Menu

Discussion Forum

New Discussion
Filter

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.

Frequency sweep for boundary load but except for gravity

Posted Jan 29, 2020, 1:16 a.m. EST Structural & Acoustics, Structural Mechanics Version 5.3 4 Replies

Igju Jeon
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hi all,

The situation that I want to simulate is this: There is a cylinder, and gravity is applied. Then I give boundary load on the floor of the cylinder. I want to change the frequency of boundary load, then get the frequency response. Without gravity, it is very easy-I can just use frequency sweep, but if I add gravity the frequency sweep is also applied to gravity. So I want to sweep the frequency only for the boudnary load not for gravity.

Any comment will be helpful. Thank you.

4 Replies Last Post Feb 6, 2020, 1:18 a.m. EST
Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 4 years ago Jan 29, 2020, 12:29 p.m. EST

Hi,

The first important question is whether the gravity load affects the dynamics at all, that is if stress stiffening effects are important. I this is the case, you should user the Frequency Domain, Prestressed study type.

If not, you can still use this study type, but it may be overkill. If it is a pure superposition, then you can run a separate stationary study for the gravity load, and sum the results during result presentation. The postprocessing is however more convenient when you use the Frequency Domain, Prestressed study.

In either case, note that there is a phase angle involved in the frequency domain results, so a pure summation may not be sufficient. The worst-case sum may not occur at zero phase angle.

Regards,
Henrik

-------------------
Henrik Sönnerlind
COMSOL
Hi, The first important question is whether the gravity load affects the dynamics at all, that is if stress stiffening effects are important. I this is the case, you should user the Frequency Domain, Prestressed study type. If not, you can still use this study type, but it may be overkill. If it is a pure superposition, then you can run a separate stationary study for the gravity load, and sum the results during result presentation. The postprocessing is however more convenient when you use the Frequency Domain, Prestressed study. In either case, note that there is a phase angle involved in the frequency domain results, so a pure summation may not be sufficient. The worst-case sum may not occur at zero phase angle. Regards, Henrik
Igju Jeon
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 4 years ago Feb 4, 2020, 9:18 p.m. EST
Updated: 4 years ago Feb 4, 2020, 9:25 p.m. EST

Hi,

Thanks for your quick response and your comment helps me a lot. I learned how to simulate the situation I want using prestressed frequency domain study.

But I have a problem again, which is about the evaluated value. I want to get the information of deformed geometry, particularly the value of deformed area. I refered https://www.comsol.com/forum/thread/206041/deformed-surface-area-calculation, but this is just effective in stationary study, and give undeformed area for frequency domain study( it is regardless of prestressed or not). So, I want to get the value of deformed area when I use the prestressed frequency domain study.

I did a simple test, assuming a cylinder with radius of 200 mm and height of 100 mm. I fix the floor of it, and give boundary load to top of cylidner. Then I did "frequency domain" study with 10 frequencies but the area calculation of side of cylinder (following the link above) gave me undeformed area for all frequencies. If I fix the floor and just give gravity then study "stationary", it gives me perfect area calculation of side of cylinder.

Any comment will be helpful. Thank you.

Regards, Igju

Hi, Thanks for your quick response and your comment helps me a lot. I learned how to simulate the situation I want using prestressed frequency domain study. But I have a problem again, which is about the evaluated value. I want to get the information of deformed geometry, particularly the value of deformed area. I refered [https://www.comsol.com/forum/thread/206041/deformed-surface-area-calculation](http://), but this is just effective in stationary study, and give undeformed area for frequency domain study( it is regardless of prestressed or not). So, **I want to get the value of deformed area when I use the prestressed frequency domain study.** I did a simple test, assuming a cylinder with radius of 200 mm and height of 100 mm. I fix the floor of it, and give boundary load to top of cylidner. Then I did "frequency domain" study with 10 frequencies but the area calculation of side of cylinder (following the link above) gave me undeformed area for all frequencies. If I fix the floor and just give gravity then study "stationary", it gives me perfect area calculation of side of cylinder. Any comment will be helpful. Thank you. Regards, Igju
Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 4 years ago Feb 5, 2020, 12:56 p.m. EST

Hi,

Since frequency domain is a perturbation study, having implied linearity, areas should not change appreciably. Thus, the constant area is consistent with the approximation level. The spatial frame does not differ from the material frame (undeformed), so the method you are referring to (area integral of '1' computed in the spatial frame) is not relevant.

If you still want to compute the area in frequency domain, you must be aware of that it changes dependent on where in the cycle you are evaluating it. There are terms varying both with the frequency and with the double frequency, though the latter will disappear with a proper linearization.

To compute the area for a linear (or linearized) study, you must resort to expressions that make use of the displacement explicitly. One way of doing that is to use Stokes' theorem, and convert the area integral of '1' into a line integral of something containing the coordinates. The current area can be then be computed by adding the displacements to the coordinates.

Regards,
Henrik

-------------------
Henrik Sönnerlind
COMSOL
Hi, Since frequency domain is a perturbation study, having implied linearity, areas should not change appreciably. Thus, the constant area is consistent with the approximation level. The spatial frame does not differ from the material frame (undeformed), so the method you are referring to (area integral of '1' computed in the spatial frame) is not relevant. If you still want to compute the area in frequency domain, you must be aware of that it changes dependent on where in the cycle you are evaluating it. There are terms varying both with the frequency and with the double frequency, though the latter will disappear with a proper linearization. To compute the area for a linear (or linearized) study, you must resort to expressions that make use of the displacement explicitly. One way of doing that is to use Stokes' theorem, and convert the area integral of '1' into a line integral of something containing the coordinates. The current area can be then be computed by adding the displacements to the coordinates. Regards, Henrik
Igju Jeon
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 4 years ago Feb 6, 2020, 1:18 a.m. EST

Hi,

Thanks for your comment. It stops my wandering and reduces the time.

I have each question to your two replies.

My situation in detail is that there is a cylinder on the slightly vibrating table. The cylinder deforms due to two forces, which is gravity and force from table. I want to get the information of deformed area of the cylinder. Since the force from the table has certain frequency spectrum, I want to study the effect of table force frequency on the cylider.

  1. Question to first reply I simulate this situation by fixing the floor and give gravity to the domain, then give boundary load to the floor with harmonic perturbation. Then I did prestressed frequency domain study but as you know, it does not give result well, because the floor is already fixed and harmonic load to floor does affect nothing. Is there any good way to simulate my situation?

    1. Question to first reply Thanks to your comment, I get the reason why frequency domain give me undeformed area. But because of my poor understanding, I cannot get the way that you suggested. So can you give more detail about how to get the deformed area in frequency domain study, or show some tutorials or manuals dealing with deformed area(length) problems?

    I really hope my explanation will be conveyed well despite my poor English. Any comment will help a lot. Thank you

    Best regard, Igju

Hi, Thanks for your comment. It stops my wandering and reduces the time. I have each question to your two replies. My situation in detail is that there is a cylinder on the slightly vibrating table. The cylinder deforms due to two forces, which is gravity and force from table. I want to get the information of deformed area of the cylinder. Since the force from the table has certain frequency spectrum, I want to study the effect of table force frequency on the cylider. 1. Question to first reply I simulate this situation by fixing the floor and give gravity to the domain, then give boundary load to the floor with harmonic perturbation. Then I did prestressed frequency domain study but as you know, it does not give result well, because the floor is already fixed and harmonic load to floor does affect nothing. Is there any good way to simulate my situation? 2. Question to first reply Thanks to your comment, I get the reason why frequency domain give me undeformed area. But because of my poor understanding, I cannot get the way that you suggested. So can you give more detail about how to get the deformed area in frequency domain study, or show some tutorials or manuals dealing with deformed area(length) problems? I really hope my explanation will be conveyed well despite my poor English. Any comment will help a lot. Thank you Best regard, Igju

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.

Suggested Content