New approaches for equalizing the granulate size and bulk density in mechanical recycling using heuristic approaches based on specific data analyses

Exactly defined and constant granulates become more and more important in recycling business. The material is very often mixed with virgin granulate, sold on the world trade market and manufactures must handle different material streams without process changes. For solving that issue, a new regulation model for pelletizing systems, based on specific data analyses, was developed. In comparison to standard regulation models, available on the market, the developed model was generated by using symbolic regression based on genetic programming and focuses on the combination of actual process parameters. The model was generated and tested with a broad range of in-house, post-industrial and post-consumer materials. It turned out that the new model enables a constant granulate size without remarkable changes during the production process and leads to a substantial equalization of the produced granulate size.