AI for Metallurgy - Challenges and Opportunities
主讲嘉宾：英国皇家工程院院士、英国莱斯特大学教授Hongbiao Dong
Guest Speaker: Prof. Hongbiao Dong，Fellow of Royal Academy of Engineering，professor 偶of University of Leicester
讲座时间：2024年8月6日 14:30-16:00
Date & Time: 14:30-16:00, Aug.6, 2024
讲座地点：徐祖耀楼姚征报告厅500号
Location: Yiucheng Lecture Hall 500, Xu Zuyao Building
邀请人：李军副教授
Inviter: Assoc.Prof. Jun Li
主讲嘉宾介绍/ Biography：
Professor Hongbiao Dong is a Fellow of Royal Academy of Engineering, the President of Association of British Chinese Professors (ABCP), an outstanding engineer who has made pioneering and significant contributions in numerical modelling of metal processing and digital transformation of metal industry. He is the Deputy Head and Director of Research of the School of Engineering at Leicester, a Director of EPSRC Doctoral Training Centre in Innovative Metal Processing at Leicester. He is internationally renowned for modelling work and its application in metals and manufacturing. His research aims to bring knowledge-inspired decision making to the production routes of high value-added engineering structures, such as aero-engine components, deep-sea oil and gas transport systems. He successfully led a major European project on modelling of advanced welding. He chaired STFC-ISIS Engineering User Committee and a member of Steering Committee of the European Materials Modelling Council, a recipient of the Metrology for World Class Manufacturing award and a Royal Society Industry Fellow at Rolls-Royce Precision Casting Facility. His work has been exploited in various sectors, including in aerospace, energy, transport and steel sectors. He has played a key role in establishing international partnerships between UK and China, India and South Africa.
报告内容概要/Abstract：
The application of AI is a key enabling factor of the digital transformation of metal industry. However, this application has significant implementation challenges. This talk will use a case study of BOF steelmaking process to analyses these challenges and proposes ways of addressing them. Among the main identified challenges are high dimensionality, data scarcity, validity and the difficulty in characterizing abnormal events/conditions as well as modeling processes, which require the creative use of different learning paradigms as well as incorporating physics-based models and principal in the data analysis process, which can make the learning process more efficient. Other challenges are related to the need of developing reliable in-line sensors, adopting interoperability data models and tools, and implementing the continuous measurement of critical variables.