A simulated dye method for flow visualization with a computational model for blood flow.

A numerical dye method for the visualization of unsteady three-dimensional flow calculations is introduced by coupling the unsteady convection-diffusion equation to the Navier-Stokes equation for mass and momentum. This system of equations is descretized using a finite volume projection-like algorithm with generalized coordinates and overset grids. A powerful pressure prediction method is used to accelerate the convergence of the Pressure Poisson equation. To demonstrate the visualization technique, blood flow through the aortic arch region and the three main arterial branches is computed using various Womersley numbers. In this technique, parcels of fluid are followed in time as a function of the cardiac cycle without having to track individual particles, which in turn aids us to better understand some important aspects of the three-dimensionality of the developing unsteady flow. Using this numerical dye method we analyze the strength of the cross flow during the cardiac cycle, the relationship between the penetration of blood into the aortic branches from its relative position in the ascending aortic region and the effects of the Womersley parameter. This technique can be very useful in the design and development of stents where the topology of the device would require understanding where the blood emanating from the heart ends up at the end of the cardiac cycle. Moreover, this method could be useful in investigating the influence of flow and geometry on the local introduction of medication.

Unsteady and three-dimensional simulation of blood flow in the human aortic arch.

A three-dimensional and pulsatile blood flow in a human aortic arch and its three major branches has been studied numerically for a peak Reynolds number of 2500 and a frequency (or Womersley) parameter of 10. The simulation geometry was derived from the three-dimensional reconstruction of a series of two-dimensional slices obtained in vivo using CAT scan imaging on a human aorta. The numerical simulations were obtained using a projection method, and a finite-volume formulation of the Navier-Stokes equations was used on a system of overset grids. Our results demonstrate that the primary flow velocity is skewed towards the inner aortic wall in the ascending aorta, but this skewness shifts to the outer wall in the descending thoracic aorta. Within the arch branches, the flow velocities were skewed to the distal walls with flow reversal along the proximal walls. Extensive secondary flow motion was observed in the aorta, and the structure of these secondary flows was influenced considerably by the presence of the branches. Within the aorta, wall shear stresses were highly dynamic, but were generally high along the outer wall in the vicinity of the branches and low along the inner wall, particularly in the descending thoracic aorta. Within the branches, the shear stresses were considerably higher along the distal walls than along the proximal walls. Wall pressure was low along the inner aortic wall and high around the branches and along the outer wall in the ascending thoracic aorta. Comparison of our numerical results with the localization of early atherosclerotic lesions broadly suggests preferential development of these lesions in regions of extrema (either maxima or minima) in wall shear stress and pressure.

Crossflow filtration in suspension-feeding fishes.

Rows of comb-like or tufted gill rakers in the oral cavity of suspension-feeding fishes (for example, herring, anchovies and tilapia) have been thought to serve as (1) non-porous barriers that direct particle-laden water to the sticky oral roof, where particles are retained as water exits from the oral cavity, (2) conventional dead-end filters that sieve particles from water exiting between rakers, or (3) sticky filters that retain particles encountered by a hydrosol filtration mechanism. Here we present data from computational fluid dynamics and video endoscopy in suspension-feeding fish indicating that the rakers of three distantly related species function instead as a crossflow filter. Particles are concentrated inside the oral cavity as filtrate exits between the rakers, but particles are not retained on the rakers. Instead, the high-velocity crossflow along the rakers carries particles away from the raker surfaces and transports the particles towards the oesophagus. This crossflow prevents particles from clogging the gaps between the rakers, and solves the mystery of particle transport from the rakers to the oesophagus.

Computational fluid dynamics in the oral cavity of ram suspension-feeding fishes.

We have modeled steady, three-dimensional flow with a no-slip boundary condition in cylindrical and conical oral cavities possessing vertical or slanted branchial slits. These numerical simulations illustrate the transport of food particles toward the esophagus, as well as the velocity profiles of water exiting the oral cavity via the branchial slits. The maximum and average velocities are highest for flow exiting the most posterior branchial slit. The highest volume flow rates also occur in the most posterior slit for the cylindrical simulations, but occur in the most anterior slit for the conical simulations. Along the midline, there is a pronounced bilaterally symmetrical vortex in the posterodorsal region of the cylindrical and conical oral cavities and a second bilaterally symmetrical vortex in the posteroventral region of the cylinder. Particles entrained in the vortices will recirculate in the posterior oral cavity, increasing the probability of encounter with sticky, mucus-covered surfaces such as the oral roof, gill arches, or gill rakers. The posterodorsal vortex could serve to concentrate particles near the entrances of the epibranchial organs. The ventral vortex could be involved in sequestering dense inorganic particles that sink toward the floor of the oral cavity. All vortices are absent in the conical simulation with vertical branchial slits, indicating that the slanted branchial slits between the gill arches are responsible for the formation of the vortex in the conical oral cavity. Experiments using in vivo flow visualization techniques are needed to determine whether ram suspension feeders, pump suspension feeders, and non-suspension-feeding fishes possess vortices in the posterior oral cavity that contribute to particle transport, food particle encounter with sticky surfaces, and inorganic particle rejection.

Computational study of the effect of geometric and flow parameters on the steady flow field at the rabbit aorto-celiac bifurcation.

Arterial hemodynamic forces may play a role in the localization of early atherosclerotic lesions. We have been developing numerical techniques based on overset or "Chimera" type formulations to solve the Navier-Stokes equations in complex geometries simulating arterial bifurcations. This paper presents three-dimensional steady flow computations in a model of the rabbit aorto-celiac bifurcation. The computational methods were validated by comparing the numerical results to previously-obtained flow visualization data. Once validated, the numerical algorithms were used to investigate the sensitivity of the computed flow field and resulting wall shear stress distribution to various geometric and hemodynamic parameters. The results demonstrated that a decrease in the extent of aortic taper downstream of the celiac artery induced looping fluid motion along the lateral walls of the aorta and shifted the peak wall shear stress from downstream of the celiac artery to upstream. Increasing the flow Reynolds number led to a sharp increase in spatial gradients of wall shear stress. The flow field was highly sensitive to the flow division ratio, i.e., the fraction of total flow rate that enters the celiac artery, with larger values of this ratio leading to the occurrence of flow separation along the dorsal wall of the aorta. Finally, skewness of the inlet velocity profile had a profound impact on the wall shear stress distribution near the celiac artery. While not physiological due to the assumption of steady flow, these results provide valuable insight into the fluid physics at geometries simulating arterial bifurcations.

High venous urea concentrations in the opposite arm. A consequence of hemodialysis-induced compartment disequilibrium.

Resistance to urea diffusion among body fluid compartments diminishes the therapeutic effectiveness of hemodialysis. Cell membrane or capillary wall resistance is thought to be responsible for hemodialysis-induced urea disequilibrium. The authors examined the possibility that reduced blood flow might contribute to urea disequilibrium in the arm opposite the blood access site. Blood samples were taken simultaneously from a vein in the arm opposite the access site and from the arterial port after occluding the access graft between the needle sites for 1 min. Venous urea nitrogen levels from the opposite arm averaged 10% higher after 5 min, 26% higher after 60 min, and 36% higher after 120 min of dialysis. A three-compartment model of urea kinetics that includes a blood flow term accurately predicted all measured urea nitrogen concentrations in both arms. These data suggest that the opposite arm often behaves as a compartment with high resistance to urea diffusion. Slow diffusion from this compartment is partially due to reduced blood flow/compartment volume, and results in a delayed fall in venous blood urea nitrogen (BUN).