Multiscale space–time fluid–structure interaction techniques

We present the multiscale space–time techniques we have developed for fluid–structure interaction (FSI) computations. Some of these techniques are multiscale in the way the time integration is performed (i.e. temporally multiscale), some are multiscale in the way the spatial discretization is done (i.e. spatially multiscale), and some are in the context of the sequentially-coupled FSI (SCFSI) techniques developed by the Team for Advanced Flow Simulation and Modeling (T \bigstar AFSM){({\rm T} \bigstar {\rm AFSM})}. In the multiscale SCFSI technique, the FSI computational effort is reduced at the stage we do not need it and the accuracy of the fluid mechanics (or structural mechanics) computation is increased at the stage we need accurate, detailed flow (or structure) computation. As ways of increasing the computational accuracy when or where needed, and beyond just increasing the mesh refinement or decreasing the time-step size, we propose switching to more accurate versions of the Deforming-Spatial-Domain/Stabilized Space–Time (DSD/SST) formulation, using more polynomial power for the basis functions of the spatial discretization or time integration, and using an advanced turbulence model. Specifically, for more polynomial power in time integration, we propose to use NURBS, and as an advanced turbulence model to be used with the DSD/SST formulation, we introduce a space–time version of the residual-based variational multiscale method. We present a number of test computations showing the performance of the multiscale space–time techniques we are proposing. We also present a stability and accuracy analysis for the higher-accuracy versions of the DSD/SST formulation.     

Wall shear stress calculations in space–time finite element computation of arterial fluid–structure interactions

The stabilized space–time fluid–structure interaction (SSTFSI) technique was applied to arterial FSI problems soon after its development by the Team for Advanced Flow Simulation and Modeling. The SSTFSI technique is based on the Deforming-Spatial-Domain/Stabilized Space–Time (DSD/SST) formulation and is supplemented with a number of special techniques developed for arterial FSI. The special techniques developed in the recent past include a recipe for pre-FSI computations that improve the convergence of the FSI computations, using an estimated zero-pressure arterial geometry, Sequentially Coupled Arterial FSI technique, using layers of refined fluid mechanics mesh near the arterial walls, and a special mapping technique for specifying the velocity profile at inflow boundaries with non-circular shape. In this paper we introduce some additional special techniques, related to the projection of fluid–structure interface stresses, calculation of the wall shear stress (WSS), and calculation of the oscillatory shear index. In the test computations reported here, we focus on WSS calculations in FSI modeling of a patient-specific middle cerebral artery segment with aneurysm. Two different structural mechanics meshes and three different fluid mechanics meshes are tested to investigate the influence of mesh refinement on the WSS calculations.     

Fluid–structure Interaction Modeling of Aneurysmal Conditions with High and Normal Blood Pressures

Hemodynamic factors like the wall shear stress play an important role in cardiovascular diseases. To investigate the influence of hemodynamic factors in blood vessels, the authors have developed a numerical fluid–structure interaction (FSI) analysis technique. The objective is to use numerical simulation as an effective tool to predict phenomena in a living human body. We applied the technique to a patient-specific arterial model, and with that we showed the effect of wall deformation on the WSS distribution. In this paper, we compute the interaction between the blood flow and the arterial wall for a patient-specific cerebral aneurysm with various hemodynamic conditions, such as hypertension. We particularly focus on the effects of hypertensive blood pressure on the interaction and the WSS, because hypertension is reported to be a risk factor in rupture of aneurysms. We also aim to show the possibility of FSI computations with hemodynamic conditions representing those risk factors in cardiovascular disease. The simulations show that the transient behavior of the interaction under hypertensive blood pressure is significantly different from the interaction under normal blood pressure. The transient behavior of the blood-flow velocity, and the resulting WSS and the mechanical stress in the aneurysmal wall, are significantly affected by hypertension. The results imply that hypertension affects the growth of an aneurysm and the damage in arterial tissues.     

Interface projection techniques for fluid–structure interaction modeling with moving-mesh methods

The stabilized space–time fluid–structure interaction (SSTFSI) technique developed by the Team for Advanced Flow Simulation and Modeling (T★AFSM) was applied to a number of 3D examples, including arterial fluid mechanics and parachute aerodynamics. Here we focus on the interface projection techniques that were developed as supplementary methods targeting the computational challenges associated with the geometric complexities of the fluid–structure interface. Although these supplementary techniques were developed in conjunction with the SSTFSI method and in the context of air–fabric interactions, they can also be used in conjunction with other moving-mesh methods, such as the Arbitrary Lagrangian–Eulerian (ALE) method, and in the context of other classes of FSI applications. The supplementary techniques currently consist of using split nodal values for pressure at the edges of the fabric and incompatible meshes at the air–fabric interfaces, the FSI Geometric Smoothing Technique (FSI-GST), and the Homogenized Modeling of Geometric Porosity (HMGP). Using split nodal values for pressure at the edges and incompatible meshes at the interfaces stabilizes the structural response at the edges of the membrane used in modeling the fabric. With the FSI-GST, the fluid mechanics mesh is sheltered from the consequences of the geometric complexity of the structure. With the HMGP, we bypass the intractable complexities of the geometric porosity by approximating it with an “equivalent”, locally-varying fabric porosity. As test cases demonstrating how the interface projection techniques work, we compute the air–fabric interactions of windsocks, sails and ringsail parachutes.     

A parallel sparse algorithm targeting arterial fluid mechanics computations

Iterative solution of large sparse nonsymmetric linear equation systems is one of the numerical challenges in arterial fluid–structure interaction computations. This is because the fluid mechanics parts of the fluid + structure block of the equation system that needs to be solved at every nonlinear iteration of each time step corresponds to incompressible flow, the computational domains include slender parts, and accurate wall shear stress calculations require boundary layer mesh refinement near the arterial walls. We propose a hybrid parallel sparse algorithm, domain-decomposing parallel solver (DDPS), to address this challenge. As the test case, we use a fluid mechanics equation system generated by starting with an arterial shape and flow field coming from an FSI computation and performing two time steps of fluid mechanics computation with a prescribed arterial shape change, also coming from the FSI computation. We show how the DDPS algorithm performs in solving the equation system and demonstrate the scalability of the algorithm.     

FSI analysis of the blood flow and geometrical characteristics in the thoracic aorta

We present a fluid–structure interaction (FSI) analysis of the blood flow and geometrical characteristics in the thoracic aorta. The FSI is handled with the sequentially-coupled arterial FSI technique. The fluid mechanics equations are solved with the ST-VMS method, which is the variational multiscale version of the deforming-spatial-domain/stabilized space–time (DSD/SST) method. We focus on the relationship between the centerline geometry of the aorta and the flow field, which influences the wall shear stress distribution. The centerlines of the aorta models we use in our analysis are extracted from the CT scans, and we assume a constant diameter. Torsion-free model geometries are generated by projecting the original centerline to its averaged plane of curvature. The flow fields for the original and projected geometries are compared to examine the influence of the torsion.     

Sequentially-coupled space–time FSI analysis of bio-inspired flapping-wing aerodynamics of an MAV

We present a sequentially-coupled space–time (ST) computational fluid–structure interaction (FSI) analysis of flapping-wing aerodynamics of a micro aerial vehicle (MAV). The wing motion and deformation data, whether prescribed fully or partially, is from an actual locust, extracted from high-speed, multi-camera video recordings of the locust in a wind tunnel. The core computational FSI technology is based on the Deforming-Spatial-Domain/Stabilized ST (DSD/SST) formulation. This is supplemented with using NURBS basis functions in temporal representation of the wing and mesh motion, and in remeshing. Here we use the version of the DSD/SST formulation derived in conjunction with the variational multiscale (VMS) method, and this version is called “DSD/SST-VMST.” The structural mechanics computations are based on the Kirchhoff–Love shell model. The sequential-coupling technique is applicable to some classes of FSI problems, especially those with temporally-periodic behavior. We show that it performs well in FSI computations of the flapping-wing aerodynamics we consider here. In addition to the straight-flight case, we analyze cases where the MAV body has rolling, pitching, or rolling and pitching motion. We study how all these influence the lift and thrust.     

Solution of linear systems in arterial fluid mechanics computations with boundary layer mesh refinement

Computation of incompressible flows in arterial fluid mechanics, especially because it involves fluid–structure interaction, poses significant numerical challenges. Iterative solution of the fluid mechanics part of the equation systems involved is one of those challenges, and we address that in this paper, with the added complication of having boundary layer mesh refinement with thin layers of elements near the arterial wall. As test case, we use matrix data from stabilized finite element computation of a bifurcating middle cerebral artery segment with aneurysm. It is well known that solving linear systems that arise in incompressible flow computations consume most of the time required by such simulations. For solving these large sparse nonsymmetric systems, we present effective preconditioning techniques appropriate for different stages of the computation over a cardiac cycle.     

On the use of conformal mapping for the computation of hydrodynamic forces acting on bodies of arbitrary shape in viscous flow. Part 1: simply connected body

Some aspects of the force and moment computations in incompressible and viscous flows are revisited. The basic idea was developed in Quartapelle and Napolitano (AIAA J. 21:991–913, 1983). They formulated the way to compute the force and moment without explicitly calculating the pressure. The principle is to project Navier–Stokes equations on a set of functions. Surprisingly these functions have a meaning in potential theory. They are precisely the solutions which give the added masses and added moment of inertia for potential flow. By revisiting this problem for two-dimensional flows in unbounded liquid, a general identity giving the added masses and added moment of inertia is formulated. To this end a conformal-mapping technique is used to transform the fluid domain. Once the potential solution has been obtained, the projection method by Quartapelle and Napolitano is implemented. In addition an a posteriori computation of the pressure is described. Applications illustrate the present study.     

Fluid–structure interaction modeling of ringsail parachutes

In this paper, we focus on fluid–structure interaction (FSI) modeling of ringsail parachutes, where the geometric complexity created by the “rings” and “sails” used in the construction of the parachute canopy poses a significant computational challenge. It is expected that NASA will be using a cluster of three ringsail parachutes, referred to as the “mains”, during the terminal descent of the Orion space vehicle. Our FSI modeling of ringsail parachutes is based on the stabilized space–time FSI (SSTFSI) technique and the interface projection techniques that address the computational challenges posed by the geometric complexities of the fluid–structure interface. Two of these interface projection techniques are the FSI Geometric Smoothing Technique and the Homogenized Modeling of Geometric Porosity. We describe the details of how we use these two supplementary techniques in FSI modeling of ringsail parachutes. In the simulations we report here, we consider a single main parachute, carrying one third of the total weight of the space vehicle. We present results from FSI modeling of offloading, which includes as a special case dropping the heat shield, and drifting under the influence of side winds.     

Stabilized space–time computation of wind-turbine rotor aerodynamics

We show how we use the Deforming-Spatial-Domain/Stabilized Space–Time (DSD/SST) formulation for accurate 3D computation of the aerodynamics of a wind-turbine rotor. As the test case, we use the NREL 5MW offshore baseline wind-turbine rotor. This class of computational problems are rather challenging, because they involve large Reynolds numbers and rotating turbulent flows, and computing the correct torque requires an accurate and meticulous numerical approach. We compute the problem with both the original version of the DSD/SST formulation and a recently introduced version with an advanced turbulence model. The DSD/SST formulation with the advanced turbulence model is a space–time version of the residual-based variational multiscale method. We compare our results to those reported recently, which were obtained with the residual-based variational multiscale Arbitrary Lagrangian–Eulerian method using NURBS for spatial discretization and which we take as the reference solution. While the original DSD/SST formulation yields torque values not far from the reference solution, the DSD/SST formulation with the variational multiscale turbulence model yields torque values very close to the reference solution.     

Role of 0D peripheral vasculature model in fluid–structure interaction modeling of aneurysms

Patient-specific simulations based on medical images such as CT and MRI offer information on the hemodynamic and wall tissue stress in patient-specific aneurysm configurations. These are considered important in predicting the rupture risk for individual aneurysms but are not possible to measure directly. In this paper, fluid–structure interaction (FSI) analyses of a cerebral aneurysm at the middle cerebral artery (MCA) bifurcation are presented. A 0D structural recursive tree model of the peripheral vasculature is incorporated and its impedance is coupled with the 3D FSI model to compute the outflow through the two branches accurately. The results are compared with FSI simulation with prescribed pressure variation at the outlets. The comparison shows that the pressure at the two outlets are nearly identical even with the peripheral vasculature model and the flow division to the two branches is nearly the same as what we see in the simulation without the peripheral vasculature model. This suggests that the role of the peripheral vasculature in FSI modeling of the MCA aneurysm is not significant.     

Dynamic and fluid–structure interaction simulations of bioprosthetic heart valves using parametric design with T-splines and Fung-type material models

This paper builds on a recently developed immersogeometric fluid–structure interaction (FSI) methodology for bioprosthetic heart valve (BHV) modeling and simulation. It enhances the proposed framework in the areas of geometry design and constitutive modeling. With these enhancements, BHV FSI simulations may be performed with greater levels of automation, robustness and physical realism. In addition, the paper presents a comparison between FSI analysis and standalone structural dynamics simulation driven by prescribed transvalvular pressure, the latter being a more common modeling choice for this class of problems. The FSI computation achieved better physiological realism in predicting the valve leaflet deformation than its standalone structural dynamics counterpart.     

A Multiscale/stabilized Formulation of the Incompressible Navier–Stokes Equations for Moving Boundary Flows and Fluid–structure Interaction

This paper presents a multiscale/stabilized finite element formulation for the incompressible Navier–Stokes equations written in an Arbitrary Lagrangian–Eulerian (ALE) frame to model flow problems that involve moving and deforming meshes. The new formulation is derived based on the variational multiscale method proposed by Hughes (Comput Methods Appl Mech Eng 127:387–401, 1995) and employed in Masud and Khurram in (Comput Methods Appl Mech Eng 193:1997–2018, 2006); Masud and Khurram in (Comput Methods Appl Mech Eng 195:1750–1777, 2006) to study advection dominated transport phenomena. A significant feature of the formulation is that the structure of the stabilization terms and the definition of the stabilization tensor appear naturally via the solution of the sub-grid scale problem. A mesh moving technique is integrated in this formulation to accommodate the motion and deformation of the computational grid, and to map the moving boundaries in a rational way. Some benchmark problems are shown, and simulations of an elastic beam undergoing large amplitude periodic oscillations in a viscous fluid domain are presented.     

Transient response of an elastic half space to normal pressure acting over a circular area on an inclined plane

An earthquake source has been simulated as a simple finite source, i.e., normal pressure acting over an inclined fault plane. The transient response of the surface displacement of an elastic half space due to the above internal source is calculated. A series of transformations, followed by the traditional Cagniard–de Hoop technique, are used to compute the transient response. Various wave arrivals are discussed. Numerical computations bring out the special character of the finite source vis-à-vis the point source. The originality of the paper lies in the fact that for the first time an exact computation of the surface response due to an inclined finite source has been computed by Cagniard’s approach.     

3D fluid–structure-contact interaction based on a combined XFEM FSI and dual mortar contact approach

Finite deformation contact of flexible solids embedded in fluid flows occurs in a wide range of engineering scenarios. We propose a novel three-dimensional finite element approach in order to tackle this problem class. The proposed method consists of a dual mortar contact formulation, which is algorithmically integrated into an eXtended finite element method (XFEM) fluid–structure interaction approach. The combined XFEM fluid–structure-contact interaction method (FSCI) allows to compute contact of arbitrarily moving and deforming structures embedded in an arbitrary flow field. In this paper, the fluid is described by instationary incompressible Navier–Stokes equations. An exact fluid–structure interface representation permits to capture flow patterns around contacting structures very accurately as well as to simulate dry contact between structures. No restrictions arise for the structural and the contact formulation. We derive a linearized monolithic system of equations, which contains the fluid formulation, the structural formulation, the contact formulation as well as the coupling conditions at the fluid–structure interface. The linearized system may be solved either by partitioned or by monolithic fluid–structure coupling algorithms. Two numerical examples are presented to illustrate the capability of the proposed fluid–structure-contact interaction approach.     

Numerical-performance studies for the stabilized space–time computation of wind-turbine rotor aerodynamics

We present our numerical-performance studies for 3D wind-turbine rotor aerodynamics computation with the deforming-spatial-domain/stabilized space–time (DSD/SST) formulation. The computation is challenging because of the large Reynolds numbers and rotating turbulent flows, and computing the correct torque requires an accurate and meticulous numerical approach. As the test case, we use the NREL 5MW offshore baseline wind-turbine rotor. We compute the problem with both the original version of the DSD/SST formulation and the version with an advanced turbulence model. The DSD/SST formulation with the turbulence model is a recently-introduced space–time version of the residual-based variational multiscale method. We include in our comparison as reference solution the results obtained with the residual-based variational multiscale Arbitrary Lagrangian–Eulerian method using NURBS for spatial discretization. We test different levels of mesh refinement and different definitions for the stabilization parameter embedded in the “least squares on incompressibility constraint” stabilization. We compare the torque values obtained.     

Mathematical foundations of the immersed finite element method

In this paper, we propose an immersed solid system (ISS) method to efficiently treat the fluid–structure interaction (FSI) problems. Augmenting a fluid in the moving solid domain, we introduce a volumetric force to obtain the correct dynamics for both the fluid and the structure. We further define an Euler–Lagrange mapping to describe the motion of the immersed solid. A weak formulation (WF) is then constructed and shown to be equivalent to both the FSI and the ISS under certain regularity assumptions. The weak formulation (WF) may be computed numerically by an implicit algorithm with the finite element method, and the streamline upwind/Petrov–Galerkin method. Compared with the successful immersed boundary method (IBM) by Peskin and co-workers (J Comput Phys 160:705–719, 2000; Acta Numerica 11:479–517, 2002; SIAM J Sci Stat Comput 13(6):1361–1376, 1992) the ISS method applies to more general geometries with non-uniform grids and avoids the inaccuracy in approximating the Dirac delta function     

Computation of residence time in the simulation of pulsatile ventricular assist devices

A continuum-based model of particle residence time for moving-domain fluid mechanics and fluid–structure interaction (FSI) computations is proposed, analyzed, and applied to the simulation of an adult pulsatile ventricular assist device (PVAD). Residence time is a quantity of clinical interest for blood pumps because it correlates with thrombotic risk. The proposed technique may be easily implemented in any flow or FSI solver. In the context of PVADs the results of the model may be used to assess how efficiently the pump moves the blood through its interior. Three scalar measures of particle residence time are also proposed. These scalar quantities may be used in the PVAD design with the goal of reducing thrombotic risk.