Abstract:
The small intestine is the primary site of enzymatic digestion and nutrient absorption in humans. Intestinal contractions facilitate digesta mixing, transport and contact with the absorptive surfaces. In silico models of intestinal motility such as computational fluid dynamics (CFD) models provide a better understanding of the fluid flow patterns found within the duodenum compared to existing in vitro simulator studies and in vivo studies. Existing CFD models of intestinal motility have used idealised contraction patterns and/or simplified geometries. This thesis develops an anatomically realistic model of flow and mixing in the first segment of the small intestine - the duodenum. For validation, an experimental study of peristaltic flow in a C-shaped compliant tube representing the duodenum was initially conducted. A bench-top model comprising of a silicone tube filled with a glycerol-water mixture deformed by a rotating roller was created. Particle image velocimetry (PIV) was used to image flow patterns for deformations approximating conditions in the duodenum. Reversed flow was present underneath the roller with fluid moving opposite to the direction of the peristaltic wave propagation. Deformations of the tube were imaged and used to construct a CFD model of flow with moving boundaries. The PIV and CFD vorticity and velocity fields were qualitatively similar. Using the same numerical techniques from the validation study, an anatomically realistic model was constructed. The duodenum geometry was obtained from the visible human male dataset, and the deformation patterns were derived from electrophysiological simulations of slow wave propagation. Parameters including the amplitude of contraction (10-50% reduction of radius), rheology of the digesta (Newtonian vs non-Newtonian power-law fluid) and contraction patterns (retrograde and colliding waves) were altered in order to study their effects on mixing. Increases in contraction amplitude resulted in increased transport and mixing of digesta, whilst lower levels of mixing was observed with a more viscous non-Newtonian digesta. Retrograde contractions travelling from the jejunum towards the inlet had a negligible impact on the flow and mixing, whilst colliding wave-fronts significantly increased (up to 2.6 times) the amount of intestinal mixing relative to the normal slow wave patterns. Statistical shape analysis was conducted on duodenum geometries to assess the anatomical variations in duodenum shape within a population. CFD simulations were conducted on four selected geometries; two extreme geometries and the geometry most similar to the mean, as well as the mean geometry, to assess the differences in flow and mixing between individuals. No differences in the overall flow patterns were observed between the different geometries. Two times the mean mixing was observed in the most extreme geometry. The computational frameworks developed in this thesis provide new tools for understanding the mixing and nutrient absorption patterns under both normal and diseased conditions.