mid2SUPAERO/RESUMEaPAPER

TOPCFAO

Opened this issue · 0 comments

How TOPOPT can discuss with Fiber Orientation (ALM) ?

Link

https://www.mdpi.com/2079-6439/7/2/14

Author and organization

by Delin Jiang 1, Robert Hoglund 2 and Douglas E. Smith 1,*
1 Mechanical Engineering Department, Baylor University, Waco, TX 76798, USA
2 Altair Engineering, Troy, MI 48083, USA

Year/publication/conference/workshop

Fibers 2019, 7(2), 14; https://doi.org/10.3390/fib7020014

Abstract

This paper presents a topology optimization approach for continuous fiber angle optimization (CFAO), which computes the best layout and orientation of fiber reinforcement for AM structures. Statically loaded structures are designed for minimum compliance where the adjoint variable method is used to compute design derivatives, and a sensitivity filter is employed to reduce the checkerboard effect. The nature of the layer-by-layer approach in AM is given special consideration in the algorithm presented. Examples are provided to demonstrate the applicability of the method in both two and three dimensions.

Innovation/contribution. What is new in this paper?

The classical Compliance minimisation is updated with local fibre orientation (each pixel)

Problem

Short description of method used

Of course sensitivities are "analytical"

Sensitivities

and a filter is used through iterations

Filter

Experiment and result

I will extract one nice result on this test case (cantilever beam)

fibers-07-00014-g009-550
Figure 9. Cantilever beam with point load, printed in three different directions.

According to the printing direction you will obtain 3 cases:

fibers-07-00014-g010
Figure 10. Isometric views of optimal element density values for three different print directions: Cases 1, 2, and 3. Elements with ρe≤0.5 are not shown.

and of course 3 results, I'll briefly comment the case 2

Results through layers 1 to 6

fibers-07-00014-g014-550
Figure 14. Optimized material distribution and material orientation, layer-by-layer plots for Case 2. Layer 1 corresponds to the bottom layer, and layer 10 corresponds to the top layer.

For information the front view

fibers-07-00014-g013-550

Figure 13. Topology result for Case 2 (front view). Elements with ρe≤ 0.5 are not shown.

What is discussed in this paper (discussion section)? What is the future work?

The 3D example problems illustrated the significant difference in minimum compliance that can be obtained based on print direction. It was also shown that the case where the plane of material alignment was parallel with force direction gave the lowest compliance. Furthermore, printing the structure at different plane directions gave very different compliance results, which is important for designers when considering the potential loading scenarios. In our example, the minimum compliance was reduced by 63% by selecting a different print orientation. Finally, fiber orientation generally followed the outer contour of the dense material region for each layer.

Your comments

Is it fully compatible with top88 framework and so toggp?

In Case 1, material angle design variables allowed the preferred material direction to rotate in a plane that had a larger effect on the structure’s minimum compliance. However, for Case 2 and Case 3, the plane of material angle was normal to the direction of the force, resulting in an optimization that yielded less improvement in terms of the compliance. These results show that the relationship between the applied loads and print plane orientation has a significant effect on the outcome of the topology optimization.

Experimental test available

MechTest

As shown in Table 2, samples optimized and printed with a horizontal bead were 15.5% stiffer than samples optimized and printed with a predominant vertical bead. This result is to be expected since horizontal stiffening is more efficient than vertical stiffening under the given bending load. Table 2 also shows that CFAO optimized MBB beams performed better than either horizontal or vertical optimized and printed beams. It is seen that the CFAO beam was 29.9% stiffer than the vertical optimized and printed structure and 12.4% stiffer than the horizontal optimized and printed structure.

Table2

Paper you should read next

Nice paper, with code available ? inside these thesis
https://baylor-ir.tdl.org/handle/2104/10163
https://baylor-ir.tdl.org/handle/2104/9821

Results can be compared with AFP from Alicia Kim and Enrico's work, Gabriele + Bhat's work, and Edouard's PHD