Optimising 3D bioprinter nozzle design through in silico modelling

3D bioprinting is a flourishing technology, in addressing tissue-engineering construct manufacturing related challenges. However, the realization of the potential of this developing technique is hindered by multiple technical hurdles which restrict the printability and cell survivability. In addition, commonly employed experimental trial and error approaches are time consuming and resource intensive, especially when a new material needs to be printed. To address these issues, computational or in silico modelling of specific parts of the system can be a viable option to optimize the relevant design as well as the printing and material parameters.
Shear stress is proven to be a crucial factor for cell survivability in extrusion-based 3D bioprinting. Here, we sought to provide the appropriate choice for nozzle design in order to minimize the maximum shear stress occurring in the nozzle during bioprinting. We have modelled three widely used natural and synthetic shear-thinning hydrogel materials, namely alginate, alginate-gelatine and pluronic F127 (PF127) in two different nozzle configurations (conical and blunted). The model started with varying all the design parameters in the range relevant to practical application, using space-filling latin hypercube sampling (LHS) and running computational fluid dynamics (CFD) models to obtain flow profile and shear stress responses for each design. The outcomes from 1200 different in silico tested combinations are fitted into a machine learning method, known as Gaussian process to obtain the response of individual design parameters on the maximum shear stress generated in the hydrogel. It is found that the lower nozzle length and nozzle exit radius are the most important parameters for blunted nozzle designs whereas for conical designs middle and exit radii of the nozzle are crucial factors influencing shear stress. In addition, shear-thinning material properties were also shown to have important effects. In summary, we demonstrate the efficacy of CFD and ML based in silico modelling as a feasible pathway to overcome costly experimental trial and errors. This will help in optimising printing parameters, quantification and cost reduction for the development of new bio-printable materials.