Furniture: Objects or Structures?
The majority of people, as well as designers and manufacturers, tend to regard furniture principally as objects and not as structures that carry loads.
Even in the case when purely structural parts exist, i.e. parts that do not contribute to the actual functionality of the object but instead transfer loads, their design is either not separated from that of the functional part or it is dramatically simplified in favor of the manufacturing cost.
In some sense, this approach is justified since it is not reasonable to undertake a detailed structural analysis for each product we construct, which would probably increase its cost to a non-rational level. Manufacturers prefer to over-dimensionalize the structural parts using their experience or even experiment themselves via a trial and error approach, instead of relying to structural analysis techniques for the design.
In my opinion, this dominance of manufacturability over the aesthetics derived from the structural design is often the culprit for spending a great part of our lifes living next to products that are not attractive and emotionally indifferent. Separating the purely structural parts of furniture, then trying to elaborate on their shape via Computational Morphogenesis techniques could yield aesthetically interesting designs, having an emotional impact on our everyday life. Living next to objects that we appreciate or admire, thanks to a certain sense of appropriateness that their structural form conveys, may worths the additional cost to undertake.
Already pioneer designers and engineers (Joris Laarman, Morphing Matter Lab, etc.) have started experimenting towards this direction, producing interesting results. I strongly believe that the continuous progress on additive manufacturing, as well as the popularization of Computational Morphogenesis techniques, will turn such an approach all the more afordable and will attract the interest of more furniture designers.
Biomimetic design
via
CAE-based Computational MorphoGenesis
Biomimicry stands for the consious emulation of nature's genius. It has become a topic of increasing interest in a variety of fields (material science, artificial intelligence, etc.), since it permits to simulate the evolutionary rules of nature and use them to solve problems.
In nature, shape and structure are inseparable, since they are fused under the evolutionary processes. In the recent past, several works have been presented around the biological growth rules for natural structures. Analyzing the shape of trees, living shells, bones, etc. with respect to the principal loads they carry during their lifetime, researchers seem to convergence to the following conclusions:
-
Natural structures tend to evolve towards an homogeneous stress level at their boundary. This means that the "stress defines the shape", which is the inverse from the standard human design process where "the shape defines the stress".
-
For some natural structures, not only their extenal shape, but also their internal meso-scale structure, adapts according to the stress level that is exercised on it. For bones, this procedure is known as "adaptive bone mineralization".
To my knowledge, the first architect and engineer that used optimization techniques to define the structural shape was Alexandre Gustave Eiffel.
The parabolic shape of the world-known Eiffel Tower in Paris has been optimized for wind and self-weight loads using simplified models, since no computational techniques existed at that time.
Moreover, Eiffel got inspired by nature to define hierarchical structures that correspond to different scales in nature. Seen from a distance, it is difficult to observe the internal structure of the principal features defining the shape. However, approaching the Tower one can distinguish three distinct levels of stell bars (truss) that result in a very efficient use of matter.
Computational MorphoGenesis can be used to ressemble the evolutionary processes described above.
-
Constructing appropriately the optimization problem to solve, the iso-stressed condition on the boundary of the shape can be achieved. Indeed, choosing to minimize the structural compliance, i.e. the work produced by external loads, is known to result in shapes with constant internal elastic energy on their optimizable boundaries.
-
Moreover, instead of choosing an homogeneous material for the domain, one may choose an infill material that corresponds to a meso-scale structure. This infill material may also be subjected to optimization, contributing further to structural efficiency. Using 3d printing techniques, we are able to realize such designs with satisfying precision.
University of Pittsburgh | Computational Mechanics Group
Aesthetics of biomimetic design
Different aspects concerning the aesthetical part of structurally optimized structures have been exposed through time.
The majority of structural optimization supporters highlight the intrinsic value of visualizing the flow of forces. Contrary to the extreme homogeneity that dominates contemporary architecture, the exposure of the stress trajectory instantly attracts the human interest. Since we are biased towards regularity and uniformity, "natural design" (i.e. designs that mimic evolutionary processes in nature) seems unnatural.
Moreover, biomimetic design is usually identified under organic structures with smooth curved features and sophisticated networks. This is not an exception to the previously presented rules. Nature works under simple rules, but in complicated ways. Indeed, when several loadcases are taken into consideration in topology optimization problems, optimized structures quite often dispose an organic shape. Although such a shape is difficult to interprete even for experienced engineers, it emits a sense of awe to its spectators. A first explanation for this feeling can be the admiration for suceeding to decode nature. A second one, is that we admire nature for what it has suceeded without using consciousness.
Finally, we shall not neglect the implicit message on the preservation of natural ressources conveyed by lightweight optimized structures.
Numerical experimentation
As part of the Structural Optimization team at ANSYS Inc., I firmly believe that our software has now reached the necessary level of maturity to democratize Topology Optimization among designers and engineers, especially during the early stages of design conception.
Both Discovery Live and ANSYS Mechanical provide a user-friendly environment for the set-up of the problem and a complete workflow until
3d-printing (STL file) or CAD reconstruction of the optimized part, under minimal interference of the user, which facilitates the design exploration.
In the sequel, I present some personal examples produced via ANSYS software.
Exhibition Table
Suppose that you search to design a table to be used during exhibitions, i.e. an exclusive piece intended to catch the eye of the spectator.
In that case, the cost of production may be of minor importance and the designer disposes the freedom to create under minimal constraints.
Using Topology Optimization to inspire for an efficient structural system, his main concerns limit to the definition of the design-space and the corresponding structural problem, i.e.the loads and the kinematic constraints.
For example, consider the case depicted in the figure below. One quarter of the total design-space has been considered to account for double symmetry in the X and Y directions. The displacement is fixed at a small part of the bottom face and a uniform pressure is applied on the top surface, to imitate the structural behavior of a table under vertical loads.
Figure: design-space, loads and kinematic constraints for the design of a doubly-symmetric table.
Solving the Topology Optimization problem of creating a rigid structure under some predefined volume of material in ANSYS Mechanical, provides the result shown in the figure below. As one can observe in the video of the optimization process, starting from the full-domain, the shape automatically evolves in an iterative manner until converging to an optimized configuration.
The optimized shape resembles to a bridge structure, which is to be expected since the structural functionality is indeed quite similar!
However, people tend to be surprised with the result since they are used to focus on the functional part of a table instead of its structural system.
Figure: optimized shape (top-left); shape evolution process (top-right) and complete facetized design (bottom).
Modifying any aspect of the previous setting, thus altering the optimization problem, can result in a quite different optimized structure.
In fact, this is neither surprising, nor should one find it perturbing! On the contrary, when the goal is to inspire from the results of structural optimization, this characteristics dictates the designer to experiment with the available parameters.
For example, consider the previous problem but this time fix the displacement all along the side as shown in the next figure.
The optimization result differs from the previous one, placing more material at the center part of the table, as expected.
Chair
A well-known example among those experimenting on the methodology of topology optimization for furniture design, is the design of a chair.
Below, one can recognize such a chair from its purely functional parts, whereas the structural part reminds the organic shapes of living structures.
TV Table
Here is another example of how your boring TV-table could look like if it has been structurally optimized!
Figure: design-space and regions of fixed-displacement.
Figure: optimized design.
Saving material: shelves
Shelves are typical examples of excessive material waste in the design of furniture.
In fact, the structural behavior of shelves is similar to that of plates for buildings. While it is extremely costly to manufacture a plate with ribs, as shown in the sequel, it should be affordable to produce shelves, whose optimized design permits to save a great amount of material while keeping almost the same structural performance.
Next, we present such an optimized design of a double-symmetric, simply-supported shelve at its the bottom corners.
Figure: loads and kinematic constraints for a typical shelve.
Figure: optimized design, featuring ribs that visualize the flow of forces.
At that point, we shall agree that Computational MorphoGenesis cannot substitute designers!
It rather constitutes a valuable tool and methodology in their effort to improve their designs and inspire new shapes.
As Prof. Dr. Eng. Mutsuro Sasaki correctly points out:
“The method is not important, is the motivation that is important. It is a tool you need to control, if you don’t know the architectural
rules, no matter what, you get a Frankenstein.”
In collaboration with the greek Architect Engineer and Designer Mavridis Georgios, we have started our own experimentation on the topic,
in the framework of the CREA CENTERS project.