Applying feedback control to improve 3D printing quality

Abstract

Small-scale 3D-printing has become a common appearance. Today there are many companies specialized in selling and designing small-scale 3D-printers . Applications range from creating custom-made toys to creating spare parts for the 3D-printer itself as well as medical applications such as skull parts, knee replacements and hip replacements. The research and development start-up company HB|3D uses a 6 degrees-of-freedom robotic arm combined with a plasticating single-screw extruder and heated hose to create large-scale thermoplast products. One challenge is to scale up the 3D-printing process while ensuring a high-quality product.

Feedback control looks promising in increasing the dimensional accuracy of large-scale printed objects by influencing the screw speed to minimize flow disruptions. The goal is to create a simulation of a model predictive controller and show that an output flow trajectory can be tracked within 1 percent accuracy.

The flow response is modeled using a commercially available computational fluid dynamics software package (ANSYS FLUENT Academic Research version 18.1). This computationally intensive model is simplified using system identification techniques; resulting in a wiener model.

Finally, a simulation of using a model predictive controller to track an output flow reference signal is then performed.