LIFETIME FORECASTING FOR HEMISPHERICAL RESONATOR GYROSCOPE WITH WAVELET ANALYSIS-BASED GM(1,1)
Yunxia Zhang1, Chenglong Dai2, Jifeng Cui2
1De Network and Information Center, Southeast University, Nanjing 210018, China
2College of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China
Because of high cost and small batch of spacecraft like flywheel and gyroscope, how to estimate their reliability and lifetime becomes a tough task. A method to predict the lifetime of hemispherical resonator gyroscope (HRG) is put forward in this paper. This method utilizes grey correlation and mean absolute percent error (MAPE) to estimate the reliability of predictive data sequence. For reducing noise, Daubechies wavelet is used to decompose and reconstruct the test data in the paper as well. After pre-processing, predictive data sequences are gained by using GM(1,1) prediction model and then according to grey correlation and MAPE of each predictive data sequence, the threshold value meets conditions can be gained. Finally, the lifetime of HRG is predicted with using the threshold value. In this paper, the method is applied to the data of one type of HRG provided by a research institute in China and the result shows the gyroscope can normally run 4780 days at least, namely about 13.10 years.