What does Diadem use for Integration of acceleration to velocity? I am trying to integrate an acceleration signal from a uniaxial accelerometer.

[Dancer]

I have tried to integrate this accelerometer signal in Analysis/Basic Math. The result I get back is certainly not velocity. 

Any idea what the problem might be?

Thanks 

[Song_D]

Dancer,

DIAdem uses the trapezoid rule to compute integrals. Here's a few things you can check.

1) Are you sure your acceleration have the correct unit? Your accelerometer probably produces a signal in the mini volts range, you need to convert that to m/s^2 or others that suit your application.

2) You might want to check the x values for your integration. DIAdem uses the differences between x values as dt. If your x values are just 1,2,3,4..., your integration results are not going to be correct.

Message Edited by Song D on 04-08-2008 11:30 AM

[Brad_Turpin]

Hi Dancer,

Also note that when you integrate you will also need to add a constant value equal to the starting velocity of the first time interval.  This information is not contained in the acceleration data.  If you know that your starting velocity is zero, then you can ignore this step.

Brad Turpin
DIAdem Product Support Engineer
National Instruments

[preston_johnson]

the sound and vibration measurement suite provides integration functions specifically designed for integrating acceleration to velocity and displacement.  These measures handle all the conversion from voltage to g to in/sec, to mil etc.  Further, the use a high pass filter in front of the integration part to remove the DC offset, which if there is any, your integration will tend towards infinity. 

You might consider using a LabVIEW step with the tools from our Sound and Vibration toolkit.  But do insure there is a high pass filter in front of your integration step and that all of your units are flowing thru the equations appropriately. 

[Dancer]

Hi

Thanks to all.  These are all great suggestions for me to try. 

Will let the forum know of my progress.

Thanks!

Dancer

[hosss]

Hi all,

I want to double integrate the acceleration to displacement. I do all the procedure you discussed here, but the results that I get dont look right. I do the high pass filtering to remove very low frequencies and etc. Is there any other thing I might miss?

Moreover, I see in some references discussing the direct integration in time domain to get velocity and displacement doesn't give accurate results. Alternatively they suggest doing FFT and then doing the integration. I dont know how I should go with this method. wondering if anyone has an experience with this ? 

Thanks