INVITED SPEAKER

Zhenglei He is currently an Assistant Professor of Automation and Intelligent Manufacturing at the State Key Laboratory of Pulp and Paper Engineering, South China University of Technology, China. He holds a Ph.D. degree of Computer Engineering, Automation and Signal Processing from University of Lille, France.

He has published more than 30 papers in SCIE peer-reviewed international journals and proceedings of international conferences. He contributed 4 chapters with books published by Springer Nature and CRC Press, Taylor & Francis Group. He is the board member of Advanced Materials & Sustainable Manufacturing, and the guest editor of Applied Science, International Journal of Computational Intelligent Systems, and Journal of Smart Environments and Green Computing etc. He has chaired / co-chaired special sections of International conferences of FLINS 2022, DSBFI2023, ISKE2023, GCPC 2023 etc. He has co-supervised 13 postgraduate students.

Title: Reducing Carbon Emission of Papermaking Wastewater treatment Process through Data-driven Models and Multi-agent Reinforcement Learning

Abstract: 

Reducing Carbon Emission of Papermaking Wastewater treatment Process through Data-driven Models and Multi-agent Reinforcement Learning and abstract: Due to different sources and the water-using habits, the wastewater of papermaking fluctuates sharply over time. Quality control of the effluent in the papermaking wastewater treatment process (PWTP) is rather challenging and costly. Concerns over PWTP's environmental effects are becoming more prevalent, particularly with regard to its greenhouse gas (GHG) emissions. This research established a multi-agent deep reinforcement learning framework to simultaneously optimize process cost, energy consumption, and GHG emission in the PWTP, subject to the effluent quality. This allowed for the realization of the economic, energy-efficient, and environmentally friendly objectives. The biological treatment process of wastewater used in papermaking was simulated using Benchmark Simulation Model No. 1 (BSM1). The model system was confirmed using real data acquired from the papermaking industry, and the default wastewater influence data from the BSM manual was utilized for training. The outcomes show that the proposed method outperforms conventional techniques in efficiently identifying the best control strategies for a variety of targets.