Steel Production Efficiency Improvements by Digitalization

Steel is one of the most important engineering and construction materials and plays a critical role in the overall economic development. However, iron and steel industry is one of the most energy-intensive industries which accounts for >20% of global industrial energy use and about a quarter of industrial CO2 emissions in the world. Energy efficiency improvements are one driver for production optimizations, but efficiency improvements in steel production can be reached in several ways like equipment availability improvements as well as yield increases or optimized production schedules. Using digitalization for these improvements in steel production requires reliable data pipelines from the sensors and automation systems to the applications. An additional data collection and storage layer between the IT and OT systems helps to process these data and inserts them in a compressed form in a data warehouse. The applications can access the data in that additional layer through a generic API. This data warehouse, which unifies data from different data sources, is the key enabler for the successful and quick deployment of value adding applications. Our integrated energy platform has the capabilities to manage and optimize energy consumption as well as to track the carbon footprint on a product level. It helps to reduce overall consume of resources by process efficiency modeling, simulation, and forecasting. Process data are used to train AI/ML models which can be utilized in the area of asset optimization. Especially in the current steel market situation, the availability of the equipment is critical for steel producers. Forecasting of the time-to-failure of critical components helps to reduce the unplanned downtimes, to extend the lifetime of the assets, and to avoid catastrophic failures. Pattern in the data of hundreds of signals indicates upcoming equipment failures in early stages and helps to trace down the root cause of the issue. This is has successfully been developed for several use cases for electric arc furnaces, e.g. for refractory relines, taphole failures, and panel failures. Despite the fact that digitalization drives and enables these efficiency improvements and data are the foundation for the value of these applications, the understanding of the physics and processes in metallurgy as well as of automation are required for a successful application in the field. Especially when the data basis is small or not even existing, the physical models and measurement capabilities play an inherent role to close that gap. In this way, we could develop smart applications, which, e.g. optimize mechanical properties, reduce the amount of virgin material consumption by increased consume of recycled material and the optimization of heat treatment times and temperatures for reduced energy waste. These types of applications get more and more accepted in the steel industry as they have shown their value for the producers. The use of these applications starts usually for small and isolated processes, but the future is heading towards a more holistic optimization for production areas, entire plants, and even enterprises which will increase the impact of AI/ML for this industry for the next years significantly.