Title: Shape Optimization Problem for Incompressible Viscous Flow
1Shape Optimization Problem for Incompressible
Viscous Flow Masaki Tashiro, Mutsuto Kawahara
Department of Civil Engineering, Chuo
University, Japan Department of Civil
Engineering, Chuo University, Japan 5th
ICHE2002 September,2002
2- Introduction
- Lagrange multiplier method
- Gradient method
- Numerical example
- Conclusion and Future work
3What is the shape which has minimum drag force?
4Incompressible Stokes equation
5APPROXIMATION
PSPG formulation
The weighiting residiual equation of the PSPG
formulation can be denoted by adding the
stability term of the Petrov-Galerkin formulation
to the Galerkin formulation.
The stability parameter is defined in each
element.
6APPROXIMATION
Finite element equation
7FORMULATION
Performance function
Fluid forces are directly applied to the
performance function.
F1 Drag forces F2 Lift forces
8The volume is being kept in each minimization
iteration step. To keep the volume of the body is
equal to keep the volume of the whole analytical
domain.
9FORMULATION
Lagrange multiplier method
Shape optimization problem with constraint
condition can be transformed to the minimization
problem of the extended performance function.
10FORMULATION
The stationary condition of the extended
performance function can be derived from the
first variation of that.
11FORMULATION
Adjoint equation
When the state and adjoint values are solved, the
optimal condition of this problem is given as
follows.
12FORMULATION
Gradient of the extended performance function
13MINIMIZATION
Gradient algorithm
New shape is found by deforming the surface
according to
14THE PROCESS OF OPTIMALSHAPE
15NUMERICAL EXAMPLE
Objective Drag minimization problem Constraint
conditionArea of body is constant Initial
shape Circle
16NUMERICL EXAMPLE
Finite element mesh of optimal shape
Pressure distribution of optimal shape
17NUMERICL EXAMPLE
Vorticity
Normal derivative
18NUMERICL EXAMPLE
Iteration history
19- Numerical determination of minimum drag shape
- in an incompressible flow was attempted.
- The body shape that satisfied the necessary
condition was obtained. - The present method can be applied to other flow..