Hello Mr. Rivai,

For the first question:

Automatic Feature Recognition (AFR) is the key bridge linking CAD and CAPP (Computer aided Process Planning) systems. The STEP file (AP203) contains low level data such as geometric and topological entities (lines, circles, vertices, faces…), however, CAPP systems don’t understand these data, they cannot be processes directly because CAPP systems need information about manufacturing features. So the first thing to do in CAPP is to recognize manufacturing features from low geometric and topological data contained in STEP file.

A manufacturing feature can be Surfacic of Volumetric. The Surfacic feature is the trace let by the cutting edge of a cutting tool on the surface of the workpiece after machining. The Volumetic feature, which is much used, refers to a material removal volume, for which a cutting tool and a machining operation are well known. Which means that when a manufacturing feature is detected (for example a hole), its cutting tool is known (a drill) adding to its operation (drilling).

For the second question:

The output of feature recognition are manufacturing features: holes, grooves, chamfers, rounds, cylinders, shoulders….each feature has its own dimensions, its locations, adding to the appropriate machining operation. These data can be used then by CAPP and CAM systems.

Hope I answered to your questions. Do not hesitate if you have any other questions, we are here for that.

Best regards

In the CAD-CAM workflow, the piece to manufacture often comes from a 3D scan technology producing big point clouds that later need to be meshed (in a reversed-engineering fashion).

These meshes, when done correctly, are watertight 2D manifolds without any explicit feature lines. For example, even the edge between two perpendicular planes wound not be that easy to recognize from the local geometric properties of the mesh along these triangles due to noise and defects. Identifying these feature lines is often done semi-automatically in order to generate a proper STEP file used on the rest of the manufacturing workflow.

Dear Martin,

Honestly, i have never heard about meshing in automatic feature recognition (AFR), maybe they are used in a field other CAD/CAPP/CAM. The AFR approaches that i know are rules-based approaches, cell decomposition approaches , hint based approaches , syntactic pattern approaches, neutral network and fuzzy logic approaches, and hybrid ones. Here we are talking about the neutral file STEP, which is taken as input to feature recognition systems. As far as I'm concerned, STL files describe parts in term of meshes, STEP file need not to be meshed, the data in STEP are geometric and topological entities that are exploited to detect manufacturing features. For example, in turned workpieces, a plan bounded by two edge loops that shares it's inner loop with  a cylinder can almost be a shoulder. When features are recognized, they can be used by CAPP systems.

Best regards  

Dear Jaider,

My answer might have been oriented towards my previous 5 years experience working for Innovmetric on the IMInspect software.  We use to reverse engineer objects and parts scanned with lasers and extract feature curves that could be used to build curve network for NURBS surfaces. Sorry if it didn't exactly fit the question.

Bruno

http://www.innovmetric.com/fr

Dear MArtin,

Oh yes, great, i knew that you answer was correct but in a specific field that i ignore. Very interesting is the website, thank you, you let me know new things :). Keep in touch

Faithfully 

Hi Martin and Jaider, thank you for your explanations. 

I'm doing my research in Design to Cost, as a part of the research task-1 is to develop a manufacturing process configurator - aim is to map the manufacturing features automatically to decide the process. Right know I'm plan to try on Visual basics. Do you have any other opinion?