Add Thesis

How can artificial intelligence based sorting solutions support the realisation of circular economy and closed loop recycling of scrap tyres?

A case study at Ragn-Sells

Written by S. Kilander

Paper category

Master Thesis


Business Administration>Supply Chain & Logistics




Master Thesis: Circular economy As population, consumption and key raw materials increase, a transition to a sustainable economic paradigm is required (Klum and Rockström, 2012; Jesus and Mendonça, 2018). This includes our use of materials already circulating in our society (European Parliament, 2018; European Commission, 2015; Ellen MacArthur Foundation, 2019b; Ragn-Sells, n.d.(b)). The adjusted circular business model will keep materials in the system after the end of the first life cycle, and then reuse them as much as possible at the highest possible waste level (European Commission, 2014; Alan MacArthur Foundation, 2019b; Dunås) , 2018; Larshans and Kihl, 2018). For an overview of the circular economy business model, see Figure 1. Higher material circulation is the core principle of the circular economy, which redefines growth and decouples the growing society from the traditional linear model; procurement, design, production, consumption and disposal. Materials need to be fully utilized, reused and recycled before degraded recycling, incineration or landfill, and then reused (Jørgensen and Pedersen, 2018; Ellen MacArthur Foundation, 2019b). The circular economy is based on three principles; design waste and pollution, use materials and products for as long as possible, and regenerate natural systems (Allen MacArthur Foundation, 2019b; European Commission, 2015). Talking about renting instead of buying is also a cyclical business model, which still provides actual services. The results show that the traditional linear model is the reason for the increase in resource abundance, and the limit that the biosphere can support will soon be reached (Klum & Rockström, 2012; Sarati, 2017). This is the driving force behind many initiatives to develop new business models. Circular economy is one of the most successful frameworks for this integration, combining economic and environmental sustainability (Murray et al., 2015). The Ellen MacArthur Foundation (2013) pointed out that there are 24 ecosystem services that are exhausted or unsustainable. The circular economy aims to consume natural resources responsibly without compromising quality. The circular economy is not a promoter of recycling. It is one of the most unfavorable steps in the circular value chain. This cannot be overemphasized (Ellen MacArthur Foundation, 2019b; McKinsey Podcast, 2016). Refurbishment, reuse and secondary use are examples of end-of-life scenarios. They are all more beneficial than recycling and can save greater value of products or materials because they are more recycled in the cycle and are closer to the original product or material. Value (Lansink, 1979). 1.6 New technologies and sorting possibilities With the new development of artificial intelligence, machine learning and autonomous sorting technology, downstream customers' ability to perceive valuable components has improved, which helps to distinguish materials with different substances and characteristics. On the one hand, it can be used to achieve higher circulation of key raw materials and to classify more specific substreams. When the substances in certain products change, this kind of automatic sorting is also required, which causes problems for recycling plants. In this paper, the possibility of using image recognition to classify mixed-wear tires and the possibility of accelerating the transition to a circular economy is evaluated. 1.6.1 New technological solutions of machine learning and artificial intelligence are important driving forces to transform the linear value chain into a cycle (Preston, 2012). Deep technical knowledge will create more abstract and complex products, but it can also create products that have been carefully thought out and produced to last longer, promote refurbishment and extend service life. The possibilities are endless. 3D printing a new tread instead of replacing the entire tire is an option. The other is to design a multi-layered pedal, fill the hollow layer, a new pedal appears, and then remove the bottom layer. Rapid and large-scale solutions have exploded and can accelerate the transformation of the circular economy, making it more efficient than ever before (Ellen MacArthur Foundation, 2019a). Machine learning and deep learning are the two main subsets of artificial intelligence, which are widely used in our daily lives (MathWorks, undated). One of the pioneers of machine learning, Arthur Samuel (1959), explained it as "a field of study that allows computers to learn without explicit programming." Like humans, artificial intelligence can learn to reason, and the more experience and data the computer exposes, the better the prediction and interpretation. Thanks to the development of artificial intelligence, computers can understand patterns beyond human capabilities, promote our understanding and provide tools to accelerate the understanding of complex and rich data (Alan MacArthur Foundation, 2019a). The core of this research is to study how artificial intelligence-based sorting solutions can improve downstream sorting of worn tires, especially when using machine learning and image recognition. The idea behind image recognition is to collect enough data for the computer to understand and find patterns in the presented images, which are obtained by using computer algorithms. Read Less