Study on Processing Effect Prediction System of AFM for Injector Hole of Twin Flapper-Nozzle Valve

The processing effect prediction system is developed in order to solve the diffculty of selecting processing parameters in Abrasive Flow Machinig (AFM) for the injector hole of twin flapper-nozzle servo valve. In this paper, the relationship between major processing parameters and machining quality is analysed firstly. Then, a prediction model is created by RBF neural network and trained by the effective sample data which are collected from the properly designed grinding experiments. Finally, based on this prediction model, the processing effect prediction system is realized to validate the model and the experimental results show that the processing effect prediction model has high forecast accuracy and can provides reliable reference for the selection of process parameters.