Generative Manufacturing – The True Digital Twin

moruzzi nasa 3d

This post is also available in: Italiano

Generative Design is a reality that is destined to revolutionise the approach to design and the way it operates.

The most common example to explain this concept is the formal optimisation of the chair. After “learning” a great number of chairs, a certain system is able to simulate many design solutions, thanks to artificial intelligence. We no longer start from a formal objective. We rather think of a result that matches a specific functional objective. The resulting design is often something unusual compared to the traditional concept of the chair, recalling rather complex organic shapes that could not be achieved without the 3D printing.

generative design autodesk 3d

The infographic compares the potential between the “traditional” design process and the one assisted by artificial intelligence, on which generative design software is based (credits: Autodesk)

This will be possible thanks to the huge amount of Cloud data that generative design systems will be able to process and simulate more and more accurately. But what characteristics must these data have? During the Redshift Live event in Milan, we had the opportunity to meet Massimiliano Moruzzi, a researcher at the Autodesk Centre for Computational Sciences, based in San Francisco (USA). Massimiliano illustrated the research that his group is carrying out in the context of a new discipline: Generative Manufacturing.

(FLT) – Redshift Live’s main event is called “Autonomous Manufacturing and Smart Materials”, essentially the two basic ingredients of Generative Manufacturing.

(MM) – I propose a very simple reflection to the public of Redshift Live. If we used 3D printers to do the same things we have been doing with traditional production methods only to take advantage of their direct benefits, we would probably largely underestimate the potential of Additive Manufacturing. I do not mean to diminish the merits of current applications, I am just saying that we have to start thinking of all potential developments, that is, conceiving a manufacturing system that can regulate itself autonomously, according to the environment, in the same way a natural process does.

(FLT) – I remember one of your famous articles, published on TechCrunch; there is a sentence which probably shows all the force of this concept: It’s as if materials are becoming not just smart, but “alive”.

(MM) – Of course. With new composite materials, it is possible to create systems that on top of carrying out monitoring functions, thanks to their sensors, are also able to adapt to the conditions of the environment around them. This is the most interesting aspect in terms of design. Some materials, like concrete, are heavy, easily degradable and scarcely eco-sustainable; it makes no sense to use them for the constructions of the future. Today we can produce light and economical materials that respect the environment and have active properties, such as the ability to repair themselves thanks to microcells equipped with special resins. This would increase the average lifespan of buildings, making them even more efficient in terms of eco-sustainability and safety.

Applications of this kind are already at an advanced research stage in aeronautics. If the sensors on the covering layer diagnose a crack, let’s say on the wing, the system can repair it autonomously. Airbus has long been active on this project in collaboration with the University of Bristol. We could name many others.

(FLT) – What is the relationship between Generative Manufacturing and Generative Design, and to what extent the former is functional to the latter?

(MM) – Generative Manufacturing and Generative Design together define what in my opinion should be the real Digital Twin. The current perception of the digital twin is mainly linked to a 3D model that reproduces the shape and characteristics of a physical counterpart. The software performs simulations based on this data. However accurate it may be, a digital twin conceived with this logic is inevitably limiting. It can be a first step in the digital concept, but it definitely cannot be considered an ending point, so long as we speak of true Generative Design. In fact, the generative system would be limited to all the instructions it was given during the learning phase. No matter how complete these instructions are, they will never be able to perceive and simulate the behaviour of a material in nature. The only way to overcome this limit is that information comes from the material itself, thanks to its ability to “feel” and react actively to environmental circumstances.

In my opinion, Generative Design can work on a real Digital Twin as long as it is able to synthesise and simulate all the formal and material properties of the latter. There must be no difference between the real model and the digital model. The latter cannot be a partial reproduction of the former.

Through Machine Learning we are able to instruct systems with data of any kind, including data from smart materials or “live” materials – if this definition can help illustrate the concept.

(FLT) – Once again, nature becomes the most natural source of inspiration, if you pardon the pun.

(MM) – Of course, and this is not something new. We have been working on this subject for many years: there is only one problem. So far, we started from the tree. Instead, it is necessary to start from the seed. In nature the tree is not predetermined, it assumes decisive characteristics according to the environment in which it grows. The same thing can happen in design, opening up immense creative potentials, in terms of personalising the product as much as possible. The coming of additive Manufacturing allows us to see things in a new light. Before, we could not do this.

(FLT) – Additive Manufacturing makes it possible to reproduce almost any form existing in nature, but to what extent can composite materials venture?

(MM) – It’s hard to say exactly but, surely, in the future, the applicative possibilities will always be greater, thanks to research. I would like to stress the fact that today it is possible to give shape to and to experience almost any idea; this is why innovation should not miss this opportunity at all. In the industrial sector, some standards tend to prevail. If we always stick to established patterns, we can improve a product; as for innovating, in the true sense of the word, it really becomes difficult.

(FLT) – There is a very interesting research case I am thinking about: car wiring, generated in 3D printing along with body parts. We could apply the same logic to build home systems. When will we see this kind of application?

(MM) – We are already experimenting with cars. For now, I cannot disclose many details because of the project’s confidentiality but I can give a broad outline to understand the concept on which we are working. Instead of making the structures and wiring separately, it is possible to optimise the number of components by printing them both in 3D. We simply drew inspiration from the nervous system of the human body, which is similar both in the form and in the way machines work.

materiali smart 3d protocube

A graphene “nervous system” of a car, printed in 3D. Graphene is a material with a huge potential, as it has the properties of conducting and shielding. It has an enormous mechanical resistance and can be used as a composite. When combined with materials
acting as “Functional Agents”, it allows the creation of real sentient systems… similar to the behavioural logic of a person’s neurons. (credits, see notes)

(FLT) – Cognitive systems represent another extraordinarily suggestive example: they are able to recognise and analyse complex systems like the human organism, detecting its biometric parameters

(MM) – In this kind of application, in addition to the intelligence of the computational system that receives the data from the sensors, Augmented Reality helps us to visualise the data, for example, inside a car or a house. The environments we live in can become smart. A concrete case comes, once again, from the automotive sector. Paulo Gameiro, with his HARKEN project, used the smart fabrics of the car seats to monitor in real time the heartbeat and the breathing of the driver. The goal in this case is to create a warning system to prevent drowsy driving. Based on this principle, we can develop any application at the service of safety, with very positive effects also in social terms.

generative manufacturing harken 3D protocube

The Harken smart seat project, designed by Paulo Gameiro, allows the monitoring of the driver’s biometric parameters in a non-intrusive manner. This application can tremendously innovate the field of road safety, preventing distractions, drowsy driving and other possible causes of accidents (credits: see notes)

(FLT) – Buildings’ smart skins that would adjust autonomously according to the mood and comfort of their inhabitants: it is a cognitive concept, too. Quite futuristic but certainly suggestive, above all, for the advantages that could derive from it.

(MM) – The development of these technologies raises ethical, legal and privacy issues and many other aspects related to the biomedical debate. I believe that we have to stay positive. Take, for example, the advantages of using sentient materials in the early diagnosis of particular pathologies. This is not something far in the future.

Cognitive robots can prevent diseases and save lives, provided that we can implement them in the environments in which we live. We are talking about solutions that science fiction cinema has been picturing for many years. The technology to make this real is right here. Research must absolutely go in this direction. We must not look for a resounding discovery but start to develop feasible applications that even a child could understand how important and useful they are.

generative manufactuing auto deploy 3d

SMART SKINS allow the creation of cognitive shells that can receive and process a big variety of data.
DEPLOYABLE MICRO FACTORY – A factory concept made up of a single robot and a “synthetic brain”. Together, they create free-form structures thanks to additive manufacturing. The synthetic brain, thanks to its neural algorithms, assumes a behaviour similar to that of a person, with the advantage of being entirely digital. A system of this kind is able to manage any product in its entirely digital version (a digital twin), replicating it in the physical world at any time, in any place, thanks to the robot and 3D printing (credits: see notes)


Where not otherwise specified, the images are taken from the presentation “The Future of Making Things” by Massimiliano Moruzzi. The rights of the images belong to their respective owners.

In the cover image, Massimiliano Moruzzi is at work in the experimental laboratories of NASA Swamp Works, where construction systems on Mars are being developed by recovering raw materials directly on the red planet. This result is made possible thanks to systems based on AI, Robotics and 3D Printing.

This post is also available in: Italiano

Related Articles

About author View all posts Author website

Francesco La Trofa

Architect and journalist with 20 years’ experience in 3D technologies.
Consultant to public entities and 3D businesses for aspects relating to design and communications.
Head of editorial content at and co-founder of Digital Drawing Days, the only event of its kind in Italy.
Actively involved in research and teaching at Milan Polytechnic.
Edits 3D STORIES for Protocube Reply.