Learning Forum

[Mahavirsinh Zala]

Hello, 

I am currently in the process of trying to find a way to effectively use ANSYS for 2D patient specific geometries derived from 3D scans. Early work has shown that simple shapes provide accurate and comparable results to their 3D counterparts. The discrepancy is created with more complex shapes as the axissymetric solver can not be used. The differences being observed in Velocities and Wall Shear Stress. Has anybody tried this yet or have had any progress made towards this?

[Federico]

Hello, 

I don't know what your application is but in general 2D analyses are most relevant when there is some level of symmetry, which allows for reducing the order of the system.

The farther you move away from this assumption, the more 3-dimensional effects and other interactions may come into play, reducing the accuracy of a 2D model.

[Rob]

2d was commonly used in the 90s and into the 00s because of a lack of compute power, it's not as common now for the reasons you mention. What has your literature survey turned up? 

[Mahavirsinh Zala]

 

Literature has shown that from a CT Scan we can create an accurate patient specific 3D model of large vessels such as the Aorta. From this 3D Model it is possible to find many accurate readings per patient. The issue is that the 3D models can take a while to create and run. So I was looking into creating an accurate 2D model which would provide feasible results for fluid flow within the Aorta. Like mentioned it was easier working with symmetric models such as a pipe and a bifurcation, but is it possible to do so with asymetric models? For simple models I used the axisymmetric solver which is why it was possible, when using the planar solver I have been unable to get the desired results. 

 

[Rob]

We can, and it's been doable (and increasing common) since the early-mid 00s. 

However, as the shape becomes complex 2d approximations become less useful, and given the modern compute power the 3d model is relatively quick. There are also very significant 3d effects in the flows at the bifurcations. If you use the Materialise Mimics/Magics or Synopys Simpleware tools into Fluent Meshing the time to build the models is relatively quick especially for the aorta as it's a fairly straight pipe.