How to get velocity and position from vibration acceleration data

[Berin]

Hello, 

I'm working on a project that requires acceleration, velocity, and position data over time. There's no problem getting the acceleration data, it's converting it into velocity and position that's the problem. There's an NI link for a solution (http://decibel.ni.com/content/docs/DOC-4308, see block diagram below) but it doesn't work. 

As u can see in the picture below the velocity and position graphs don't look like the acceleration graph at all and i know they should look alike.

Any ideas on how to fix this

Thanks in advance,

Berin.

[Bob_Schor]

---

@Berin wrote:

Hello, 

I'm working on a project that requires acceleration, velocity, and position data over time. There's no problem getting the acceleration data, it's converting it into velocity and position that's the problem. As u can see in the picture below the velocity and position graphs don't look like the acceleration graph at all and i know they should look alike.

---

Do you know the (mathematical) relationship between acceleration, velocity, and position?  If you have a constant non-zero acceleration, the velocity and position graphs will certainly not look like the acceleration graph!

Bob Schor

[Berin]

Hello Bob, thank you for replying

I am very well aware of the mathematical relationship between acceleration, velocity, and position but however my acceleration signal (shown in picture above) is not constant its a damped sinusoidal signal and as far as i can remember the acceleration, velocity and position graphs of a sinusoidal signals should look alike. 

[crossrulz]

---

@Berin wrote:

my acceleration signal (shown in picture above) is not constant its a damped sinusoidal signal and as far as i can remember the acceleration, velocity and position graphs of a sinusoidal signals should look alike.

---

If your acceleration curve was centered around 0, I would be more inclined to agree.  But it is almost always negative.  So I would expect the velocity to keep going more negative.  Position looks fine based on what I see in the velocity curve.

Could you post your code?  Perhaps you missed a simple setting.

EDIT:  After looking a little closer, I think you have your graphs backwards.  The integration of what you have as the "Position" should look a lot like the "Velocity".  In other words, what you have as "Position" is what looks more like the "Acceleration" and visa versa.

[comrade]

It is in any case kind of baffling, that with the acceleration being almost constantly non-zero (not constant, but never zero!), velocity and position should not level out to be zero. With the roughly linearly increasing negative acceleration from t=1:14 onwards, velocity should be changing quadratically and position to the third power in the negatives.

[Berin · Accepted solution]

Thank you so much for replying but i recently found a solution, my signal is from measuring vibrations of a spring with kistler 8203a50 acceleration sensor and all i needed to do go get the expected graphs  is apply a 10Hz Highpass filter to the acceleration signal and use time domain integration functions from NI sound and vibration measurement suite. 

Thank you again for responding. 

[ductuan.mta]

i find it interesting. Could you send this its source to me (ductuan.mta@gmail.com)

thank you in advance