Acoustic emission monitoring to optimize the end-of-cast in blast furnaces

Abstract

When tapping a blast furnace, a break-through of gas through the taphole at the end of a cast needs to be prevented, both to preserve a healthy state of the taphole and to prevent gas and dust from escaping into the environment. In this paper, two acoustic techniques are presented that can be used to prevent gas emissions from the taphole at the end of a cast. The first approach is by using the spectral analysis of the recorded data, the second approach uses a neural network to recognize the popping sounds that announce the slag-gas interface is approaching the taphole level. It was found that with both methods the end of the cast is detected about 4-2 min before the cast is ended.