Research · Compiling intent to fabrication

Process modeling and control

Functionally graded, intent-driven design needs calibrated process models and controls, not only material models, so a designer can ask for a density, hardness, or stiffness and get the process settings that produce it.

The key idea. Functionally graded, intent-driven design requires calibration of both material models, which relate a material mix to its properties, and process models and controls, which relate printing parameters such as temperature, speed, flow, and nozzle height to the material that is actually deposited. A foaming filament expands more when printed hotter, and a nozzle raised above the part lays down coils instead of lines. With these effects modeled and controlled, a designer can ask for a density, hardness, or stiffness, and the compiler chooses the process settings that produce it.

Temperature changes the material. The same foaming filament extruded at two nozzle temperatures expands by a different amount at each.

Foaming filaments

Foaming filaments, such as varioShore TPU and LW-PLA, contain additives that expand with heat, so nozzle temperature controls how much each extruded line foams. I characterized how wall expansion, density, and Shore hardness change with temperature, and how much the flow must be reduced to keep line width constant. Inverting these models turns a requested density or hardness field into temperature and flow fields that the gradient-informed slicer and the slicer-project compiler can print.

Six panels of plots of wall thickness, flow multiplier, density, and hardness against nozzle temperature, and two SEM images of printed TPU
Calibrated models. (a) Wall expansion versus temperature for varioShore TPU and LW-PLA. (b) Flow compensation computed from that expansion. (c) Compensated walls stay near the 0.45 mm target. (d) Density and (e) Shore A hardness versus temperature. (f) SEM images of unfoamed and foamed TPU.
A blue printed bar above a plot where measured Shore hardness follows the designed linear gradient
Hardness inverse design. A 65A to 85A design, FFF-printed in foaming TPU. Measured hardness follows the design with a mean absolute error of 0.5 Shore A.
A blue printed bar balanced on a stand, with marks showing its geometric center and its center of mass 18 millimeters away along a 200 millimeter length
Density inverse design. A 200 mm bar, FFF-printed in foaming TPU, designed to place its center of mass (b) 18 mm from its geometric center (a). The printed bar balances at the designed point.

Viscous thread instability

When a nozzle is raised above the part and extrudes faster than it moves, the molten thread buckles into coils. Lipton and Lipson showed that this viscous thread instability (VTI) can print foams whose stiffness can be tuned over a wide range [1]. I adapted the technique as a spatial field in our slicer: extrusion rate, print speed, and nozzle offset are planned region by region from an OpenVCAD design, and I characterized how coil density responds to them. Coil density, and with it stiffness, can then be graded through a part.

Diagram comparing normal printing with a nozzle at layer height and viscous thread instability with an extra offset, where the extruded thread bends and coils
How VTI works. Raising the nozzle and extruding faster than it moves makes the thread coil. Coil density is set by extrusion rate, print speed, and the extra nozzle offset. Adapted from Lipton and Lipson (2016).
A teal printed coil foam with looser coils on the left labeled softer and tighter coils on the right labeled stiffer
Variable-stiffness foam. Looser coils are softer and tighter coils are stiffer.
Printing coils. VTI printing up close.
Coil density as a design field. The printer switches from tighter to looser coiling where the OpenVCAD design asks for it.
A smoothly graded insole. Coil density follows a smooth gradient designed in OpenVCAD.

[1] J. I. Lipton and H. Lipson, "3D Printing Variable Stiffness Foams Using Viscous Thread Instability," Scientific Reports 6, 29996 (2016). doi:10.1038/srep29996

Publications

  1. Functionally Grading the Slicing Process by Compiling Design Intent into Slicer Projects
    Functionally Grading the Slicing Process by Compiling Design Intent into Slicer Projects
    Charles Wade, Devon Beck, and Robert MacCurdy
    SCFProceedings of the 11th ACM Symposium on Computational Fabrication (SCF ’26), 2026
  2. Implicit Toolpath Generation for Functionally Graded Additive Manufacturing via Gradient-Informed Slicing
    Implicit Toolpath Generation for Functionally Graded Additive Manufacturing via Gradient-Informed Slicing
    Charles Wade, Devon Beck, and Robert MacCurdy
    Additive Manufacturing 112, 104963, 2025