3DS.COM © Dassault Systèmes | Confidential Information | 4/9/2015 | ref.: 3DS_Document_2014 FLOW TOPOLOGY OPTIMIZATION OF A TURBO CHARGERS INFLOW DUCT Dr. Jens Iseler [email protected] 17.03.2015
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014 FLOW TOPOLOGY
OPTIMIZATION OF A TURBO
CHARGERS INFLOW DUCT
Dr. Jens Iseler
17.03.2015
2
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014 1 Problem Description
2 Method Description
3 Topology Optimization
4 Shape Optimization
5 Summary
Overview
5
3
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Problem Description
Provide compressor inflow duct with:
low pressure loss
increased flow uniformity
Procedure:
Topology optimization (draft design)
Reconstruction via CAD tool
Shape optimization
Existing inflow duct
4
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Applied Workflow
Topology optimization based on
optimality criteria
Reconstruction of obtained design
Shape optimization based on
adjoint solution
5
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Topology Optimization For CFD ProblemsOptimization problem is based on the
(meshed) available design space
initi
al d
esig
n
6
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Optimization problem is based on the
(meshed) available design space
Geometric variation is achieved by
sedimenting individual cells
initi
al d
esig
npo
ssib
le v
aria
nts
Topology Optimization For CFD Problems
7
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Optimization problem is based on the
(meshed) available design space
Geometric variation is achieved by
sedimenting individual cells
An individual design proposal can be
derived based on the collectivity of all free
(= non-sedimented) cells
Optimization represents coupled run. New
design available after 1 single run
initi
al d
esig
npo
ssib
le v
aria
nts
Topology Optimization For CFD Problems
8
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Tosca Fluid Optimization – Sedimentation Process
9
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Tosca Fluid Setup
Optimization Setup
Optimization approach: Optimality criteria
Elimination of recirculation zones
Defined iteration number: 20000
Mesh with 400 K elements
Simulation time: 20.0 hrs. with 4 CPU Initial geometry Design space vs.
existing design
Reference streamline
10
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Tosca Fluid Setup
Optimization Setup
Optimization approach: Optimality criteria
Elimination of recirculation zones
Defined iteration number: 20000
Mesh with 400 K elements
Simulation time: 20.0 hrs. with 4 CPU
Postprocessing
Extraction of areas with low velocities
Smoothing of the remaining geometry
Initial geometry Design space vs.
existing design
Reference streamline
11
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Tosca Fluid Design
Fully 3D design proposal
Cross section area bigger compared to existing design
Section plane
12
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Tosca Fluid Design
Fully 3D design proposal
Cross section area bigger compared to existing design
Result: Overall smaller velocities reduced pressure
drop likely
Section plane
Tosca Fluid design
Existing design
Velocity distribution
13
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Why?
Obtained design fully 3D
Manufacturing constraints must be considered
Design may contain rough areas
Reconstruction of Optimized Design
14
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Why?
Obtained design fully 3D
Manufacturing constraints must be considered
Design may contain rough areas
Strategy:
Maintain global shape of optimized design
Adjust shape locally (geometric constraints, spikes)
Reconstruction of Optimized Design
15
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Reconstruction of Optimized Design
16
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Reconstruction of Optimized Design
17
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Reconstruction of Optimized Design
18
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Reconstruction of Optimized Design
19
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Reconstruction of Optimized Design
20
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Flow Performance – Existing Design
Cone outlet plane
21
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Flow Performance – Optimized Design
Cone outlet plane
22
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Comparison Optimized Design – Existing Design Total pressure loss + uniformity
Optimized
design
Existing design
Total pressure
loss
426 Pa 495 Pa
Uniformity 0.982 0.978
Optimized design Existing design
23
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Shape Optimization With STAR-CCM+®
Idea:
Topology optimization delivers design based on
empiric optimality criteria
Further potential of improvement by usage of
gradient method
Efficient solving of gradients/sensitivities by means
of adjoint solution
24
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Shape Optimization With STAR-CCM+ Idea:
Topology optimization delivers design based on
empiric optimality criteria
Further potential of improvement by usage of
gradient method
Efficient solving of gradients/sensitivities by means
of adjoint solution
Strategy:
Define a closed loop including adjoint solver and
morphing module
Deformation dependent on computed sensitivities
Initial geometry = topology optimized design
25
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Shape Optimization With STAR-CCM+
Pressure drop: -45%
26
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Shape Optimization With STAR-CCM+ Objective function: Pressure drop
Creation of morphing boxes via Lattice
Overall 500 control points
STAR-CCM+ macro:
Calls primal solver, adjoint solver and morpher for
predefined number of loops
Considers maximum allowed deformation through
scaling of sensitivities
Gradual adjustment of scale factor dependent on
behavior of objective function
27
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Shape Optimization With STAR-CCM+
Topology
optimization
Shape
optimization
Total pressure
loss
426 Pa 389 Pa
Uniformity 0.982 0.983
28
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Shape Optimization With STAR-CCM+
Topology optimized
design
Shape optimized
design
29
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Workflow Based on Non-Parametric Optimization
Topology optimization based on optimality criteria
Reconstruction of obtained design
Shape optimization based on adjoint solution
30
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Summary Objective: Duct Flow with low pressure drop and increased uniformity
Tosca Fluid topology optimization based on available design space
Reconstructed optimized design reveals a significant total pressure loss
reduction (-14%) and increased uniformity (from 0.978 to 0.982)
STAR-CCM+ macro developed to run primal solver, adjoint solver and morpher
for predefined number of cycles. Considers maximum allowed deformation
through scaling of sensitivities
STAR-CCM+ shape optimization based on adjoint solution led to further
reduction of the total pressure loss (-7.5%) and slightly increased uniformity
31
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
32
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Shape Optimization With STAR-CCM+
Pressure drop: -45%
33
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Shape Optimization With STAR-CCM+
34
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
Shape Optimization With STAR-CCM+
35
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
CFD setup
Topology-Optimization
Solver: STAR-CCM+ 9.06.009
Physics:
Incompressible
Steady
k-ε turbulence, All y+
Boundary conditions:
Inlet: Stagnation inlet
Outlet: Mass flow inlet
Verification
Solver: STAR-CCM+ 9.06.009
Physics:
Incompressible
Steady
k-ε turbulence, All y+
Boundary conditions:
Inlet: Massflow inlet
Outlet: Pressure outlet
36
3DS
.CO
M©
Das
saul
tSys
tèm
es| C
onfid
entia
l Inf
orm
atio
n | 4
/9/2
015
| ref
.: 3D
S_D
ocum
ent_
2014
CFD setup
Topology-Optimization
Solver: STAR-CCM+ 9.06.009
Physics:
Incompressible
Steady
k-ε turbulence, All y+
Boundary conditions:
Inlet: Stagnation inlet
Outlet: Mass flow inlet
Verification
Solver: STAR-CCM+ 9.06.009
Physics:
Incompressible
Steady
k-ε turbulence, All y+
Boundary conditions:
Inlet: Massflow inlet
Outlet: Pressure outlet
Related Documents
