TRAINS
Textile Recycling supported by Artificial Intelligence in Spinning
Scheda del progetto
Kiwifarm s.r.l.
The TRAINS project aims to develop a Digital Twin based on Artificial Intelligence capable of improving the development process of recycled yarns and automatically estimating their color. The DT will support process decisions, to accelerate and streamline the development of yarns from recycled fibres. Its development will lead to the definition of AI algorithms capable of predicting the color of yarns starting from certain types of input fibers. There will be a significant reduction in the number of expensive physical experiments, because it will be possible to preliminarily measure impacts with a set of product and process indicators. The Marchi & Fildi company will carry out the application coordination of the project, while Kiwifarm will lead the design of the experiments for the training and validation of the Digital Twin's AI.
Contacts:
Luca Cinguino
The challenge is linked to the possibility of obtaining the target quality of the blends to be spun through a smaller number of prototypes, thanks to the availability of a Digital Twin capable of simulating with sufficient accuracy the final color tone obtainable from mixing recycled fibers of various shades of color.
Today the definition of mixtures is carried out by highly specialized technicians and requires a large number of prototypes. The availability of a Digital Twin can bring significant savings in costs, resources consumed and transition times, increasing the level of maturation of the technology.
The AI-based Digital Twin represents an innovation in the textile sector globally. There are no Digital Twins that are applied or applicable in the intermediate links of the textile supply chain such as spinning. Much less that they are applied or applicable to the development of recycled yarns. The complexity of this prediction has led to the choice of Artificial Intelligence as a simulation technology. The colors of the material entering and exiting the process will be objectively measurable, so there will be the possibility of creating well-structured datasets for training the Digital Twin.
The digital model developed can considerably reduce the quantity of yarn prototypes necessary to obtain the desired colour, with proportional impacts on efficiency (material and energy savings), on sustainability (less consumption of raw materials, lower CO2 emissions) , on circularity and on the ability to respond promptly to market demands. The main benefits for consumers are linked to the project's ability to accelerate the development of a commercial offer of circular textile products.