Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations MintJulep on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
wrap a curve
- Thread starter helpersnx
- Start date
JohnRBaker
Mechanical
Note that Wrap Curve also works with Conical as well as Cylindrical faces.
What I've done on occasion, if the final model is not overly complex, is to create a sort of 'armature' model using Cylindrical and Conical faces which closely approximates the shape the actual surface. I then Wrap my curve(s) onto this 'armature' and then project the resulting curve(s) normal to the face of the actual surface.
Another approach which sometimes can be useful is if the curves are simple and you are able to sort of 'sketch' them using a Spline, then it may be possible to draw them directly on the surface model using the 'Curve on Surface' function along with perhaps 'Offset in Face'.
John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA
To an Engineer, the glass is twice as big as it needs to be.
What I've done on occasion, if the final model is not overly complex, is to create a sort of 'armature' model using Cylindrical and Conical faces which closely approximates the shape the actual surface. I then Wrap my curve(s) onto this 'armature' and then project the resulting curve(s) normal to the face of the actual surface.
Another approach which sometimes can be useful is if the curves are simple and you are able to sort of 'sketch' them using a Spline, then it may be possible to draw them directly on the surface model using the 'Curve on Surface' function along with perhaps 'Offset in Face'.
John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA
To an Engineer, the glass is twice as big as it needs to be.
FrankSwinks
Mechanical
It is feasible to do something that may or may not be along the lines that you are looking for using programming.
Consider first a case which can be determined analytically. Assume we have a spherical cap (radius R1) and a circle (radius R2) which is to be wrapped onto the sphere. That is the wrapped figure will still be a circle of (wrapped) radius R2. Now to flatten out the sphere (over the circle region) the flattened circle radius is R3. The value of R3 is obtained using
R1*theta = R2
where R2 is the length of the sphere segment subtended by the angle theta (in radians)
and
R3 = R1*sin(theta)
or simply
R3 = R1*sin(R2/R1)
To apply this as a program we start with a figure central point and project radial lines to the required surface(s). Then determining a length along the projected spline equal to the figure line length we get a point on the surface along the projected line (spline). Repeating this an appropiate number of times and creating a spline through these points gives a reasonable approximation. It is probably necessary to project the resulting spline(s) onto the surface(s).
Frank Swinkels
Consider first a case which can be determined analytically. Assume we have a spherical cap (radius R1) and a circle (radius R2) which is to be wrapped onto the sphere. That is the wrapped figure will still be a circle of (wrapped) radius R2. Now to flatten out the sphere (over the circle region) the flattened circle radius is R3. The value of R3 is obtained using
R1*theta = R2
where R2 is the length of the sphere segment subtended by the angle theta (in radians)
and
R3 = R1*sin(theta)
or simply
R3 = R1*sin(R2/R1)
To apply this as a program we start with a figure central point and project radial lines to the required surface(s). Then determining a length along the projected spline equal to the figure line length we get a point on the surface along the projected line (spline). Repeating this an appropiate number of times and creating a spline through these points gives a reasonable approximation. It is probably necessary to project the resulting spline(s) onto the surface(s).
Frank Swinkels
- Thread starter
- #4
JohnRBaker
Mechanical
What version of NX are you using?
John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA
To an Engineer, the glass is twice as big as it needs to be.
John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA
To an Engineer, the glass is twice as big as it needs to be.
- Thread starter
- #6
JohnRBaker
Mechanical
Let me work on this a bit.
John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA
To an Engineer, the glass is twice as big as it needs to be.
John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA
To an Engineer, the glass is twice as big as it needs to be.
FrankSwinks
Mechanical
Attached is your part file with one possible solution. It realy depends on how strain can occur on the wrapping.
What I did was to modify the sketch to initially ingnore the blends. I then created two centre lines (the cross).
My next assumption is that the edge adjacent to the datum plane is first attached to the face. After creating the cyan splines I created the corner blends by creating an approriate oriented plane to which I projected the edge curves. I then created the fillet and then projected the blends back to the surfaces. Finally the edge curves have to be trimmed back to the projected blends.
More detail information can be provided if necessary. It really depends on how strain of the wrapping is allowed.
Frank Swinkels
The part is NX7.5
What I did was to modify the sketch to initially ingnore the blends. I then created two centre lines (the cross).
My next assumption is that the edge adjacent to the datum plane is first attached to the face. After creating the cyan splines I created the corner blends by creating an approriate oriented plane to which I projected the edge curves. I then created the fillet and then projected the blends back to the surfaces. Finally the edge curves have to be trimmed back to the projected blends.
More detail information can be provided if necessary. It really depends on how strain of the wrapping is allowed.
Frank Swinkels
The part is NX7.5
JohnRBaker
Mechanical
OK, here's my solution. As I originally suggested, I created some additional faces and curves and then projected them back to the original model. You should be able to go though the model step-by-step to see my workflow. I'll admit that it's a bit convoluted, but the I think the results will be very close to what you're looking for.
John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA
To an Engineer, the glass is twice as big as it needs to be.
John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA
To an Engineer, the glass is twice as big as it needs to be.
- Thread starter
- #10
Thank you , John and Frank,
that are possible workarounds to solve this problem, I feel not very convident. I thought I missed any function or option in Nx.
FYI, I found metaform can do it,too, but require extra license.
I'll see if I can sell this to my colleagues.
regards
that are possible workarounds to solve this problem, I feel not very convident. I thought I missed any function or option in Nx.
FYI, I found metaform can do it,too, but require extra license.
I'll see if I can sell this to my colleagues.
regards
We do this all the time where I'm at, because we have to show seat sensors wrapped onto the foam of a seat.
1) Extract the surfaces that you'll be mapping to.
2) Use the plane of your section to create a rectangle that encompasses your section, and is parallel/perpendicular with it.
3) Intersect the surfaces with the box.
4) Do section curves through the trimmed surface sheet, at regular intervals (perhaps adjust later based on surface quality, and how your system behaves).
5) Use those sections to create a single Through Curves surface.
6) This is the 'wrapping' step. With Project Curve, use the Projection Option of 'Equal Arc Length'. This will maintain the x and y lengths of your original section, as it maps down to the single surface (it only works with a single surface - I'd love UG a lot more if it could go to a surface sheet). The reason for the box, and the surface method, was to keep the u and v lines perp to each other, so that it maps correctly. This should work great for your task, but it can easily fail if you're mapping to something with a lot of sharp direction changes.
1) Extract the surfaces that you'll be mapping to.
2) Use the plane of your section to create a rectangle that encompasses your section, and is parallel/perpendicular with it.
3) Intersect the surfaces with the box.
4) Do section curves through the trimmed surface sheet, at regular intervals (perhaps adjust later based on surface quality, and how your system behaves).
5) Use those sections to create a single Through Curves surface.
6) This is the 'wrapping' step. With Project Curve, use the Projection Option of 'Equal Arc Length'. This will maintain the x and y lengths of your original section, as it maps down to the single surface (it only works with a single surface - I'd love UG a lot more if it could go to a surface sheet). The reason for the box, and the surface method, was to keep the u and v lines perp to each other, so that it maps correctly. This should work great for your task, but it can easily fail if you're mapping to something with a lot of sharp direction changes.
There is a feature that can do the required.
The sad part is that there is very thin documentation about it, - and it is tricky to use due to that one has to trial and error parameters / tolerances to get the correct result. - Wrong values = long calculation time It's called Metaform and exists in NX75 in the Sheet Metal Application. ( previous versions in the Modeling application) It seems to be supposed to only form/unform sheet metal shapes ( sheets and solids) but it will also form/unform curves onto faces. ( multiple faces.) It requires that the wrapped model ( - the curves) has a thickness, so either one sets some material parameters such that the thickness is "ignored" or one sets a very thin thickness. One also has to have a starting face to wrap from. ( datum plane not sufficient.)
As said, it is trial and error to find the optimal solution.
I played around a little w the supplied part, see attached.
If somebody has a tutorial on this feature, it would be helpful.
There is a newer colleague-feature called one step formability-something, but i don't know if it can handle curves.
The sad part is that there is very thin documentation about it, - and it is tricky to use due to that one has to trial and error parameters / tolerances to get the correct result. - Wrong values = long calculation time It's called Metaform and exists in NX75 in the Sheet Metal Application. ( previous versions in the Modeling application) It seems to be supposed to only form/unform sheet metal shapes ( sheets and solids) but it will also form/unform curves onto faces. ( multiple faces.) It requires that the wrapped model ( - the curves) has a thickness, so either one sets some material parameters such that the thickness is "ignored" or one sets a very thin thickness. One also has to have a starting face to wrap from. ( datum plane not sufficient.)
As said, it is trial and error to find the optimal solution.
I played around a little w the supplied part, see attached.
If somebody has a tutorial on this feature, it would be helpful.
There is a newer colleague-feature called one step formability-something, but i don't know if it can handle curves.
- Status