Sorting and Separation

Challenges for the automated sorting of post-consumer textiles using standard NIR spectroscopy

Saal B
Donnerstag, 12.09.2024, 14:20 - 14:45 Uhr

Sorting of the textile waste is one of the essential steps for successful recycling of post-consumer textiles. The objective of this work is to present the results of various experimental analyses to identify the challenges that can affect the sorting efficiency when sorting using standard NIR spectroscopy. Analyzed were the identification (sorting) limits of post-consumer textiles, focusing on fiber composition and other influencing factors, i.e., accessories presence, different colors, material thickness and structure, and how they can influence the sorting quality and efficiency.

Hana Stipanovic (Montanuniversitaet Leoben)
In order to perform high-quality recycling of post-consumer textiles high quality sorting is vital step. With the growing amounts of textile waste and manual sorting reaching its limits, automated sorting is in a growing phase, and although various sorting technologies enable automated sorting, sorting using standard NIR spectroscopy has become a widespread technology for automated sorting. Numerous experiments were performed to identify the challenges and limits of standard NIR spectroscopy in sorting post-consumer textiles. The fiber composition is already known to be a challenge for the automated sorting using standard NIR spectroscopy, and the performed experiments presented the results that can help with identifying where the limits are. The results of the experiments showed the difficulties when separating different post-consumer textiles made of natural fibers, namely between linen and cotton, as well within wool, merino wool and cashmir wool. Especially challenging was identifying fibers presented in small amounts, like elastane and viscose, in cotton-rich post-consumer textile blends, where the limiting content, which still enables the identification of elastane and viscose, was detected. The presence of accessories in and on post-consumer textiles was analyzed. To list the most often found accessories, inspect if it is possible to identify them using standard NIR spectroscopy and how it could affect the sorting steps. Additionally, experiments were performed to identify the possible influence of different colors, material thickness and structure on the identification, showing particular challenges and limits when identifying black and grey textiles depending on the different coloring techniques, and post-consumer textiles with low thickness. Constant experiments are performed in house, and progress is made in identifying the challenges and creating the foundation for developing new solutions for improving automated sorting; therefore, the chance exists to identify more challenges that are to be presented, which are not yet part of this abstract.