Elastic effects on the hydrodynamic instability of inviscid parallel shear flows are investigated through a linear stability analysis. We focus on the upper convected Maxwell model in the limit of infinite Weissenberg and Reynolds numbers. Specifically, we study the effects of elasticity on the instability of a few classes of simple parallel flows, specifically plane Poiseuille and Couette flows, the hyperbolic-tangent shear layer and the Bickley jet. The equation for stability is derived and solved numerically using the Chebyshev collocation spectral method. This algorithm is computationally efficient and accurate in reproducing the eigenvalues. We consider flows bounded by walls as well as flows bounded by free surfaces. In the inviscid, nonelastic case all the flows we study are unstable for free surfaces. In the case of wall bounded flow, there are instabilities in the shear layer and Bickley jet flows. In all cases, the effect of elasticity is to reduce and ultimately suppress the inviscid instability. The numerical solutions are compared with the analysis of the long wave limit and excellent agreement is shown between the analytical and the numerical solutions. We found flows which are long wave stable, but nevertheless unstable to wave numbers in a certain finite range. While elasticity is ultimately stabilizing, this effect is not monotone; there are instances where a small amount of elasticity actually destabilizes the flow. The linear stability in the short wave limit of shear flows bounded by two parallel free surfaces is investigated. Unlike the plane Couette flow which has no short wave instability, we show that plane Poiseuille flow has two unstable eigenmodes localized near the free surfaces which can be combined into an even and an odd eigenfunctions. The derivation of the asymptotics of these modes shows that our numerical eigenvalues are in agreement with the analytic formula and that the difference between the two eigenvalues tends to zero exponentially with the wavenumber. / Ph. D.

Uma das mais recentes abordagens propostas na literatura para tratar o problema do alto número de Weissenberg (We) é a Formulação Log-Conformação (FLC). Nesta formulação, a equação constitutiva viscoelástica utilizada é reescrita em termos de uma variável Ψ, que é o logaritmo do tensor conformação. Apesar do potencial de aplicação da FLC, pouca atenção tem sido dirigida para análise da acurácia da solução obtida para o campo de tensões quando se utiliza esta formulação. Assim, o objetivo do presente trabalho foi estudar a acurácia da solução obtida pela FLC na análise de escoamentos de fluidos viscoelásticos usando duas geometrias padrão de estudo: placas paralelas e cavidade quadrada com tampa móvel. Primeiramente, a FLC foi implementada no pacote de CFD OpenFOAM. Em seguida foram verificados os limites do número de Weissenberg na formulação numérica padrão (Welim,P), onde para a geometria de placas paralelas foi encontrado Welim,P = 0,3 e para a geometria da cavidade quadrada com tampa móvel foi encontrado Welim,P = 0,8. Depois o código implementado foi aplicado em ambas as geometrias, comparando-se a solução obtida pela FLC com aquela da formulação padrão na faixa de We < Welim,P. Os resultados obtidos na geometria de placas paralelas apresentaram boa concordância com a solução padrão e solução analítica. Para a geometria da cavidade quadrada com tampa móvel, que não possui solução analítica, boa concordância dos resultados também foi observada em comparação com a solução padrão. Posteriormente foram comparados os resultados obtidos pela FLC na faixa de We > Welim,P. Na geometria de placas paralelas, além da boa concordância com a solução analítica, obteve-se convergência em todos os casos estudados neste trabalho, com o maior número de Weissenberg utilizado sendo igual a 8 Os resultados da geometria da cavidade quadrada com tampa móvel também apresentaram boa concordância em comparação com dados da literatura, porém a convergência foi obtida até para We = 2. Com respeito à comparação das formulações numéricas com a solução analítica, feita apenas na geometria de placas paralelas, foi observado um erro máximo de 7,57% na solução padrão e de 12,33% na FLC. Em relação à análise da qualidade das tensões usando os resíduos da equação constitutiva viscoelástica como critério de acurácia, foi verificado nas duas geometrias que os valores de tensão obtidos usando a FLC são menos acurados que aqueles obtidos pela formulação explícita no tensor das tensões nos casos em que esta última converge. Também foi observado que a acurácia diminui com o aumento do We. Esse efeito pôde ser melhor notado na geometria de placas paralelas. Uma razão para a perda de acurácia da tensão provavelmente ocorre devido à natureza matemática da transformação algébrica inversa de Ψxx para τxx. O novo solver implementado neste trabalho apresentou convergência e soluções corretas para as duas geometrias, logo foi implementado corretamente. Ele também potencializa o solver de partida viscoelastiFluidFoam ao estender simulações para uma faixa maior do número de Weissenberg. / A recent approach proposed in the literature to deal with the High Weissenberg Number Problem is the Log-Conformation formulation (LCF). In this formulation the viscoelastic constitutive equation is rewritten in terms of the logarithm of the conformation tensor Ψ. Despite the great potential application of the LCF, little attention has been given in the literature to the accuracy of the obtained stress fields. The purpose of this work was to study the solution obtained by LCF in the analysis of viscoelastic flows using two benchmark geometries: parallel plates and lid driven cavity. Firstly, the LCF was implemented in the OpenFOAM CFD package. Then, the limits of Weissenberg number for the standard numerical formulation (Welim,P) were verified, obtaining Welim,P = 0.3 for the parallel plates and Welim,P = 0.8 for the lid driven cavity. When comparing the solution obtained by the LCF with that of the standard formulation in a range of We < Welim,P, the results obtained for the parallel plates geometry showed good agreement with the standard solution and the analytical solution. For the lid driven cavity geometry, for which there is not analytical solution, good agreement with the standard solution was also observed. For We > Welim,P in the parallel plates geometry, in addition to the good agreement with the analytical solution, it was possible to obtain convergence in all the cases studied in this work, with the largest number of Weissenberg used being equal to 8 The results of the lid driven cavity geometry also presented good agreement in comparison with literature data, but convergence was obtained up to We = 2. With respect to the comparison of the numerical formulations with the analytical solution for the parallel plates geometry, a maximum error of 7.57% was observed in the standard solution and of 12.33% in the LCF. When using the residues of the viscoelastic constitutive equation as a criterion of accuracy, it was verified that for the two geometries the stress values obtained using the LCF were less accurate than those obtained by the explicit formulation in the stress tensor. It has also been observed that accuracy decreases with increasing of We. One reason for the loss of stress accuracy probably occurs because of the mathematical nature of the inverse algebraic transformation from Ψxx to τxx. The new solver implemented in this work presented convergence and correct solutions for the two geometries, so it was implemented correctly. It also potentiates the viscoelastiFluidFoam starting solver by extending simulations to a larger range of Weissenberg number.

Uma das mais recentes abordagens propostas na literatura para tratar o problema do alto número de Weissenberg (We) é a Formulação Log-Conformação (FLC). Nesta formulação, a equação constitutiva viscoelástica utilizada é reescrita em termos de uma variável Ψ, que é o logaritmo do tensor conformação. Apesar do potencial de aplicação da FLC, pouca atenção tem sido dirigida para análise da acurácia da solução obtida para o campo de tensões quando se utiliza esta formulação. Assim, o objetivo do presente trabalho foi estudar a acurácia da solução obtida pela FLC na análise de escoamentos de fluidos viscoelásticos usando duas geometrias padrão de estudo: placas paralelas e cavidade quadrada com tampa móvel. Primeiramente, a FLC foi implementada no pacote de CFD OpenFOAM. Em seguida foram verificados os limites do número de Weissenberg na formulação numérica padrão (Welim,P), onde para a geometria de placas paralelas foi encontrado Welim,P = 0,3 e para a geometria da cavidade quadrada com tampa móvel foi encontrado Welim,P = 0,8. Depois o código implementado foi aplicado em ambas as geometrias, comparando-se a solução obtida pela FLC com aquela da formulação padrão na faixa de We < Welim,P. Os resultados obtidos na geometria de placas paralelas apresentaram boa concordância com a solução padrão e solução analítica. Para a geometria da cavidade quadrada com tampa móvel, que não possui solução analítica, boa concordância dos resultados também foi observada em comparação com a solução padrão. Posteriormente foram comparados os resultados obtidos pela FLC na faixa de We > Welim,P. Na geometria de placas paralelas, além da boa concordância com a solução analítica, obteve-se convergência em todos os casos estudados neste trabalho, com o maior número de Weissenberg utilizado sendo igual a 8 Os resultados da geometria da cavidade quadrada com tampa móvel também apresentaram boa concordância em comparação com dados da literatura, porém a convergência foi obtida até para We = 2. Com respeito à comparação das formulações numéricas com a solução analítica, feita apenas na geometria de placas paralelas, foi observado um erro máximo de 7,57% na solução padrão e de 12,33% na FLC. Em relação à análise da qualidade das tensões usando os resíduos da equação constitutiva viscoelástica como critério de acurácia, foi verificado nas duas geometrias que os valores de tensão obtidos usando a FLC são menos acurados que aqueles obtidos pela formulação explícita no tensor das tensões nos casos em que esta última converge. Também foi observado que a acurácia diminui com o aumento do We. Esse efeito pôde ser melhor notado na geometria de placas paralelas. Uma razão para a perda de acurácia da tensão provavelmente ocorre devido à natureza matemática da transformação algébrica inversa de Ψxx para τxx. O novo solver implementado neste trabalho apresentou convergência e soluções corretas para as duas geometrias, logo foi implementado corretamente. Ele também potencializa o solver de partida viscoelastiFluidFoam ao estender simulações para uma faixa maior do número de Weissenberg. / A recent approach proposed in the literature to deal with the High Weissenberg Number Problem is the Log-Conformation formulation (LCF). In this formulation the viscoelastic constitutive equation is rewritten in terms of the logarithm of the conformation tensor Ψ. Despite the great potential application of the LCF, little attention has been given in the literature to the accuracy of the obtained stress fields. The purpose of this work was to study the solution obtained by LCF in the analysis of viscoelastic flows using two benchmark geometries: parallel plates and lid driven cavity. Firstly, the LCF was implemented in the OpenFOAM CFD package. Then, the limits of Weissenberg number for the standard numerical formulation (Welim,P) were verified, obtaining Welim,P = 0.3 for the parallel plates and Welim,P = 0.8 for the lid driven cavity. When comparing the solution obtained by the LCF with that of the standard formulation in a range of We < Welim,P, the results obtained for the parallel plates geometry showed good agreement with the standard solution and the analytical solution. For the lid driven cavity geometry, for which there is not analytical solution, good agreement with the standard solution was also observed. For We > Welim,P in the parallel plates geometry, in addition to the good agreement with the analytical solution, it was possible to obtain convergence in all the cases studied in this work, with the largest number of Weissenberg used being equal to 8 The results of the lid driven cavity geometry also presented good agreement in comparison with literature data, but convergence was obtained up to We = 2. With respect to the comparison of the numerical formulations with the analytical solution for the parallel plates geometry, a maximum error of 7.57% was observed in the standard solution and of 12.33% in the LCF. When using the residues of the viscoelastic constitutive equation as a criterion of accuracy, it was verified that for the two geometries the stress values obtained using the LCF were less accurate than those obtained by the explicit formulation in the stress tensor. It has also been observed that accuracy decreases with increasing of We. One reason for the loss of stress accuracy probably occurs because of the mathematical nature of the inverse algebraic transformation from Ψxx to τxx. The new solver implemented in this work presented convergence and correct solutions for the two geometries, so it was implemented correctly. It also potentiates the viscoelastiFluidFoam starting solver by extending simulations to a larger range of Weissenberg number.

Uma das mais recentes abordagens propostas na literatura para tratar o problema do alto número de Weissenberg (We) é a Formulação Log-Conformação (FLC). Nesta formulação, a equação constitutiva viscoelástica utilizada é reescrita em termos de uma variável Ψ, que é o logaritmo do tensor conformação. Apesar do potencial de aplicação da FLC, pouca atenção tem sido dirigida para análise da acurácia da solução obtida para o campo de tensões quando se utiliza esta formulação. Assim, o objetivo do presente trabalho foi estudar a acurácia da solução obtida pela FLC na análise de escoamentos de fluidos viscoelásticos usando duas geometrias padrão de estudo: placas paralelas e cavidade quadrada com tampa móvel. Primeiramente, a FLC foi implementada no pacote de CFD OpenFOAM. Em seguida foram verificados os limites do número de Weissenberg na formulação numérica padrão (Welim,P), onde para a geometria de placas paralelas foi encontrado Welim,P = 0,3 e para a geometria da cavidade quadrada com tampa móvel foi encontrado Welim,P = 0,8. Depois o código implementado foi aplicado em ambas as geometrias, comparando-se a solução obtida pela FLC com aquela da formulação padrão na faixa de We < Welim,P. Os resultados obtidos na geometria de placas paralelas apresentaram boa concordância com a solução padrão e solução analítica. Para a geometria da cavidade quadrada com tampa móvel, que não possui solução analítica, boa concordância dos resultados também foi observada em comparação com a solução padrão. Posteriormente foram comparados os resultados obtidos pela FLC na faixa de We > Welim,P. Na geometria de placas paralelas, além da boa concordância com a solução analítica, obteve-se convergência em todos os casos estudados neste trabalho, com o maior número de Weissenberg utilizado sendo igual a 8 Os resultados da geometria da cavidade quadrada com tampa móvel também apresentaram boa concordância em comparação com dados da literatura, porém a convergência foi obtida até para We = 2. Com respeito à comparação das formulações numéricas com a solução analítica, feita apenas na geometria de placas paralelas, foi observado um erro máximo de 7,57% na solução padrão e de 12,33% na FLC. Em relação à análise da qualidade das tensões usando os resíduos da equação constitutiva viscoelástica como critério de acurácia, foi verificado nas duas geometrias que os valores de tensão obtidos usando a FLC são menos acurados que aqueles obtidos pela formulação explícita no tensor das tensões nos casos em que esta última converge. Também foi observado que a acurácia diminui com o aumento do We. Esse efeito pôde ser melhor notado na geometria de placas paralelas. Uma razão para a perda de acurácia da tensão provavelmente ocorre devido à natureza matemática da transformação algébrica inversa de Ψxx para τxx. O novo solver implementado neste trabalho apresentou convergência e soluções corretas para as duas geometrias, logo foi implementado corretamente. Ele também potencializa o solver de partida viscoelastiFluidFoam ao estender simulações para uma faixa maior do número de Weissenberg. / A recent approach proposed in the literature to deal with the High Weissenberg Number Problem is the Log-Conformation formulation (LCF). In this formulation the viscoelastic constitutive equation is rewritten in terms of the logarithm of the conformation tensor Ψ. Despite the great potential application of the LCF, little attention has been given in the literature to the accuracy of the obtained stress fields. The purpose of this work was to study the solution obtained by LCF in the analysis of viscoelastic flows using two benchmark geometries: parallel plates and lid driven cavity. Firstly, the LCF was implemented in the OpenFOAM CFD package. Then, the limits of Weissenberg number for the standard numerical formulation (Welim,P) were verified, obtaining Welim,P = 0.3 for the parallel plates and Welim,P = 0.8 for the lid driven cavity. When comparing the solution obtained by the LCF with that of the standard formulation in a range of We < Welim,P, the results obtained for the parallel plates geometry showed good agreement with the standard solution and the analytical solution. For the lid driven cavity geometry, for which there is not analytical solution, good agreement with the standard solution was also observed. For We > Welim,P in the parallel plates geometry, in addition to the good agreement with the analytical solution, it was possible to obtain convergence in all the cases studied in this work, with the largest number of Weissenberg used being equal to 8 The results of the lid driven cavity geometry also presented good agreement in comparison with literature data, but convergence was obtained up to We = 2. With respect to the comparison of the numerical formulations with the analytical solution for the parallel plates geometry, a maximum error of 7.57% was observed in the standard solution and of 12.33% in the LCF. When using the residues of the viscoelastic constitutive equation as a criterion of accuracy, it was verified that for the two geometries the stress values obtained using the LCF were less accurate than those obtained by the explicit formulation in the stress tensor. It has also been observed that accuracy decreases with increasing of We. One reason for the loss of stress accuracy probably occurs because of the mathematical nature of the inverse algebraic transformation from Ψxx to τxx. The new solver implemented in this work presented convergence and correct solutions for the two geometries, so it was implemented correctly. It also potentiates the viscoelastiFluidFoam starting solver by extending simulations to a larger range of Weissenberg number.

It is well known that when a flexible or flexibly-mounted structure is placed perpendicular to the flow of a Newtonian fluid, it can oscillate due to the shedding of separated vortices at high Reynolds numbers. If the same flexible object is placed in non-Newtonian flows, however, the structure's response is still unknown. The main objective of this thesis is to introduce a new field of viscoelastic fluid-structure interactions by showing that the elastic instabilities that occur in the flow of viscoelastic fluids can drive the motion of a flexible structure placed in its path. Unlike Newtonian fluids, the flow of viscoelastic fluids can become unstable at infinitesimal Reynolds numbers due to the onset of a purely elastic flow instability. This instability occurs in the absence of nonlinear effects of fluid inertia and the Reynolds number of the flows studied here are in the order of 10-4. When such an elastic flow instability occurs in the vicinity of a flexible structure, the fluctuating fluid forces exerted on the structure grow large enough to cause a structural instability which in turn feeds back into the fluid resulting in a flow instability. Nonlinear periodic oscillations of the flexible structure are observed which have been found to be coupled to the time-dependent growth and decay of viscoelastic stresses in the wake of the structure. Presented in this thesis are the results of an investigation of the interaction occurring in the flow of a viscoelastic wormlike micelle solution past a flexible rectangular sheet. The structural geometries studied include: flexible sheet inclinations at 20°, 45° and 90° and flexible sheet widths of 5mm and 2.5mm. By varying the flow velocity, the response of the flexible sheet has been characterized in terms of amplitude and frequency of oscillations. Steady and dynamic shear rheology and filament stretching extensional rheology measurements are conducted in order to characterize the viscoelastic wormlike micelle solution. Bright field images show the deformation of the flexible sheet during an unstable oscillation while flow-induced birefringence images highlight the viscoleastic fluid stresses produced in the wake of the flexible sheet.

Recently, a new kind of turbulence has been discovered in the flow of concentrated polymer melts and solutions. These flows, known as purely elastic flows, become unstable when the elastic forces are stronger than the viscous forces. This contrasts with Newtonian turbulence, a more familiar regime where the fluid inertia dominates. While there is little understanding of purely elastic turbulence, there is a well-established dynamical systems approach to the transition from laminar flow to Newtonian turbulence. In this project, I apply this approach to purely elastic flows. Laminar flows are characterised by ordered, locally-parallel streamlines of fluid, with only diffusive mixing perpendicular to the flow direction. In contrast, turbulent flows are in a state of continuous instability: tiny differences in the location of fluid elements upstream make a large difference to their later locations downstream. The emerging understanding of the transition from a laminar to turbulent flow is in terms of exact coherent structures (ECS) — patterns of the flow that occur near to the transition to turbulence. The problem I address in this thesis is how to predict when a purely elastic flow will become unstable and when it will transition to turbulence. I consider a variety of flows and examine the purely elastic instabilities that arise. This prepares the ground for the identification of a three-dimensional steady state solution to the equations, corresponding to an exact coherent structure. I have organised my research primarily around obtaining a purely elastic exact coherent structure, however, solving this problem requires a very accurate prediction of the exact solution to the equations of motion. In Chapter 2 I start from a Newtonian ECS (travelling wave solutions in two-dimensional flow) and attempt to connect it to the purely elastic regime. Although I found no such connection, the results corroborate other evidence on the effect of elasticity on travelling waves in Poiseuille flow. The Newtonian plane Couette ECS is sustained by the Kelvin-Helmholtz instability. I discover a purely elastic counterpart of this mechanism in Chapter 3, and explore the non-linear evolution of this instability in Chapter 4. In Chapter 5 I turn to a slightly different problem, a (previously unexplained) instability in a purely elastic oscillatory shear flow. My numerical analysis supports the experimental evidence for instability of this flow, and relates it to the instability described in Chapter 3. In Chapter 6 I discover a self-sustaining flow, and discuss how it may lead to a purely elastic 3D exact coherent structure.

The polymers are integral part of vast number of products used in day to day life due to their anomalous viscoelastic behaviour. The remarkable flow behaviour exhibited by the polymeric fluids including rod climbing, extrudate swell, tube-less siphon, viscoelastic jet, elastic recoil and sharkskin instability is attributed to the complex microstructures in the polymeric liquids that arise due to the interactions of long chain polymer molecules with each other and with the surrounding fluid particles. The significance of polymer in transportation, packaging, pharmaceutical, chemical, biomedical, textiles, food and polymer processing industries highlights the requirement to comprehend the complex rheology of polymeric fluids. First, we investigate the flow features exhibited by different shear thinning vis-coelastic fluids in rectangular cavities over a wide range of depth to width ratio. We have developed a viscoelastic flow solver in order to perform numerical simulations of highly elastic flow of viscoelastic fluids. In particular, we discuss the simulations of flows of constant viscosity Boger and shear thinning viscoelastic fluids in the complex flow problems using different constitutive equations. The effects of elasticity and inertia on the flow behaviour of two shear thinning vis-coelastic fluids modeled using Giesekus and linear PTT constitutive equations in rectangular cavities is studied. The size of the primary eddies and critical aspect ratio over which the corner eddies merge to yield a second primary eddy in deep cavities is discussed. We demonstrate that the flow in the shallow and deep cavities can be characterized using Weissenberg number, defined based on the shear rate, and Deborah number, specified based on the convective time scale, respectively. The study of flow in driven cavities is important in understanding of the mixing process during synthesis of blends and composites. Next, we study two phase polymeric flow in confined geometries. Nowadays, polymer processing industries prefer to develop newer polymer with the desired material properties mechanically by mixing and blending of different polymer components instead of chemically synthesizing fresh polymer. The microstructure of blends and emulsions following drop deformation, breakup and coalescence during mixing determines its macroscopic interfacial rheology. We developed a two phase viscoelastic flow solver using volume conserving sharp interface volume-of-fluid (VOF) method for studying the dynamics of single droplet subjected to the complex flow fields. We investigated the effects of drop and matrix viscoelasticity on the motion and deformation of a droplet suspended in a fully developed channel flow. The flow behaviour exhibited by Newtonian-Newtonian, viscoelastic-Newtonian, Newtonian-viscoelastic and viscoelastic-viscoelastic drop-matrix systems is presented. The difference in the drop dynamics due to presence of constant viscosity Boger fluid and shear thinning viscoelastic fluid is represented using FENE-CR and linear PTT constitutive equations, respectively. The presence of shear thinning viscoelastic fluid either in the drop or the matrix phase suppresses the drop deformation due to stronger influence of matrix viscoelasticity as compared to the drop elasticity. The shear thinning viscoelastic drop-matrix system further restricts the drop deformation and it displays non-monotonic de-formation. The constant viscosity Boger fluid droplet curbs the drop deformation and exhibits flow dynamics identical to the shear thinning viscoelastic droplet, thus indicating that the nature of the drop viscoelasticity has little influence on the flow behaviour. The matrix viscoelasticity due to Boger fluid increases drop deformation and displays non-monotonic deformation. The drop deformation is further enhanced in the case of Boger fluid in viscoelastic drop-matrix system. Interestingly, the pressure drop due to the presence of viscoelastic drop in a Newtonian matrix is lower than the single phase flow of Newtonian fluid. We also discuss the effects of inertia, surface tension, drop to matrix viscosity ratio and the drop size on these drop-matrix systems. Finally, we investigate the emulsion rheology by studying the motion of a droplet in the square lid driven cavity flow. The viscoelastic effects due to constant viscosity Boger fluid and shear thinning viscoelastic fluid are illustrated using FENECR and Giesekus rheological relations, respectively. The presence of viscoelasticity either in drop or matrix phase boosts the drop deformation with the drop viscoelasticity displaying intense deformation. The drop dynamics due to the droplet viscoelasticity is observed to be independent of the nature of vis-coelastic fluid. The shear thinning viscoelastic matrix has a stronger influence on the drop deformation and orientation compared to the Boger fluid matrix. The different blood components, cells and many materials of industrial importance are viscoelastic in nature. Thus, the present study has significant applications in medical diagnostics, drug delivery, manufacturing and processing industries, study of biological flows, pharmaceutical research and development of lab-on-chip devices.