Go Down

Topic: Structural Engineering Project Help (Read 1 time) previous topic - next topic

kanwar88

For my Master's thesis I am conducting creep rupture testing on reinforced concrete specimens, these tests will last anywhere from 3 to 416 days. I will be running 12 independent tests at any given time and need a way to collect all the load data from all the test rigs. Each one of the 12 test rigs will have 8 strain gauges. The strain gauges output very accurate readings and they need to be recorded over the course of the testing period.

I have yet to work with Arduino chips, but have been doing a lot of research in hopes that it could help me complete my thesis a lot faster. I've only taken basic ECE courses and so I am having difficulty understanding all the schematics of test setups. Anyways, my question to you guys is:

-What is the best way (most reliable) to pull data from 12*8=96 strain gauges?
-Is there a particular board that will work best?
-Should I have 12 separate boards as some tests may end earlier than others (creep rupture is fairly un predictable) or something like a mega board?
-I understand I may need a voltage amplifier, as the voltage supplied by the strain gauges is very small or nil.
-Finally, will the boards need to be connected to a computer at all times, or can they save data etc. (how to I put the data on a computer for analysis usb or ethernet shield)?

I don't expect anyone to do the legwork for me, but I would greatly appreciate any advice and help to get me started. Thank you in advance

Kanwar

Nick Gammon

Do you have a link to the strain gauges? It is hard to answer without knowing a bit more.

Also more physical details. How far apart are the sensors? How often do you need to check each one (eg. once a second?)

My initial reaction would be to have one board per rig, and run the strain gauge input through a multiplexer, however without knowing more about the strain gauge that is just a guess. You can make up (or buy) quite cheap boards that are not a full Arduino board, if price is a concern, however even 12 x $30 would probably not break the bank, compared to all the other stuff it sounds like you would have.

Quote

-Finally, will the boards need to be connected to a computer at all times, or can they save data etc. (how to I put the data on a computer for analysis usb or ethernet shield)?


I would write to a connected SD card. Then you don't need to worry about keeping one or more computers up all the time. I did something a little similar to detect temperature and humidity:

http://www.gammon.com.au/forum/?id=12106
Please post technical questions on the forum, not by personal message. Thanks!

More info:
http://www.gammon.com.au/electronics

PeterH


I would write to a connected SD card.


It sounds as if the data is going to be difficult to recreate so I would go for a belt-and-braces approach - store it locally and also transmit to a central recording station. It would be a great shame if it turned out at the end of the year that a strain gauge was faulty or the SD card hadn't initialised correctly and you didn't have any data. If you store the data centrally you can watch the live values as the experiment progresses and confirm that it's being collected and contains the evidence you're hoping for.
I only provide help via the forum - please do not contact me for private consultancy.

Nick Gammon

I would have gone around once a week and replaced the 12 SD cards (or once a day if you are worried), downloaded the data, and put them back next "cycle". That way you only lose a week or day's data. So you only need 24 cards, the current ones and the alternate ones.
Please post technical questions on the forum, not by personal message. Thanks!

More info:
http://www.gammon.com.au/electronics

kanwar88

The strain guages being used are similar to http://www.kyowa-ei.co.jp/eng/product/strain_gages/gages/kfml.html
They may vary a bit, but the output would be the same.

The gauges will be 1-5 cm apart running the length of Glass Fibre Reinforcement.

For the shorter duration tests, like 3 days - 1 week, the frequency would be higher for data collection (1/sec to 1/min). Having a life feed would be nice, knowing that everything is working and the range/resolution is good. But for long duration tests, 30days to 416 days, its okay if the sensors are offline for a bit (even a day) if it means I can pull data etc.

Also, what about resolution, is there any worry that the strain may go beyond the 1024 bit resolution? There is a way to scale resolution I believe, correct me if I'm wrong.

Budget should not be an issue, we're spending tens of thousands on hydraulic jacks, arduino (automated data acquisition) would just make my like infinitely easier as our lab isn't set up to record data for that long of durations without running computers (which ties up resources).

Thank you for the responses everyone.

Nick Gammon

Judging by the datasheet the gauges have a resistance output. So you just set up a suitable voltage divider to convert that to 0 to 5V.

A normal Arduino has 6 analog inputs (A0 to A5) which are really one ADC converter switched between each input. So you could take 6 readings in fairly quick succession. Nominally a reading takes 104 ┬ÁS, but you would need to add time to save the data somewhere, process it, etc.

You could conceivably run an 8-port multiplexer to increase the number of inputs slightly, or more simply, just get a few more Arduinos.

I would be doing a bit of testing to make sure everything is working as intended. You can go into a low-power mode (SLEEP_MODE_ADC) to reduce noise on the ADC input. This involves slightly more complex code as you have to read asynchronously. Example here:

http://www.gammon.com.au/forum/?id=11488&reply=5#reply5

You might want to consider four-terminal sensing:

http://en.wikipedia.org/wiki/Four-terminal_sensing

That actually mentions using it for "precision wiring configurations for strain gauges".

Otherwise the resistance in the cable run from the gauge to the Arduino might influence the results.
Please post technical questions on the forum, not by personal message. Thanks!

More info:
http://www.gammon.com.au/electronics

kanwar88

How would I go about setting up a four point terminal? This is something I haven't seen before.

Nick Gammon

I'm not sure for the Arduino, I must admit. Maybe someone else can post about that. This page describes something similar:

http://wiki.analog.com/university/contest/design/submissions/team_enigma

Quote

The primary goal of this project is to design a low cost 1)Arduino add-on module, or 2)"shield", to enable students, hobbyists, and professionals alike to implement precision measuring systems as easily as possible.
...
Applications of the Shield

...
    Strain gauges


You may not need to bother if you can precisely measure the resistance of the wire from the Arduino to the strain gauge, and allow for it (or calibrate it and allow for some calibration offset). However if the resistance of the gauge is low enough it might help with accuracy.
Please post technical questions on the forum, not by personal message. Thanks!

More info:
http://www.gammon.com.au/electronics

groundfungus

#8
Aug 30, 2013, 03:23 am Last Edit: Aug 30, 2013, 03:57 am by groundfungus Reason: 1
A four wire configuration uses a power supply with sense terminals.  There are two wires running to the positive side of the bridge (strain gages are usually wired in a bridge configuration) and  two wires run to the negative side.  One of the positive and one negative wire carry current for the bridge and the other two wires connect to the sense terminals on the power supply so that it sees and can adjust the voltage right at the gage.  Or you can measure the excitation voltage and use it in your calculation,  Four (sometimes call 6 wire) is used a lot with transducers (load, pressure cells).

kanwar88

Thanks for your reply, that makes a bit more sense, but again I'm new to this field so a lot of the ideas I have a hard time picturing. Is this 4 point terminal a pre fabricated item I can buy, if so from where, or does it require me to setup from scratch like the schematics found on wikipedia with voltage meters and amp meters? 

Nick Gammon

You basically run four wires. Read the Wikipedia link to see why.
Please post technical questions on the forum, not by personal message. Thanks!

More info:
http://www.gammon.com.au/electronics

groundfungus

#11
Aug 30, 2013, 04:35 am Last Edit: Aug 30, 2013, 04:53 am by groundfungus Reason: 1
Each strain gage should be wire into a Wheatstone bridge.  This will allow you to measure both tension and compression loads.  The bridge also allows you to balance the bridge (zero output for zero load) and apply shunt calibration.  Shunt calibration consist of putting a known and repeatable imbalance on the bridge to calibrate measuring equipment.  Check Omega.com, http://vishaypg.com/micro-measurements/list/product-11042/they have bridge completion networks that you can wire your gages to and have a complete bridge.  From what I have read you will be using a quarter bride setup (one active gage).  Will you need temperature compensation?  

kanwar88

I don't need to adjust for temperature. Thank you for your response that is very helpful. I will look up how to connect strain gauges to a bridge right now.

kanwar88

#13
Aug 30, 2013, 05:01 am Last Edit: Aug 30, 2013, 05:03 am by kanwar88 Reason: 1
The wheatstone bridge will take the place of a voltage divider, correct?

groundfungus

#14
Aug 30, 2013, 05:06 am Last Edit: Aug 30, 2013, 05:14 am by groundfungus Reason: 1
Here is a link to a pdf from Omega about the wheatstone bridge and applying gages
http://www.omega.com/toc_asp/frameset.html?book=Pressure&file=strain-gage-position_REF

and one from Vishay for a lot of good strain gage info.
http://vishaypg.com/micro-measurements/stress-analysis-strain-gages/appnotes-list/
You would connect your amp to the bridge.  

Go Up