The next generation of design

When the term artificial intelligence (AI) is uttered we tend to think of super smart computers with malevolent intentions such as HAL in the film 2001, but there are many more everyday circumstances in which these advances are being applied. AI is going to impact areas of life not immediately obvious, but one area where there is tremendous scope for this technology is manufacturing.

Combined with ever more agile engineering advances such as 5-axis machining and additive manufacturing, AI has the potential to provide highly flexible production methods able to produce components and products in much more fluid and non-geometric forms than ever before. This also means components greatly optimised to provide the required mechanical and operational functions.

At software developer Autodesk, the futurists believe this will have a revolutionary impact on production throughout workflow, from the first initial designs down to the factory floor.

Francesco Iorio, head of computational science research at Autodesk, says: “AI can actually blend the two sides of producing parts; the design side (CAD) element and manufacturing (CAM). It can work from the design intent through the design workflow and into manufacturing. At the moment, there is a significant divide between design engineering and the factory floor. Often this means that CAD designs and CAM programs have to be modified many times until they make it to production. This is because they have to ultimately satisfy the capabilities that the manufacturing machines can provide. The process engineer can help, but it would be better if the machine itself is the final depository of that knowledge.”

Mr Iorio believes that in the not too distant future, production machines will be able to learn as each new part or task comes along in the same way as humans do. The machines will be smarter, self-programming, interact more easily with humans and become highly flexible work cells able to offer more bespoke manufacturing.

Computer designed solutions

A major advance of AI under development at Autodesk is ‘Generative Design’ – an AI influenced process that has the potential to come up with thousands of design solutions for components. Autodesk describes it as “mimicking evolution’s approach to design”.

“What we are trying to do with generative design is truly bring the aspect of ‘design’ into the CAD acronym,” Mr Iorio explains. “For many years, Autodesk has helped refine the abilities that people have to find solutions to design problems. What we actually provide is an extremely sophisticated electronic drawing board. The design solution is based on the human designer’s experience, skills, invention etc. which they formulate in their head. When they open up our design software tool, the program itself has no way of knowing if it is about to draw something as simple as a shoe or as complex as a jet engine.”

Mr Iorio says that generative design starts from the basis of design intent (i.e. what a part or design has to achieve) but can then generate many alternatives to come up with the desired outcome.

“We are communicating to the software what the objective of the design problem is that we are trying to solve. When this can be communicated with data, processes and instructions that a machine can understand, it can come up with many different solutions that vastly exceed the parameters of traditional design optimisation.

“Generative design is not just a design tool; it is also a decision-making tool, enabling designers to handle a number of ‘what if’ scenarios. So, what if in six months’ time a new material is invented that improves one of the stiffness properties of the design, or the cost of a particular part of your supply chain or production method massively varies. This could mean suggesting different processes for a more cost-effective outcome.”

With generative design, the design intent is the starting point for whatever solution is best for a particular set of circumstances.

Mr Iorio says: “The design intent is captured as a document or file, but a document that machines can understand. Experience shows that when it comes to real life design solutions there’s never an absolute ‘best’ solution. The best is always going to be a compromise dictated by an ever-changing environment. Generative design gives the user an understanding of why to make one choice over another and what are the trade-offs. Is it cost, speed of manufacturing, supply chain continuity or a new material development?

“Furthermore, this knowledge will grow in the system over time through the use of AI so in the future people will subscribe to Autodesk in order to gain access to that knowledge.”

With such massive amounts of data being managed, updated and then made accessible to users the only way to practically run generative design is through the cloud.

Adding it up

The main impetus for developing generative design is because of the developments in additive manufacturing (AM) or 3D printing which offers much greater flexibility in material combinations and design options.

“Additive manufacturing introduces flexibility in terms of how the geometry and shapes of parts is possible. However, one human mind cannot explore all the possibilities and freedoms that AM offers. This is especially the case when we look at parts that are intended to be multi-functional, for example, you could create a structural member of an airplane that also works as an antenna. This requires material options and production methods that might be beyond one person to know totally.”

This kind of rapid design creation is particularly useful when dealing with composite materials increasingly found in industries such as aerospace, automotive and construction.

“Using all those composite combinations at the moment is not as good as it could be because it is impossible for one human being to hold all that manufacturing and material knowledge.”

As design options become numerous this means manufacturing needs to become more flexible also. Manufacturing cells could include both additive, subtractive and moulding processes to produce more complete parts.

“With these modern production methods, you could create a bespoke part even as part of the production line. Let’s take the example of the car door. At present this is an assembly of many parts sourced from the supply chain. For example, the wire harness or loom is a separate part fitted to the door. In future that part could be directly imbedded into the door, so as the door is printed the wires can be printed integrally at the same time. When you have that amount of flexibility with the manufacturing process, customisation of the end product becomes so much easier.”

The smarter way

These kinds of developments naturally go hand-in-hand with concepts such as Industry 4.0 and the Smart Factory.

“These advances are about interconnection, analytics, risk control and prediction, so Industry 4.0 will be an integral part in the future of generative design,” Mr Iorio says. “The Smart Factory is realisation of the full potential of generative design. i.e. an entirely agile manufacturing centre.”

So does this mean an end to more traditional manufacturing methods and the role of design engineers?

“There will always be compromises that mean a particular manufacturing method is the best option,” Mr Iorio affirms. “Generative design at present is more likely to be used for making parts subtractively or by injection moulding, fabrication or casting but it may also suggest additive processes in future.

He concludes: “We believe a revolution is in the making. Generative design will remove trivial designs that might appear on a design engineer’s drawing board. No one should ever have to design a simple part such as a bracket ever again. Generative design will augment the capabilities that human designers have because it can provide information on types of designs that would otherwise be beyond most people’s comprehension.”

AUTODESK www.autodesk.com/solutions/generative-design