Stream lines around the hull calculation algorithm.

[Alexanov]

Looking for algorithm of preliminary stream lines calculations. Is it actually possible without full scale CFD calculations?

 

[DCockey]

Looking for algorithm of preliminary stream lines calculations. Is it actually possible without full scale CFD calculations?

Depends on what you mean by CFD. It is possible to calculate streamlines using a potential flow panel method which includes free surface effects. These can take less computational time than CFD which uses meshes in space but still require still require siginficant computations. The only such software I'm aware of is the FSAero version of VSAero. It is not inexpensive and I don't know of any open source versions or similar. My impression is that mesh type CFD is used almost exclusively for calculating flow around surface vessels.

 

[Alexanov]

Depends on what you mean by CFD. It is possible to calculate streamlines using a potential flow panel method which includes free surface effects. These can take less computational time than CFD which uses meshes in space but still require still require siginficant computations. The only such software I'm aware of is the FSAero version of VSAero. It is not inexpensive and I don't know of any open source versions or similar. My impression is that mesh type CFD is used almost exclusively for calculating flow around surface vessels.

Thank you. CFD is a best, but take too much time. Idea was to get quick respond to shape modification.

 

[DCockey]

The free surface effects greatly complicate the use of potential flow panel codes. If the free surface effects can be ignored, either because the vessel is moving slow enough that waves can be neglected and/or the area of interest is sufficiently far below the surface, then a 3D panel code such as VSAERO can be used. Such panel codes still require a lot of computations and some expertise to use. But then there is always the question of how much different would the results be with free surface effects.

 

[jehardiman]

Looking for algorithm of preliminary stream lines calculations. Is it actually possible without full scale CFD calculations?

Yes, you can do it the way it has always been done by using sources and sinks. See the 1945 text Hydrodynamics by Sir Horace Lamb. FWIW, most modern CFD is just all the simple elements in Hydrodynamics combined on a scale that renders doing it by hand (or even on the mainframes we had in the early 1980's) impossible in a lifetime. Having a program grid, select, and place all the different sources, sinks, and vortices has been a godsend to those of us who had to toil doing it by hand. Other than that, most of it is just straightforward brute force calculation....and hope that the program can handle the occasional singularity.

 

[DCockey]

Yes, you can do it the way it has always been done by using sources and sinks. See the 1945 text Hydrodynamics by Sir Horace Lamb. FWIW, most modern CFD is just all the simple elements in Hydrodynamics combined on a scale that renders doing it by hand (or even on the mainframes we had in the early 1980's) impossible in a lifetime. Having a program grid, select, and place all the different sources, sinks, and vortices has been a godsend to those of us who had to toil doing it by hand. Other than that, most of it is just straightforward brute force calculation....and hope that the program can handle the occasional singularity.

Singularity based methods are fundamentally different than the space grid based methods which what is usually meant by "CFD". The singularity based methods typically solve for the solution only on the boundaries while the space grid based methods solve over the entire flow field.

The singularity based methods, such as panel methods, have equations which relate every location on the body surface to every other point on the body surface. If free surfaces are involved they become much more complicated than if the body is in an infinite or semi-infinite with rigid boundary space. Singularity based methods are limited to potential flow (viscous effects are negligable) though they are sometimes linked to boundary layer calculations.

"CFD" commonly refers to space grid based methods do not use singularities. Instead they use equations which relate the flow at each location to the flow at adjacent locations. The flow is solved for everywhere in the flow field. The equations solved can be those for potential flow, Euler equations for inviscid compressible flow (not applicable for watercraft), or the Navier Stokes equations, typically with some form of turbulence modeling. There are several methods for including free surfaces. Considerable computer power is needed for other than the most trivial examples.

As computer power has increased exponentially with time there has been a shift from the use of singularity based methods to space grid based methods. My 2012 single processor chip desktop computer which cost around $1200 has over ten times the computational power of the Cray supercomputer I used in the mid-1980s. More recent desktop computers have considerably more power.

 

[jehardiman]

Singularity based methods are fundamentally different than the space grid based methods which what is usually meant by "CFD". The singularity based methods typically solve for the solution only on the boundaries while the space grid based methods solve over the entire flow field. …..<snip>…..As computer power has increased exponentially with time there has been a shift from the use of singularity based methods to space grid based methods. My 2012 single processor chip desktop computer which cost around $1200 has over ten times the computational power of the Cray supercomputer I used in the mid-1980s. More recent desktop computers have considerably more power.

Yes, we all get that, but what Alexander asked for was a method to determine "preliminary stream lines". For this, inviscid, incompressible, and irrotational is acceptable, and the potential stream lines are acceptable, and was the way to do it way back when. While RANS gives a complete solution, the check on the solutions has to start somewhere as most people are not born hydrodynamicists who can visualize flow in their head.

 

[DCockey]

Yes, we all get that, but what Alexander asked for was a method to determine "preliminary stream lines". For this, inviscid, incompressible, and irrotational is acceptable, and the potential stream lines are acceptable, and was the way to do it way back when. While RANS gives a complete solution, the check on the solutions has to start somewhere as most people are not born hydrodynamicists who can visualize flow in their head.

Space grid methods can solve for inviscid, incompressible, invisicid flow(aka incompressible potential flow). They do not have to solve for turbulent viscous flow using RANS equations.

 

[tspeer]

Looking for algorithm of preliminary stream lines calculations. Is it actually possible without full scale CFD calculations?

A panel code is the simplest calculation that can give you streamlines. Check out Aero Logic's CMARC. It's a complete solution, with Loftsman to loft the geometry and generate grids, CMARC to solve for the flow, and Postmarc to display the results. Postmarc can also calculate the on-body streamlines you're looking for. No free surface effects, though.

 

[Alexanov]

A panel code is the simplest calculation that can give you streamlines. Check out Aero Logic's CMARC. It's a complete solution, with Loftsman to loft the geometry and generate grids, CMARC to solve for the flow, and Postmarc to display the results. Postmarc can also calculate the on-body streamlines you're looking for. No free surface effects, though.

Interesting. Thank you!

 

[philSweet]

If you have the streamlines for a hull, and want to see how a local change in hull shape changes them, that is fundamentally possible, but not easy.

This document has an excellent section on the numerical treatment of the local flow potential and local velocities. It's a free PDF download from Springer.

The author, Noblesse from Taylor Model Basin, Has dozens of such papers on potential flow related to surface ships, and numerical methods of exploiting them.

https://www.researchgate.net/public..._of_sources_on_a_steadily_advancing_ship_hull

 

[DCockey]

Phil, thanks for the link.

 

[Alexanov]

If you have the streamlines for a hull, and want to see how a local change in hull shape changes them, that is fundamentally possible, but not easy.

This document has an excellent section on the numerical treatment of the local flow potential and local velocities. It's a free PDF download from Springer.

The author, Noblesse from Taylor Model Basin, Has dozens of such papers on potential flow related to surface ships, and numerical methods of exploiting them.

https://www.researchgate.net/public..._of_sources_on_a_steadily_advancing_ship_hull

Very useful document.