torsional vibration problem

[rob768]

Hi All,

I am moddelling the drive of an aluminium mill, using mass elements (rotary inertia) and torsional springs between them, much like a conventional torsional vibration calculation. However, instead of steady-state conditions, i want to calculate the response to the torque that is created when the aluminium enters the rolls, which creates a sudden load change in a very brief moment. We measured an increase from zero to 10 kNm in 0.02 seconds, after which the installation shows resonance and the vibation dies out due to the damping present in the system.
Does anyone have an idea how to model the sudden load change? How do i define the load? how do i define the time spectrum?
Any ideas are welcome.

 

[mastroit2002]

Your problem is really interesting. It seems that you are looking for the time domain simulation of a structure as a response of step increment of load (10 kNm in 0.02 seconds).

This behaviour seems due to milling operation starting: this load change could be simulated with an impulsive load change. The load can be a step like force applied on one end of your torsional beam.

With a calculation by hand you could stimate the first torsional frequency and dumping value and relaxation time.

I hope this could be of some help.

 

[Drej]

The very first thing I would do would be to build a model of your mill system (using torsional springs and point masses as you've described) and then carry out a modal analysis to assess the dynamic characteristics of the system. Subsequently you can do a number of things here:

1/ run a time history (and hence transient analysis) on this model to obtain the response or
2/ run the job in the frequency domain (FFT your data to produce a spectrum and then run an harmonic analysis using the FFTd data).

If it is a shock load (load applied over very short time span) I would try inputting a small time history and run a transient analysis (make sure the time integration is switched on to capture inertia effects and that rotary inertia is active in the mass elements). This would be a reasonably inexpensive analysis since you would have only a small DOF in the model. How to define the load/spectrum depends on the FE package you use, but is usually very easy to do.


------------
See faq569-1083 for details on how to make best use of Eng-Tips.com

 

[rob768]

I have calculation model (and natural frequencies). It certianly is a shock load, and occurs when the ingot enters the rolls. I want to do a transient, but don't know how to model it correctly.

 

[Drej]

As I said, unless you specify which package are you using, no one can give you any specific help.


------------
See faq569-1083 for details on how to make best use of Eng-Tips.com

 

[rob768]

The package is Femap as a pre- and postprocessor, and CAEfem is solver