This thesis deals with numerical simulation of turbulentflows in geometrically simple cases. Both plane and cylindricalgeometries are used. The simplicity of the geometry allows theuse of spectral methods which yield a very high accuracy usingrelatively few grid points. A spectral method for planegeometries is implemented on a parallel computer. Thetransitional Reynolds number for plane Couette flow is verifiedto be about 360, in accordance with earlier findings. TurbulentCouette flow at twice the transitional Reynolds number isstudied and the findings of large scale structures in earlierstudies of Couette flow are substantiated. These largestructures are shown to be of limited extent and give anintegral length scale of six half channel heights, or abouteight times larger than in pressure-driven channel flow.Despite this, they contain only about 10 \% of the turbulentenergy. This is demonstrated by applying a very smallstabilising rotation, which almost eliminates the largestructures. A comparison of the Reynolds stress budget is madewith a boundary layer flow, and it is shown that the near-wallvalues in Couette flow are comparable with high-Reynolds numberboundary layer flow. A new spectrally accurate algorithm isdeveloped and implemented for cylindrical geometries andverified by studying the evolution of eigenmodes for both pipeflow and annular pipe flow. This algorithm is a generalisationof the algorithm used in the plane channel geometry. It usesFourier transforms in two homogeneous directions and Chebyshevpolynomials in the third, wall-normal, direction. TheNavier--Stokes equations are solved with a velocity-vorticityformulation, thereby avoiding the difficulty of solving for thepressure. The time advancement scheme used is a mixedimplicit/explicit second order scheme. The coupling between twovelocity components, arising from the cylindrical coordinates,is treated by introducing two new components and solving forthem, instead of the original velocity components. TheChebyshev integration method and the Chebyshev tau method isboth implemented and compared for the pipe flow case.

As the demand of oil and gas is increasing, using the existing reservoirs more efficiently as well as searching for new reservoirs is mandatory. Most undiscovered reservoirs are in deep or ultra-deep offshore locations, where drilling to such targets are very difficult with the available fluid circulation technology, since there exists a significant frictional pressure loss due to extreme long wellbores. In order to reduce the frictional pressure losses inside the drillstring, frictional drag reducers are used. Frictional drag reducers are mostly high molecular weight linear polymer molecules and can be used with water or hydrocarbon based solvents. The system used in this study is Baroid EZ-Mud water solutions. Baroid EZ-Mud is a liquid polymer emulsion containing partially hydrolyzed polyacrylamide / polyacrylate (PHPA) co-polymer. This study aims to observe the performance of EZ-Mud as a frictional drag reducer. For this purpose, a flow loop that consisted of a circular pipe where the frictional pressure losses can be observed under various flow rates and concentrations is developed. Pipe flow experiments were performed using water-based mud generated using different concentrations of Baroid EZ-Mud at different flow rates. Differential pressure values were recorded for each run. Rheological properties of each mud sample were determined using Fann (Couette) viscometer in order to determine the theoretical frictional pressure losses. Theoretical and measured frictional pressure losses were compared. Results showed that, as the concentration of EZ-Mud was increased, considerable frictional drag reduction as high as 60% was observed. Based on the experimental data obtained from the flow loop using EZ-Mud with different concentrations, a friction factor correlation as a function of Reynolds Number and EZ-Mud concentration is developed. The proposed correlation performance was also compared with the existing correlations from the literature. It has been observed that, frictional pressure losses using the developed friction factor could be estimated within an error range of maximum 15 %, whereas, the existing models could not predict frictional pressure losses as accurate as the proposed model.

<p>This thesis deals with numerical simulation of turbulentflows in geometrically simple cases. Both plane and cylindricalgeometries are used. The simplicity of the geometry allows theuse of spectral methods which yield a very high accuracy usingrelatively few grid points. A spectral method for planegeometries is implemented on a parallel computer. Thetransitional Reynolds number for plane Couette flow is verifiedto be about 360, in accordance with earlier findings. TurbulentCouette flow at twice the transitional Reynolds number isstudied and the findings of large scale structures in earlierstudies of Couette flow are substantiated. These largestructures are shown to be of limited extent and give anintegral length scale of six half channel heights, or abouteight times larger than in pressure-driven channel flow.Despite this, they contain only about 10 \% of the turbulentenergy. This is demonstrated by applying a very smallstabilising rotation, which almost eliminates the largestructures. A comparison of the Reynolds stress budget is madewith a boundary layer flow, and it is shown that the near-wallvalues in Couette flow are comparable with high-Reynolds numberboundary layer flow. A new spectrally accurate algorithm isdeveloped and implemented for cylindrical geometries andverified by studying the evolution of eigenmodes for both pipeflow and annular pipe flow. This algorithm is a generalisationof the algorithm used in the plane channel geometry. It usesFourier transforms in two homogeneous directions and Chebyshevpolynomials in the third, wall-normal, direction. TheNavier--Stokes equations are solved with a velocity-vorticityformulation, thereby avoiding the difficulty of solving for thepressure. The time advancement scheme used is a mixedimplicit/explicit second order scheme. The coupling between twovelocity components, arising from the cylindrical coordinates,is treated by introducing two new components and solving forthem, instead of the original velocity components. TheChebyshev integration method and the Chebyshev tau method isboth implemented and compared for the pipe flow case.</p>

<p>Non-pulsatile and pulsatile flow in bent pipes and radial turbine has been assessed with numerical simulations. The flow field in a single bent pipe has been computed with different turbulence modelling approaches. A comparison with measured data shows that Implicit Large Eddy Simulation (ILES) gives the best agreement in terms of mean flow quantities. All computations with the different turbulence models qualitatively capture the so called Dean vortices. The Dean vortices are a pair of counter-rotating vortices that are created in the bend, due to inertial effects in combination with a radial pressure gradient. The pulsatile flow in a double bent pipe has also been considered. In the first bend, the Dean vortices are formed and in the second bend a swirling motion is created, which will together with the Dean vortices create a complex flow field downstream of the second bend. The strength of these structures will vary with the amplitude of the axial flow. For pulsatile flow, a phase shift between the velocity and the pressure occurs and the phase shift is not constant during the pulse depending on the balance between the different terms in the Navier- Stokes equations.</p><p>The performance of a radial turbocharger turbine working under both non-pulsatile and pulsatile flow conditions has also been investigated by using ILES. To assess the effect of pulsatile inflow conditions on the turbine performance, three different cases have been considered with different frequencies and amplitude of the mass flow pulse and different rotational speeds of the turbine wheel. The results show that the turbine cannot be treated as being quasi-stationary; for example, the shaft power varies with varying frequency of the pulses for the same amplitude of mass flow. The pulsatile flow also implies that the incidence angle of the flow into the turbine wheel varies during the pulse. For the worst case, the relative incidence angle varies from approximately −80° to +60°. A phase shift between the pressure and the mass flow at the inlet and the shaft torque also occurs. This phase shift increases with increasing frequency, which affects the accuracy of the results from 1-D models based on turbine maps measured under non-pulsatile conditions.</p><p>For a turbocharger working under internal combustion engine conditions, the flow into the turbine is pulsatile and there are also unsteady secondary flow components, depending on the geometry of the exhaust manifold situated upstream of the turbine. Therefore, the effects of different perturbations at the inflow conditions on the turbine performance have been assessed. For the different cases both turbulent fluctuations and different secondary flow structures are added to the inlet velocity. The results show that a non-disturbed inlet flow gives the best performance, while an inflow condition with a certain large scale eddy in combination with turbulence has the largest negative effect on the shaft power output.</p>

<p>1D engine simulations of turbocharged engines are difficult to <!-- @page { size: 21cm 29.7cm; margin: 2cm } P { margin-bottom: 0.21cm } --></p><p>Techniques that can increase the SI- engine efficiency while keeping the emissions very low is to reduce the engine displacement volume combined with a charging system. Advanced systems are needed for an effective boosting of the engine and today 1D engine simulation tools are often used for their optimization.</p><p>This thesis concerns 1D engine simulation of a turbocharged SI engine and the introduction of CFD computations on components as a way to assess inaccuracies in the 1D model.</p><p>1D engine simulations have been performed on a turbocharged SI engine and the results have been validated by on-engine measurements in test cell. The operating points considered have been in the engine’s low speed and load region, with the turbocharger’s waste-gate closed.</p><p>The instantaneous on-engine turbine efficiency was calculated for two different turbochargers based on high frequency measurements in test cell. Unfortunately the instantaneous mass flow rates and temperatures directly upstream and downstream of the turbine could not be measured and simulated values from the calibrated engine model were used. The on-engine turbine efficiency was compared with the efficiency computed by the 1D code using steady flow data to describe the turbine performance.</p><p>The results show that the on-engine turbine efficiency shows a hysteretic effect over the exhaust pulse so that the discrepancy between measured and quasi-steady values increases for decreasing mass flow rate after a pulse peak.</p><p>Flow modeling in pipe geometries that can be representative to those of an exhaust manifold, single bent pipes and double bent pipes and also the outer runners of an exhaust manifold, have been computed in both 1D and 3D under steady and pulsating flow conditions. The results have been compared in terms of pressure losses.</p><p>The results show that calculated pressure gradient for a straight pipe under steady flow is similar using either 1D or 3D computations. The calculated pressure drop over a bend is clearly higher1D engine simulations of turbocharged engines are difficult to <!-- @page { size: 21cm 29.7cm; margin: 2cm } P { margin-bottom: 0.21cm } -->using 1D computations compared to 3D computations, both for steady and pulsating flow. Also, the slow decay of the secondary flow structure that develops over a bend, gives a higher pressure gradient in the 3D calculations compared to the 1D calculation in the straight pipe parts downstream of a bend.</p><p> </p>

No description available.

This paper presents modifications to two widely used numerical groundwater flow models in an effort to improve upon the interaction between a well of finite length and conductivity with the surrounding formation. The first objective is to discard the common assumptions about flux- or head-based boundary conditions along the well screen by coupling pipe flow hydraulics and groundwater flow. The second objective is to avoid restricting the wellbore hydraulics to a single flow regime. Five flow regimes (laminar through rough-turbulent), based on Reynolds number and pipe roughness, are considered. The modifications are integrated into the highly versatile, well-documented and well-tested models HydroGeoSphere (finite-element/finite-difference) and USGS MODFLOW (finite-difference). Verification of the algorithm and code and is performed by comparing results to: 1) the idealized, analytical Theis solution; 2) the original, unmodified code; and 3) the results of a third party numerical solution that also accounts for variable frictional wellbore losses. Results highlight the inadequacy of either a uniform flux or a uniform head assumption along the wellbore. The solution also tends to produce much steeper hydraulic gradients in those portions of the aquifer nearest the pump intake than have previously been predicted. Systems most affected by in-well hydraulic losses include those for which well screen is long, pumping rate is large, pipe diameter is small, pipe roughness is large (either through design or aging) and aquifer conductivity is high. Improved modeling of the non-linear hydraulic conditions within the well screen can particularly influence the interpretation of wellbore flowmeter and tracer tests, leading to more precise knowledge of the variation of local aquifer hydraulic conductivity along well screens. Aquifer drawdown curves, solute transport and inflow velocities will also be influenced, which can impact capture zones and remediation costs. Given that the solution is incorporated within the HydroGeoSphere and MODFLOW models, it presents the additional advantage over existing approaches of offering a wide range of modeling capabilities, such as three-dimensional flow, arbitrary well inclination and surface-subsurface flow integration. / Thesis (Master, Civil Engineering) -- Queen's University, 2008-01-04 17:27:50.629

Since the discovery of the drag reduction effects of even small amount of macromolecules in solutions in turbulent pipe flows, there have been many experimental and theoretical studies in order to understand mechanisms behind this phenomenon. Theories have been proposed based on the observations on the change in the characteristics of the turbulent flow near the pipe wall where friction of the momentum transfer between the flow and the conduit takes place. In this study drag reduction in fully developed turbulent pipe flow with four concentrations (200 to 500 wppm) of low molecular weight Sodium Carboxymethylcellulose (CMC) in aqueous solutions was investigated experimentally. Drag reduction was determined by pressure drop measurements. In order to observe the impact of the presence of CMC on the flow, Ultrasound Doppler Velocimetry (UDV) was employed to monitor the instantaneous velocity distributions. UDV is a non-invasive technique allowing one to obtain quick velocity profiles. Experimental measurements were used to calculate Fanning friction factor and radial distributions of the axial time-averaged velocity, velocity fluctuation (turbulent intensity) and eddy viscosity. The drag reduction level was determined through the Fanning friction factor versus Reynolds number data. Velocity data could be obtained as close as 3 mm to the wall by UDV. Two impacts of increasing CMC concentration on the flow field, hence pressure drop, were observed. The first effect was the decrease of the mean velocity gradient especially near the wall with increasing polymer amount which in turn gave rise to lower friction factor or pressure drop. In addition smaller eddy viscosities were obtained in the flow. The second impact of the polymer addition was on the velocity fluctuation or turbulent intensity variation along the radial distribution. An increasing trend in turbulence intensity in the turbulent core with polymer addition was observed. This was in agreement with the earlier studies in which similar turbulence behavior was observed in addition to the suppression of the turbulent intensities near the wall

Theoretical and experimental studies of particle deposition in turbulent pipe flow have been carried out for over forty years, but some of the most important transport mechanisms are still not well understood. The first part of this thesis is concerned with the calculation of particle density when using Lagrangian methods to predict inertial particle transport in two-dimensional laminar fluid flows. Traditionally, Lagrangian calculations involve integrating the particle equations of motion along particle pathlines, and the particle density is obtained by applying a statistical averaging procedure to those pathlines which intersect a particular computational grid cell. Unfortunately, extremely large numbers of particles are required to reduce the statistical errors to acceptable levels, and this makes the method computationally expensive. Recently, the Full Lagrangian approach has been developed, which allows the direct calculation of the particle density along particle pathlines. This method had previously been applied only to simple analytical flow fields. The application of the method to CFD generated fluid velocity fields was shown to be possible, and the results obtained using the Full Lagrangian approach were compared to those from a traditional Lagrangian approach. It was found that better quality solutions could be obtained with the use of far fewer particle pathlines. An analysis of the manner in which the Full Lagrangian approach deals with particles whose paths cross each other (and the resulting discontinuity in particle density) was also undertaken, and this illustrates the sophistication of the method. The second part of the thesis comprises an experimental and theoretical study of the deposition of small particles in turbulent flows by thermophoresis. Thermophoresis is the phenomenon whereby small particles suspended in a gas in which there exists a temperature gradient experience a force in the direction opposite from that of the temperature gradient. Previous researchers have attempted to impose a radial temperature difference in pipe flow experiments, but have not yet succeeded in attaining a constant thermophoretic force along the length of the pipe. This limits the accuracy and usefulness of the data for the validation of theoretical expressions for the thermophoretic fluxes. An experimental rig has been designed to achieve a constant thermophoretic force. This was done by using an annular geometry with a cold inner wall and hot outer wall. The particle size was varied and the deposition flux was measured for turbulent flow with three temperature differences. The deposition fluxes for small particles were found to be independent of dimensionless particle size, with each increase in temperature difference resulting in an increase in magnitude of the flux. Evidence of a thermophoresis-turbulence coupling was found for intermediate-sized particles, and large particles were not influenced by thermophoresis. A theory of particle deposition, developed for the case of turbulent pipe flow, was modified to study flow in a turbulent annulus, so that theoretical expressions for the thermophoretic fluxes could be included and compared with the experimental results. Agreement with experimental data was quite good, but some deficiencies in a widely used theoretical expression for the thermophoretic flux were exposed. An alternative expression was used, which gave much better agreement with the experimental data, and the mechanisms behind the thermophoresis-turbulence coupling were also investigated. The validation of this expression for the thermophoretic force will allow its inclusion in numerical studies of particle deposition in more complex geometries.

Slush is a two phase flow of solid particles (crystals) and liquid at the triple point temperature, and constitutes an appealing alternative to liquid propellants for space launchers. The crystals give to the mixture higher density and lower specific enthalpy than liquid, enabling reduced tank volume storage and larger fuel holding time. However, the presence of solid crystals significantly modifies the thermo-hydraulics of the fuel transport, and requires novel predictive tools and diagnostic techniques for efficiently exploiting slush propellants. This thesis contributes to both aspects. In particular, this work studied the flow pressure losses and the heat transfer of solid-liquid mixtures in pipelines, combining experimental and numerical methods. Hydraulic and thermal flow features were analyzed separately with substitute mixtures chosen to mimic the behavior of slush flows in engine fuel feed systems. A dedicated facility was designed and built. The pipeline mounted conventional probes for pressure, temperature and mass flow rate measurements. Moreover, a capacitance-based density meter was developed and validated to measure the mixture's solid content. Optical flow visualization and image processing routines were combined to retrieve particulate phase distribution and velocity fields. The experimental work was complemented with 3D Unsteady Reynolds Averaged Navier Stokes simulations in OpenFOAM. The simulations coupled the Euler-Euler approach with the granular kinetic theory for the treatment of the solid dispersed phase. The model was validated with the experimental results on the pressure drop, heat transfer and solid volume fraction.The resulting physical insights and the proposed empirical correlations on the pressure drop and heat transfer in solid-liquid flows contribute to move a step forward towards slush propelled space launchers. / Doctorat en Sciences de l'ingénieur et technologie / info:eu-repo/semantics/nonPublished