Art-to-part manufacturing

In our lab, we believe it is nearly impossible for a human to create the best possible approaches to machining discrete parts. When it comes to milling a part, there are many decisions that process engineers make. Decisions such as what tools to use, whether they should use ceramic inserts or carbide tools, what coatings to use, should alternative tool geometries be considered, what are the best feeds and speeds for this part, how to achieve maximum productivity while maintaining dynamic stability and etc.

Some of these decisions are shown in the figure below:  

current-approach-to-machining

Out lab has a philosophy that when it comes to machining a discrete part that the part should dictate the decisions related to machining it. The overall objective here is to create an art-to-part approach where we leverage all the research about machining stability and productivity and wear and allow the part to pick tools from the vast tool libraries to produce the best path with respect to weights on cost, speed and quality. We imagine a world where a user only has to upload the part file to an art-to-part software solution that produces the best manufacturing strategy based on the specific assets available and user preferences. 

future-approach-to-machining

Unfortunately our current systems don’t even integrate recent (and not-so-recent) research results. Here is a summary of where we are with respect to tool path planning: 

  • We have the research results and should have the software to allow us to continuously vary the feed rate and depth of cut based on forces and machining stability but we really don’t.
  • These tools have not made their way into the most commonly used software tools we use to machine parts.
  • This is because nearly all software tools are based on two geometric surface modelling kernels that do not lend themselves to this type of modelling or analysis. 

If you take a look at the pyramid below, you’ll see that nearly everything we use in CAD/CAM systems rely’s on just a few geometric modelling kernels kernel

We believe that these geometric modelling kernels dramatically limit how quickly and well we can integrate research results. If we can change the underlying geometric approach that we can integrate research advances much more easily by using, for example, a truly volumetric approach instead of the current surface modelling approach. Click on the menu items to learn more about our voxel-based approach to automating the art-to-part discrete machining process.