Is there anyway to convert vibration acceleration data into velocity ?

How can I convert the vibration acceleration data collected in real time to velocity ? I know the relation is to integrate the data but my system requires real time velocity monitoring. I appreciate your reply.

Thanks in advance

[Walter Roberson]

Variable time step or fixed time step?

[Ella]

variable time step The acceleration data is acquired in real time

Answer 1

[Ella]

I already found the answer so I'll just post here if anyone need.

velocity = cumtrapz(acceleration-mean(acceleration));

-- The acceleration signal is subtracted from its mean() value in order to remove the DC components and prevent the unnecessary ramp in the velocity signal.

-- cumtrapz() Matlab function is used to integrate the acceleration data in real time using trapezoidal rule.

[sercan demirkiran]

Hello Ella,

I would like to ask how it worked. Because, My data is random vibration data which is acquired on a bearing. When I applied your method 'velocity = cumtrapz(acceleration-mean(acceleration))' and ramp was still there. In some blogs people mentioned about using high pass (which I have used 5 Hz) filter to prevent ramping in time data. I have tried, 5 Hz High pass filter then applied cumtrapz. Ramping is not existing anymore.

Answer 2

[Bora Eryilmaz]

The new convertVibration function in MATLAB R2024a release of the Predictive Maintenance Toolbox lets you compute baseline-corrected and filtered acceleration, velocity, and displacement signals from vibration measurements using a single sensor output from either an accelerometer, velocity sensor, or displacement sensor.

[Qiaohong]

this convertVibration function says nothing about the unit.

Is the vibration unit m/s^2, then get velocity in m/s, displacment in m, or other units are associated?

[Bora Eryilmaz]

Since the relation between acceleration, velocity, and displacement is either integrations or differentiations, the units do not really matter. For example, when going from acceleration to velocity, if your acceleration was measured in m/s^2, the resulting velocity would be m/s; if your input acceleration was in in/s^2, the resulting velocity will be in in/s. Otherwise, there are no unit conversions in the function.

[Qiaohong]

The proposed convertVibration function is designed to align with ISO standards that focus on velocity and displacement. However, the function lacks clarity regarding the units of input and output. Without knowing the specific units of velocity, aligning with ISO standards becomes challenging.

Furthermore, the acceleration could be measured in g, mg, or m/s². If the unit is in g, it cannot be directly converted to m/s without proper scaling. Failing to apply the correct scaling could lead to errors of up to an order of magnitude.

If m/s^2 results into m/s (actually i am not sure if this is right), then how the g results into? If someone input the g-unit acceleration into the funciton, he may think it is still output m/s, but apparently it is not. Then the result condition according to iso are quite different.