Solum: Sorted construction waste optimized with robots

Today, construction waste and sorting it is a challenge. Waste is not always sorted efficiently at sorting facilities, and due to limited space, it is often not possible to sort the waste at the construction site. Solum near Roskilde has taken the initiative to solve this problem by using advanced robotics and expertise from DTU, TRE Consulting Engineers, and Roskilde Municipality through a knowledge bridge project.

Around 10% of construction waste is recycled, but the potential is greater. It just requires more and better sorting. As the first in Denmark, Solum has invested in an automatic robot system that can scan and sort waste. With the help of experts from DTU, Solum has studied and created knowledge about robot sorting, including sorting rates and documentation.

The results have been impressive, as the robot system can sort mixed construction waste almost six times better than regular sorting.

6 loads were sorted conventionally by grab and then by robot sorting. After each sorting, the sorted fractions were weighed to determine the sorting rate. The trial shows that the sorting rate is 6 times higher with robotic sorting – Final report

Using AI and cameras, the robot identifies materials such as wood, concrete, and metal. With an arm, the robot sorts the waste into different fractions, which can then be recycled and sold.

Proven environmental benefits

Using an LCA calculation model, it was calculated that robotic sorting can reduce CO₂ emissions by approximately 150 kg per ton of waste compared to conventional sorting. This result emphasizes not only the effectiveness of robotic sorting but also its potential to contribute to a more sustainable future.

The project partners have all contributed knowledge and insights that have helped optimize Solum’s processes and facilities.

If we don’t have data, we can’t improve our solution – The information that the knowledge bridge project gives us is important as it can tell us how much better robotic sorting is than manual/traditional sorting – Anne Dorthe, Development Director, Solum A/S

DTU has also presented Solum with several calculations that can validate and point out potential challenges in LCA analysis. This data will help Solum better understand its solution options and further development.

Further development

The results from the project have been good for Solum and the partners. Therefore, they also plan to use the project results to further develop the robotic sorting, and a repeat of the project is already planned to assess further improvements. The LCA tool, methods, lessons learned and data will form the foundation.