Creating custom 3D printing material colors using optical modeling of waste plastic

Distributed recycling and additive manufacturing (DRAM) offer a unique promise for obtaining a circular economy. To maintain or even enhance the value of common 3D printing feedstocks like polylactic acid (PLA) waste an approach to further incentivize prosumers to use recycled feedstocks is to provide something the market currently does not—custom filament colors. To enable prosumers to create custom colors from their own recycled 3D printing waste this article presents a new open-source software named SpecOptiBlend. Specifically, this study introduces a novel method for customizing color filaments by recycling waste 3D printing samples, thereby enhancing the capabilities of color 3D printing. Traditional 3D printing is limited by a narrow range of filament colors, and even multi-color printing heads can utilize only a limited number of colored filaments among the available options. The new approach here repurposes discarded prototypes and unused samples back into the printing cycle with desired colors, allowing for a broader spectrum of colors and gradients. This enables engineers and designers to create more intricate and functionally graded materials. To do this, waste plastics are quantified after processing for spectral reflectance, then Kubelka-Munk theory provides the initial estimate for color mixing. Three discrete optimization techniques are applied: Nelder-Mead, Limited-memory BFGS with bounds, and Sequential Least Squares Quadratic Programming. To determine the optimal method, assessment criteria include the application of root mean square (RMS) and the color difference (ΔE CIE-2000). Three case studies were conducted, and the Nelder-Mead method was found to provide an optimal balance between the precision of color differences and the RMS, essential for producing high-quality colors. This research has provided a free tool that will now enable prosumers to convert their plastic waste into specific custom colors to enable DRAM.