Savitsky method result understanding confusion

[javed]

So i am using the code from openplaning python: https://github.com/elcf/python-openplaning

I have an understanding that for hulls with a large deadrise the stability of the hull is supposed to increase but when i am using the program based on savitsky formulation i got a different result, an opposite one.

basically for stability i am just plotting the eigen value of the system from this code and i see that for increase in deadrise i see the eigen values shifting to RHS of the complex plane i dont understant why because i was under impression that the increase in deadrise should make the hull more stable and have the eigen values more towards LHS plane i am uploading the result as well.

 

[jehardiman]

Welcome to the forums
Have you run Lyapunov's second method on your algorythm equations to verify the form of solution you will get? Or did you just "plug and chug" and are now confused by the results? And what "stability" are you expecting? roll? trim?
Remember that the Savitsky method is about planning resistance related to trim, not about transverse stability (indeed, the models used were fixed in roll). Much like the classic calculation of GM, the vessel is initially unstable but oscillates between two final stability positions where dynamic lift is sufficient to push the vessel back to the other side (i.e. chine walking). All v-type hulls chine walk to some degree, and depending on the parameters of the hull, chine walking can become excessive

 

[baeckmo]

As far as I can see, there is nothing in that calculation procedure that refers to transverse stability, but there is a calc module dealing with porpoising, ie longitudinal trim instability. This is a dynamic phenomenon that depends on a number of factors, including deadrise, and there are specific limiting boundaries for all configurations. I strongly suggest that you read Savitsky's original papers before "pressing computer buttons".

 

[Kayakmarathon]

...basically for stability i am just plotting the eigen value of the system from this code and i see that for increase in deadrise i see the eigen values shifting to RHS of the complex plane i dont understant why because i was under impression that the increase in deadrise should make the hull more stable and have the eigen values more towards LHS plane i am uploading the result as well.

It's not as bad as you think. The eigen values near (-6, +/-5i) are the dominant response. The 30 deg deadrise has more damping and less vibration than the 10 deg deadrise; however, the 20 deg deadrise is the least stable. Apparently the study indicates the design is near a maximum instability point. Maybe re-run the test using deadrise from 5 deg to 45 deg, on 1 degree increments to narrow down the deadrise with the least stability.

 

[Zachary Peter]

From what I know, the Savitsky method has some assumptions and simplifications that might cause results that don’t always match intuition, especially with big deadrise angles. Also, eigenvalues moving right doesn’t always mean less stability in all contexts—it depends on how the system is modeled. Maybe double-check the code details or try some simpler test cases to see if the trend holds. Sometimes, numerical methods can behave unexpectedly for certain hull shapes.

 

[Jimboat]

Lateral stability is tricky to analyze. A static analysis (such as Savitsky theory) evaluates planing hull behavior using static equilibrium of forces and moments. Lateral stability is inferred from restoring stiffness and lift distribution at individual operating points, implicitly assuming that roll disturbances decay naturally. Roll inertia, energy persistence, and speed-dependent amplification mechanisms are not modeled. As a result, Savitsky-based methods cannot predict when or why lateral instability will develop as speed increases.

There is a method developed that can predict dynamic lateral instability and chine-walk specifically. Check out the full description of lateral stability analysis and Chine-Walk.
Lambda, A_DSI, and C_DSI describe different parts of lateral stability in high-speed planing hulls. Lambda represents restoring leverage from hull geometry and weight distribution, indicating how strongly the hull tends to self-right in roll. A_DSI regime (high Lambda, stiffness / restoring-dominated regime) represents the available restoring stability margin relative to a minimum required level, showing how much “reserve” stability remains as conditions change. C_DSI regime (low Lambda, damping-governed stability response regime) represents the available damping capacity relative to roll inertia and loading, indicating how effectively roll motion is suppressed.

Chine walk is defined as the condition where restoring leverage (Lambda) is weakened, the restoring stability margin (A_DSI) is reduced, and damping capacity (C_DSI) is no longer sufficient to suppress alternating roll oscillations. So, the hull is no longer primarily self-stabilizing through restoring geometry, and is instead relying on damping to maintain roll control - but damping is insufficient relative to demand.

 

[jehardiman]

Jimboat; have you or AeroMarine Research submitted that paper for peer review? As a big ship NAME PE when you say "lateral stability" it implies body stability in sway; because transverse stability already implies all stability in roll, dynamic or not, which is well trodden ground.

Just making sure everybody has the same understanding when we use the same words.

 

[Jimboat]

John, thanks for your feedback. I agree with your comments. While the performance powerboat community seems to use "lateral stability" to describe chine walk and rollover, naval architecture calls chine walk as a “transverse” (roll) instability, reserving "lateral stability" for sway. The frequent coupling, in higher velocity planing hulls, of lateral and roll dynamics does sometimes further complicate the terminology.

On another note, your NAME certification and long career in naval architecture are certainly a testament to a great professional run. Congrats!