Working Backwards with Simulink?

Question

SAMUEL SHRAGO 2015-1-23

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/171255-working-backwards-with-simulink

回答： A Jenkins 2015-1-23

Hello folks,

I am a student of mechanical engineering and am analyzing some data. The data is collected by a tachometer from a flexible shaft, whose torsional damping coefficient (and stiffness) (in addition to the damping coefficient and stiffness of several torsional bearings in the system) I am to determine. Certainly after having drawn the free body diagram, I have a differential equation to model the motion, which I can then put into SIMULINK and produce a theoretical model of the vibration.

The only problem?

One of the terms in this equation is an unknown, so I don't have a value to put into one of the GAIN boxes in my free-response SIMULINK model.

But I know what the graph SHOULD look like! The SIMULINK model only serves to be a theoretical model and should match my experimental data, which I have! I could show you a photo of it, but it's just like any exponentially decaying sinusoidal function. Heck, I have the peaks and the period and can find all attributes of that graph.

So my question is, can I use both SIMULINK and the data from the experiment to work backwards and determine what this unknown damping constant is?! I fiddled around with the model by varying the gain and checking the scope to see if the model mimics my experimental values, but it's just not a good enough way. There must be another way.

Any help on this is appreciated.