Continue to Site

Welcome to 3DCADForums

Join our CAD community forums where over 25,000 users interact to solve day to day problems and share ideas. We encourage you to visit, invite you to participate and look forward to your input and opinions. Acrobat 3D, AutoCAD, Catia, Inventor, IronCAD, Creo, Pro/ENGINEER, Solid Edge, SolidWorks, and others.

Points Cloud

Vivien Bo

New member
Solid/surfaces to cloud of points

Hello,

I'm trying to obtain a points cloud usable in Generative Shape Design from a body part (made in Part Design), how can I manage this?
So far I managed to mesh the body in Advanced Meshing Tools (which generate the desired points cloud) but the exported data contain all the mesh information (points plus mesh elements) and apparently it can't be imported in this format into Generative Shape Design.

Any idea would be much appreciated,
thank you,

Vivien

EDIT for more details :

I'm a PhD student working on (mechanical) vibration assisted drilling. Currently I'm interested in the chip geometry (before deformation), in this process it can be quite complex for many reasons (tool tip complex geometry, clearance face interference, complex tool trajectory, ... and all of these interfering with one another to create the generated surface). Two consecutive generated surfaces allow me to completely know the theoretical chip geometry (under certain hypothesis), so that is what I am trying to obtain.

I created a parametric "realistic" drilling tool (including different kinds of rectifications on the tool tip), both surface & part model : done and functional
I created a 3D parametric curve representing the vibrating tool path : done and functional
The elements above are controlled under many parameters for research purpose,

Now I tried different methods to generate the resulting surface, either by :
- extruding tool surfaces along the tool path (but as the profile cross itself along the tool path, the result can't be obtained)
- fusing consecutive tool body positions along the tool path (discrete step needs to be small, causing numerical failure in the fusing process after a few macro iteration)

none of the methods above worked.
Then I though, I could work with numerical results, that is when I decided to discretize my tool surfaces / solid into clouds of points, then duplicated (transformations and macros) these clouds of points along the tool path (with a given step), and extract the resulting front (using for instance Pareto's method) which will be the a discrete version of the desire surface. (and with a few more easy operation I can obtain a precise chip geometry, parameterized by the tool path and realistic tool geometry).

All of this to say that the only thing that is blocking me is the underlined bold step, which can be sum up to : "how to create a point of cloud based on a complex surface / body" (projecting points isn't possible, the geometry is too complex)
I focused on the meshing tool, that is indeed interesting because it allows me to create more points where I need more precision, and to force points on edges, the exported file contains all the points information (and more...), but there is no compatible export format to import with Digitized Shape Editor, which is sad because every information are here, it's just a formatting issue...

To summarize the meshing tool allows to discrete (and mesh) any element, but there is currently no way to exploit with other workbenches the points created by this method. Simply because you can't export JUST the points.
 
Last edited:
Vivien,

A cloud of points is typically created by scanning a physical object, creating thousands of points.

A cloud of points can be imported into CATIA with the Digitized Shape Editor workbench, which has tools to create curve and surface geometry from the points.

If you are trying to create surface geometry from a solid part, there are better tools for this in CATIA.
 
Thank you for your answer,

Since my objectives weren't presented, it was probably confusing as to why I would want to do that, I will go into more details into what i'm trying to achieve, it might get even more confusing, but at least all the information will be there.

First post edited (in few sec) for accuracy.
 
You are correct; I was wondering why you seemed to be working in a loop (solid-to-points-surface-solid). Your detailed explanation does help to explain your task. Thank you.

Using the meshing tools to create all the points does seem to be a good idea - except the mesher creates "Nodes" and Nodes are only useful for FEA applications. (there might be a way to convert FEA Nodes into CAD points, but I'm not aware of any way to do this)

I'm puzzled why you can't fuse the series of translated solids to create the tool envelope.

I'm also puzzled why you don't just create a circle based on the outermost point, and then extrude this circle to create the hole. Maybe this is too simple, and I probably don't fully understand your scenario.

Could you send us a picture of the tool bit to give us some understanding of the complexity of the shape you are working with.
 
Last edited:
here is the tool : http://oi57.tinypic.com/2ni6z2d.jpg
the path of any point in the tool during the vibration assisted drilling process can be described as a composition of an helical and oscillating profile along Z axis (the oscillating profile having an amplitude superior to the helical step, which make everything a lot more complicate, and is the reason to be of the vibration assisted drilling process)

To create the realistic non-deformed chip geometry (the final objective), I need to generate 2 consecutive envelopes of the tool along its tool path, and subtract one to the other, as I mentioned before the envelopes can't be generated by Catia as the solid/surface cross itself along the tool path and Catia can't manage the solution, that is why I' considering to obtain a discrete solution using points.

there might be a way to convert FEA Nodes into CAD points, but I'm not aware of any way to do this
This is exactly what I need, sadly I don't think it is directly possible, I might change the topic title into what you said,

I've decided to make a little program in Scilab/Matlab to read the mesh exported data file to extract the cloud of points (it won't be very complicated, the formating is easy), and I will finish treating the numerical results in Scilab/Matlab to obtain the chip geometry, or re-format them to be import-able under the Digitized Shape Editor.
 
Thanks for the picture - looks like a standard drill bit with a chisel tip.

I would take a different approach to your problem and just model the chip directly with some of the CATIA surfacing tools. Extract the two cutting edges of the drill bit, and Sweep them along a Helix curve based on the step (helical pitch). Use the surfaces to build the solid chip.
 
The solution you propose makes an enormous hypothesis I don't want to make : the cutting edges are the only elements removing the material

by the very nature of the process (vibrations of amplitude superior to the helical step, synchronized to the rotation, for example 1.5 oscillations/turn), there are many other surface elements that participate in the previously generated surface (clearance face interference, tool tip indentation), and thus in the chip geometry (which is primarily dependent of the vibration & machining parameters)

Not only this hypothesis isn't true (proven theoretically and experimentally), but also it is fair to say that the result for this particular process is notably different by taking this into account, and as I work on optimizing this process (including forces), these differences cannot be ignored, by a long shot. As a general example, a very thin indentation on the tool tip generate a large amount of forces, that is why my geometric simulation needs to be very accurate
I have already simulated chip geometry without interference and a simplified tool geometry (under Scilab), I'd like a more realistic simulation now, I cannot obtain results with interference all by myself (coding interference would take way too long), that is why I though about using Catia, which manage interference well enough.

(It is very constructive for me to explain this in english, I can see where I lack in details to expose my work, thank you for taking the time to read so far)
 

Articles From 3DCAD World

Sponsor

Back
Top