Cracks are a common form of damage in infrastructure, posing significant risks to both personal safety and property. Along with the development of deep learning, visual-based crack automatic detection has been widely studied. However, this task is still challenging due to complex crack topology, noisy backgrounds, unbalanced categories, etc. To address these challenges, this research proposes a novel hybrid network, named CrackNet, which leverages the strengths of both CNN and transformer. On the encoder side, CNNs are employed to extract multi-level local features, while transformers are used to model global dependencies. Additionally, a strip pooling module is introduced to suppress irrelevant regions and enhance the network’s ability to segment narrow and elongated cracks. On the decoder side, an attention-based skip connection strategy and a mixed up-sampling module are implemented to restore detailed information. Furthermore, a joint learning loss combining Dice and cross-entropy with dynamic weighting is proposed to mitigate the effects of severe class imbalance. CrackNet is trained and evaluated on three public crack datasets, and experimental results show that the proposed model outperforms several well-known deep neural networks, with a particularly noticeable improvement in recall rate.

Considering the new applications of high-pressure gas pumps in practical engineering, the gas-liquid two-phase flow in small-diameter vertical pipes of the system in this study was tested. The gas volumetric flow rate, water volumetric flow rate, and working efficiency at different submergence height, pressure, and head values were measured. The results revealed that the pressure of high-pressure gas determines the gas consumption of the device and that a linear relationship exists between them. The void fraction of the vertical pipe effectively reflects the mixing of gas and water and affects the pumping capacity of the system. The water volumetric flow rate and working efficiency of the system reach their maximum values simultaneously when the void fraction reaches β₀, which is a fixed value. A smaller diameter of the gas inlet pipe d enables the pump water efficiency to reach its peak at a lower pressure. We also present an optimization scheme for the selection of the diameter of vertical pipe D. These findings help facilitate a better understanding of the mechanism of the phase distribution and the interaction of the gas-water phases with each other in small-diameter vertical pipes and provide the basis for experiments to further improve the device performance.