Abstract

A basic oxygen furnace (BOF) is the main unit process for refining carbon steel. The aim of this work was to study the use of vibration and audio signal measurements to monitor, predict, and control the BOF process. Vibration and audio data from nearly 300 blows were collected and analyzed together with process variables. We could confirm high correlations between some of the process variables and vibration and audio measurements. Median filtered low-frequency (3–20 Hz) audio as well as X- and Z-direction acceleration root mean square (RMS) time series correlate with the off-gas temperature, although this is much more significant for the audio data. For Y-direction measurements (the upward direction) the correlation is negligible. The low-frequency audio and vibration data are likely related to the rate of decarburization. Median filtered mid-frequency (100–1000 Hz) audio as well as X-, Y-, and Z-direction acceleration RMS time series correlate with the lance height measurement during the interval 20–600 seconds from the beginning of oxygen blow. For the audio data, the correlation was high even without median filtering. We suggest that audio and vibration activity in this frequency range is possibly related to the formation of the metal–slag–gas foam and maybe even to slopping.