An Efficient Machine Learning Approach for Degradation Forecasting in AEM Water Electrolysis

2026-09-29

AI Summary

This study addresses the challenge of accurately predicting voltage degradation in anion exchange membrane (AEM) water electrolyzers by proposing a mid-term degradation curve prediction method based on heterogeneous experimental data. A rigorous training and evaluation framework tailored for industrial heterogeneous data is designed to enhance model robustness, and the predictive performance of multiple machine learning algorithms—including linear baselines, long short-term memory (LSTM) networks, and convolutional neural networks (CNNs)—is systematically compared. The results demonstrate that the proposed framework effectively achieves accurate mid-term prediction of voltage degradation in AEM electrolyzers, thereby providing a reliable data-driven solution for equipment health management.

Abstract

This study provides a data-driven analysis of a novel dataset of single-cell Anion Exchange Membrane water electrolyzers (AEMWE), operated under constant current load across multiple heterogeneous experimental campaigns. We train and evaluate a range of machine learning models with different complexity, including linear baselines, LSTMs and CNNs, to perform medium-term forecasting of the cell voltage degradation curve. The models are assessed within a rigorous training and evaluation framework specifically designed for heterogeneous industrial data.

Problem

• Anion Exchange Membrane Water Electrolysis
• Degradation Forecasting
• Voltage Degradation
• Machine Learning

Innovation

• Anion Exchange Membrane Water Electrolysis
• Degradation Forecasting
• Machine Learning
• LSTM
• Heterogeneous Industrial Data