In order to improve the precision of gas-dynamic bearing gyroscope

a new method for predicting the dynamic error of the gyro scope is proposed

which combines lubrication calculation theory

gyroscope error analysis and long short-term memory(LSTM) neural network.Firstly

the relevant mechanical parameters of the gyroscope are obtained according to the lubrication calculation and rotor dynamics equation

and the error of the gyroscope is derived from the interference torque formula.Then

the relationship between the five-degree-of-freedom motion and the output error of the gas-dynamic bearing gyroscope are established.The 6000 sets of data created under 60 working conditions are substituted into the LSTM neural network model for iterative training

and a new error compensation model is constructed.Finally

the model is tested under different working conditions to verify the precision of the error compensation model.The results show that the error compensation model established by using LSTM neural network for error prediction of five-degree-of-freedom motion is applicable for error prediction of gyroscope.The proposed method has certain theoretical guiding significance for error compensation of high-precision gyroscope.