This concept attempts to cover sustainability on a product level. The aim is to improve the environmental sustainability of individual products, in the knowledge that real sustainability is an issue that needs to be addressed at different levels and looked at systemically.
01 how THE CONCEPT WORKS
The core of the concept is a machine learning model which generates 3D data. The generated data undergoes a feedback loop by being evaluated by a Life Cycle Assessment tool (LCA) on sustainability factors. The (human) designer intervenes in the system at this point and can then continue working with the generated output. The interaction between machine and designer constitutes the creative moment.
02 DESIGNER + SYSTEM
The model described below is one way in which designers can use machines to create more sustainable products in the future. It is not intended to replace humans. Rather, the concept must be seen as an attempt to combine human capabilities with the potential of machine learning.
03 SUSTAINABILITY AS A COMPLEX SYSTEM
To what extent a product is sustainable or not depends on a wide range of factors. These factors partly influence each other. Such a complex system exceeds human capabilities. Therefore it is reasonable that designers interact with machines to create more sustainable products.
04 POTENTIAL FOR MACHINE LEARNING Sustainability - precisely sustainable product development is seen as a potential use case of machine learning. In this case, technology offers the possibility of recognizing patterns in large data sets. A trained ML model can therefore be able to evaluate which product in what extent is sustainable.
05 ML Model: 3D GAN
The machine learning model used for "AI aided design" is a "generative adversarial network" (GAN). GANs are algorithmic architectures that use two neural networks, pitting one against the other in order to generate new, synthetic instances of data that can pass for real data. GANs are used to generate images, videos or text. In this case 3D data is generated.
06 DATA SET
The supplied data set contains a large number of products of one category. Each product contains information about materials extraction, processing and manufacturing, product use, and product disposal. The shape is only one of many factors. Each product also has a "sustainability indicator", which was determined by a Life Cycle Assessment tool (see below) and indicates the degree of sustainability of a product. The aim must be to feed in a data set with as much information as possible, which is at the same time as clean in itself as possible. At the beginning, the data set only includes individual product categories. Later the system can be fed with additional data sets.
07 OUTPUT: 3D DATA
The machine learning model generates an output of 3D data. This data can be further processed in a CAD tool. To ensure that the system is constantly improving and learning, the outputs are evaluated by a LCA tool. With the resulting data the ML model is trained and improves continuously.
08 FEEDBACK LOOP: LCA TOOL To reduce the error rate and improve the system, a feedback loop is being considered. This "safety instance" is a Life Cycle Assessment tool (LCA). LCA is a systematic analysis of the environmental impacts of products throughout their entire life cycle. It attempts to map the ecological consequences / footprint of products during the entire lifecycle ("from cradle to grave").
09 INTERACTION OF "HUMAN DESIGNER" + SYSTEM The system "AI aided Design" can not exist without the human being. He enables the possibility to create innovations. The interaction between designers and the system can create a creative moment. This is where the circuit of the system is interrupted - something really innovative can be created. The designer's main task is to select the best possible solution for a specific application and to modify it if necessary.
In conclusion, the model is based on the two main components 3D GAN (machine learning model) and LCA. The LCA tool can be seen as a quality assurance tool.
It is important to note that the "human designer" is not replaced by the system. Quite the opposit: the human factor is crucial for a successful implementation.