LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

acceleration to velocity

Solved!
Go to solution

Hello,

I am trying to convert acceleration data to velocity. But I am getting ramp down when integrated(using time Domain math VI) and also a mean Offset value. I don't know the method to do it. Please explain how to correct it. Thank you.

0 Kudos
Message 1 of 10
(5,176 Views)

The offset is pretty easy to understand if you look at a single cycle. You start at V=0 and negative acceleration. The first half of the acceleration cycle is all negative, so it is building up negative velocity. The second half is positive, so it works its way back to V=0. 

 

I can't tell why it is sloping downward based on the acceleration graph you showed, it could be real, or due to weird sampling. 

0 Kudos
Message 2 of 10
(5,168 Views)

@Gregory wrote:

I can't tell why it is sloping downward based on the acceleration graph you showed, it could be real, or due to weird sampling. 


Just look at the acceleration peaks.  It is obvious that the acceleration has a negative offset.  Therefore the velocity should slope downwards over time.


GCentral
There are only two ways to tell somebody thanks: Kudos and Marked Solutions
Unofficial Forum Rules and Guidelines
"Not that we are sufficient in ourselves to claim anything as coming from us, but our sufficiency is from God" - 2 Corinthians 3:5
0 Kudos
Message 3 of 10
(5,165 Views)

crossrulz: that's true, good eyes! The max is about 32 and the min is about -34. 

 

This post belongs in the Calc 1 sub-forum 😄

0 Kudos
Message 4 of 10
(5,153 Views)

velociraptor-distraptor-timeraptor-5391844

Message 5 of 10
(5,150 Views)

Hi manideepj,

Thanks for posting I have never used these functions but played around to see if I could duplicate what was going on.

 

As crossrulz points out you have a slight negative offset and your plot is correct.  If you would like to see the results change put in a large negative offset in and you will see the velocity drop very quickly (see attached).

Thanks,

-SS

 



Download All
0 Kudos
Message 6 of 10
(5,127 Views)

Thank you for the replies and suggestions.... will have a look into it... thank you once again

0 Kudos
Message 7 of 10
(5,098 Views)
Solution
Accepted by topic author manideepj

Any DC-offset in your signal will result in ramping of your integrated signal. If you jump out of the window you'll feel the constant acceleration of gravity G and your speed will increase (ramp-up) until ... well you know how it'll end.

 

Integration of acceleration can be calculated using different methods depending on your requirements. For real world signals it is often done using analog or digital filters. Analog integrators have the advantage of being 'mathematically' correct but suffer from remaining offset problems. Digital integration can be achieve with reasonable accuracy using a simple low-pass filter as long as you signal frequency is small compared to your sample rate.

 

The attached VI compares what you can achieve when doing raw summation after removing DC-offset to what you get using a first order IIR filter.

- For raw summation you get a stationary velocity level but you may have a constant DC offset (depending on start phase of your signal)

- For low-pass filtering you'll eventually achieve a DC free velocity after a settling time. You may consider adding a high pass filter to remove DC instead of subtracting the measured DC value (especially for continuously acquired signals)

 

billede.png

 

Message 8 of 10
(5,095 Views)

Hi

We have 2 accelerometers from B&K type 4508b with a conditioning amplifier, and a DAQ system from NI. We would like to analyse the vibration characteristics of a rubber and will fix 1 accelerometer on the actuator of an instron and the other on the free end of the rubber which will hang in a system that can only move in the z axis. We will excite the system with a sine function, an ampliturde of 0.1mm and a sweep freq. from 20 to 40 hz.  

We have a nice graph from the accelerometer however: 

1. When trying to remove the ramp (by determining the least squares best fit and substracting it), the velocity and position graph are completely wrong, why? 

2. How can we remove the DC offset? 

3. Do you know how accurate the displacement will be? 

 

PS: We do not have the sound and vibration tool kit 

0 Kudos
Message 9 of 10
(4,553 Views)

Your VI doesn't show any data. Can I suggest you to acquire typical data, save it as default on your graphs and repost the VI?

 

Are you sure your ramp goes through the zero at t = 0. Also if your acceleration signal contains DC it will already be integrated to a ramp on your velocity signal, then to a parabola on your displacement signal. So a ramp may not be the right target for your fitting algorithm.

 

In general there are other ways of removing offset that may work better for you.

 

1 - You are applying a lowpass filter at 140 Hz to your acceleration data but try to also apply a high-pass filter to remove DC. You can combine the two filters by creating a bandpass filter instead of a simple lowpass filter.

 

2 - Alternatively you can measure the DC component of your original acceleration signal and subtract it from it. I'd recommend you to use the waveform measurements DC-RMS VI and eventually experiment with windowing (rectangular and/or Hanning) to achieve the best possible estimate of the DC value.

 

Good Luck!

0 Kudos
Message 10 of 10
(4,547 Views)