The principle of the Electric Arc Furnace (EAF) was invented in the end of the 19th century and put into industrial use in a steel mill in the beginning of the 20th century. Today the importance of this type of installation in the steel industry is increasing rapidly. They represent a key component in the production of recycled steel. Their importance will grow even further when the new alternative production routes, replacements for the blast furnace based route, will come into operation. Unlike this blast furnace, an EAF uses Electricity as its main power source. An other significant difference is however that an EAF operates following a batch production principle, whereas a blast furnace today is used in long stints of quasi constant production. In this electricity-based approach the production is split over multiple individual so-called “melts”, each of them with the associated preceding and subsequent mechanical operations related to filling and emptying.
As with all production processes, efficiency and availability is key for the EAF route. The installations thus have to remain in top shape all of the time and unforeseen stops need to be avoided, as these have a massive impact on energy efficiency (loss of heat) as well as subsequent process steps. Also production planning and smooth and predictable working days for the workforce are relying on a process without interruptions. Last but not least, some of the spare parts of such an oven are very expensive and/or hard to come by. Together with the sometimes complex interventions, requiring a lot of external equipment, a stop due to mechanical damage can take quite a while to come by.
some of the spare parts of such an oven are very expensive and/or hard to come by
When following up the health of a more complex asset it is essential to approach it with a Systems View. Other than more simple components such as ventilators or pumps, an EAF consists of and relies on a multitude of individual components that mutually interact, and operate at intermittent time intervals, sometimes in parallel, sometimes in series. Also the behavior and production parameters used might differ for each individual melt. Therefore such a monitoring approach should inherently take into account the various aspects of relevance in a combined manner for it to be able to provide context on the nature of the issue and its origin or rate of evolution. It should be modular in the aspects that are followed up, because each oven is a bit different. A good solution also relies on a combination of different sensor types installed on the relevant individual components or subsystems of the furnace, like the slewing bearing, busbars, tower, motors..., again for reasons of context and early detection. The solution should continuously ingest the collected raw data such as vibrations, temperatures, electrical currents, inclinations at sufficiently high frequencies... At the same time the relevant operational parameters such electrode power, slewing angle, electrode height, lid position… have to be ingested as well. This type of combined data stream forms the basis of the follow-up.
As a first stage all data has to classified, based on properties of each melt, but also on the stage of the process. As such the behavior of the different components under similar conditions can be tracked with time. All these measurements, the raw data, subsequently have to be combined and/or transformed into a series of derived descriptive parameters, preferably using continuously running automatic algorithms. A first set of higher-level parameters should be used to predict degradation and detect deviations in an early stage. Whenever such a deviation is detected, the additional, higher resolution parameters should be available to identify the exact nature and location of the issue, without requiring additional manual calculations. This should be done using a decision tree and the associated alarming module, informing the on-site experts of the situation at hand. Depending on the modules (combinations of sensors and algorithms) activated, the platform can be applied to track various aspects of relevance to the state-of-health of the asset in parallel: bearing degradation, bearing deformation, busbar mounting issues, structural changes, overloads...
By following this kind of approach one can continuously focus on increasing the smoothness of the operations, while the algorithms will pick up all aspects that might have a negative impact on availability or efficiency in the future. The perfect companion for the melt shop: a virtual engineer.