Provide a Predictive Model for the Preliminary Design of Swirl Injectors Based on Neuro-Fuzzy Model

Abstract

the effect of the internal geometry of the injector of a low-thrust monopropellant thruster on the characteristics of the outlet liquid sheet, such as the liquid sheet thickness, the spray cone angle, the average output velocity, and its mass flow rate, have been studied. For this purpose, simulation of the internal flow based on the computational fluid dynamics was performed to predict the output flow characteristics, then parametric studies were conducted to investigate the effect of geometry. In order to extract a predictive model for the preliminary design of the injector, a neural fuzzy network model was used. Using this model, the results of parametric analyzes were used to obtain a geometric specification without geometric modeling and computational fluid dynamics analysis or the use of initial design methods. The purpose of the implementation of the neural network model is to obtain an estimate of the internal geometry of the injector by entering the desired macroscopic characteristics of the output flow.