The 12th World Congress of Chemical Engineering and the 21st Asian Pacific Confederation of Chemical Engineering Congress (WCCE12 & APCChE 2025) were held in Beijing, China, from July 14 to 18, 2025. The event was hosted by the Chemical Industry and Engineering Society of China (CIESC) and co-organized by China National Petroleum Corporation (CNPC), China Petrochemical Corporation (Sinopec), and Beijing University of Chemical Technology, bringing together over 6,000 professionals from more than 50 countries under the theme “Paradigm Shifting in Chemical Engineering for Global Challenges.”
On July 17, 2025, we had the distinct honor of participating at the congress. As a Professor of Process Systems Engineering at the National Technical University of Athens, Prof. A. Kokossis was invited to present a dedicated talk on behalf of the Bioindustry 4.0 project, a cutting-edge EU-funded initiative perfecting advanced technologies to empower the European bioindustry sector.
Within the framework of the project, Prof. A. Kokossis delivered an invited oral presentation under the Session 27: Innovation and Practice of Industrial Software in Process Manufacturing.
The presentation, entitled “Bioreactor Digital Twins: Process and Biocatalyst Intelligence Using the Latest Developments in AI and ML” showcased the project’s innovative work on digital twins for bioprocessing, developed collaboratively within the Bioindustry4.0 community.
At the core of the presentation was the transformative role of Physics-Informed Neural Networks (PINNs) as a foundational framework for industrial bioreactor digital twins. By integrating data-driven machine learning techniques with first-principles models, the proposed approach demonstrated significantly enhanced predictive performance over both short- and long-term horizons, effectively addressing the challenges posed by sparse and limited experimental datasets. A key focus was the deployment of these hybrid digital twin models within Model Predictive Control (MPC) frameworks. By capturing the underlying nonlinear and multiscale dynamics of bioprocesses, the AI-enhanced controllers consistently outperformed conventional control strategies. Furthermore, the capability for online kinetic parameter estimation enables the creation of digital threads that link real-time bioprocess operation with strain and biocatalyst engineering. This methodology extends into pathway discovery and strain optimization, substantially accelerating the Design–Build–Test–Learn (DBTL) cycle.
Highlights from the talk include:
- Development of scalable, physics-informed digital twins for industrial bioreactors.
- Advanced AI-enabled process control strategies enabling self-regulated and autonomous bioreactor operation.
- Digital twinning methodologies that support biocatalyst and strain optimization, accelerating innovation in industrial biotechnology.
The approaches presented support the digital transformation of bio-based industries, promote sustainable production and resource efficiency, and facilitate the adoption of next-generation biorefinery technologies, aligning with the broader goals of Bioindustry4.0.
Participation at WCCE12 offered valuable international visibility for the project and highlighted IPSEN’s contributions to applying digital, AI-driven, and sustainable approaches in industrial biotechnology.