Another surface issue

[JNR]

OK, here is another problem for those with lots of surface experience.

We have to sweep a (planar) profile along a spline curve. Problem is that the curve does not start in the plane of the profile.

The actual problem is a crash dummy head. We are looking for head impact with some of our equipment (aircraft Head Up Display). The head trajectory data is derived from high speed video of the actual tests for an given seat/harness design using a standard 50% human test manikin. So this curve starts and ends wherever the target positioned is on the head. This is existing data. We then take the planar head profile at the highest point, construct the spline curve of the head path and sweep a surface to detect head impact. The starting location of the head is predefined also. So they rarely match up. In most cases there is very little difference, so we have just translated the curve slightly. Unfortunately we now have some where there is a large distance between the start of the curve and the plane of the profile. You can't then translate the curve, because you get a very different surface (the rotation of the head is the big problem).

Any ideas? The only thing I can think of is to make a series of derived sketches of the profile at incremental locations relative to the curve points and either the normal vector to the spline (how?) or if actual head rotation data is available (sometimes is - much easier). Then loft.

3/4 of all the Spam produced goes to Hawaii - shame that's not true of SPAM also.......

 

[TheTick]

Why not make a bridge curve tangent to your calculated path curve to make up for the distance?

If the profile plane is in the middle of the path curve, sweep the profile to the near end of the path curve (use a surface sweep),and then use the edges of the swept surface to orient and define a new section.

If the ladies don't find you handsome, they should at least find you handy.

 

[JNR]

Yeah, thought of that too. The profile is never in the middle (that Murphy guy gets around doesn't he!). If I bridge, then the profile just sweeps straight to the plane of the end of the curve first and the result for the rest of the curve is as if you just translated the profile.

The profile is about half way back in the head - close to the cg, so when targets are attached at the cg there is only the small difference I mentioned. (Oh yeah, is it obvious I don't mean a conventional "portrait profile" but the section through "ear to ear"? Just realize there might be a confusion in the use of the word "profile" with heads.) The problem with the new ones is the curve starts between the eyes. About 3.5 inches in front of the profile plane. I did a test by doing a sketch in the side view with a line at the profile and a second line joining it to the start of the curve. Then I swept this sketch and got a flat surface. The top edge of course should be the same as what we are looking for at the centerline of the desired sweep. It's shape was significantly different from translating the curve to the plane or vice-versa. Its all in the head rotation with the curve displacement as the harness starts to restrain the torso and stretch, then the seat deforms. So he eventually smacks into the glareshield pretty much face down. (Sorry - thats like the "dashboard" above the instrument panel for the non-aircraft folks.)

3/4 of all the Spam produced goes to Hawaii - shame that's not true of SPAM also.......

 

[Arlin]

If I understand you correctly (which I probably don't), you need to extrapolate the path curve to meet the profile plane.

Like Tick said, you could insert a new path segment 'bridging' the distance between the path and the profile.

Thing is, you need to be very careful about how you do the extrapolation. Curve extrapolation is a notoriously difficult thing to do, especially if there are no known governing equations for the base curve. You could try some various tools (Excel, Mathematica..) to do some extrapolation work. Or you could try to do it manually as it appears you have done with poor results.

If doing it manually, be sure C1 and C2 continuity is maintained where the bridge curve meets your actual path.

(If extrapolation were easy, you could easily use it to predict the stock markt, everyone would use it, and our economy would BLOW UP!!)

 

[JNR]

Extrapolating the curve does not work. If I pulled your nose in a straight line the results would be the same as if I pulled your ear. But if I pulled your nose in a downward curve, it gives a different surface than if I pulled your ear in the same curve due to the the rotation ofthe profile about a different point (actually an infinite series of points of course). It's hard to explain until you see it. It's kinda rotating about a longer moment arm and lagging the curve at the same time if you see what I mean. When you extrapolate the curve you merely sweep the profile up to the curve and then along it. That's the same as just translating the profile plane to the end of the curve first then adding a straight sweep backwards to the orignal plane.

3/4 of all the Spam produced goes to Hawaii - shame that's not true of SPAM also.......