A Resonant frequencies of liquid in a cylinder

AI Thread Summary
The discussion revolves around calculating the resonant frequencies of an inviscid liquid in a cylindrical tank without gravity, focusing on the effects of varying contact angles. The original poster has developed a set of equations governing the fluid dynamics but is seeking assistance to finalize the solution, particularly using spectral techniques and variational methods. Participants emphasize the importance of clarifying assumptions about the system, such as the orientation of the cylinder and the nature of the liquid's surface. The conversation also touches on related literature, including NASA's research on liquid sloshing in containers. Ultimately, the original poster resolves their issue, attributing it to a minor error, and expresses intent to share their findings once published.
member 428835
Hi PF!

I have been on and off working on a fluids problem for 2 years. I am SO close but the answer isn't coming out clean. I'll highlight the equations I solve and the technique. If you can help me finish this, I'll not only be incredibly grateful but I'll either thank you in the paper acknowledgments or if you're interested we can discuss authorship with you. I know I am SO close! Here's the problem: given an inviscid liquid in a cylindrical tank of radius 1 and height h, what are the resonant frequencies of the liquid? Gravity is off, so we also must be given the contact angle ##\alpha## of the liquid (##\alpha = 90^\circ## is very simple, and is the benchmark with which to compare for general ##\alpha##). Before continuing, since the fluid is assumed inviscid, let's call the potential field ##\phi##. Also, let's call the cylindrical container ##\Sigma##, the equilibrium interface ##\Gamma## (which will be a spherical cap, since zero gravity, unless ##\alpha = 90^\circ##, which it's then perfectly flat, which is parameterized in terms of it's surface coordinate ##s##), and let ##\gamma## define the equilibrium gas-liquid-solid line of contact, or the contact-line. Denote subscript ##n## as the normal derivative to the surface for which it is prescribed over.

The equations of motion are: $$\nabla^2 \phi = 0 \,\,\, (\Omega) $$
$$ \phi_n = 0 \,\,\, (\Sigma) $$
$$\int_\Gamma \phi_n = 0 \,\,\, (\Gamma) $$
$$\phi_n'(s) + \left( k \cot \alpha - \frac{ \overline k}{\sin\alpha} \right)\phi_n = 0 \,\,\, (\gamma) $$
$$-\Delta_\Gamma \phi_n - (k_1^{-2} + k_2^{-2})\phi_n = \lambda^2 \phi \,\,\, (\Gamma)$$

Equation 1 is continuity. Equation 2 is no penetration through the cylinder walls (sides and bottom). Equation 3 is conservation of volume. Equation 4 is a constant contact angle condition (as the surface is disturbed, the contact angle ##\alpha## never changes). Equation 5 is an energy balance between kinetic and potential energy (capillary and inertial), where ##\lambda## is the eigenvalue of the system. All ##k## values are curvatures, which I'll explain in detail for the interested reader. ##\Delta_\Gamma## is the Laplace-Beltrami operator.

The solution approach: analytically solve equations 1-3 for the ##\alpha = 90^\circ## case. This is simple to do via separation of variables. Build equation 4 into the an inverse operator (more on this for those interested). Use these solutions as basis functions for equation 5, which will be solved via the Rayleigh-Ritz variational approach.

Now there's a few tricks along the way, which I can describe in detail, and there's a paper on sessile drops which outlines the process for drops on a flat surface. Our problem behaves identically to the sessile drop.

In short, if you're good with math, familiar with spectral techniques, and like fluid dynamics, please let me know. I would seriously appreciate any help you can offer! All code I've used is in Mathematica, and I can implement any suggestions you have so you don't even have to code.
 
Last edited by a moderator:
Physics news on Phys.org
What sort of waves are we talking about?
 
tech99 said:
What sort of waves are we talking about?
Those that propagate from base-normal vibrations. So think of vibrating the cylinder up and down at various frequencies. Particular frequencies will induce resonance.
 
Take a look at NASA SP-106 - The Dynamic Behavior of Liquids in Moving Containers, available as a download at: https://ntrs.nasa.gov/api/citations/19670006555/downloads/19670006555.pdf. Part 1 of Chapter 5 is titled Simulation of Liquid Sloshing.

Chapter 8 is titled Vertical Excitation of Propellant Tanks. Plenty of math, plus some plots of results. Typical plot:
Slosh.jpg


The focus of that document is sloshing in rocket fuel tanks. I found it some years ago when I was working on ways to damp a tuned mass damper.
 
jrmichler said:
Take a look at NASA SP-106 - The Dynamic Behavior of Liquids in Moving Containers, available as a download at: https://ntrs.nasa.gov/api/citations/19670006555/downloads/19670006555.pdf. Part 1 of Chapter 5 is titled Simulation of Liquid Sloshing.

Chapter 8 is titled Vertical Excitation of Propellant Tanks. Plenty of math, plus some plots of results. Typical plot:
View attachment 298068

The focus of that document is sloshing in rocket fuel tanks. I found it some years ago when I was working on ways to damp a tuned mass damper.
I appreciate the reference. Chapter 8 looks very similar to what I'm doing, but they examine 90 degree contact angles specifically. But there are references that may be helpful. Thanks for your interest!
 
joshmccraney said:
Here's the problem: given an inviscid liquid in a cylindrical tank of radius 1 and height h, what are the resonant frequencies of the liquid? Gravity is off, ...
You have not specified your most fundamental assumptions.
1. Does the model include gravity or not ? Is there an up ?
2. Is the cylinder mounted with a vertical axis, or horizontal like a pipeline ?
3. Is the cylinder full ? or What orientation can the liquid in the cylinder take ?
 
Baluncore said:
You have not specified your most fundamental assumptions.
1. Does the model include gravity or not ? Is there an up ?
2. Is the cylinder mounted with a vertical axis, or horizontal like a pipeline ?
3. Is the cylinder full ? or What orientation can the liquid in the cylinder take ?
Didn't think I needed to specify no gravity since there's no gravity term in the governing equations I listed, but nope, no gravity, no thermo-capillary or Marangani effects, no elasticity. Just barebones as I wrote it.

The cylinder is partially full such that the top surface is free. I'm wondering though, would the contact-angle even be an issue if the cylinder was closed and full. Intuitively seems like it wouldn't appear, though now I'm curious. The oscillations are base-normal.

Do you have experience in these types of problems?
 
joshmccraney said:
Do you have experience in these types of problems?
I don't yet know what type of problem it is.
How can there be a top surface without gravity ?
It seems you have a sealed vial that contains vapour and a large drop of fluid that could be stimulated to vibrate inside the vial. The free surface is bounded only by surface tension. Minimum surface area, and the 90° contact angle, would require a single drop be at one hemisherical end of the cylinder.

joshmccraney said:
I'm wondering though, would the contact-angle even be an issue if the cylinder was closed and full.
The liquid to vapour equilibrium will be temperature dependent. The partial pressure of the vapour acts against the contact surface and determines pressure throughout the liquid. Cavitation is possible during oscillation in that system, as is a cyclic phase change at the surface.

A liquid filled closed cylinder, if it did not burst, would also store energy in the cylinder wall. The oscillation modes would be those of the loaded cylinder, not the liquid content.
 
  • Like
Likes member 428835
Baluncore said:
I don't yet know what type of problem it is.
How can there be a top surface without gravity ?
It seems you have a sealed vial that contains vapour and a large drop of fluid that could be stimulated to vibrate inside the vial. The free surface is bounded only by surface tension. Minimum surface area, and the 90° contact angle, would require a single drop be at one hemisherical end of the cylinder.
I don't understand what is meant by "How can there be a top surface without gravity ?" It's a circular cylinder opened at one end and closed at the other, so like a pool, that is not sealed, though as long as the liquid doesn't touch the top and we ignore thermo-capillary effects it really doesn't matter. The free surface is held by surface tension, as you noted.

Well actually 90 degree contact angle would give you a flat equilibrium interface. Having 0 degrees and 180 degrees gives you a concave and convex hemispherical cap.

But we digress. The problem is now just a math problem, really. The 5 equations are given in the initial post thread. The solution technique is outlined there too.
 
  • #10
Baluncore said:
You have not specified your most fundamental assumptions.
1. Does the model include gravity or not ? Is there an up ?
2. Is the cylinder mounted with a vertical axis, or horizontal like a pipeline ?
3. Is the cylinder full ? or What orientation can the liquid in the cylinder take ?
Actually, looking at my question stem, I specifically stated "gravity is off". It was even at the end of your reply when you quoted me.
 
  • #11
joshmccraney said:
Actually, looking at my question stem, I specifically stated "gravity is off". It was even at the end of your reply when you quoted me.
Yes. I never trust the OP to be definative.
The most interesting answers here are those that destroy the question.

Problems often become impossible when they are too restricted, or over defined. It is necessary to step back and appraise the situation.

When the OP math is divorced from the physical reality, and clouded by unspecified assumptions, it is very difficult to identify a parallel physical situation that might share the same math, which is the way I usually work.

You can imagine my surprise when I read in post #7 that the cylinder might be open at one end.
 
  • Like
Likes member 428835
  • #12
Baluncore said:
Yes. I never trust the OP to be definative.
The most interesting answers here are those that destroy the question.

Problems often become impossible when they are too restricted, or over defined. It is necessary to step back and appraise the situation.

When the OP math is divorced from the physical reality, and clouded by unspecified assumptions, it is very difficult to identify a parallel physical situation that might share the same math, which is the way I usually work.

You can imagine my surprise when I read in post #7 that the cylinder might be open at one end.
I can appreciate your holistic approach. As is, the problem looks pretty bare. But I think I've selected a strong parametrization for the equilibrium surface (possibly the hardest part, certainly the most creative). If you're interested I can email you more details. I have several conditions checked, and so far everything looks really good, but the wrong answer means I must be making a terrible mistake somewhere.
 
  • #13
Update:solved it. Dumb error at my end. Thanks for the references; great citations.
 
  • #14
Ok, so now I'm curious for the paper :smile:
 
  • Like
Likes member 428835
  • #15
Arjan82 said:
Ok, so now I'm curious for the paper :smile:
I'll link it once it's published (haven't yet submitted, still writing). Thanks!
 
Back
Top