Skip to main content

Real-time Evaluation and Classification of Plastics

Periodic Reporting for period 1 - RECycle Plastics (Real-time Evaluation and Classification of Plastics)

Reporting period: 2019-12-01 to 2020-11-30

Of the 8.3 billion tonnes of plastic ever produced only 0.5 bn tonnes were ever recycled of which only 0.1 bn tonnes are still in use (i.e. also not discarded). If we keep this up, by 2050 we will have dumped more than 12 billion tonnes of plastic into the environment.

The plastic recycling industry is characterized by low margins and high volumes. One way to increase recycling rates is to increase the profitability of plastic recycling, which requires the optimization of production processes and a strong demand from the market. To enable plastic recyclers to produce a stable and high quality material, they need to know the composition of their incoming and outgoing materials Polytential is developing an automated composition analysis device based on Artificial Intelligence and NIR hyperspectral imaging. A sample of such hyperspectral data is provided in the attached graph.

We will provide recyclers insight into the composition and quality of their plastic streams with higher precision than current methods, at a fraction of the price, to enable them to produce material that can be used in high grade applications. This will allow them to produce a product that can be sold at a premium, reduce the use of additives and reduce their production costs significantly by preventing machine downtime.

The main objective was creating a self-supervised learning algorithm, which can learn not only based on a known dataset, but also on unknown client data. Such algorithm could reach a higher classification accuracy and provide insight into various material properties (e.g. plastic types, additives, catalysts). Providing insight into these material properties enables a recycler to adapt its processes accordingly.

The conclusion of the action is that the envisioned innovation is deemed feasible to realise and be implemented in real life applications. Uncertainty remains with regards to its eventual economic value in an operational environment as a result of the absence of a universal set of setting parameters that yield usable results in the most common circumstances. Regardless of that, it can be concluded that the technology will be of significant value during development processes, which in turn indirectly provide value to our products in the operational environment.
Research was performed on the best approaches to tackle said algorithmic challenges and which algorithm architectures to use. Also, time was spent on the design, creation, testing and evaluating of prototypes for algorithms. Additionally, the prototyping infrastructure and pipelines used to acquire and process data were created, prototypes tested and deployed.

It turned out that random sampling with artificial neural networks, is effective at differentiating material that is known from materials that the algorithm does not know, which is considered a complex problem as such algorithms are generally deterministic and forced to choose between the available options.

Optimization of the above approach for dealing with unknown classes, such as minimizing false positives and false negatives, has been a challenge throughout the action. It has been found that there is no one setting that always delivers the best results hence it seems to be required to optimize settings per input, limiting the scalability of the technology.

After the described methods show which materials are unknown, clustering is used to map the properties of the material onto vectors which correlate to physiochemical properties of plastics, such as monomer, molecular weight, viscosity, tensile strength, and additive presence, all of which influence a plastic’s mechanical and rheological properties. Caveat is that a certain degree of uncertainty in the mapping of said properties remains, which is not easy to quantify.

The aforementioned clustering methods have also been used to train algorithms on existing datasets while omitting the data’s labelling, meaning the algorithm does not know up front which objects belong to the same classes, nor does it know which classes should exist.

The work performed as part of the action has produced some very useful and valuable results, the most noteworthy (and nonconfidential) of which are:
- A semi-selfsupervised learning algorithm has been created by enabling algorithms to show what they are uncertain about, which makes improving the algorithms a more well directed effort.
- The above has led the company’s algorithms to improve their classification accuracy from 98.7% by 99.5%, essentially reducing the errors made by 60%.
- Correlating certain rheological and mechanical properties of plastic objects to their near-infrared spectral reflectance data has been successful.
- A method has been created that is able to train algorithms without requiring upfront labelled data.

At the time of writing the action’s application, it was said to be expected that during the action, merely a prototype version of the algorithm was to be expected and that further development to reach a market ready state was to be expected for several years. Reflecting on this, it can be said that progress has been more swift than anticipated, as a part of the technology created during the action has already contributed to the improvement of market ready systems, while the remaining developments are likely to become ready for an introduction onto the market in the first half of 2021.

The results of the action are thus already being exploited in the field, by plastic recyclers, to improve their production processes by doing quality control with the technology. In the future, the technology is expected to be exploited in the form of sorting of plastic waste as well.

The results of the action shall be disseminated to a limited extent, due to their confidential nature. This very report is likely to be the only form of dissemination of the technical results. Technically speaking, however, is the process of exploiting the technology also a form of dissemination, whereas the technology interprets information, which is then redistributed to other technologies and/or people (plastic recyclers).
Progress beyond the state of the art has been made. At one end, this is compared against what is currently available on the market and at the other end, what is known in science:
- Industry: Classification of plastic objects under industrial conditions using near-infrared spectroscopy with an accuracy of 99.5% is unprecedented and it thus beyond the state of the art.
- Science: The prediction of certain material properties of plastics under industrial conditions and at a rate that is suitable for industrial implementation, using near-infrared spectroscopy is unprecedented in science (and in industry). This is far beyond the previous state of the art.

The societal implications of the action are likely going to be that the recycling of plastic packaging waste becomes more attractive for recyclers, as the market value of the produced material goes up significantly, as a result of strongly improved material properties and hence applicability in new high quality products. This in turn would create more jobs (recycling plastic is more labour intensive than the production of virgin plastic), reduce demand for oil and reduce greenhouse gas emissions.
Example of near-infrared spectroscopic recordings of plastics