Overview of 3D printers
A 3D printer is a device for creating a three-dimensional object from a digital model, whereby materials are layered to produce an object based on a computer-designed 3D model. This process is called additive manufacturing. The most common materials used are plastics, but metals, ceramics, resins, foodstuffs and even biomaterials are also used.
The way a 3D printer works is that a 3D model is first designed using computer-aided design (CAD) software, exported to a standard 3D printing format such as STL (Standard Tessellation Language), then the 3D model is analysed by the slicing software and divided into hundreds or thousands of thin horizontal layers. The slicing data is then sent to a printer, which laminates each layer with material to create the final 3D object.
Types of 3D printers include.
1. fused deposition modelling (FDM): thermal melting and lamination. Plastic filaments are heated and the molten material is stacked layer by layer, widely used in household and industrial applications.
2. SLA (Stereolithography): a light modelling method that uses UV light to selectively cure light-curing resins to form objects layer by layer, enabling high-precision modelling.
3. selective laser sintering (SLS): selective laser sintering. SLS (Selective Laser Sintering): uses a laser to sinter powder materials to form objects layer by layer and can be used for a wide variety of materials.
4. digital light processing (DLP): digital light processing. Uses a projector to cure light-curing resins, similar to SLA but with a different light source.
5. direct metal laser sintering (DMLS): direct metal laser sintering. Metal powders are sintered with a laser to create metal parts, which are used when high-strength parts are required.
6. giant 3D printers utilising liquid crystal panels: these use high-resolution liquid crystal panels to create 3D objects by selectively curing light-curing resins.
Applications include (1) prototyping used to validate and improve designs in the early stages of product development; (2) for medical applications used to manufacture custom medical devices, prostheses and implants; (3) for aerospace applications used to produce lightweight, high-strength components; (4) for automotive components. used for prototyping and manufacturing; (5) for architectural modelling and the manufacture of actual building materials; and (6) for the production of accessories and artwork.
The advantages of using 3D printers may include
- Freedom of customisation: products can be easily tailored to individual needs based on digital designs.
- Rapid prototyping: reduced time from design to manufacture, enabling rapid product development.
- Efficient use of materials: only necessary materials are used, thus minimising waste.
- Realisation of complex geometries: complex geometries that are difficult to create using traditional manufacturing methods can be created.
On the other hand, the following challenges exist
- Material constraints: the range of materials available is limited and it is sometimes difficult to find the right material for a particular application.
- Cost: industrial 3D printers are expensive, especially for high-precision applications, and operating costs are also high.
- Production speed: depending on the size and complexity of the object, production can be slow.
- Quality control: due to the stacked manufacturing process, the adhesive strength and surface finish between the layers can vary.
3D printing is seen as an innovative manufacturing technology in many fields and is a manufacturing method that has a significant impact on product development and manufacturing processes.
Combining 3D printers and generative AI
Combining such 3D printers with generative AI, which has been attracting attention in recent years, can revolutionise the manufacturing process by combining digital manufacturing and AI technology, enabling highly automated processes from design to manufacture, and the discovery of new modelling methods and designs. Specific examples of possible combinations are as follows.
1. automatic design generation: the AI can generate an optimised 3D design based on input requirements (shape, function, material, etc.) and output that design directly on the 3D moulding machine. For example, the user could input their requirements to the AI, which would then draw inspiration from existing databases and propose its own design. These would enable the creation of shapes and structures that human designers had not thought of.
2. design optimisation: a generative AI could use machine learning algorithms to optimise the performance of 3D objects, such as their mechanical properties or weight. This allows, for example, AI to automatically optimise the geometry in the design of parts with specific strength or weight reduction requirements, which can then be actually manufactured by 3D printing, a process also known as ‘topology optimisation’, which is used in the aerospace and automotive industries.
3. reverse design (reverse engineering): physical objects can be scanned and the AI can reconstruct the original design based on this data. The scanned data is processed by AI to create an optimised 3D model, which is then reproduced by a 3D printer, thereby streamlining the repair of damaged parts and the redesign of custom parts.
4. generation of complex geometries: generative AI can generate complex geometries and structures that are difficult to produce using conventional design methods, thereby enabling 3D printers to realise complex designs that cannot be manufactured using traditional methods. In particular, AI is increasingly being used in the medical sector to generate implants and prostheses customised for each patient and created by 3D printers.
5. material optimisation and tailoring: an AI can select the materials to be used and suggest optimal formulations of composite materials, thereby reducing material costs and improving product durability and flexibility when creating high-performance products with a 3D printer. For example, a generative AI can design objects using new materials and 3D print them to produce products that are stronger and more durable than before.
6. real-time quality control with AI: AI can monitor in real-time during the 3D printing process to detect and correct print progress and errors, and AI can monitor the printing process and automatically correct material output and shape deviations, enabling highly accurate modelling.
7. an AI-learning design process: the AI can learn feedback from previously generated 3D designs and moulding to further streamline the design and manufacturing process for the next time. The evolution of this ‘generative AI + 3D printer’ loop will shorten the time-consuming prototyping process and enable rapid completion of the final product.
In this way, the combination of generative AI and 3D moulding machines will promote innovation in a wide range of fields, including manufacturing, design and medicine.
Application examples of combining 3D printers and generative AI
Specific applications of the combination of 3D printers and generative AI are described below.
1. topology optimisation of parts in the aerospace industry: parts for aircraft and spacecraft must be lightweight yet strong. An attempt is being made to use generative AI to optimise the topology of parts, automatically generate designs that are both lightweight and strong, and manufacture them using a 3D printer. This is expected to result in shapes that cannot be obtained using conventional design methods, leading to improved fuel efficiency and reduced costs.
Example: Airbus has 3D printed AI-optimised metal parts to achieve weight savings and cost reductions compared to conventional parts.
2. production of custom prostheses and implants in the medical sector: the combination of generative AI and 3D printing is also being applied to the production of custom prostheses and implants for individual patients: AI generates an optimised design based on the patient’s 3D scan data and 3D printers produce implants and assistive devices. This enables precise customisation to suit the individual patient’s body shape and medical needs.
Example: the Italian company LimaCorporate uses AI to design bone implants and manufacture them with 3D printers, improving surgical success rates.
3. ingeniously designed garments and accessories in the fashion industry: the fashion industry is increasingly using 3D printers to create uniquely designed garments and accessories generated by AI. Generative AI is able to create new patterns and structures without being bound by traditional design rules, and uses 3D printers to produce fabrics and accessories directly. This facilitates the creation of custom-made and limited edition products.
Example: designer Iris van Herpen uses AI and 3D printers to produce dresses with complex geometric shapes.
4. design and manufacture of automotive parts: the automotive industry is also increasingly using generative AI to design lightweight, high-strength parts and manufacture them with 3D printers; AI learns the loads and material properties of parts and automatically generates the optimum structure, thereby improving fuel efficiency and reducing manufacturing costs. The use of generative AI is particularly important for electric vehicles (EVs), as the weight of the battery must be taken into account in the design.
Example: the Ford Motor Company 3D prints AI-optimised parts to reduce weight and improve durability.
5. auto-generated designs in the architectural sector: in the architectural sector, projects are underway where AI can automatically generate optimal designs for the site and environment, and 3D printer parts or models of these designs. In particular, complex shaped buildings and free-form interior designs are being generated by generative AI, and AI can also make buildings safer and more eco-friendly by taking into account the building’s seismic and energy efficiency.
Example: in Shanghai, China, building components designed by AI are output by 3D printers and used in practice. While enabling futuristic designs, it also reduces manufacturing costs.
6. customised manufacturing of consumer products: in the manufacture of consumer products, AI is increasingly generating personalised product designs based on individual requirements and manufacturing them using 3D printers. This allows consumers to tailor-make furniture and appliances to their own tastes.
Example: platforms such as Shapeways use generative AI and 3D printing to make it easy to create custom furniture and décor.
7. prototyping in education and research: the combination of generative AI and 3D printing is also being used in education and research, where AI is accelerating the learning process by automatically modelling designs and ideas from students and researchers, which can then be prototyped using 3D printing. AI is also a powerful tool for deepening understanding, as complex theories and models can be visualised by AI and experienced as concrete objects in 3D printing.
Example: some universities are using AI to 3D print structures needed in chemistry and physics research for experiments.
Applying generative AI and GNN to 3D printers
Graph Neural Networks (GNNs) are a type of deep learning that learns using graph structures consisting of nodes (points) and edges (lines) of data, and the combination of this technology, generative AI and 3D printers can enable complex structures and dynamic optimisation and enable new design and manufacturing processes. Specific examples of their application are described below.
1. materials design and optimisation: GNNs can treat chemical molecules and crystal structures as graphs in materials science, allowing AI to learn the properties of materials and suggest optimal designs. For example, the structure of material molecules can be represented as a graph, and the data can be used to design new composite materials or optimise materials for strength and flexibility.
Application example: a generative AI uses GNN to design a composite material suitable for a specific application, which is then manufactured using a 3D printer. For automotive and aerospace parts that require strength, heat resistance and light weight, GNNs are used to optimise the material design and 3D print the part.
2. advanced topology optimisation: topology optimisation is a technique commonly used in 3D printer-based design, where the aim is to optimise the strength and performance of a shape; by using GNNs, more complex and functional shapes can be generated than with conventional regular methods, and a generative AI can suggest their structure The system is able to.
The GNN here represents the structure as a graph, learns how each node or edge affects each other mechanically, and uses this data to automatically optimise the distribution of materials, thickness and placement of cavities.
Application example: a GNN learns the mechanically optimal shape of a machine part or building, and a generative AI proposes a new design. The 3D printer then produces a component that reduces weight while maintaining a high level of strength. 2. structurally complex shapes (e.g. beehive structures or designs with many cavities) can be automatically generated by GNNs and manufactured by 3D printers to improve the efficiency and durability of the material.
3. custom design in the medical field: the GNN can be applied to the medical field, for example in the design of custom implants and prostheses for each patient, by modelling complex structures such as bones, joints and blood vessels as graphs, learning their interactions and using this information, the generative AI can optimise the It can suggest designs, which can then be manufactured with precision using a 3D printer.
Application example: bone structure is analysed using GNN, and the generative AI automatically generates an implant with the optimum shape for the patient. It is then manufactured by a 3D printer to provide surgical implants optimised for individual patients. Learning complex biological tissue, such as the vascular structure of the brain, as a graph structure, for surgical simulation and custom design.
4. generation of complex mechanical structures and network designs: in the design of complex mechanical structures, pipelines and network structures (e.g. piping systems, cooling systems, electrical circuits, etc.), GNN can efficiently learn complex interactions and automatically generate optimal designs. Furthermore, the generative AI proposes designs based on the results, and 3D printers manufacture the actual physical parts, resulting in efficient and reliable structures.
Application example: utilising GNNs to optimise complex mechanical parts, such as the internal structure of an engine or a cooling system. Generative AI proposes new structures and the parts are manufactured using 3D printers. GNN analyses the efficiency of large piping systems and energy networks, and the generative AI proposes the best arrangement. Based on this, 3D printed components can be manufactured, facilitating their implementation in the field.
5. integration of bioprinting and AI: Bioprinting is a technology that uses cells and biomaterials to create biological tissue with a 3D printer, which uses GNNs to model the interactions between cells and the structure of the tissue as a whole, enabling the generative AI to design an optimal tissue structure based on this. This will enable the generation of advanced customised tissues in the fields of artificial organs and regenerative medicine.
Application example: a GNN analyses cell interactions, and a generative AI designs organs such as the heart and liver based on this analysis; a 3D bioprinter creates artificial organs based on these designs, which can then be applied in transplant medicine; a 3D bioprinter can be used to create artificial organs based on these designs, which can then be used in transplant medicine.
6. learning design feedback using knowledge graphs: GNNs are also suitable for managing and learning the vast amount of technical data and feedback generated during the design process as ‘knowledge graphs’. Generative AI can use these data to suggest new design processes and technical improvements, which can then be efficiently reflected by 3D printers, thereby enabling the development of more advanced products while learning from past design failures and technical improvements.
Application example: in the automotive industry, for example, GNN analyses past design processes and product reviews, and the generative AI proposes improvements based on them; the improvements are then commercialised by the 3D printer, enabling a rapid feedback loop.
reference book
Reference books are described below.
1. “Fabricated: The New World of 3D Printing” (Hod Lipson, Melba Kurman)
2. “Generative Design: Visualize, Program, and Create with JavaScript in p5.js” (Anna Carreras, Jared Tarbell)
3. “Artificial Intelligence for Robotics and Autonomous Systems” (Francisco Martín Rico)**
4. “Graph Representation Learning” (William L. Hamilton)
5. “Deep Learning on Graphs” (Yao Ma, Jiliang Tang)
3. “Graph Neural Networks: Foundations, Frontiers, and Applications” (Zonghan Wu, Shirui Pan)
コメント