1938 E Harney Street, Laramie, WY 82072
1-307-314-2891   | 

  • linked-in
  • twitter
  • youtube
  • Jeff Selden

Intelligent Slicing – Automatic Selection of 3D Print Parameters for Structural Performance

Choosing good print parameters is a requirement for successful 3D printing. Adding structural performance expectations can take this selection process from challenging to seemingly impossible. The state-of-the-art in slicers simply lacks the necessary capabilities; namely, physical simulation of the part and a recommendation system to help us find the right print parameters.

Let’s consider an example. Here is a bicycle pedal fixture that we would like to 3D print. We use the phrase “valid print parameters” to describe those that result in a part that can perform the necessary task. So, how do we find valid print parameters for the pedal fixture?

A standard way to discover usable print parameters is the build/break cycle. We print the part with a given set of parameters and then test the part. If it breaks, we tweak some of the parameters and print a new part. This part is then tested. If it breaks, we repeat the process, and so on. It is not unusual for such a cycle to contain 50 iterations until valid print parameters are discovered. Experience with 3D printing can shorten this process, but such expertise is tribal and volatile.

Simulation software exists precisely to provide solutions to such problems without the need to construct the physical models. CNC Kitchen has a couple of very nice videos (here and here) describing a workflow that incorporates the use of physics simulation software to explore structural performance. Unfortunately, the software is external to the slicer and far from a “one-click” solution, requiring a fair amount of knowledge about the use of simulation software. Moreover, as mentioned in the video, the software cannot be used to accurately model a 3D printed part, specifically, the internal structures and their corresponding anisotropy (cf. previous post). Leveraging these internal structures (to save time and material, for example) is a huge advantage in 3D printing, but only if it can be done without sacrificing the structural performance of the part.

These observations are not meant to take away from the workflows described in the videos. Instead, they are articulating an opportunity for slicer software to grow to meet the needs of those who want structural performance for their 3D printed parts in a convenient, fast, and easy-to-use package.

Enter Intelligent Slicing

Teton Simulation is developing a super-fast physics simulation tool embedded in the slicer to validate structural performance of the as-printed part and, if desired, optimize print parameters to meet the requirements set by the user.

Let’s focus on infill density for the moment. Imagine that a part is required to have a factor of safety equal to 2. We can think of this as requiring the part to be twice as strong as it needs to be in operation. What infill density should we use to obtain the desired part strength?

Suppose that we would like to see if an infill density of 20% would suffice. This value would result in relatively faster print times and lower material usage, but that’s useless if the part doesn’t perform. The validation capabilities in Intelligent Slicing allow us to leverage simulation inside the slicer to test the part virtually before having to print anything. If the results are positive, we can then print the part for testing with a measure of confidence that it will perform as desired. If the results are negative, we could change the infill density, validating performance along the way, until a suitable infill density is found.

While this is certainly much faster than building and breaking, only a single print parameter is being tuned in this case. Exploration of the larger space of possibilities when we include things like number of walls, number of top/bottom layers, layer height, layer width, material, etc., is truly daunting.

This is where the optimization capabilities of Intelligent Slicing completely change the game. Instead of manually tuning the print parameters, we simply enter the requirements for the part and then click a button. Teton’s software intelligently searches the space of possibilities to find valid print parameters that minimize time and material.

Let’s consider this approach for the pedal fixture. After setting the requirement for a factor of safety to 2 and specifying how the part will be loaded and anchored, Teton’s software can be used to validate a user-defined set of print parameters or search for optimal choices for valid print parameters. As before, we focus on changing only the infill of the model, but this time, we give the task to the optimization software. In addition to tuning the infill density, our software will also test local variations in infill density using modifier meshes. Such spatially-varying infill properties can be used to great effect for structural performance, as they allow us to target inherently weaker areas of the part.

Here are some typical results.

Model of the pedal fixture in the Cura slicer
Angled view of sliced, optimized part: The modifier mesh can be seen in the middle of the part. It is surrounded by its own walls, which also contribute to part strength.
Top view of the sliced, optimized part: The infill density inside the modifier mesh (45%) is noticeably higher than in the rest of the part (20%); a result of the localization of material to improve structural performance.

The Intelligent Slicing software has added a modifier mesh in an appropriate position to strengthen the part. The result is a set of valid print parameters with significantly less time and material usage than changing the global infill density to meet the requirements - which would be a 95% infill density (or... basically solid). 95% infill has a build time of 6:54 and uses 82 grams of material. In contrast, the print with spatially varying infill has a build time of 4:09 and uses 39 grams of material.

Concluding Remarks

Additive manufacturing is expanding at an increasing rate. It is finding its way into all sorts of markets, generating new demands in terms of materials, machines, and software. With that expansion comes the accompanying realization that a new breed of simulation tool is needed. One that integrates seamlessly with the workflow inside the slicer. One that allows users to validate the structural performance of the as-printed part during the design process. One that explores the space of print parameters for the user and offers options that meet the requirements while also minimizing time and material usage. This is Intelligent Slicing.