During my one-month research stay at the University of Pisa (UNIPI), I had the opportunity to collaborate with the Department of Civil and Industrial Engineering – DICI, an inspiring hub of expertise in process control and industrial automation. This visit marked the beginning of my research on Recursive Dynamic Mode Decomposition for Online Modeling and Predictive Maintenance in Chlor-Alkali Electrolysis, a project aimed at enhancing predictive maintenance strategies for electrolysis systems of the industrial partner.
Research and Collaboration
My time at UNIPI was centered on improving data-driven modeling techniques for impurity buildup in ion-exchange membranes. Unlike traditional approaches that focus on voltage prediction, my research integrates impurity concentration estimates, allowing for a more comprehensive understanding of membrane fouling dynamics. By leveraging Dynamic Mode Decomposition (DMD), I explored how real-time laboratory measurements can be used to update predictive models continuously, ensuring accurate long-term performance assessment.
Throughout my stay, I engaged in fruitful discussions with faculty and researchers, refining my methodology and receiving valuable feedback on my approach. I also contributed to the SIPPY Python package, improving its usability and standardizing interfaces for dynamic system identification—efforts that resulted in over 21,000 lines of code modifications to enhance clarity for users and developers alike.
Presentation and Impact
As part of my visit, I presented my findings to the faculty and students at UNIPI. This session not only showcased my progress but also sparked insightful discussions on how my methodology could be extended to other industrial applications. The feedback from this session has significantly shaped my next research steps, particularly in refining the model’s capability for real-time closed-loop predictions.
Beyond research, the stay provided an invaluable cultural and professional experience. The historical city of Pisa, with its rich academic heritage and vibrant research community, offered the perfect setting for innovation and collaboration. Exploring Tuscany’s landscape, architecture, and cuisine added a refreshing balance to my intense research schedule.
I extend my gratitude to Professor Gabriele Pannocchia from UNIPI and Professor Miroslav Fikar from the STUBA for facilitating this opportunity. Their guidance and support have been instrumental in making this research stay both productive and memorable.
This experience has laid the foundation for future collaborations, and I look forward to building upon the insights gained during my time in Pisa. The advancements in soft-sensing methodologies and predictive maintenance frameworks developed here will undoubtedly contribute to more reliable and efficient industrial electrolysis systems.
Text/Photo: Marek Wadinger (STUBA)
This project has received funding from the European Union’s Horizon under grant no. 101079342 (Fostering Opportunities Towards Slovak Excellence in Advanced Control for Smart Industries).
0 Comments