Computational models for the extrusion-based additive manufacturing

The Fourth Industrial Revolution (4IR) has introduced new business models focused on consumers and product customization. As a natural consequence, both the quantity of the service provided and the added value have increased. Additive Manufacturing (AM) is one of the nine pillars of 4IR; it enables the production of small batches of customized and lightweight components on demand. In addition, it plays a key role in sustainability, as it offers opportunities to minimize wastes, energy consumption and use eco-friendly materials. In the framework of sustainability, the transportation emissions are lowered by this decentralized and flexible production. Furthermore, a second key pillar of 4IR is Digital Twin (DT), which indicates a virtual simulation of a real-world machine, product or complex system; in general, DT is based on data collected through a complex network of sensors, to better analyze the behavior of real systems. However, computer simulations are gaining increasing attention, because of the possibility to analyze quantities that cannot be measured directly and gain a deep insight of physical processes occurring in AM. On the one hand, analytical methods allow for a closed form solution of a given problem with many assumptions. On the other hand, fully numerical methods describe more complex scenarios, but they can be computationally expensive. An intermediate solution is given by the semi-analytical models. In the present work, different Material Extrusion (MEX) processes have been studied by means of both semi-analytical and fully numerical methods. In the first part of the dissertation the screw-based MEX, based on the processing of pelletized thermoplastics has been reviewed and studied mathematically; for the first time, a complete model aiming at coupling the screw-barrel and deposited layer dynamics has been formulated. The influence of process parameters on important flow characteristics (i.e., mass flow rate, pressure and melting profile) has been explored. In the second part, generalized methods to solve the flow in: i) straight cylindrical ducts and ii) tapered cone geometries, usually found as internal geometrical features in the nozzles used in all MEX processes, have been formulated. The first method has been validated with respect to literature data, while the latter has been compared with real force measurement by means of a silicone MEX custom-made setup. Then, a model for the counterpressure arising under the extrusion head has been developed and validated using the same MEX setup; the composite model (made up of nozzle flow and deposited bead models) showed an accuracy up to 99.7[%] in predicting overall printing force; this paves the way for the application of the proposed methods in other MEX processes modeling, control and optimization.