Abstract:

This seminar introduces a project that applies data analytics and machine learning techniques to monitor, evaluate, and optimize the operational performance of an electrolyzer system. The core methodology involves the processing and exploratory analysis of operational data to identify correlations between key process variables and to develop models capable of assessing equipment health. To achieve this, the study will utilize historical and real-time operational datasets, primarily including electrical variables (cell/stack voltage, applied current, and power consumption), thermodynamic process parameters (operating temperature, system pressure, and water flow rate), and production metrics (hydrogen yield). Ultimately, the project expects to deliver clear degradation curves to understand efficiency loss over time, anomaly detection capabilities for predictive maintenance, and data-driven recommendations for optimal operating windows. These outcomes will maximize hydrogen production efficiency while minimizing energy costs and equipment wear, transforming raw data into actionable intelligence for operational decision-making.

Speaker: 

João Victor Thomas Feyh is a Chemical Engineer graduated from the State University of Santa Catarina (UDESC). He holds a Master’s degree in Chemical Engineering, where he focused on hydrogen production from Steam Reforming of Ethanol. Currently a PhD candidate at Unicamp, he is completing a research period as a visiting scholar at THI (Technische Hochschule Ingolstadt). His doctoral work centers on developing data-driven Digital Twin models for electrolyzers. Through his research, João seeks to understand how these systems age over time and extract crucial metrics, such as the Levelized Cost of Hydrogen (LCoH) and Life Cycle Assessment (LCA), to drive the future of sustainable green energy.

Link to registration: https://docs.google.com/forms/d/e/1FAIpQLSf-76vSixnuC48vDZGlw7vdpxMV32NCig6nRad5KtWAQBLufA/viewform