interested to get a measure of phase difference between current and voltage 'exported' by by 10 kW Bergey. I have a dual strace digital oscilloscope that ought to help but using a shunt (like that in our original posting for current measurement) would require differnetial voltage measurement to get A, whereas a oascilloscope has single-ended voltage measurement (one end grounded). Given that the voltage ouput on the Gridtek10 might go up to 300 or 400 VAC, how would you suggest I do this?
I understand that phase difference between A and V is useful for designing an inverter for MPPT. I presume I would need a high voltage differential amplifier that would not itself induce another phase shift.
Isn't there a way to do this with X-Y plots on an Oscope? Its been a long time since I took my University electronics course.

[Mark Parsons]

Greetings Mel,

My old 20MHz Tektronics dual trace analog Oscilloscope would provide the X vrs Y channel plot you wish to observe volts and amps phase difference. Set each channel for AC coupling. Connect one channel to shunt, the other to voltage. Use a 10:1 attenuating probe for the voltage. Adjust gain for each channel to show a common amplitude on the screen and then switch to X vrs Y mode. The phase difference shows up as anything but a straight line. The more round the plot the higher the phase difference.

What you are plotting is essentially power factor (cosine of true to reactive power). This would change with inductive or capacitive loads. How is this used for an MPPT algorithm?

Regards,
Mark

[Mel Tyree]

Yup, I have one of those too, cost me $21K when new in 1982. But it finally died two months ago. I bought some replacement part for the Tekronics on Ebay 4 years ago but can't figure where I put them! So I recently bought another digital scope but have to read the manual to see if it can do the job (I spent just $2K? last year for this 150 MHz scope, EZ Digital Co, Korea; this unit is lighter than the boat-anchor-Tektronics). I used to do research on ultrasonic accoustic emissions so that is why I started working with digital scopes.
How this relates to MPPT, I don't know. I will get back to you on that. The people who make the Aurora inverters are working with me to see if they can get one working on my Bergey 10 kW turbine. They have requested that I measure these values.

[Rob Beckers]

The relation to MPPT is that one sets actual power in the MPPT table. Measuring voltage and current only tells you power if both are in phase. In case there's a phase angle then power becomes P = V * I * cos(angle). So, to get points for populating the MPPT table power is needed. From the equation it also follows that if the phase angle is small it can just be ignored, and knowing V and I for various frequencies is sufficient.

-RoB-

P.S. To be exact, the MPPT table in the Auroras is set for output power. So, strictly speaking, measuring V and I coming from the turbine gives input power. Also measuring either power or V and I at the GridTek output would be great, but it shouldn't be too hard to figure a certain efficiency and derive the power curve (with output power) from the GridTek input measurements.

[Mel Tyree]

Hi folks, (This applies to datal logging too! so see my approach)
I have fulfilled a promise to myself to do a ‘scientific’ study of the performance of my Geothermal Heat Pump (GHP) which provides heat, A/C, and some hot water in my Net Zero Energy House. WHY? Because I could not find a heating installer in Clinton County, NY, where my home is located, who would advise me to use one or believed one would actually work at all … or cost-effectively if it did work.
So I have done measurements and written a report close to ‘peer reviewed’ scientific standards, but am not sure anyone will read it! I am a scientist with a very high citation index (upper 0.5% highly-cited in my very narrow field), so I am used to being read but have my doubts about this one. So if at least two people read it I will be happy! Let me know if you do crack a page. It is ‘published’ on the Home page of my SWIEP site;
www.ualberta.ca/~mtyree/SWIEP
Happy reading,
Mel Tyree, Moderator of SWIEP, Dept Renewable Resources, University of Alberta, Edmonton, Canada.
PS- I will double post this on small-wind-turbine sites, the reason being that my turbine provides all the power need for the GHP… maybe that will get me my two readers. It also provides an example of the level of technology I am using to study my turbine.

[Rob Beckers]

Mel, do you have a direct link to the article? I don't see it on your homepage, nor in the 'publications' page.

Ground source heat pumps are something I'm most definitely interested in, at a personal level in fact. Our house used to have a first generation ground source heat pump (water source actually, using two separate wells). Being first generation it performed poorly and was replaced by a gas furnace about 10 years ago when natural gas became available in this area. We bought the house 3 years ago, and the current crop of heat pumps is supposed to be 400 - 500% efficient (ie. 1kWh of electricity gets one 4kWh of heat). That makes running a heat pump half the price of heating with gas, and since we already have the two wells it would be relatively cheap to go back to a heat pump. So, I've been looking at them. Others in the neighborhood have gone back to new heat pumps and reactions here sofar are positive. Nobody has measured efficiency though, and scientific data would be very welcome.

-RoB-

[Mel Tyree]

Rob,
Yes, the link is at the top edge of the photo of my home. I should have said that.
sorry,
Mel

[Mel Tyree]

Rob Becker and I are working on connecting a 10 kW Bergey to 2 Aurora inverters. He and I have been puzzling over some data & since he has already stated in another thread that we are working together on this project, I hope he doesn't mind if I open the question up to a wider group.

The attached Excel file summarizes some measurements I have made with a digital oscilloscope and digital VOM to get Hz and RMS voltage at the point where the Bergey Generator meets the input of the GridTek10. The voltages and Hz were measured between two of the 3-phase lines.
OBSERVATIONS:
(1) When unloaded (GridTek10 off) the Vrms is a linear function of Hz and a nice sine wave
(2) When loaded (GridTek10 on) the Vrms is a non-linear function of Hz and is a modified square wave ... the shape of which changes with Hz. (See xls file.)
QUESTION:
(1) What is the inverter doing to distort the voltage waveform?
(2) Is it 'valid' to measure RMS voltage with a digital VOM on a modified square wave? My tentative answer is YES??? Here is the reason: The Fourier Theorem says that a square wave (or any wave form) can be described as the sum of sine waves of differing amplitude and frequency so the VOM ought to measure a correct RMS voltage for a family of sine waves.
(3) Can I use this RMS value to estimate power input? On a 3-phase circuit the power flowing = SquareRoot(3)*RMS Current*RMS Voltage; well this is from memory and I will check but whatever it is: it is a formula.
Now: I have a CT coil to measure current with the following specs: 10 Hz to 200 Hz and NON-sine wave. Seems good. So final question
(4) Can I use the same power function as I quote from memory above to compute power with the values from the CT coil and VOM (collected with a data logger of course)??

I will be asking other people (maybe Paul Gipe or some of the NREL group) but thought I would start here to see where I get.
Why do I care? I want to measure the performance of the GridTek10 first so I can tell if the Aurora works better or worse.

[Rob Beckers]

1) Those flat tops are the DC rail voltage after the rectifier. Clearly the GridTek inverter does not use PFC (Power Factor Correction), resulting in high harmonic distortion of the input. The flat shelves around V=0 are a bit tougher to explain, my guess would be that they're the interaction between the inductive source and capacitive load.

2) If your Volt meter is a true RMS meter it should work fine for any waveform (within the frequency limits of the meter). If it's a non-RMS meter then it will only give RMS for sine-wave input signals (those are just peak detectors that use a factor to 'correct' for RMS value).

3) Yes you can, as long as there is not too much of a phase angle between current and voltage. Officially there's a cos(angle) term in that equation.

4) Yes, with the caveat mentioned in (3).

-RoB-

[Robin Gilks]

So I have done measurements and written a report close to ‘peer reviewed’ scientific standards, but am not sure anyone will read it! I am a scientist with a very high citation index (upper 0.5% highly-cited in my very narrow field), so I am used to being read but have my doubts about this one. So if at least two people read it I will be happy! Let me know if you do crack a page. It is ‘published’ on the Home page of my SWIEP site;
www.ualberta.ca/~mtyree/SWIEP
Happy reading

Looking at the diagram of the connection between the source well and the dump well, I wondered how long the pipe was going into the dump well. I can't help thinking that a relatively small pump would be required if the water flow could be made to siphon by having a long pipe going down into the dump well and not just pouring the water in from the top.

Enough pumping to overcome friction and provide for the correct flow rate, which presumably would be determined by the diameter of the smallest pipe in the flow path, and allowing for the draw-down of the source well is all that is required.

Would welcome your thoughts on this...

[Mel Tyree]

The pipe goes about 5 or 10 feet under the surface. I was advised NOT to dump at the surface because the water-fall resulting would generate air bubbles (embolisms) that will work their way into the aquifer and plug the flow. When you pump from the source well the water level drops below the top of the aquifer and in the dump well the water level rises above the top of the aquifer. So there can be no siphon effect, the level is in the wrong direction.
I use siphon principles in my research all the time. Having a pipe going deeper does not help. If it did, you could generate hydraulic power for free! Try it to confirm. I tried it about 40 years ago just to make sure I understood the physics.

[Mel Tyree]

I found this inexpensive $25 module for a puter A/D interface which includes free software package:
http://www.dataq.com/194.htm

Stew Corman from sunny Endicott

Hi Stew,
I was looking at this old message to see if this cheap Dataq might provide digital output support. The unit above has one digital output line but not for general use.
Does anyone have a cheap alternative?
I want to be able to remote reset my GridTek10 inverter. The manual reset button is a normally-closed switch so reset involves a brief open-circuit condition. So my idea is to put a NC-solinoid-driven relay switch into the circuit. In the old days I could just write a bit to my parallel port and use that to activate a solinoid driven relay switch thru a couple of external components. I have done this, for example, to turn pumps on and off. But computers these days do not have parallel ports any more.
I already have energy monitors on my GridTek10 so I know when it needs a manual reset and I can use the Windows Remote Desktop Connection to access the computer with the monitors to run 'reset software', the problem is to locate reset hardware that can drive a solinoid switch.
Does anyone know of a cheap solution that uses a USB port or USB to RS232 port? Perhaps something that allows a computer to remotely turn on and off lights in a home???
I have seen USB-to-printer-port hardware but they work ONLY thru parallel port printer drivers. The hardware does not allow for user-written code.
Thanks,
Mel