Lewandowski Transverse Stability Papers

[AltamiroAJ]

Hello,

I have a similar project, despite not working on it on an open-source way, that I am developing as conclusion work from my bachelors degree. On it I implemented a number of the same solutions you did, but I also did Lewandowski(1996) implementation, to analyses transverse instabilities.

I would love to collaborate on your project, do you see that those implementations I have would be a good addition ?

My implementation is not on Python, but I would love to rewrite it to get rid of some bugs and help you with this project.

[elcf]

Hello,
Yes please feel free to submit a pull request, the goal is to get all those empirical methods available for use, so the Lewandowski would be a great addition.

[elcf]

Hi @AltamiroAJ, I've been working on and off on Lewandowski now for quite a bit, have it mostly finished. The issue I'm having with Lewandowski's method is that the equations seem to have some typos and some are inconsistent between references. Unfortunately, the derivations are not included in the papers, so I've been trying to decipher them, but that has just brought up new questions. It seems prudent we try to replicate at least Figure 2 of Lewandowski 1997, to make sure the equations are all "correct". Do you know if you can replicate Figure 2 with your Lewandowski code? That in itself is some work since we'll have to look over "reference [15]" to figure out the entries---I still need to do this on my end too.

[AltamiroAJ]

Hi @AltamiroAJ, I've been working on and off on Lewandowski now for quite a bit, have it mostly finished. The issue I'm having with Lewandowski's method is that the equations seem to have some typos and some are inconsistent between references. Unfortunately, the derivations are not included in the papers, so I've been trying to decipher them, but that has just brought up new questions. It seems prudent we try to replicate at least Figure 2 of Lewandowski 1997, to make sure the equations are all "correct". Do you know if you can replicate Figure 2 with your Lewandowski code? That in itself is some work since we'll have to look over "reference [15]" to figure out the entries---I still need to do this on my end too.

Hello there, sorry for not being able to follow up with my proposition to help back in 21. I will boot it up the old hard drive and run the code to report back.

[AltamiroAJ]

Hi @AltamiroAJ, I've been working on and off on Lewandowski now for quite a bit, have it mostly finished. The issue I'm having with Lewandowski's method is that the equations seem to have some typos and some are inconsistent between references. Unfortunately, the derivations are not included in the papers, so I've been trying to decipher them, but that has just brought up new questions. It seems prudent we try to replicate at least Figure 2 of Lewandowski 1997, to make sure the equations are all "correct". Do you know if you can replicate Figure 2 with your Lewandowski code? That in itself is some work since we'll have to look over "reference [15]" to figure out the entries---I still need to do this on my end too.

Also, about the derivations, have you checked his book? If I am not mistaken, one of the chapters on his book, Marine Craft: Maneuvering and Seakeeping, had some equations in greater detail. I implemented a bit of time ago so I don't quite remember, but I remember using the book together with his articles to be able to achieve the final result.

[elcf]

No worries, I know that it gets tricky doing the code implementations.

That's the issue, the book and the paper have some equations that are slightly different. Here's my summary of the discrepancies I've found:

1. Book N_phi_dot and '97 N_r: Term signs are different between the two, and the book has a 6.714 constant while the '97 is 0.08289. I'm assuming that the constant from '97 is the correct, but the signage I'm less certain. I have to run some example cases to see which makes more sense.
2. Both book and '97 Y_r, the Cv^2 term shouldn't be there, another colleague mentioned this to me, and I confirmed it by reverse engineering the equation with Smiley '52.
3. Book K_phi_dot and '97 K_p, the '97 paper divides the last term by U, while the book does not. I've yet to try to work backwards to see which is "correct.".
4. Y_phi: I'm not sure this deviation is entirely correct. This is derived in the '96 paper, and I've attached my attempt at derivation following Lewandowski's approach. Nevertheless, this equation will not directly show up in the OpenPlaning code, because the numerical derivative is used instead. Derivation: Yphi Derivation (rev 1).pdf. However, if the attach approach is indeed the correct approach, this brings the question on if the "correct" equation also matches well the experimental results. This is something I've yet to look into.

I wished I could re-derive the equations in the '97 paper, but I've had a hard time reverse engineering what Lewandowski did in some of his equations. I'm curious on the mathematical form of Y_r and N_r. I'm not sure how Lewandowski got these forms, as the Smiley reference has a steady-state yaw angle. My attempts at re-deriving them come up with a x_cg term.

[elcf]

I've added a new branch dynroll, that shows what I have so far.

[AltamiroAJ]

I've added a new branch dynroll, that shows what I have so far.

As soon as I can I will take a look, busy with test season in my Master's right now.

[elcf]

No worries, take your time. Good luck on the semester!