Intersection of two hyperboloids

Hi,

I have a question about intersection of two hyperboloids. I can easily plot them with plot3 or surf so I can clearly see the intersection of those hyperboloids(circle) but I need somehow to get only the intersection function. By that I mean values in matrix. Does anyone know how to do that ? I didnt find any solution anywhere and I couldnt come with anything good myself...

"Long story short" I need something like this http://www.mathworks.com/matlabcentral/fileexchange/11837 but in 3D for somethink else than just simple intersection of planes...

Thanks, Peter

[Azzi Abdelmalek]

Post the code that plots those curves

Answer 1

[Petr]

Anyway...I did it by cumulation in the end...I counted the errors and it is acceptible in some conditions...

Thanks...

Answer 2

[Matt J]

You could try looping over planar cross sections of the hyperboloids and apply the FEX file you referenced to their 2D cross sections. I.e., find the intersection of the 2D curves in each planar cross sections and make a cumulative record all the intersections in all the cross-sections.

You would want to loop over planes that are as perpendicular as possible (not parallel) to the plane of the intersecting circle.

Answer 3

[Petr]

I probably should have said that I could do it as a cumulation for example in Z plane row by row...but that takes about a 2-3 min...

I just need something quicker because I need to do that 2-3 times...

So you see that 10 min for result is quite a lot...

[Matt J]

Is there a reason you don't just find the intersection analytically?

Answer 4

[Petr]

Yes...those equations do not have analytical solution...and numerical it always exceeds limit...plus when there is a low of values in my case circle it is difficult...I have to do it very precisely with high accuracy...but even then it is analytically unsolvable...

That's why I whink is easier to "plot" them and then somehow determine the intersection...

[Matt J]

I can't see why the analytical solution is difficult. Are the hyperboloids rotated/disaligned with the coordinate axes? If analytical solution is difficult, then how do you already know the intersection is a circle?

If you need high accuracy, I don't see how the plotting tool will help. The intersection obtained from discrete plot samples has to use some manner of interpolation, and that will cost you accuracy.