Fluids & Structural Mechanics > CM Projects > Biological/Biomedical Flows (BBF)

Images | Animations (Below)

The CM Division is active in several fluid-mechanics research programs related to human systems. One recent activity involves the use of modern medical imaging technology and three-dimensional, time accurate multi-phase CFD to model oxygen uptake and particle deposition in the human respiratory system. This effort is being advanced in collaboration with the Center for Medical Diagnostic Systems and Visualization (Germany), and Image Processing Laboratory, Institute of Anatomy and Cell Biology (Germany). A physically accurate representation of the trachea and most of the convective regime bronchi (up to generation 13) are obtained using high-resolution computed tomography (HRCT) of a rubber cast of an adult human tracheobronchial tree, Figure 1. A triangulated model of this data is used as a bounding surface for hybrid unstructured meshes generated using the ICEM-CFD 4.2 package. A parallel, n-fluid CFD solver developed at ARL, NPHASE, is employed in the analyses. The differential models employed are homogeneous n-species (n-gas constituents) and ensemble averaged n-fluid (gas+particle) systems. Specialized boundary conditions have been developed and applied at each of the more than 700 bronchi termini. These boundary conditions account for effective volume variation during a breathing cycle, O2-CO2 exchange, and diffusion range deposition. Convective and diffusive particle deposition models are also developed and applied for the resolved bronchi. Figure 2 shows predicted oxygen concentration in the wall-adjacent finite volumes at a particular inhalation timestep. In Figure 3, contours of normalized relative helicity at a cut just downstream of the first bronchi branch illustrate the secondary flows that are present. Figure 4 shows a composite plot of contours of 16 mm particle deposition in the upper 6 bronchi, and quantitative deposition efficiency predictions across a range of particle sizes.

The goal of this research is to evolve a complete simulation system for the human respiratory system, including detailed CFD analysis of the near-field external environment, gas-masks, inhalers, oral/nasal cavities and lung, including improved sub-grid deposition and pulmonary exchange modeling. This system will benefit: 1) medical practitioners in the assessment of respiratory performance of unhealthy individuals before, during and after treatment, 2) pulmonary drug delivery and protection gear system designers, 3) regulatory and WMD response agencies that need to assess short- and longer-term exposure to various environmental contaminants and toxins.

Another research project involves the modeling of the micro-fluid mechanics of white blood cell and cancer cell interactions with one another and blood vessel walls. Overset, unstructured, moving, and deforming meshes are employed. 6DOF dynamics and shell structural mechanics modeling are coupled to the time-accurate fluid mechanics simulations. In addition to these modern CFD technologies, the modeling of the complex bio-chemistry of cell-collision and adhesion and the validation of the model with micro-experimental measurements are important components of this research, some elements of which are shown in Figure 5.

The division is involved in the hydrodynamic design and blood damage assessment for several cardiovascular assist devices, Figure 6. From the standpoint of hydrodynamic design, the CM division is able to draw on its extensive experience in the detailed design analysis of complex, multistage axial, mixed, and centrifugal pumps. Division personnel use structured and unstructured grids as well as overset grid methods as appropriate to handle the complex and moving geometries that are often required. For cardiovascular assist devices, the potential for blood damage adds additional analysis requirements. Computational models for hemolysis, thrombosis, and platelet activation hold huge potential to help decrease design time, offer a controlled alternative to delicate in vitro testing, and can help to minimize the need for animal and human testing as well as to improve the likelihood of success when such tests are ultimately conducted. In the process of supporting the design of actual devices, CM division personnel are performing critical evaluations of the current state of blood damage modeling as well as extending current models for the evaluation of platelet activation and for the evaluation of hemolysis in unsteady flows.

CFD of Blood Pumps
Members of the CM division have performed design analyses of several artificial heart assist devices, including both axial and positive displacement devices. CM engineers have worked in conjunction with researchers from the Pennsylvania State Hershey Medical Center and Pennsylvania State Bioengineering Department on the design of a small (50 cc), positive displacement device for use in smaller adult and adolescent patients. A sample positive displacement device is shown in Figure 7.

The geometry was meshed using the unstructured grid generation capability of Gridgen (Figure 8). The standard resolution mesh contained approximately 1 million points, whereas resolved meshes contained approximately 3.5 million points. The CFD analyses were performed using the commercial solver AcuSolve from AcuSim Software. The unsteady analyses performed a minimum of 200 time steps per device beat and were carried out for a minimum of three beat cycles.

A large number of design variants and operating conditions were examined. The operating conditions and boundary conditions were chosen to match in vitro experiments performed in the Bioengineering Department at the Pennsylvania State University. The CFD results were found to compare well with PIV flow velocity measurements.

The unsteady piston motion was modeled by compressing the CFD mesh in the chamber region (Animation 1). The flow field was then plotted at slices normal to the piston face which corresponded to PIV measurement planes. The in-plane velocity magnitude (x and y components only) were plotted in Animations 2 and 3. Animation 3 also includes a line plot of the instantaneous chamber flow rate. A rotational pattern is established within the device with jets observable at the mitral and aortic ports during diastole (filling) and systole (ejection) respectively.

Hydrodynamic stresses are an important fluid dynamic parameter influencing the amount of blood damage occurring within an artificial blood pump. Large stresses can lead to red blood cell damage (hemolysis) and platelet activation. Low wall stresses and/or flow stagnation can promote clot formation, growth, and embolization (thrombosis). Animation 4 shows the computed scalar wall shear rates observed in the blood pump as a means to assess device thrombus potential.

Images
image
View Image (71kb)

Figure 1:

High-resolution computed tomography (HRCT) of a rubber cast of an adult human tracheobronchial tree
image
View Image (72kb)

Figure 2:

Predicted oxygen concentration in the wall-adjacent finite volumes at a particular inhalation timestep
image
View Image (75kb)

Figure 3:

Contours of normalized relative helicity at a cut just downstream of the first bronchi branch illustrate the secondary flows that are present
image
View Image (84kb)

Figure 4:

A composite plot of contours of 16 mm particle deposition in the upper 6 bronchi, and quantitative deposition efficiency predictions across a range of particle sizes
image
View Image (280kb)

Figure 5:

Elements of the modeling of the complex bio-chemistry of cell-collision and adhesion and the validation of the model with micro-experimental measurements.
image
View Image (152kb)

Figure 6:

Cardiovascular assist device
image
View Image (128kb)

Figure 7:

Sample positive displacement heart pump design. The pump inlet (mitral) port is located on the right and the outlet (aortic) port on the left.


View Image (222kb)

Figure 8:

Top and front view of the unstructured mesh used for the unsteady CFD analysis (~1,000,000 nodal points).


Animations
image
View Animation
(Flash Player 607kb)

Animation 1:

Unsteady mesh compression demonstrating piston model.

image
View Animation
(Flash Player 549kb)

Animation 2:

Pulsatile in-plane velocity magnitude contours showing the jets and rotational pattern formed during a beat.

image
View Animation
(Flash Player 573kb)

Animation 3:

Second example of the in-plane velocity magnitude contours along with a plot of the corresponding device flow rate.

image
View Animation
(Flash Player 513kb)

Animation 4:

Scalar wall strain rates observed during a beat cycle.