Monitoring the Reliability of Two-Phase Flows using Machine Learning Strategies

(27:57 + Q&A) Yoonjin Won, Associate Professor, University of California, Irvine
From the 2024 IEEE Symposium on Reliability for Electronics and Photonics Packaging
Summary: Two-phase flow reliability refers to the consistent and safe operation of systems that involve the simultaneous movement of two distinct phases, such as liquid and vapor. In many industrial processes, such as electronics cooling, nuclear reactors, and energy conversion processes, understanding and monitoring the dynamics of two-phase flows is critical for maintaining system performance and safety. Traditional methods of monitoring and analyzing two-phase flows often rely on physical models and empirical correlations, which may not account for the complex and dynamic nature of these systems. This work explores the application of machine learning strategies to enhance the reliability monitoring of two-phase flows. We present a framework that integrates data-driven approaches, to analyze flow patterns, detect anomalies, and predict system behaviors. By leveraging historical data from sensors and operational parameters, our model improves the accuracy of flow characterization and failure prediction. The results demonstrate that machine learning techniques can significantly enhance the monitoring capabilities of two-phase flows, leading to improved decision-making and proactive maintenance strategies. This study underscores the potential of advanced analytics in transforming the management of two-phase flow systems and offers insights into future research directions for integrating machine learning in industrial applications to ensure their safe operation.

For additional talks from this REPP, or earlier ones, please visit https://attend.ieee.org/repp