University of Twente Student Theses
Tuning liquefier dynamics : a strategy for accuracy improvement in extrusion based additive manufacturing with PEEK
Holtrup, R. (2019) Tuning liquefier dynamics : a strategy for accuracy improvement in extrusion based additive manufacturing with PEEK.
|
PDF
35MB |
| Abstract: | Extrusion based 3D printing is an additive manufacturing technique in which a material is robotically dispensed through a nozzle or orifice on prescribed locations. Bond 3D printing develops a new system that uses PEEK as the feeding material and has formulated three main challenges within their extrusion based printing process: enhancement of lead time, strength and accuracy (personal communication, 2017). The focus of this study was set on the enhancement of material dispensing accuracy for small single line width features with PEEK. Bond3D states that due to the high performance applications inevitable small features show up in applications such as small tubes in manifolds, porosity in implants and thin walls in lightweight structures (personal communication, 2017). Additionally, the lack of accuracy and discontinuities can cause weaknesses in the final products (Gibson, Rosen, & Stucker, 2009, p. 155). To enhance this accuracy, two approaches were employed: an empirical approach and an analytic model approach. The first approach studies the real-world effects and is well suited to find local optima, while the latter one provides better fundamental support and can often be used to extrapolate results to experimentally unverified testing conditions. For both approaches, a custom-built 3d printer prototype was used to print single wall test geometries with abrupt velocity changes. Accuracy was assessed visually. In the empirical study, individual or combination of process settings were varied, or accuracy enhancement “tricks” were applied (priming, coasting, retracts, wipes), after which the accuracy of the product was determined. However, it resulted that the effects of changing the process variables, such as line height, segment length and setting and coasting length were highly interdependent and setting up the experiments to obtain predictable results proofed to be challenging. Still, some qualitative results were found. Experiments that used retraction of the plunger, decreased travel time of the print head and thereby the time that oozing can take place and experiments which used coasting, led to the highest vertical stop-straightness. Wiping in some cases enhanced the vertical straightness of the printed structures during starts, but increased the risk of smearing, resulting in completely inaccurate prints. Furthermore, in all experiment setups, the line width of deposited strand deviated strongly from the expected line width, which indicated that there were dynamic effects that affected the final accuracy. As such, the system dynamics were analysed through a system identification consisting of two experiment types that characterised the dynamics: a step response and frequency response experiment. They characterise the response of the flow rate to the changes in plunger velocity. The responses were assessed by measurements of the oozed length and the resulting line width. It was found that the compressibility of the feeding material and its thermal expansion were the factors dominating dynamic effects. The dynamics were controlled using a controller making real-time system adjustments. The systems dynamics showed strong non-linear behaviour, while the implemented controller assumed linearity. This complicated extrapolating the accuracy as a function of plunger velocity and controller settings. The strong non-linearity was explained by the fact that the stiffness of the feeding material is a function of the consumed volume. The decreasing fluidic restriction value of the nozzle for increasing the flow rates was explained by shear thinning behaviour of the molten polymer. Thermal oozing was explained by the difference in average melt room temperature that is dependent both on time and temperature. The first two effects were described by analytic expressions and also an expression of the maximum thermally induced oozing volume was given. The Boles model of flow through a converging die using power law viscosity functions that were derived from the Cross viscosity model of PEEK was proven suitable to predict the steady-state pressure drop accurately for two different nozzles for PEEK within the studied flow regime. The model are powerful tools in the design and optimisation of 3D printing nozzles, materials and tuning and development of a numerical model, however the effect on accuracy in real-world printing conditions needs to be proven in a future study as well as the development of the melt room temperature over time and extrusion rate. Keywords Additive Manufacturing, Accuracy, PEEK material extrusion |
| Item Type: | Essay (Master) |
| Clients: | Bond High Performance 3D Technology BV, Enschede, The Netherlands |
| Faculty: | ET: Engineering Technology |
| Programme: | Industrial Design Engineering MSc (66955) |
| Link to this item: | https://purl.utwente.nl/essays/93386 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page