Compared to ore-based steel manufacturing, steel created using recycled scrap has approximately 78% fewer carbon emissions per tonne of steel(1). The UK’s demand for steel scrap is expected to triple by 2050(2), due to increased EAF production, which primarily utilises scrap to manufacture new steel products.
Steel is one of the most recyclable materials in the world, and the UK generates about 10-11 million tonnes of scrap steel annually. However, the recycling process has historically been inefficient, leading to material waste, higher energy consumption and challenges in sorting different grades of steel. Emerging technologies, such as artificial intelligence (AI) and robotics, have the potential to solve these difficulties, by providing optimisation, absolute precision and real-time decision making.
Traditional Methods of Steel Recycling
Steel recycling generally involves several key stages: collection, sorting, shredding and re-melting. Used steel products are collected as scrap and transported to recycling facilities, before being sorted into various grades based on composition, size and contaminants. After sorting, the steel is shredded into smaller pieces, melted, refined and recast into new steel products.
Opportunities for Optimisation
Consistent and Accurate Sorting: Traditional sorting methods rely heavily on manual labour or mechanical separation systems that may not accurately distinguish between different types of steel or remove contaminants effectively. This can result in a lower quality of recycled steel which requires additional processing.
However, one of the most transformative applications of AI in steel recycling is the use of machine vision systems for sorting and classification. Using cameras and sensors to scan and analyse materials on a conveyor belt, machine vision can distinguish different grades of steel based on their visual and chemical properties and identify contaminants with high levels of accuracy.
Building on the capabilities of AI-powered machine vision, robotic sorting systems can take automation to the next level by picking and placing metals into designated areas based on their classification. Companies like ZenRobotics are at the forefront of this technology, helping recycling facilities to increase their efficiency by up to 20 times(3) compared to manual processing.
Process Optimisation: Recycling steel requires significant amounts of energy, particularly in the melting phase. AI has the potential to optimise energy usage by monitoring furnace temperatures and adjusting energy inputs based on real-time data.
Additionally, AI can use predictive maintenance to prevent equipment failures through sensors embedded in recycling equipment. By detecting subtle changes in vibration, temperature or pressure, AI can predict when a piece of equipment is likely to fail or require servicing.
Material Recovery: In traditional recycling processes, some steel is lost due to inefficiencies in shredding, sorting or melting. AI can help to reduce material loss by improving the accuracy of sorting and providing real-time feedback on material quality throughout the recycling process. AI’s ability to process and analyse large datasets allows it to provide optimisation in ways that were unrealistic to achieve previously. Using historical data on steel scrap composition, recovery rates and processing outcomes, AI can identify patterns and even make recommendations on how to improve recovery strategies.
Case Studies and Industry Applications
ZenRobotics and Skrotfrag: The Skrotfrag Group is one of Sweden’s largest scrap and metal recycling companies and is at the forefront of integrating AI into its recycling operations. The company was among the first to introduce robotic sorting technology to metal, purchasing a sorting line in 2019 from ZenRobotics(4). Using ZenBrain, an AI software, the sorting robots are able to analyse sensor data in real-time to identify and sort scrap metal. The software can also be trained to identify specific objects or contaminants and eliminates the potential hazards and inefficiencies of manual scrap sorting.
FeroLabs: FeroLabs, a startup in New York City, has pioneered an exciting new application for AI in steel recycling – using software to create customised ‘green’ recipes for recycled steel. Since every batch of scrap steel has a different chemical composition, additional materials must be added to ensure the final product meets industry standards. This is typically a lengthy and costly process, which often results in an excess use of raw materials and, by extension, carbon emissions.
However, by training AI to analyse historical and real-time data and identify opportunities for improvement, the software is able to recommend the exact amount of materials needed for each batch of steel, effectively eliminating the use of mined ingredients in steel production by up to 34%(3).
The Future of AI and Robotics in Steel Recycling
One common challenge among companies looking to implement AI or robotics solutions is the high initial setup costs. This may be a barrier to adoption for companies operating with tighter margins, although the long-term savings in energy, maintenance and increased efficiency often outweigh the investment cost.
The introduction of AI also requires retraining of workers to maintain and operate the new systems. While AI can reduce the need for manual labour in some tasks, there will be increased demand for skilled workers, and therefore, investing in workforce training is an essential step for any business implementing AI technology.
There is tremendous potential for AI and robotics in the future of steel recycling, however, as these technologies are quickly evolving, and recycling facilities will benefit from more sophisticated machine learning algorithms and improved picking and sorting capabilities. Scalable AI tools, such as cloud-based data platforms and modular AI systems, will allow smaller facilities to benefit from data-driven insights and real-time process optimisation without the burden of extensive upfront investments.
In time, AI-driven predictive analytics may even enable steel manufacturers to better anticipate fluctuations in scrap availability, steel demand and market conditions. This will further optimise the recycling process and allow manufacturers to recommend adjustments to production schedules in order to reduce waste, lower costs and ensure that supply meets demand.
Author: Shirley Carruthers - Content Creator at ParkerSteel
References:
How AI and Robotics Are Transforming the Steel Scrap Industry
First published on 20/09/24