A new technique makes the construction of complicated materials simpler.

Researchers may swiftly create a variety of cellular metamaterial structures with distinct mechanical properties using a new, user-friendly interface.

A novel method unifies several cellular metamaterial building pieces into a single graph-based representation. Using this, it is possible to create an intuitive interface that makes it simple to model metamaterials, modify their structures, and simulate their properties. Finding materials with unique, desirable attribute combinations is an ongoing effort for engineers. A material that is porous and friendly to biomechanics may be utilized for bone implants, for instance, or ultra-strong, lightweight materials might be employed to increase the fuel efficiency of vehicles and airplanes. Artificial structures made of units, or cells, that repeat are known as cellular meta materials. These objectives can be accomplished with the aid of cellular metamaterials, which are synthetic structures made up of units, or cells, that repeat in different patterns. But it might be challenging to predict which cellular configuration would result in the desired qualities. There are countless configurations to take into account, even if one concentrates on constructions built of simpler building components like linked beams or thin plates. Therefore, only a small portion of the conceivably conceivable cellular metamaterials can be manually explored by engineers. A computer method created by researchers from MIT and the Institute of Science and Technology Austria makes it simpler for users to quickly design a metamaterial cell from any of those smaller building components and then assess the properties of the finished metamaterial. Their method, similar to a specialized computer-aided design (CAD) system for metamaterials, enables an engineer to swiftly model even very complicated metamaterials and experiment with designs that could have otherwise taken days to construct. Since all the necessary building elements are available to the user, they are also able to explore the whole space of potential metamaterial shapes thanks to the user-friendly interface.

"We developed a representation that can encompass all of the many shapes in which engineers have historically showed interest. According to Liane Makatura, a graduate student at MIT studying electrical engineering and computer science and the co-lead author of a paper on this method, "since you can design them all the same way, it implies you can switch between them more fluidly. Together with co-lead author Bohan Wang, a postdoc at MIT, senior author Wojciech Matusik, a professor of electrical engineering and computer science at MIT who directs the Computational Design and Fabrication Group within the MIT Computer Science and Artificial Intelligence Laboratory, and Yi-Lu Chen, a graduate student at the Institute of Science and Technology Austria (ISTA), they wrote the paper. At SIGGRAPH, the study will be presented

A united approach

A scientist usually starts by selecting a representation that will be utilized to explain her prospective designs while creating a cellular metamaterial. The collection of shapes that can be explored depends on this decision. She might opt for a method that uses numerous interconnected beams to depict metamaterials, for example. This, however, hinders her from investigating metamaterials based on different elements, like thin plates or three-dimensional objects like spheres. These forms are provided via various representations, but there isn't yet a single approach that can be used to describe all shapes.You restrict your exploration and impose a bias based on your intuition by pre-selecting a particular subspace. While this can be helpful, intuition can be wrong, and for your specific application, it might have also been worthwhile to explore some of the other shapes, according to Makatura.She took a step back and carefully investigated several metamaterials with her team. They saw that the shapes that make up the overall structure could be simply described by lower-dimensional shapes; for example, a beam could be reduced to a line or a thin shell to a flat surface.They also observed that cellular metamaterials frequently contain symmetries, necessitating the representation of only a small portion of the entire structure. By rotating and mirroring the first component, the rest can be put together. We came to the conclusion that cellular metamaterials might be effectively represented as a graph structure by integrating those two observations, the author writes.Users construct a metamaterial skeleton using their graph-based representation by using construction blocks made of vertices and edges. For instance, one places a vertex at each end point of the beam and connects them with a line to construct a beam structure.

When the user uses a function over that line to set the beam's thickness, the thickness of the beam can be changed such that some parts are thicker than others. Similar steps apply when dealing with surfaces; first, the user marks the vertices of the most crucial features before selecting a solution that infers the rest of the surface.

Users can also quickly build a triply periodic minimum surface (TPMS), a highly complex type of metamaterial, using these simple-to-use solvers. These structures are extraordinarily strong, but their typical development is difficult and prone to mistakes. "With our illustration, you can begin fusing these shapes. You might find fascinating qualities in a unit cell that has both a TPMS structure and a beam structure. But those combinations haven't really been thoroughly studied up until now," the author claims. The system outputs the full graph-based approach, including all the vertices, edges, solvers, transformations, and thickening operations the user used to arrive at the final structure. Designers can preview the present structure at any stage of construction and immediately forecast specific characteristics, such as stiffness, within the user interface. Once an acceptable design is found, the user can iteratively adjust various parameters and reevaluate the result.

An approachable framework

The researchers' technique was utilized to recreate structures that covered a wide range of distinctive metamaterial classes. Each metamaterial structure was created in a matter of seconds after the skeletons were created. Additionally, they developed automatic exploration algorithms, giving each one a set of guidelines before releasing it into their system. In one experiment, a computer program returned more than 1,000 potential truss-based constructions in approximately an hour. The team also carried out a user study with 10 people who had minimal prior modeling experience with metamaterials. All six of the offered structures could be successfully modeled by the users, and most of them thought that the procedural graph representation facilitated the task. "Our representation increases peoples' access to various structures. Our satisfaction with users' capacity to produce TPMS was particularly high. Even for specialists, it is typically difficult to generate these intricate structures. Nevertheless, out of all six structures in our investigation, one TPMS had the shortest average modeling time, which surprised and excite the researcher. The researchers intend to improve their method in the future by including more intricate skeleton thickening techniques, allowing the system to mimic a larger range of shapes. Additionally, they aim to keep investigating the application of autonomous generation methods. Long-term plans call for the use of this system for inverse design, in which one specifies the required material qualities and then employs an algorithm to identify the best metamaterial structure.

The Defense Advanced Research Projects Agency (DARPA), the MIT Morningside Academy Design Fellowship, the National Science Foundation Graduate Research Fellowship, an ERC Consolidator Grant, and the New Sat project all contributed to the funding of this study.