Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

sine excitation

Please login with a confirmed email address before reporting spam

Hi all.

I am performing a 2D frequency analysis in the Structural Mechanics Module.

I want to apply a prescribed displacement on the left boundary and to see which waves propagate inside the medium. The prescribed displacement is a sine wave of 5 cycles with a certain frequency.

Could anyone tell me how to apply this boundary condition?

If I perform a time-dependent analysis, I can define a piecewise function using t (time) as argument, but this doesn't work in the frequency analysis.

I want to perform a frequency analysis mainly for two reasons:

1. The medium is a thin layer and the time-dependent analysis doesn't give the expected solution. The resulting waves are undersampled even if I use the proper time-step and maximum element size.

2. I would like to avoid reflections from the boundaries and PML doesn't work in the time-domain.

Any suggestion is more than welcome!!!

Best regards
Ramona

4 Replies Last Post 7 giu 2013, 10:23 GMT-4

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 giu 2013, 03:01 GMT-4
Hi Ramona,
first, I think you should apply your sine wave as force as follow: sin(2*Pi*f[1/s]*t), unless you want to modulate your excited wave with Hanning window then you will need to multiply with Hanning function.
second, transient analysis(time-dependent) is more suitable than frequency analysis your case especially for thin layer elements where the expected wave is type of guided waves like Lamb wave.

Good Luck
Hi Ramona, first, I think you should apply your sine wave as force as follow: sin(2*Pi*f[1/s]*t), unless you want to modulate your excited wave with Hanning window then you will need to multiply with Hanning function. second, transient analysis(time-dependent) is more suitable than frequency analysis your case especially for thin layer elements where the expected wave is type of guided waves like Lamb wave. Good Luck

Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 giu 2013, 04:03 GMT-4
Hi,

Note that in frequency domain, the time dependence is implicit. This means that if you write a load as 17[N], it actually means 17[N]*cos(2*pi*f).

In frequency domain, the assumption is however that it is a steady state situation, which contradicts your statement about "5 cycles".

In a time domain analysis the 'Low-Reflectiing Boundary' boundary condition can be used to suppress reflections of outgoing waves. It was introduced in version 4.3

Regards,
Henrik
Hi, Note that in frequency domain, the time dependence is implicit. This means that if you write a load as 17[N], it actually means 17[N]*cos(2*pi*f). In frequency domain, the assumption is however that it is a steady state situation, which contradicts your statement about "5 cycles". In a time domain analysis the 'Low-Reflectiing Boundary' boundary condition can be used to suppress reflections of outgoing waves. It was introduced in version 4.3 Regards, Henrik

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 giu 2013, 10:19 GMT-4
Hi Nadon,

thanks for your suggestions.

I apply my sine wave as a fixed displacement as follow: A[m]*sin(2*Pi*f[1/s]*t). I want to simulate a rigid body vibrating in touch with the thin layer.

If I apply only one pulse (see the attached figure called one_pulse) I can see this pulse propagating inside the thin layer (I have ploted the y-displacement in function of time for different distances from the source). If I apply a burst of 5 cycles (I have changed only the number of pulses) I can't see these 5 cycles propagating inside the medium. It looks like a bad sampling (see the attached figure burst_5cycles).

For both simulations I have used a time step of 1/(20*freq) and a maximum element size of wavelength/20.

Many thanks.
Ramona
Hi Nadon, thanks for your suggestions. I apply my sine wave as a fixed displacement as follow: A[m]*sin(2*Pi*f[1/s]*t). I want to simulate a rigid body vibrating in touch with the thin layer. If I apply only one pulse (see the attached figure called one_pulse) I can see this pulse propagating inside the thin layer (I have ploted the y-displacement in function of time for different distances from the source). If I apply a burst of 5 cycles (I have changed only the number of pulses) I can't see these 5 cycles propagating inside the medium. It looks like a bad sampling (see the attached figure burst_5cycles). For both simulations I have used a time step of 1/(20*freq) and a maximum element size of wavelength/20. Many thanks. Ramona


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 giu 2013, 10:23 GMT-4
Dear Henrik,

thanks for the suggestions about the frequency domain.

I am mainly interested in the time-dependent analysis.

I wanted to perform the frequency analysis only in order to use the PML. I am using the Structural Mechanics Module of Comsol 4.0, can you suggest a technique to apply non-reflecting boundary conditions in the time-dependent analysis?

I would like to understand which wave modes propoagate inside the medium. I was thinking to perform an eigenfrequency analysis but in this case the bodt is free and without load. I would like to apply a source of displacement and constraints on the edges.

Many thanks.
Ramona
Dear Henrik, thanks for the suggestions about the frequency domain. I am mainly interested in the time-dependent analysis. I wanted to perform the frequency analysis only in order to use the PML. I am using the Structural Mechanics Module of Comsol 4.0, can you suggest a technique to apply non-reflecting boundary conditions in the time-dependent analysis? I would like to understand which wave modes propoagate inside the medium. I was thinking to perform an eigenfrequency analysis but in this case the bodt is free and without load. I would like to apply a source of displacement and constraints on the edges. Many thanks. Ramona

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.