search

ThorstenHellert / SC

6 stars 4 forks source link

SCupdateSupport assumes support and bend axes are the same #40

Open oscarxblanco opened 6 months ago

oscarxblanco commented 6 months ago

Dear Thorsten,

the SCupdateSupport.m adds up the supports and magnets offsets. This is valid only when both systems, bending and support, are the same or connected by translations.

There might be cases where the bending system is rotated with respect to the support. For example, there are two or more bendings one on each side of the girder. In that case the radial and longitudinal directions change in the girder and at least one of the bends will be rotated wrt to the girder system, or both. Or another example where the bending is on one side of the girder and all elements are aligned around the girder axis.

If the total bending angle per support is small the axis rotation might be negligible. However, for a more general implementation I think one would need to project the girder offsets to the bending system before getting the transformation with SCgetTransformation.

I guess that such projections will need the survey data, so, one could ask the SC user to either ask to provide the girder-to-element rotations, or, calculate the survey and figure out the girder-to-element rotation inside SCupdateSupport.m.

Best regards, o

ThorstenHellert commented 5 months ago

Dear Oscar,

That is correct. I was trying to workout a solution for the general case with arbitrary bending magnets but it turns out that this is quite complex with arbitrary magnets. I would suggest if you need the precise misalignments to not use any girder misalignments but calculate the corresponding magnet misalignments offline and then write the magnet offsets and rolls in SC directly. That being said, if you’re interested in working out a general solution I’m very happy to assist you. Cheers,

Thorsten

On Jan 24, 2024, at 04:33, oscarxblanco @.***> wrote:

Dear Thorsten,

the SCupdateSupport.m adds up the supports and magnets offsets. This is valid only when both systems, bending and support, are the same or connected by translations.

There might be cases where the bending system is rotated with respect to the support. For example, there are two or more bendings one on each side of the girder. In that case the radial and longitudinal directions change in the girder and at least one of the bends will be rotated wrt to the girder system, or both. Or another example where the bending is on one side of the girder and all elements are aligned around the girder axis.

If the total bending angle per support is small the axis rotation might be negligible. However, for a more general implementation I think one would need to project the girder offsets to the bending system before getting the transformation with SCgetTransformation.

I guess that such projections will need the survey data, so, one could ask the SC user to either ask to provide the girder-to-element rotations, or, calculate the survey and figure out the girder-to-element rotation inside SCupdateSupport.m.

Best regards, o

— Reply to this email directly, view it on GitHub https://github.com/ThorstenHellert/SC/issues/40, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEHUUFTRCNXJKUEUXG5MVOTYQDINDAVCNFSM6AAAAABCINOQYOVHI2DSMVQWIX3LMV43ASLTON2WKOZSGA4TOOBTGEZDINA. You are receiving this because you are subscribed to this thread.