article · Fluids
Identifying multiphase flow regimes in vertical pipelines is important for safe and efficient operations in the oil and gas sector. Traditional assessment techniques often suffer from subjectivity, particularly in transitional flow regimes. To address this, a hybrid deep learning model integrates a convolutional neural network for spatial features, a transformer network for long-range dependencies, and a physics-encoding network that embeds physical constraints. The combined system was trained and evaluated on an experimental dataset containing 2,131 wire mesh sensor images annotated via a semi-automated workflow. The resulting hybrid network attained a 95.91 percent test accuracy and a macro F1-score of 0.96, outperforming four comparative models. Analysis shows the convolutional component provides the primary discriminative signal, while the transformer and physics-inspired branches deliver consistent refinements. The physics-based component acts as a spatial regulariser, improving model interpretability.
Detecting flow patterns accurately inside vertical pipelines helps prevent operational hazards and inefficiencies in fluid transport systems. By embedding physical rules alongside spatial and sequential data processing, computational tools can classify complex fluid mixtures more objectively. This provides pipeline operators with more dependable, interpretable diagnostic data to maintain steady flow conditions and reduce the risks associated with subjective manual interpretation.
The technology is aimed at pipeline monitoring and flow management within the oil and gas sector. It could enable automated diagnostic software for industrial piping networks, particularly in identifying difficult transitional flows. The framework represents applied research tested on experimental sensor images, indicating it is at an intermediate development stage and requires further integration with field instrumentation before reaching commercial use.
AI-generated from the published abstract. Always read the original work before citing.
Accurate identification of multiphase flow patterns in vertical pipelines is critical for operational safety and efficiency in the oil and gas industry. Yet, conventional methods struggle with subjectivity and transitional regimes. This study develops and integrates three neural network architectures: a convolutional neural network (CNN) for spatial features, a transformer neural network (TNN) for long-range dependencies, and a physics-encoding network (PEN) for embedding physical constraints. These are combined into a hybrid framework trained on an experimental dataset of 2131 images from a wire mesh sensor, annotated using a semi-automated pipeline. Results show the hybrid model achieved 95.91% test accuracy with a macro F1-score of 0.96, the highest of the four models evaluated, with its main advantage in transitional regimes. A multi-seed ablation shows that the convolutional branch provides the dominant discriminative signal, while the transformer and physics-inspired branches added complementary improvements that are consistent across runs. This hybridisation mitigates individual model weaknesses, with the physics-inspired branch acting as a spatial regulariser that improves interpretability, providing a robust and objective tool for reliable pipeline monitoring.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.3390/fluids11090210
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
