Abstract:
The process of forming ferroalloys in submerged arc furnaces requires high temperatures. These are achieved by the intensive use of electrical energy, which is converted into thermal energy inside the furnace, using Søderberg-type electrodes. The material that feeds the furnace and is used to form the electrode is called electrode paste and is available in cylindrical, block, or briquette shapes. When the paste is introduced into the process, it is subject to conditions that may create asymmetries that can affect the formation of the electrode and cause it to break. This issue is difficult for operating teams to detect due to the number of level measurements taken at various points. In view of this, there is interest in solutions capable of detecting the disturbance, since it has a negative impact on the formation of electrodes and, consequently, on the production of ferroalloys. In this context, the use of expert systems presents itself as an alternative to aggregate and structure operational knowledge to infer the occurrence of asymmetry confidently and automatically. Considering this, the paper presents a fuzzy expert system approach for inferring this potential risk in the formation of the Søderberg electrode.