A Component-level Resistance Characteristic Prediction Method for Marine Gas Turbine Intake Systems Based on a Neural Network Model

In order to improve the accuracy and efficiency of resistance characteristic prediction of marine gas turbine intake system, a resistance characteristic prediction method based on component-level numerical simulation and neural network model is proposed in this paper. Aiming at the problems of high calculation cost and lack of flow field details in the traditional overall model, the intake system is divided into four parts: intake cabin, filter, muffler and intake shaft. Based on this, grid independence verification and numerical simulation are carried out respectively. The feasibility of the component-level calculation method is verified (total pressure loss relative error < 10 %, mass flow error < 0.1 %). Through the above component-level calculation method, this paper calculates the relationship between wind speed and total pressure loss under different wind directions. The BP neural network optimized by genetic algorithm is used to construct the total pressure loss surrogate model. This study provides a feasible solution for the performance prediction of marine gas turbine intake system, which has good regression and generalization performance. This study holds significance for the efficient and precise prediction of resistance characteristics in marine gas turbines.