yade-users team mailing list archive

[Robert Caulk <question660585@xxxxxxxxxxxxxxxxxxxxx>]

Question #660585 on Yade changed:
https://answers.launchpad.net/yade/+question/660585

Robert Caulk posted a new comment:
OK, my testing leads me to believe that the deformations are not large
enough in a hydraulic fracture simulation to change the Delaunay
triangulation. Therefore, it should be safe to update positions the
normal way. I've found that the remesh interval needs to be much smaller
than if we do not update the particle positions (1/5th-ish). So I guess
simulation speed was the only real issue, but the slow down is not due
to setting positions, but instead due to the more frequent remesh
requirement (and frequent solver building associated with it).

In comparison, the hydraulic fracture behavior difference is pronounced
for fixed particle positions vs updated positions. For updated
positions, the rate of injection pressure drop is about half of the rate
observed for fixed particle positions. Further, the crack area is
greater for the fixed positions by 5-10% compared to the updated
positions. So the fracture is propagating faster at lower injection
pressures for the fixed positions, and the fracture is propagating
slower at higher injection pressures with updated positions. This makes
sense, considering we aren't considering the impact of the changing void
volume on pressures for the fractured cells when we fix positions. When
we update positions, these void volumes increase in fractured cells, the
pressures decrease slightly due to fluid compressibility, and the pore
pressure at the tip will not be as high, leading to lower fluid induced
forces and a slower propagation of the fracture. For fixed positions, a
fractured cells void volume remains the same for the duration of the
simulation. So as soon as the fluid pressure increases in a fractured
cell, the pressure raises more quickly and the fluid induced forces
follow, leading to quicker fracturing, and the fluid moves more quickly
to the next cell. For updated positions, the fracture is propagating
slower, the pressure is not relieved as quickly by the fast permeability
changes enabled by fractured cells. In other words, the slow process of
opening the fracture is accommodated when we update positions, whereas
fixed particle positions yield stiffer and less accommodating fracture
propagation behavior.

-- 
You received this question notification because your team yade-users is
an answer contact for Yade.