Multi-objective optimization of protonic ceramic electrolysis cells based on a deep neural network surrogate model

Protonic ceramic electrolysis cell (PCEC) stands out as a promising device to realize large-scale green hydrogen production. This research is dedicated to advancing the optimization of PCEC, specifically targeting key performance indicators including voltage, current density, and Faradaic efficiency (FE). The central aim is the expeditious determination of optimal trade-off points that harmonize electrochemical performance and FE. To achieve this, a comprehensive framework is proposed, integrating three distinct methodologies: Multiphysics model, deep neural network, and multi-objective optimization algorithms. The investigation reveals the leakage current along the thickness of the cell is significant relative to its length. Furthermore, an increase in current density from 0.4 A cm-2 to 0.8 A cm-2 results in a reduction of FE and the uniformity of FE by 21.3% and 8.8%, respectively. The identified optimal point at 0.78 A m- 2, 600 degrees C delivers a 12.8% improvement in performance compared to the base case. The primary contribution of this work includes introducing a novel framework which substantially accelerates the optimization of PCEC as well as highlighting the importance of addressing leakage current issues during PCEC operation. The proposed framework has broader applicability for addressing optimization problems of PCEC and advancing other clean energy technologies.