Task 5: Intelligent Steelmaking

Many existing process models for steel production do not allow for process alignment and are too complex for meaningful real-time predictive use.

The primary aim of Task 5 is to take a different view on steel production in its entirety by not seeking to improve product qualities in one step alone but by focusing on decreasing energy usage and building links over the entire process chain.

This will be achieved by developing and optimising process-level models supported by experimental verification, analysing process data, and benchmarking current process routes to quantitatively assess how efficiently industry currently uses energy and materials.

Opportunities

The project officially began in October 2020. The initial task aims haven’t changed significantly, although, during discussions with industry, one or two extra ideas and opportunities arose.

For example, a focus on the life cycle of refractory materials wasn’t originally in the plans, but after discussions with industrial partners, it was added to the list of potential investigations.

Another change was an increased focus on the use of methods from artificial intelligence such as machine learning for regression, as opposed to improving physics-based modelling approaches.

Key findings

Collection and analysis of steelmaking data to allow material and energy efficiencies to be calculated. Methodologies are being developed/adapted to be able to turn these data into meaningful conclusions and suggestions for efficiency gains.

Development of a hybrid model to combine the accuracy of physics-based simulations with the speed of machine learning methods. This forms a template to be used for many process simulations across the entire steel manufacturing route.

Industrial applications

This methodology can be used to compare the resource efficiency of various individual components in a particular process step and thus identify overall potential efficiency gains – for example, a decrease in raw material input, an increase in yield and the recovery of by-products.

With the model being a template for the steel industry, it allows for synergies and knowledge transfer between processes. The model can also, once verified with specific process data, be used as a platform for the process behaviour under worst-case scenarios. This could become a digital testing ground to develop guidelines for a “what-if scenario” in relation to equipment failures.

Coming up

Completion of the material and resource efficiency data analysis and implementation of suitable methodologies.

Seek funding to widen the applicability of the model for processes across the wider steel production route.

The effects of electrification

The work being carried out at Sheffield as part of this task is in conjunction with Liberty Steel, so the findings are of course directly relevant to electric steelmakers.

One of the planned EngD projects associated with this task is going to look at the increased use of scrap within BF/BOS steelmaking, and how these changes might affect the product design for an electric steelmaking route compared with an integrated route and what would be required for the design of steel products to make the re-use, recovery and recycling of them at end of life much easier in terms of exploitation in the steelmaking processes.

The modelling template developed at Warwick is expected to become a template for many steel-related processes, ranging from predicting the temperature distribution in blast furnace stoves to optimising the energy use vs. oxidation losses in a reheat furnace.

Electrification of the steel production in this framework becomes another process that is very welcome to trigger further improvements on the physics part of the general modelling template with knowledge, specific to electric arc furnaces and resistive/inductive heating techniques.