Modal Analysis

I am currently doing a college project in which I have to find the natural frequency of a disc of about 250mm in diameter. For doing this I mounted this disc to an electrodynamic shaker and had done a frequency sweep from 5hz to 500hz.
I have attached 4 strain gauges to the component and 2 accelerometers to the fixture. Around 250 hz I am getting a sudden rise in strain data reading ( A peak) also there is a loud sound for 2 seconds.

I have also performed Modal Analysis Simulation in Ansys and have found the first 3 modes at 1) 150.11 2) 150.88 3)260hz
Experimental analysis revealed the first mode at 250hz but the simulation says 150hz, there is a 100hz of difference which is huge.

Q1) Do I have to give any pre-stress condition in simulation so that the value will match? If yes how? I have tried static structural analysis results giving input to the setup of the modal and I tried giving outward pointing remote forces at an equidistant location 15mm away from the disc but it didn’t help.

Q2) In the simulation the difference between the first two modes around 0.7hz can it be true in reality ? the mode shape is transverse in nature with 2 nodes and 2 antinodes all being diametrically opposite respectively and the second mode is also same transverse vibration with 90-degree phase shift but the third mode is symmetric axial in nature.

Q3) for mode shape visualisation I have a stroboscope but since I am performing a frequency sweep I have to continuously change the flashing pulse rate by rotating the dial on stroboscpe so it becomes very tricky and the resonance time frame is soo short that I am not able to tune it into right flashing frequency (rpm) . How to pre-set the flashing rate so that it will correspond to 150 and then 250 hz for second mode ?

Sir, Please feel free to share your thoughts .

Hi, do you have a link to your SimScale project by any chance?


You have to consider experienced mechanical engineers.

I have only done it on ANSYS

Hi @svekhande. That seems interesting to me. The difference might be caused by a difference in mass distribution, or in the boundary conditions (maybe the strain-gauge or accelometers are causing damping in the experiment that is not seen in the analysis?)

I also generally use Ansys and mostly work on modal and Response Spectrum analysis, and am curious by your results. Please get in touch with me if you want, I’d love to disucss what you are doing and help figure out your issue (and learn from it).

I have only done it on Catia