Big Data Handling in Process Surveillance and Quality Control of Secondary Metallurgical Processes

Abstract

The Ruhrstahl–Heraeus (RH) degassing plant is essential for producing ultraclean steel, but its vacuum‐based nature is hiding the process from scrutiny. Despite recording the many inputs and outputs of process control and measurements surrounding the plant, the utilization of produced data is comparatively low. Data handling is a challenge, especially considering correlations of process control data with indicators such as actual steel homogeneity and cleanliness after casting. Here in this study, approaches to improving sensor data quality and maximizing the utilization of the abundance of data obtained during the RH process are presented. Time series data from process control measurements like off‐gas composition or chamber pressure, even though highly connected to specific domain knowledge, can be tackled with machine‐learning methods, such as recognition of characteristic curve sections or categorization of curves with clustering algorithms. Visual surveillance offers the possibility of image analyses of the surface flow observation inside the vacuum chamber with different approaches such as image classification or feature detection algorithms. Herein, a digitalization and data analysis strategy developed for an improved understanding of the RH process, considering linear as well as integral data from the plant, are presented.