How to get rid of spurious high frequency oscillations in output acceleration in Abaqus Explicit?

01 February 2022 2 9K Report

I am running an analysis in Abaqus Explicit in which an impact is inputted onto a face of a cylinder and the acceleration is measured at the opposite end of the cylinder. When plotting the output acceleration, I am reading high frequency oscillations in the data ranging from 50 - 200 kHz. The impact data is approximately a haversine of frequency 2 kHz so these high frequency oscillations should be spurious. The impact data was measured experimentally. Mesh elements are quad, structured elements. The material in question is aluminum with 65.7 GPa elastic modulus, 2700 kg/m^3 density, and 0.33 poissons ratio. The following are things I have tried:

Reducing mesh size seems to increase the amplitude of these oscillations.

Decreasing time step seems to help but does not completely solve the problem. The smallest time step I tried (and that Abaqus would finish the analysis) was 0.5 nanoseconds.

Replacing the impact data with a "pure" 2kHz haversine did not solve the problem.

Using a smoothing factor for the impact data did not have an effect.

Using different mesh element types did not seem to have an effect. I tried using C3D8R (8 node linear quad brick), C3D4 (4 node linear tetrahedron), and C3D10M (10 node quadratic order tetrahedron).

Introducing damping did not have an effect in the ranges of damping values that would allow the analysis to finish.

The high frequency oscillations appeared when the aspect ratio was both large and small (L/d = 20 and 1/3) so Pochhammer-Chree dispersion should not be at least solely responsible.

Four pictures are attached. The figure with the line in black is the acceleration measurement on the center of the face on which the force is applied. In this figure, the pure haversine force input was used and the cylinder was 1 inch in diameter and 12 inches in length.

The figure with the line in green is the acceleration measurement on the center of the face on which the force is applied, with a pure haversine input, and the cylinder was 40 inches in diameter and 12 inches in length.

The figure with the red line is the acceleration measurement of the center of the face on which the force is applied, measured impact data, and the cylinder was 40 inches in diameter and 12 inches in length.

The figure with the blue line is the acceleration measurement on the outside of the cylinder 1/4 of the length of the cylinder away from the face on which the force was applied. A pure haversine was used here and the cylinder was 1 inch in diameter and 12 inches in length.

Please comment if you have a suggestion for me to try or a solution to my issue.

Pavlo Krot

Depending on cylinder size, you may excite higher modes of vibrations. Actually, haversine is a non-linear function and has many harmonics in its spectral composition, which could contribute to different modes of cylinder vibration.

Sascha Eisenträger

Hi,

first you should conduct an FFT of your excitation signal and determine which frequency range you are actually exciting. As mentioned by Pavlo, it might be possible that you actually excite these higher order modes. Depending on the frequency content of the signal you can determine the correct spatial discretisation and a suitable time step (although this would be in most cases dictated by the stability limit of the central difference method).

Normally, using a small amount of stiffness proportional damping only should do the job (if the oscillations are spurious). Since this kind of damping is linear in the frequency, it should help to get rid of the high frequency content (a bit like a low pass filter).

Regarding the Abaqus element library, I would suggest to turn off all controls such as hour-glass and so forth to test the standard linear elements. Also a reduced integration is not always the best.

Typically, I would suggest high order elements (spectral elements, Article Mass lumping techniques in the spectral element method: On t...

) for these simulations, but they are not really available in commercial software.

If at all possible, try to run the same simulations in Abaqus/Standard with quadratic elements and compare results. Here, Abaqus uses the HHT-alpha method with numerical dissipation.

Cheers,

Sascha

Do you think can be any Uranium bearing rocks in Eastern part of Iran and western part of Afghanistan?

Do you think can be any diamond bearing rocks in Eastern part of Iran and western part of Afghanistan?

What is the difference between mathematical R^4 space and physical 4D unit space?

If Banks do not provide credit facility, what are the options available for FPOs and impact on producer’s income?

Controlling for pupil light reflex when analyzing pupil size time course?

What are a “Farmers Producer Organization” (FPO) and its essential features?

Strugglling with m6A dot blot any suugesstion ?

Do interactions between biosphere, carbon cycle, & water cycle impact global warming & interaction between atmosphere & hydrosphere?

How to get moment output in Abaqus Standart?

How is energy cycled through the Earth's climate system and how do matter cycle and energy flow through the rock cycle?

How to get moment output in Abaqus Standart?

How to fix errors in my heat transfer steel structure with reinforced concrete slab model Abaqus?

Please explain how the plastic input value should be considered from the true stress-strain curve for the bilinear elastoplastic material model ?

Challenges Faced by Freshers in IT Sector Recruitment?

Simulation of metal drawing by Abaqus with UMAT?

Why running a restart analysis in Abaqus in Ubuntu OS gives an error as attached when running the same job in Windows doesn't give any error and runs?

How Can I Apply Hashin Damage Properties to Solid Elements in Abaqus Without Using VUMAT?

What are the most effective strategies for integrating sustainability in your expertise?

I encountered a situation where the 16s PCR amplification worked perfectly with the DNA from the same sample, but using the same sample Don't working?

Can you suggest reliable procedures to get displacements from accelerations in frequency domain ?