Internal flow characteristics of a tank with multiple jet inflows and liquid sloshing.

This article focuses on numerical investigations of flow field characteristics in a multi-inlet aquaculture tank, emphasizing the coupled effects of multiple jet inflows and liquid sloshing under surge excitation. Using the Reynolds Stress Model combined with the Volume of Fluid method and Adaptive Mesh Refinement, the study systematically analyzes how inlet pipe diameter, inflow rate, and immersed depth influence velocity distribution, flow uniformity, vortex structures, and pressure fields within the tank. Additionally, a hybrid Proper Orthogonal Decomposition (POD) and Long Short-Term Memory (LSTM) neural network framework was developed to efficiently predict the flow field evolution, demonstrating high accuracy compared to CFD simulations. The findings reveal that inlet pipe diameter significantly affects velocity magnitude and uniformity, while sloshing impacts become pronounced at lower liquid filling levels, particularly near the free surface. The validated numerical approach and POD-LSTM model provide valuable tools for optimizing aquaculture tank design and operation in offshore environments.