Soft Sensors and Diagnostic Models Using Real-Time Data of Coke-Making at Tata Steel

• Published in: Transactions of the Indian Institute of Metals 2024, Vol.77 (8), p.2191-2196
• Main Authors: Raj, Sristy, Ganguly, Adity, Bhushan, Ashutosh, Shankar, Amitabh, Sen, Subhadra
• Format: Article
• Language: English
• Subjects: Chemistry and Materials Science; Coke ovens; Corrosion and Coatings; Levers; Materials Science; Metallic Materials; Original Article; Parameter identification; Raw materials; Sensors; Steel industry; Steel making; Tribology
• ISSN: 0972-2815; 0975-1645
• DOI: 10.1007/s12666-024-03309-9

Constant generation of massive data in steel industry offers huge opportunity to enable recognition of phenomena involved and identification of levers for influencing the process towards higher efficiency. Through few illustrative examples from coke-making processes, this paper attempts to show how process visualization and diagnostics through soft sensors has contributed to get to the current level of understanding and quantification. Thinking through basic phenomena and having it visualized and quantified using data has helped get a grip on understanding how raw materials and process conditions influence performance outcomes viz. productivity, efficiency and quality. Various terms have been used to describe the approach—‘derived parameters’, ‘soft sensors’ or even empirical models. Further, process performance gets diagnosed faster and sometimes anticipate helps in initiating corrective action through identified control levers. Together they represent an approach which adds on to the efforts of sensor development and analytical modelling to size up performance of coke ovens.