Modeling and Control of Transient Mixing Flow for Direct Ink Write Additive Manufacturing

Abstract

Direct ink write (DIW) additive manufacturing (AM) is a material extrusion process characterized by depositing viscous liquids through fine nozzle tips. The process is used to make three-dimensional objects with a wide range of materials. In this dissertation, a new impeller spiral static mixer (SSM) is designed and tested, and a transient fluid model and control method for DIW of two-part silicone with mixing is developed. An important aspect of the DIW of two-part silicone is in-situ mixing using static mixing, enabling continuous printing mixed silicone. A new SSM design, called the Impeller SSM (ISSM), inspired by centrifugal pump impeller blades and fabricated by AM is presented. The pressure drop reduction and mixing of an ISSM is compared to the standard SSM, both are measured experimentally and validated by computational fluid dynamics analysis. Compared to a standard SSM of the same size, the ISSM demonstrated a pressure drop and power reduction up to 18.2%. Experimental results also show the ability of AM to fabricate the custom ISSM without using costly fabrication techniques. Using an understanding of the in-situ mixing from the ISSM, the transient flow inside a DIW system is characterized using the continuity and momentum equations. New frictional models describing fluid flow for a viscous non-Newtonian fluid through the combined ISSM and tapered nozzle are created for the momentum equation. The continuity and momentum equations describing a DIW system are numerically solved using the CM. The transient response of the DIW output volumetric flowrate in the CM model is validated using a doppler volumetric flowsensor and two pressure sensors. CM is also used to predict the corner swelling of a 90-degree corner DIW tool path with accelerations of 100, 250, 500, 1000, 1500, and 2000 mm/s2. The predicted corner swelling is matched with the actual corner swelling found using image processing of a 90-degree corner. Across the tested accelerations the corner swell ranged from 0.76 to 0.37 mm, matching CM predictions. Demonstrating that the CM can accurately predict the transient response of the DIW volumetric flowrate. With the validated CM model, the transient fluid deposition for DIW is controlled using feedforward error correction control (FECC). FECC combines the trapezoidal motion planning, CM, machine learning, and iterative linear quadratic regulator (iLQR) controller to create new extrusion flow paths to improve the deposition accuracy for DIW. FECC is applied to two tool paths: a 90-degree corner and a U-turn. With FECC, the 2-norm error between the output volumetric flow rate and desired volumetric flow rate of the 90-degree turn is reduced from 0.32 to 0.16 mL/min, while the measured size of the 90-degree corner swell was reduced from 0.63±0.03 mm to 0.48±0.03 mm. For the U-turn, the 2-norm error between the output volumetric flow rate and desired volumetric flow rate is reduced from 0.43 to 0.18 mL/min and the measured width was reduced from 0.98±0.04 mm to 0.82±0.03 mm. The total reduction in the deposition error was 25-40%. The FECC tool paths were used with a test part containing 5000 90-degree turns and 8500 U-turns. With FECC, the test part had significant improvements to reduce bulging at the corners, material build up at the edges of infill, and gaps in the infill. This study demonstrates that FECC can correct errors in DIW deposition and be applied to improve the part quality.