WiKant-Knowledge-based production of profiles

Optimising production through automation of the entire process chain
Partners
  • Metaflex Kanttechnik GmbH
Funding
Land NÖ FTI DIgitalisierungscall
Runtime
09/01/2018 – 08/31/2020
Status
current

Digitising production

Due to increasing digitalization it is now possible to digitally map the entire supply chain in the production of metal profiles, from configuration and administration to production. However, since there are a large number of production machines from different manufacturers and ages on the market, the transmission of automatically producible design data of profiles represents a major challenge. 

Automating the production of profiles

The aim of the WiKant project is to report information from successfully produced profiles on clearly defined production machines. In the future, this data will be collected, evaluated and fed back into the system for further production orders. This unique knowledge-based and rule-controlled system will make it possible to determine during the configuration whether the connected production machine can produce the desired part in the currently designed variant. This allows for the first time to automate the entire process chain – as manual testing or the production of sample parts to determine the feasibility of producing a configured part is no longer necessary. In addition to saving time for production companies, this also reduces raw materials consumption as fewer defective parts and fewer sample parts are produced.

Implementing a knowledge-based system

In the course of this project, such a knowledge-based and rule-controlled system is to be designed, prototypically implemented, and tested at Metaflex Kanttechnik GmbH using a modern production machine. It is planned that in the future new customers using different production machines can be connected to the system and jointly benefit from the knowledge and experience of other manufacturers, saving time and resources. 

Further, the project develops a uniform data model for profile programs as well as a framework for rule-based learning, which stores and allows for searching information of both successful and non-functional edge programs for specific production machines. In addition, data of previously produced profiles are collected and new production programs are developed and tested to enable wider implementation of the knowledge-based system in the near future.