Düsseldorf, 15. März 2021

Der Cracks Preventer

Neue Predictive Quality Applikation revolutioniert Prozessoptimierung in Stahlwerken

The Cracks Preventer from SMS digital combines artificial intelligence, metallurgical expertise, and data know-how. This intelligent combination predicts multiple defects and associated strand breakage during the casting process and suggests countermeasures in real-time. It focuses on preventing the cracks in casting with necessary actions making it an essential part of the predictive quality system. 

One of the latest topics in every industry is predictive quality, optimizing steel production processes as far as possible to ensure defect-free production. Surface, geometry, and product properties are the key components of predictive quality. The Cracks Preventer from SMS digital is an essential tool in the surface section of predictive quality, contributing to saving valuable resources and costs. Cracks and the associated breakouts during the casting process often mean significant financial losses in the steel industry. The Cracks Preventer application addresses this problem and offers an innovative solution to reducing a wide variety of defect types. By analyzing data from the entire process chain in real-time, the Cracks Preventer provides real-time countermeasures in the form of specific recommendations to prevent defects in good time. Multiple interconnected machine learning models enable "end-to-end" data processing and ensure targeted defect prevention.

Continuous casting of slabs

Currently, the application offers solutions to defect types that cover longitudinal facial cracks, edge cracks, corner cracks, welded cracks, slivers, and laminations, as well as the breakouts that can occur because of these defects. These types of defects are often accompanied by strand breakage in casting plants, which can lead to large-scale production losses and tremendous economic damage. The Cracks Preventer can significantly reduce both defects and breakouts with the aid of predictive models. Machine learning models analyze each defect and perform automatic root cause analysis (ARCA), providing suggestions for suitable countermeasures to prevent these cracks. Finally, the operator can take the actions depending on the different recommendations to reduce defects and breakouts. The application can cut potential losses up to 50 percent. In close collaboration between data scientists, metallurgists, and process experts, the Cracks Preventer is set to revolutionize the steel industry in the area of process optimization.

SMSgroup Image SMSgroup Image SMSgroup Image
Example of warning of predicted LFC (longitudinal facial crack)
SMSgroup Image SMSgroup Image SMSgroup Image
Example of warning of predicted BO (breakout)

Each steel mill is very much different in its process, practices, and defects itself. The application is specially developed to pay attention to those differences and make it highly customizable. It can be implemented in any steel plant without any major architectural modifications. Closed-loop deployment processes allow real-time solutions to real-time problems and also guaranteeing higher productivity of the entire casting process. At this stage, the Cracks Preventer is focused on optimizing the casting process of both flat and long steel products.

The Cracks Preventer is characterized above all to provide maximum freedom and flexibility. This enables the wide range of services to be geared heavily towards the customer, who has the freedom to chose pre-customized whole package or any single defect. The Cracks Preventer also offers the options to be installed on a local machine on-site or as a cloud-based solution within a software-as-a-service (SaaS) model. Users also benefit from the choice of a performance-based pricing model or a predefined final price.