Yash Sanjay shewale
Under the Supervision of
Dr Prakhar Gupta
Indian Institute of Technology Hyderabad
Thermo-fluid Engineering
Mechanical and Aerospace Engineering
March 1, 2023
Important terms 2
▶ Soft Active Rods : Materials capable of bending and changing shapes in
response to various stimuli
▶ Generalize Newtonian fluid : Shear stress is a function of shear rate
a. Power Law fluid
b. cross Fluid
c. Carreau–Yasuda fluid
▶ Non sharable rods: cross-section of the rod dose not change
▶ Stokesian regime:A regime where the inertial effects of the fluid is neglected
.
Title 3
A finite element method for simulating soft active non-shearable rods
immersed in generalized Newtonian fluids
Roberto Federico Ausas , Cristian Guillermo Gebhardt, Gustavo Carlos Buscaglia
Problems with soft materials during simulations 4
▶ Attachment of solid body to immovable boundary is generally absent
▶ Large deformation and displacement of the body
▶ Similar Problems that we face in aerodynamics
▶ Fluid-structure interaction problems
Problem Statement 5
▶ A quite general model of rod-like solids is provided by the Cosserat theory of
rods
▶ This model is immersed into the generalised Newtonian fluid
Caption
Important equations 6
▶ Energy equation
RS
1. E = 12 0 [Cϵ ϵ2 + Ck ko 2 ]
ϵ = local measure of elongation change
k0 = spontanious curvature
E= Energy of the rod
▶ Governing equations
1. σ = −pI + 2µϵ(u)
2. ∇ ∗ u = 0
σ=Shear sterss
p=pressure
µ= dynamic viscosity
ϵ=strain rate function of velocity
Neumarical Experiments 7
A. Roll-up and roll-out of a flexible rod
▶ k0 = −0.999 ∗ 2π if t < 100
=0 otherwise
Cϵ = 90 Ck =0.0225
▶ Meshing:
A. Rod: 4 Hermite elements
B. Fluid: mesh has elements of characteristic size h = 1/100 close to the rod’s
boundary
▶ Boundary condition:
A. Both lateral wall are open(traction free)
B. Top and bottom wall are subjected to zero velocity boundary condition
Results 8
Shape evolution and pressure field for the roll-up and release of a pre-stressed rod immersed in a
viscous fluid. The instants of the frames above are t = 0, 0.296, 1.698, 4.089, 11.42, 19.42, 37.42,
67.42, 99.99, 100, 100.07, 100.2, 100.5, 103.0, 107.2, 109.0, 111.0, 119.0, respectively
Validation
▶ Perform same experiment for
K0 = −π
keeping all boundary condition same.
▶ Conclusion:
A.The convergence rate for the
curvature error e is O(H 2 )
B.The positional error ep is O(H 4 )
Error vs H(Hermite elements)
Numerical Experiment 10
B. Large deflection of a flexible beam in a channel
▶ a. Left end of rod is clipped at (x1 , x2 )=(0,3/2)=q
b. right end of rod is free
c. Spontaneous Curvature: k0 (s) = 2π(1 + tan(4s))
▶ Boundary Condition:
a. At clipped side dq
ds = 0
b. From 0 ≤ x2 ≤ 3−e 2 u(x1 = 0, x2 ) = ((3 − e)x2 2 − 2x2 4 , 0)
After dt = 12sec
c. The bottom wall (x2 = 0), the right wall (x1 = 3) and the part of the left
wall above the beam (x1 = 0, x2 > 3 + e/2), are permanently subject
to zero-velocity boundary conditions
d. The top wall (x2 = 3) is an outlet boundary through which the fluid is free
to leave the domain.
Numerical Experiments 11
Results
Shape, contours of pressure field and velocity vectors for the cantilever rod subject to a pre-stress
and a velocity field switched on from time t = 12.
Numerical Experiment 12
C. Swimming of a finite-length flexible rod
▶ a. Spontaneous curvature
k0 (s, t) = 20 sin(4π(s − 2t))
▶ b. Boundary Conditions
Both lateral walls are open and the bottom and top wall are subject to
zero-velocity boundary conditions.
▶ c. Mesh
1. Rod: discretised in 8 elements
2. Fluid: triangular mesh with h= 0.01 close to swimmer body
Numerical Experiments 13
Result:
counter of velocity contour of pressure
different times (from left to right, t = 0.0, 1.0, 2.4, 3.2).
Impact of discritize parameters 14
A. Rod is discretise for H = 1/4, 1/6, 1/8, 1/12 keeping fluid mesh
constant i.e.h=0.01
t=1sec t=1.8sec t=3.2sec
Conclusion
Deformation of the rod is well capture by the mesh consisting at least 8 elements
Impact of discretise parameters 15
B. different Fluid refinements i.e. h=1/50, 1/100, 1/200 keeping H =
1/8
t=1sec t=1.8sec t=3.2sec
Conclusion
From here we can conclude that finer mesh gives us better results
Impact of discretise parameters 16
Comparing displacement vs time for different H
“If you have knowledge, let others
light their candles in it.”
Margaret Fuller
Thank You!
