Where is it best to place a strain gauge to measure the twist of a beam? And what type of gauge is best?

In my application, I have a small metal part which experiences a twisting and bending force. The beam part is 8mm wide, 9mm long, and 1.3mm thick. It is fixed at the two screw holes, and a force is applied to the little flat rectangular patch at the end of the cylinder.

Twisting beam

As you can see in the image, most of the stress occurs around the root of the beam, but I'm not quite sure which areas on the surface are experiencing compression, and which are experiencing tension.

  • $\begingroup$ So which results are shown as positive and negative - don’t you have the numeric results available? Normally the image colors are based on result value ranges... $\endgroup$
    – Solar Mike
    Commented Jul 30, 2018 at 10:58
  • $\begingroup$ To clarify, what do you need the strain gauges for? To validate your FEM or just to determine what's under tension or compression? $\endgroup$
    – Wasabi
    Commented Jul 30, 2018 at 13:47
  • $\begingroup$ @Wasabi - I'm making a load cell for a specific application. I cannot use an off-the-shelf load cell. $\endgroup$ Commented Jul 30, 2018 at 16:24

1 Answer 1


Generally you are going to want to place the strain gages in the locations with the highest strain, i.e. the red spots. In addition to knowing whether the locations are in tension or compression, you will also want to know the direction of the stress. i.e. which direction should you align the gage? For this you will need to plot directional stresses. Alternatively, depending on your analysis package you could ask it to plot the vector principle stress. In which case you will want to align the gage in the direction of 1st principle stress (or maybe 3rd).

A caveat though is that the locations with the highest stress usually also have the highest gradient. That means if you place the gage a tiny bit left or right of where you wanted it, or if your analysis was off a little bit, then you aren't going to get quite what you expected. You could also look for regions with medium high strain, but low gradients. i.e. the orange areas. That should get you a value that is easier to compare to the analysis. Whether this matters to you depends on your application.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.