Novel Methodology for Failure Detection in Gas Separation Membranes

This study introduces a new methodology for detecting failures in gas separation membranes, with the goal of improving reliability and reducing maintenance costs. This project is performed at Bright Renewables, using historic data collected from its many installations. By analyzing this historic field data, a model was developed to identify potential failures, significantly reducing the detection time. This faster detection could lead to earlier interventions, facilitating improved maintenance planning and reduced downtime. The analysis highlights a positive business case achieved through more timely maintenance actions. The model not only improves failure detection but also links each identified fouling mechanism to its root cause, providing deeper information for better management. Additionally, the proposed methodology integrates data analysis with operational decision-making, potentially extending the useful life of the membranes and optimizing resource allocation. While further testing in real-time environments is necessary to fully validate these findings, the results suggest that this approach can provide substantial benefits in terms of cost efficiency, system availability, and overall maintenance effectiveness.