Research on balancing method of single-stage rotor blade mass moment based on pointer network

Abstract

The blades of the aero-engine single-stage rotor are manufactured inevitably with quality errors that will pose a great impact on the unbalance of the assembled single-stage rotor. Therefore, to reduce the initial residual unbalance of the assembled single-stage rotor of the blades, this paper establishes a pointer network optimization model to predict the reasonable sequence of the blades and controls the absolute value of the sum of vector errors of the single-stage rotor blade mass moment within the specified permissible values by learning the data relationship between the input mass moments and the sequence of the output blades. The trained model can be used to solve other problems with the same distribution of data without retraining. The simulation and validation results show that the research method can predict the blade assembly sequence efficiently and greatly improve the blade assembly accuracy, which is of guiding significance to the aero-engine assembly.