October 30, 2017 | Author: Anonymous | Category: N/A
Matthijs Langelaar Adjust part itself. 3. Add support structures. Topology optimization. Additive ......
29-3-2016
Topology Optimization for Additive Manufacturing State of the Art and Challenges
Matthijs Langelaar Structural Optimization & Mechanics Delft University of Technology
[email protected] Additive World Conference 2016
1
29-3-2016
2
29-3-2016
Additive manufacturing: focus on design
AM enables the fabrication of “almost any” design.
So.. what design to make?
From functionality to product Desired functionality
Final component
Topology optimization
Postmachining
Concept geometry
Detailed design
Additive manufacturing
3
29-3-2016
Aligned advantages Topology Optimization
Additive Manufacturing
• Design freedom: part performance not limited by imagination of designer
• Design freedom: relatively few shape restrictions, ‘complexity for free’
• Time to market: fast, nearly automated design process
• Time to market: no tooling needed, on-demand production
• Customization: tailored designs for specific requirements
• Customization: produce many different part at once
SLM limitation: critical overhang angle
Clijsters et al, 2012
4
29-3-2016
Existing solutions to overhang problem 1. Adjust part orientation 2. Adjust part itself 3. Add support structures
Design for manufacturing Desired functionality
Final component
Topology optimization
Postmachining
Concept geometry
Detailed design
Additive manufacturing
5
29-3-2016
Topology Optimization for Additive Manufacturing • Aim: include overhang restrictions in topology optimization • Benefits: • No need for support structures: less material usage • Less pre-processing for AM Matthijs Langelaar
• Less post-machining: faster production, lower costs Structural Optimization & Mechanics Delft University of Technology
[email protected] Additive World Conference 2016
Outline • Motivation • Brief introduction to topology optimization • Print-ready topology optimization • Approach • Simplified AM process model
• Examples • Next steps • Concluding remarks
6
29-3-2016
Topology optimization: generating the best material distribution What shape to use?
Where to place material?
bracket design domain topology optimization result post-processed final design
Topology optimization process 1. Define problem: - Objective, constraints - Domain, boundary conditions - Loadcases
2. Discretize and parameterize material distribution 3. Optimize material distribution for best performance
Maximize stiffness Use only 50% material
i
Load
4. Evaluate / fine-tune result (postprocessing, shape optimization)
7
29-3-2016
Topology optimization loop
New Values of the density variables
Component analysis (FEA)
New Values of the objective and constraints
New Gradient information (design sensitivity) Optimization algorithm
Example: compliant mechanism design
• Maximize desired motion • Sufficient stiffness in other directions
8
29-3-2016
Outline • Motivation • Brief introduction to topology optimization • Print-ready topology optimization • Approach • Simplified AM process model
• Examples • Next steps • Concluding remarks
Current practice
9
29-3-2016
Print-ready topology optimization
Build direction
Comparison
10
29-3-2016
Previous approaches • Automatic post-processing
Leary et al., 2014
• Suppressing overhang using filtering techniques
Gaynor and Guest, 2014
Previous attempts @ TU Delft • Filter-based approach
Serphos, 2014
• Boundary angle constraints
Driessen, 2015
• Boundary angle constraints with level sets
Van de Ven, 2015
11
29-3-2016
Topology optimization for print-ready designs
Component analysis (FEA)
Values of the density variables
Values of the objective and constraints
Gradient information (design sensitivity) Optimization algorithm
Topology optimization for print-ready designs
printed design
Component analysis (FEA)
Printing process simulation
blueprint design
Values of the objective and constraints
Gradient information (design sensitivity) Optimization algorithm
12
29-3-2016
AM process model
F Build direction
45 critical overhang angle assumed
AM process model formulation
0.5 0.8 0.5 Build direction
0.3
0.0
support max 1 , 2 , 3
print min blueprint , support
13
29-3-2016
AM process model formulation
support max 1 , 2 , 3
print min blueprint , support
Langelaar, 2016, in review
Topology optimization for print-ready designs
printed design
Component analysis (FEA)
Printing process simulation
blueprint design
Values of the objective and constraints
Gradient information (design sensitivity) Optimization algorithm
14
29-3-2016
AM process model: implementation • min/max operations are not differentiable: replace by smooth approximations
• Layer-by-layer processing: printing simulation in build direction, sensitivity analysis in reverse direction
• Computational cost: very minor (1%)
Outline • Motivation • Brief introduction to topology optimization • Print-ready topology optimization • Approach • Simplified AM process model
• Examples • Next steps • Concluding remarks
15
29-3-2016
Examples • Process simulation test • 2D validation • 3D validation
Process simulation test
Printed design (ideal) Build direction
Blueprint design • Solid parts fully correct • Light gray parts gradually fade out
Printed design (smoothed process model)
16
29-3-2016
2D validation: test problem
N E
W S
• Maximize stiffness • 50% material
2D validation: printability of reference design
17
29-3-2016
Topology optimization for AM
Printable, self-supporting designs achieving near-ideal performance
100%
94%
90%
100.0%
99%
18
29-3-2016
Higher resolution test
100%
93%
94%
100%
98%
3D validation
• Maximize stiffness • 30% material • 6 orientations
F
Langelaar, 2016, in review
19
29-3-2016
Reference design
Printability of reference design
Build direction
As printed
20
29-3-2016
Topology optimization for AM
Build direction
100% printable
Different part orientation
Build direction
F
21
29-3-2016
Design with ‘support structures’
Different orientations, different designs
99%
93%
101%
100%
100% 93%
102%
22
29-3-2016
Beyond beams
Reference
Printable
Printable
Build direction
Outline • Motivation • Brief introduction to topology optimization • Print-ready topology optimization • Approach • Simplified AM process model
• Examples • Next steps • Concluding remarks
23
29-3-2016
Limitations
Current approach fast and effective, but:
• Based on structured, regular mesh • Fixed 45 critical angle • Limited to 6 main build directions • Only fully supported designs, no control over performance vs. support structure cost • No consideration of stress, distortion, overheating
Current developments • Formulation that • works on arbitrary meshes • for any critical angle • and any build direction
Emiel van de Ven
• Formulation that allows tradeoff solutions between support structure cost and part performance • Development of more advanced thermomechanical AM process models Marius Knol, Can Ayas
24
29-3-2016
Concluding remarks • Topology optimization can provide the designs needed to fully benefit from AM freedom • Topology optimization for AM generates fully printable optimized designs: this eliminates the need and costs of part redesign, supports, postprocessing • Including AM restrictions can maintain high design performance • Methods hopefully soon adopted by commercial software companies
Topology Optimization for Additive Manufacturing
Matthijs Langelaar Structural Optimization & Mechanics Delft University of Technology
[email protected] Additive World Conference 2016
25
