Deep learning-based quantitative sand particle measurement of slug flow via Bayesian-optimized CNN-LSTM architecture

Key Points

• Quantitative sand particle measurements improve accuracy in slug flow assessment, enhancing flow management strategies.
• The method utilizes a Bayesian-optimized CNN-LSTM architecture to achieve robust particle analysis.
• Implementation within simulated slug flows aids in understanding complex multi-phase interactions, facilitating development.
• Findings indicate that accurate particle measurements may enable better predictions in industrial applications, optimizing resource usage.