The three-phase flow in a aeration tank driven by an inverted umbrella aerator is relatively complex, including the processes of the hydraulic jump, air entrainment, and sludge particle sedimentation. A three-phase flow test bench for an inverted umbrella aerator is established for studying its influence on aeration performance. The experiment mainly studies the changed law of aeration performance under different immersion depths or sludge concentrations and measures the flow rate and sludge concentrations in the aeration tank in different working conditions. The results are as follows. (1) The total oxygen transfer coefficient, standard oxygenation capacity, and standard power efficiency increase with the increase in rotational speed. The total oxygen transfer coefficient and standard-charge oxygen capacity first increase and then decrease with the decrease in immersion depth, reaching a maximum at −20 mm immersion depth. The standard dynamic efficiency has a similar trend and reaches a maximum at −8 mm immersion depth. (2) In the aeration tank, the flow velocity near the impeller is faster and has greater turbulence. The shallow water is more profoundly affected by the impeller compared with the deeper water. (3) The shallow-water sludge varies greatly, and the deep-water sludge is distributed uniformly when the inverted umbrella aerator works stably.
16 June 2025 in Nuclear Engineering and Technology
Cavitation-induced degradation is a major failure mechanism in centrifugal pumps. To characterize the degradation trajectory of cavitation damage and assess its severity, this study proposes a deep learning-based method for predicting the progression of cavitation damage in centrifugal pumps. A dedicated experimental platform was developed to simulate severe cavitation scenarios, and multi-channel vibration acceleration signals are synchronously collected from multiple measurement points. A sliding window approach is used to segment the time-series data. A convolutional neural network (CNN) extracts rich, high-dimensional spatial features, while an attention mechanism dynamically assigns weights to different sensor channels to enable multi-source data fusion. Subsequently, a BiLSTM network captures temporal dependencies in both forward and backward directions, improving prediction accuracy. Experimental results demonstrate that the proposed model effectively captures the degradation trajectory of cavitation damage, achieving a mean squared error (MSE) of 0.0098.