Abstract：The frequency domain three-probe method is a common method for separating spindle rotation errors. Its error separation accuracy is greatly affected by the noise in the measured signal. Inappropriate noise reduction methods will distort the test results. To this end, a spindle rotation error separation and noise reduction method based on LSTM-GRU neural network is proposed.First, a test system is built using the optimized sensor angle by genetic algorithms and the spindle rotation error signal is acquired.Then, the Kalman filter is configured to reduce the noise of the three sensor signals, the synchronous rotation error and asynchronous rotation error are separated by the three-probe method in frequency domain.Finally, the LSTM-GRU model is used to reduce the noise of synchronous and asynchronous rotation error respectively.The noise reduction results of the LSTM-GRU model are compared with the results of LSTM-LSTM model、Kalman filtering and Wavelet threshold denoise methods.The Allan variance is calculated to evaluate the noise reduction effect of different methods.The experimental result shows that after noise reduction using the LSTM-GRU model, the Allan variance of the synchronous rotation error is 2.014×10-8mm2 and the Allan variance of the asynchronous rotation error is 3.967×10-8mm2, which are both less than the results of Kalman filtering and Wavelet threshold noise reduction.The noise reduction effect of the LSTM-GRU model is optimal.The asynchronous rotation error of the spindle at the test speed of 6000r/min is 2.42μm and the asynchronous rotation error is 3.21μm, which meets the actual situation.
迟玉伦,李希铭,朱文博,余建华. 基于LSTM-GRU神经网络的机床主轴回转误差分离降噪研究[J]. 计量学报, 2024, 45(11): 1615-1625.
CHI Yulun,LI Ximing,ZHU Wenbo,YU Jianhua. Research on the Separation and Noise Reduction Method of Spindle Radial Rotation Error Based on LSTM-GRU Neural Network. Acta Metrologica Sinica, 2024, 45(11): 1615-1625.
HIRONAO A, NARUO O. Three-point method roundness measuring machine ［J］. Mechanical Research, 1978, 30(6): 753-759.
WANG S H. Research on on-line dynamic testing technology of high-speed and high-precision spindle rotational error［D］. Guangzhou: Guangdong University of Technology, 2006.
ISO 230-7 (2015) Test code for machine tools. Part 7: Geometric accuracy of axes of rotation, International Standardization Organization［S］. Geneva, Switzerland: International Organization of Standardization, 2015.
［4］
LEE J, GAO W, SHIMIZU Y, et al. Spindle Error Motion Measurement of a Large Precision Roll Lathe ［J］. International Journal of Precision Engineering and Manufacturing, 2012, 13(6): 861-867.
［5］
SHI S, ZHANG H, QU J, et al. Measurement uncertainty propagation in spindle error separation techniques-Investigation by means of stochastic spectral method ［J］. International Journal of Machine Tools and Manufacture, 2019, 141: 36-45.
TOGUEM TAGNE S C, VISSIERE A, DAMAK M, et al. An advanced Fourier-based separation method for spindle error motion identification ［J］. Precision Engineering, 2021, 74: 334-346.
［12］
CASTRO H. F. F. A method for evaluating spindle rotation errors of machine tools using a laser interferometer ［J］. Measurement, 2008, 41(5): 526-537.
GU Q T. Using the kalman filter approach for the evaluation of roundness errors ［J］. Acta Metrologica Sinica, 1990,11(2): 124-129
FUJIMAKI K, SASE H, MITSUI K. Effects of sensor noise in digital signal processing of the three-point method ［J］. Measurement Science and Technology, 2008, 19(1): 1-9.
FENG M, DU D Y, WANG X J, et al. Research on precision spindle rotational error and stiffness testing technology ［J］. Journal of Mechanical Engineering, 2021, 57(13): 18-25.
［19］
MOHAMED A. El-BRAWANY, DINA A I, et al. Artificial intelligence-based data-driven prognostics in industry: A survey ［J］. Computers Industrial Engineering, 2023, 184: 1-13.
［22］
HAN S P, MENG Z, ZHANG X C, et al. Hybrid Deep Recurrent Neural Networks for Noise Reduction of MEMS-IMU with Static and Dynamic Conditions ［J］. Micromachines, 2021, 12(2): 214-237.
［25］
XIANG K, WANG W, QIU R, et al. A T-Type Capacitive Sensor Capable of Measuring 5-DOF Error Motions of Precision Spindles ［J］. Sensors, 2017, 17(9): 1975-2000.
［26］
YIN T, TO S, DU H, et al. Effects of wheel spindle error motion on surface generation in grinding ［J］. International Journal of Mechanical Sciences, 2022, 218: 1-19.
ZOU X F, LU X Y. Estimate method of MEMS gyroscope performance based on allan variance ［J］. Micronanoelectronic Technology, 2010, 47(8): 490-493.
WU X J, GUI Y H, WANG A C, et al. Vehicle trajectory prediction based on LSTM-GRU integrating dropout and attention mechanism ［J］. Journal of Hunan University(Natural Science), 2023, 50(4): 65-75.
［2］
FANG Z X, CHEN Z Z, TAO F, et al. Simultaneous calibration of probe parameters and location errors of rotary axes on multi-axis CNC machines by using a sphere ［J］. Measurement, 2022, 188: 1-15.
［9］
CHEN Y, ZHAO X, GAO W, et al. A novel multi-probe method for separating spindle radial error from artifact roundness error ［J］. The International Journal of Advanced Manufacturing Technology, 2017, 93(4): 623-634.
CAPPA S, REYNAERTS D, AL-BENDER F. A sub-nanometre spindle error motion separation technique ［J］. Precision Engineering, 2014, 38(3): 458-471.
［30］
BAI J, WANG Y Z, WANG X H, et al. Three-Probe Error Separation with Chromatic Confocal Sensors for Roundness Measurement ［J］. Nanomanufacturing and Metrology, 2021, 4(4): 247-255.
MA J X, ZHAO W H, ZHANG G B. Research Status and Analyses on Accuracy Retentivity of Domestic CNC Machine Tools ［J］. China Mechanical Engineering, 2015, 26(22): 3108-3115.
［8］
王少蘅. 高速高精密主轴回转误差在线动态测试技术研究［D］. 广州: 广东工业大学, 2006
ZHOU J Q, ZHANG R, LING M X. Research on online testing technology of machine tool spindle rotary error based on three-point method ［J］. Computer Measurement & Control, 2018, 26(3): 58-61.
JING S H, YUN W G, HAN S P. Gyroscope Noise Reduction Method Based on DeepCirculation Neural Networks ［J］. Aerospace Control and Application, 2020, 46(5): 65-72.
FENG M, ZHOU C Y, ZHANG K, et al. Systematic noise separation technique in slewing error testing ［J］. Journal of Beijing University of Aeronautics and Astronautics, 2020, 46(4): 666-673.
［18］
ANANDAN P K, OZDOGANLAR B O. A multi-orientation error separation technique for spindle metrology of miniature ultra-high-speed spindles ［J］. Precision Engineering, 2016, 43: 119-131.
［20］
YANG H, ZHANG J, LI S, et al. Bi-direction hierarchical LSTM with spatial-temporal attention for action recognition. ［J］. Journal of Intelligent and Fuzzy Systems, 2019, 36(1): 775-786.
［24］
DING F, LUO X, CHANG W, et al. In Situ Measurement of Spindle Radial and Tilt Error Motions by Complementary Multi-probe Method ［J］. Nanomanufacturing and Metrology, 2019, 2(4): 225-234.
2024, Vol. 45 
