Blast Furnace Tuyere Monitoring Powered by Artificial Intelligence
(Room 317)
16 May 22
11:00 AM
-
11:30 AM
The Paul Wurth Tuyere Phenomena Detection System (TPDS) is a camera-based system for continuous monitoring of the blast furnace (BF) tuyere, injection lance, injectant and raceway zone widely installed on high-injection rate blast furnaces. Recent advances in machine learning, in particular in the domain of machine vision, allow improving TPDS substantially by introducing novel monitoring and anomaly detection methods applied to multi-modal data acquired from the BF. This work presents how these artificial intelligence techniques enable robust automated detection of region of interests, continuous injection status classification or detection of hazardous process conditions such as tuyere blockages.