Extrusion processes are used in various industries. Extruders are used to shape plastic articles, for example in the production of sealing systems for car bodies. In addition to shaping, extruders are also used for chemical processing and modification of materials. The food industry uses this principle to produce artificial meat substitutes, for example. In plastics manufacturing, this principle is used to produce rubber compounds such as thermoplastic elastomers. Single screws or twin screws are often used in both industries, which convey a material through different zones in the elongated extruder. During the conveying section, the materials are heated, compressed, degassed or mixed with other materials.
There are often non-linear relationships between the individual process characteristics in the extruder and the output variables such as structure and chemical properties of the extrudate. Variances
in the input materials or changes in the ambient conditions make it necessary to constantly monitor the process and perform adjustments of process parameters. Today, the subjective experience of the employee is relied upon. Non-linear relationships between different characteristics are often difficult for people to grasp. This leads to the process often being adapted on the basis of a try & error procedure.
This challenges leads to some unintended effects for the manufacturing company:
- Scrap costs: Until the right process parameters are identified the process is producing scrap
- Machinery capacity: Until good quality is produced the machine capacity is lost
- Stress of the operator: Between failure identification and finding the right process parameters, the operator usually is under stress

With ExtruAI, we will resolve these challenges. With the help of a scalable assistant, we will give the employee on the extrusion line recommendations on how to readjust the process. This will support the employee's existing knowledge and minimise the number of necessary iterations. The plant operator benefits from less waste and increased machine capacity. Even less experienced employees can regulate the process independently with ExtruAI, which simplifies personnel allocation.
On a technical side, ExtruAI will use a prescriptive Artificial Intelligence, which consists of three components:
- Standard scheme for the processing of extrusion data: To enable a scalability a standard scheme to provide data out of the extrusion process will be created.
- Self-learning prediction model for modelling the process state: A standard model approach that learns the process behaviour delivered by the defined standard scheme.

Please contact us for license offering.
DatenBerg GmbH
Haid-und-Neu-Straße 7
76131 Karlsruhe

[email protected]

We want to say thank you to the whole KITT4SME team - here you find the official website.
Here you can find the KITT4SME platform for application hosting and managing.
This work was supported by the H 2020 project "Platform enable KITs of Artificial Intelligence for an Easy Uptake of SMEs (KITT 4 SME)" under GA 952119.