The effect of waveform shape and extension of negative flowrate are studied at two Reynolds numbers (Re = 80, Re = 200) on stented channels by looking at the kinematics and dynamics of the flow. The waveforms are reconstructed by Fourier decomposition of a waveform corresponding to left anterior descending (LAD) of coronary artery. The stents are modeled by an immersed boundary method. Two stent geometries are created which are the idealizations of two clinical ones. The first geometry is an idealization of XIENCE V stent which is called Lambda stent and the second one is the idealization of Endeavor stent and is called X stent. The former has larger inter-strut spacing and smaller thickness than the latter one. The use of immersed boundary method for creating the stents is validated, and the spatial resolution requirements are determined. The shape of the waveform is changed by systematically filtering out the higher modes of Fourier decomposition and the negative flowrate extension is reduced by shifting the waveform to the positive flowrate. The presence of the stent causes vortical structures to be created between stent struts. These vortices are migrating to the centre of the channel and disappear. It is observed that the confined geometric feature of X stent and its larger thickness, leads to larger areas of flow recirculation which causes smaller wall-shear-stress parameters with respect to Lambda stent and more deviation of the flow from a healthy vessel. The importance of the convective terms of the Navier-Stokes equations was studied at the two Reynolds numbers for both stents. It is observed that at high Reynolds number (Re = 200), the convective terms play significant role throughout the waveform cycle while at low Reynolds number (Re = 80), the effect of convective terms become negligible during negative flowrate. Moreover the convective terms become more significant for flow in a channel with Lambda stent than X stent due to the specific shape and size of the stents. The kinematics of the flow corresponds to the study of vortex timing. It was found that this timing is mainly affected by the waveform and Reynolds number rather than the stent geometry. The time at which vortex creation occurs is coincident with the time at which wall shear stress changes its sign in an unstented channel. Therefore the analytical solution of unsteady channel flow can be used as a tool for analysing the kinematics of the flow, / Thesis (Master, Mechanical and Materials Engineering) -- Queen's University, 2012-06-07 17:41:52.501

An investigation of preconditioning techniques is presented for a Newton--Krylov algorithm that is used for the computation of steady, compressible, high Reynolds number flows about airfoils. A second-order centred-difference method is used to discretize the compressible Navier--Stokes (NS) equations that govern the fluid flow. The one-equation Spalart--Allmaras turbulence model is used. The discretized equations are solved using Newton's method and the generalized minimal residual (GMRES) Krylov subspace method is used to approximately solve the linear system. These preconditioning techniques are first applied to the solution of the discretized steady convection-diffusion equation. Various orderings, iterative block incomplete LU (BILU) preconditioning and multigrid preconditioning are explored. The baseline preconditioner is a BILU factorization of a lower-order discretization of the system matrix in the Newton linearization. An ordering based on the minimum discarded fill (MDF) ordering is developed and compared to the widely popular reverse Cuthill--McKee ordering. An evolutionary algorithm is used to investigate and enhance this ordering. For the convection-diffusion equation, the MDF-based ordering performs well and RCM is superior for the NS equations. Experiments for inviscid, laminar and turbulent cases are presented to show the effectiveness of iterative BILU preconditioning in terms of reducing the number of GMRES iterations, and hence the memory requirements of the Newton--Krylov algorithm. Multigrid preconditioning also reduces the number of GMRES iterations. The framework for the iterative BILU and BILU-smoothed multigrid preconditioning algorithms is presented in detail.

An investigation of preconditioning techniques is presented for a Newton--Krylov algorithm that is used for the computation of steady, compressible, high Reynolds number flows about airfoils. A second-order centred-difference method is used to discretize the compressible Navier--Stokes (NS) equations that govern the fluid flow. The one-equation Spalart--Allmaras turbulence model is used. The discretized equations are solved using Newton's method and the generalized minimal residual (GMRES) Krylov subspace method is used to approximately solve the linear system. These preconditioning techniques are first applied to the solution of the discretized steady convection-diffusion equation. Various orderings, iterative block incomplete LU (BILU) preconditioning and multigrid preconditioning are explored. The baseline preconditioner is a BILU factorization of a lower-order discretization of the system matrix in the Newton linearization. An ordering based on the minimum discarded fill (MDF) ordering is developed and compared to the widely popular reverse Cuthill--McKee ordering. An evolutionary algorithm is used to investigate and enhance this ordering. For the convection-diffusion equation, the MDF-based ordering performs well and RCM is superior for the NS equations. Experiments for inviscid, laminar and turbulent cases are presented to show the effectiveness of iterative BILU preconditioning in terms of reducing the number of GMRES iterations, and hence the memory requirements of the Newton--Krylov algorithm. Multigrid preconditioning also reduces the number of GMRES iterations. The framework for the iterative BILU and BILU-smoothed multigrid preconditioning algorithms is presented in detail.

Magnetic resonance imaging is often used as a source for reconstructing vascular anatomy for the purpose of computational fluid dynamics (CFD) analysis. We recently observed large discrepancies in such “image-based” CFD models of the normal common carotid artery (CCA) derived from contrast enhanced MR angiography (CEMRA). A novel quantitative comparison of velocity profile shape of N=20 cases revealed an average 25% overestimation of velocities by CFD, attributed to a corresponding underestimation of lumen area in the CEMRA-derived geometries. We hypothesized that this was due to blurring of edges in the images caused by dilution of contrast agent during the relatively long elliptic centric CEMRA acquisitions, and confirmed this with MRI simulations. CFD simulations incorporating realistic inlet velocity profiles and non-Newtonian rheology had a negligible effect on velocity profile skewing, suggesting a role for other sources of error or modeling assumptions.

Magnetic resonance imaging is often used as a source for reconstructing vascular anatomy for the purpose of computational fluid dynamics (CFD) analysis. We recently observed large discrepancies in such “image-based” CFD models of the normal common carotid artery (CCA) derived from contrast enhanced MR angiography (CEMRA). A novel quantitative comparison of velocity profile shape of N=20 cases revealed an average 25% overestimation of velocities by CFD, attributed to a corresponding underestimation of lumen area in the CEMRA-derived geometries. We hypothesized that this was due to blurring of edges in the images caused by dilution of contrast agent during the relatively long elliptic centric CEMRA acquisitions, and confirmed this with MRI simulations. CFD simulations incorporating realistic inlet velocity profiles and non-Newtonian rheology had a negligible effect on velocity profile skewing, suggesting a role for other sources of error or modeling assumptions.

The Circle of Willis (CoW) is a ring-like arterial structure forming the major anastomotic connection between arterial supply systems in the brain, and is responsible for the distribution of oxygenated blood throughout the cerebral mass. Among the general population, only approximately 50% have a complete CoW, where absent or hypoplastic vessels are common among a multitude of possible anatomical variations, reducing the degree to which blood may be rerouted. While an individual with one of these variations may under normal circumstances suffer no ill effects, there are certain pathological conditions which can present a risk to the person's health and increase the possibility of suffering an ischaemic stroke when compounded with an anatomical variation. This body of work presents techniques for generating 3D models of the cerebral vasculature using magnetic resonance imaging (MRI) and performing computational fluid dynamics (CFD) simulations in order to simulate the flow patterns throughout a circle of Willis. Incorporated with the simulations is a mathematical model of the cerebral autoregulation mechanism, simulating the ability of the smaller arteries and arterioles in the brain to either constrict or dilate in response to alterations in cerebral blood flow, thereby altering the cerebrovascular resistance of each major brain territory and regulating the amount of blood flow within a physiological range of cerebral perfusion pressure. The CFD simulations have the ability to predict the amount of collateral flow rerouted via the communicating arteries in response to a stenosis or occlusion, and the major objective of this study has been the investigation of how anatomical variations of the circle of Willis affect the capacity to provide this collateral flow. Initial work began with the development of three idealized models of common anatomical variations, created using computer aided design software (CAD) and based on the results of MRI scans. The research then shifted to developing a technique whereby patient specific models of the circle of Willis could be directly segmented from the MRI data. As a result of this shift, an interactive GUI-based tool was developed for the processing of the MRI datasets, allowing for rapid data enhancement and creation of a surface topology representing the arterial wall of the circle of Willis, suitable for a CFD simulation. The results of both sets of simulations illustrate that there exist a number of variables associated with a patients circle of Willis geometry, such as cerebral blood flow and combinations and degrees of stenosis, implying that the initial goal of drawing generalized conclusions was perhaps flawed. Instead, a crucial outcome of this body of work is that the future research should be directed toward extending the physiological complexity of both the geometry and the autoregulation model, with the intention of a patient specific application rather than producing large datasets with which to make broad generalizations.

Atherogenesis is the leading cause of death in the developed world, and is putting considerable monetary pressure on health systems the world over. Although the risk factors are well understood, unfortunately, the initiation and development of this disease still remains relatively poorly understood, but it is becoming increasingly identifiable as a dysfunction of the endothelial cells that line the walls of arteries. The prevailing haemodynamic environment plays an important role in the focal nature of atherosclerosis to very specific regions of the human vasculature. Disturbed haemodynamics lead to very low wall shear stress, and inhibit the transport of important blood borne chemicals. The present study models, both computationally and mathematically, the transport and hydrolysis of important blood borne adneosine nucleotides in physiologically relevant arterial geometries. In depth analysis into the factors that affect the transport of these low diffusion coefficient species is undertaken. A mathematical model of the complex underlying endothelial cell dynamics is utilised to model production of key intracellular molecules that have been implicated into the complex initiation processes of atherosclerosis; hence regions of the vasculature can be identified as being 'hot spots' for atherogenesis. This model is linked into CFD software allowing for the assessment of how 3D low yields and mass transfer affect the underlying cell signalling. Three studies are undertaken to further understand nucleotide variations at the endothelium and to understand factors involved in determining the underlying cell dynamics. The major focus of the first two studies is geometric variations. This is primarily due to the plethora of evidence implicating the geometry of the human vasculature, hence the haemodynamics, as an influential factor in atherosclerosis initiation. The final model looks at a physiologically realistic geometry to provide a more realistic reproduction of the in vivo environment.

The use of washable baghouses for fines collection in milk powder plants has been investigated. The main aim of this study was to increase the fundamental understanding of both operation and design of washable baghouses for application in milk powder plants. This work has focussed on the industrial scale. Industrial plant operating data has been collected, plant designs compared and analyses conducted on powder produced at the industrial scale. The amount of powder that becomes fines, the small size fraction of powder entrained in spray dryer outlet air streams, has been shown to be significantly greater than the traditionally vague estimate of 10 % to 20 %. The ratio of fines flows to total powder flows ranged from 49 ± 8 % to 86 ± 2 % depending on the powder type and plant operating conditions. A simple yet reliable method was developed to quantify fines flows based on measured powder size distributions of samples taken from around the plant. These estimates were supported by readings from an online optical scintillation instrument, which was shown to be capable of measuring fines flows at concentrations approximately four times the supposed maximum stated by the instrument’s manufacturer. Observations in another part of this work supported previous Fonterra observations showing that the amount of bulk fat in skim milk powder (SMP) has a large influence on the baghouse differential pressure. Fines flows measured by the optical scintillation instrument and analysis of other plant operating data showed that a change in bulk fat in SMP does not appear to cause any change in fines concentration. Observations of the surface of SMP by scanning electron microscopy, and electron spectroscopy for chemical analysis, both showed that fat is over-represented on the surface of the particles, and that only small increases in the bulk fat content are required to cause large increases in the surface fat coverage. It is hypothesised that increased fat on the surface of particles increases the clumping of SMP before deposition on the bags. Consequently, the powder forms more porous cakes and is less likely to penetrate into the interior of the filter bags, which also makes it easier to pulse clean powder from the filter bags. Therefore, the baghouse differential pressure is reduced. The design of pulse-jet baghouses from the literature was found to rely heavily on the authors past experience and approach, giving rise to large variation in recommended values of the key design parameters. A procedure for determining the optimal combination of these parameters was developed. This procedure showed that the main Fonterra washable baghouses are far from optimal because of their high air-to-cloth ratios, long bags and high elutriation and annular velocities. This procedure also showed that the Fonterra vibrating fluid bed washable baghouses are much closer to the optimum, which is the probable reason these washable baghouses have had almost no operational issues. Observations of the movement of the bags from below showed significant movement for bags near the inlet of the baghouse, indicating that this was the probable cause of the high bag damage in this zone. It is suggested that increasing the outer gap (distance between the baghouse wall and the bag on the edge of the bag bundle) be investigated further in an attempt to slow the annular velocity around the edge of the bag bundle and reduce bag movement. It is also recommended that stainless steel inspection hatches installed in the wall of a baghouse for this research, be included in all current and future washable baghouses because use of these hatches reduced the overall clean-in-place turn around time by 20 %. Computational fluid dynamics simulations of the air flow patterns within the Clandeboye Dryer 2 chamber were carried out using a commercial code CFX10.0. These simulations are possibly the first to include the influence of a spray dryer’s internal fluid bed airflow on the flow patterns within a spray dryer. As expected, the simulations showed the main air jet oscillated and precessed about the central axis with no apparent distinct frequency. In turn, the recirculation zones between the main jet and the chamber walls fluctuated in size. Different fluid bed flows within the industrial range had only a local influence on the air only flow field by reducing the length of the main jet. A different outlet boundary condition (including a flow resistance representing the baghouse) also appeared to have little influence on the overall flow field. Good agreement was found between the movements of the main jet via simulations and from telltale tufts installed in the plant dryer. This supported other indications that the simulations were an accurate representation of the actual flows. It was concluded that this project achieved its main aim of improving the fundamental understanding of washable baghouse operation and design, especially for application in milk powder plants. Also this project, as well as a change in production schedules, has helped to reduce downtime associated with the washable baghouses in the Fonterra Clandeboye Dryer 2 plant by an estimated 50 hours per annum.

No description available.

No description available.