Numerical Simulation and Artificial Neural Network Modeling of Cavitation on 2D Tandem Hydrofoils Arrangement

Abstract In the current research, we used a parametric approach to investigate cavitation with tandem-arranged Clark Y-11.7% hydrofoils. We simulated the tandem-arranged hydrofoil system using ANSYS FLUENT with the Zwart cavitation model combined with the Transition SST turbulence model. To substantiate the accuracy and reliability of the numerical model, computations were performed for an individual Clark Y-11.7% hydrofoil and compared to experimental results. The following parameters were studied in tandem arrangement: cavitation number( σ ), horizontal direction spacing ratio between two hydrofoils( x ), vertical direction spacing ratio between two hydrofoils( y ), angle of attack (AOA) of the upstream hydrofoil ( α 1 ), and angle of the downstream hydrofoil relative to the upstream hydrofoil ( α 2 ). The lift coefficient ( C L ) was used to estimate due to the acceptable error. An optimal artificial neural network (ANN) model was trained based on the five-parameter combination from numerical simulations. The model’s weights and biases were presented, providing a prediction method for hydrofoils in tandem arrangement with cavitation. A global sensitivity analysis based on the trained model is implemented, which reveals the AOA of the downstream hydrofoil ( α 2 ), the horizontal spacing ratio ( x ), and the AOA of the upstream hydrofoil ( α 1 ) have a main effect on the performance of the tandem arrangement.