Page 27 - ITA Journal 3-2018

ef ciency is always an issue, especially for a process-related energy-intensive company such as an iron foundry. Professor Dierk Hartmann, Kempten University is working on an optimised solution for the Adam Hönig iron foundry. The foundry uses barcodes that are scanned by employees on their smartphones and transferred to a database. In this way, new process parameters can be added to the production areas and the production process can be tracked. The aim is to improve energy and resource ef ciency by reducing overproduction of liquid metal
“Foundries are experienced in dealing with data-driven business models”, says Heinz Nelissen, Pres- ident of GIFA 2019 and NEWCAST as well as Managing Director of Vesuvius GmbH, Foseco Foundry Division in Borken. Approaches related to machine-to-machine communication, automation and robot use, computer-aided tech- nologies, and product and process development will therefore also be a focus at GIFA 2019.
How Industry 4.0 can look in prac- tice can be seen at Karl Casper Guss in Pforzheim. The foundry produces a wide range of hand- moulded parts with unit weights from 100 kg up to 9.5 t. In order to be able to react quickly to changing customer requirements while guaranteeing high produc- tion reliability and quality at the same time, Casper Guss relies on an integrated Industry 4.0 solu- tion with three pillars:
1. Interlinking of all operating equipment
2. Planning and control of pro- cesses with 100% traceability through the ERP system
3. As an interface to the extranet, a web portal that gives cus-
tomers access to production information.
Linking of all systems from end to end makes it possible to plan individual orders directly, as Managing Director Felix Casper describes. The ERP system auto- matically checks feasibility upon receipt of the order, thus ensur- ing a high level of adherence to delivery dates. Feedback from all production steps improves throughput and increases quality. Using the web portal, custom-ers can call up production informa- tion on their orders from the extranet and directly enter addi- tions as well as changes to dates or quantities. “Interlinking of the customer systems with our own systems leads to faster and more reliable processing of orders”, summarises Casper.
FeroLabs: Industry 4.0 for steel production
Agile companies from the start-up scene also have their eye on tapping new business areas with digitalisation. Digital technol- ogy opens the door to potential disruptors in the metal industry as well. Voice control via mobile phones along with face recog- nition in social media such as Facebook, Amazon, Google and Apple is nding its way into the steel industry, thanks to clever company founders: machine learning is one of the most suc- cessful subareas of arti cial intelligence. While self-learning algorithms were mainly a topic of academic research until a few years ago, today they are increas- ingly making their way into our everyday lives as well as industry.
“With Fero Software, our custom- ers are able to better understand their production process and thus increase their pro tability”, says
Tim Eschert. As an application engineer at FeroLabs in Düssel- dorf, the industrial engineer with a master’s degree from RWTH Aachen University is something like the vicegere of the New York start-up in Germany. FeroLabs uses what are known as statistical machine learning (ML) methods. Eschert sees them as a bridge between conventional analysis methods such as Six Sigma, which have so far been used in produc- tion, and the modern technology of machine learning. “The area of statistical ML combines these two elds, and we at Fero are proud to be the rst to bring these methods from academic research into industrial manufacturing”, says the FeroLabs manager.
In the area of steel, the start-up has applied and researched the use of its software in various applications, as Eschert explains. These include, for example, reduc- tion of surface oxidation (cinder) and the prediction of material properties in a hot wide strip mill, quality improvement in a tube mill, detection of inclusion defects in a wire mill and opti- misation of alloy usage in a rod mill. “At present, the application of alloy optimisation is the area in which we are most advanced in terms of implementation and scaling, but the other areas are already fully operational”, reports Eschert.
Steel customers include Gerdau in Brazil, which produces steel on the scale of ThyssenKrupp and is known for its good quality. For FeroLabs, the focus there is mainly on reducing production costs by means of ML while maintain- ing the same quality. As Eschert reports, FeroLabs combined data from several different databases to get a complete picture of the
Technical Papers
ITAtube Journal No3/October 2018
27