We are delighted to announce that the first version of our new artificial intelligence and machine learning system has now been developed, utilising many years’ of stored historical formulation data. The testing of the first version, to identify the effectiveness and operational performance, will be conducted across the latter months of 2021 using ‘real world’ situations.
There is wide variation in incoming recycled plastics (due to variation in first life product polymers being recycled at a specific time), meaning meeting customer specifications can be challenging as every batch of material requires a unique formulation.
Our unique formulations are derived experimentally, to meet the plastic’s physical specifications (e.g. impact strength) and colour. This is determined through the know-how of our highly experienced laboratory technicians but can still be significantly time consuming.
Project Objective
As such, we wanted to develop a sophisticated artificial intelligence system that will accurately formulate polymer blends to the required physical colour specifications of a repeatable consistent plastic compound grade, using the most efficient and lowest cost option of the ever-changing incoming recycled plastics.
What are the positive outcomes of this project?
• A reduction in resources required for trial-and error iterations
• A reduction in waste materials, through a greater degree of accuracy
• An increase in the proportion of lower grade recycled plastics utilised
Our sincere thanks to our funding body,
WRAP and our project partners,
The Technology Research Centre.