Research On Thermal Error Measurement And Modeling Of Gear Measurement Center

Gear measurement center(GMC)is the mainstream equipment of gear measurement.The GMC is gradually applied to the production site,with the improvement of gear measurement accuracy and quality control requirements,.Unlike the constant temperature in the measurement room,the ambient temperature of the production site changes more violently and the temperature distribution is more complex,which will cause complex thermal errors of the GMC,and have a greater impact on the measurement results measurements of the GMC.In this paper,a system is designed for the thermal error of GMC in the actual production site environment.The temperature field of GMC and the thermal errors of key parts are measured at room temperature.Based on the measured data,mathematical models between temperature and thermal errors are established,which lays a foundation for thermal error compensation of GMC.The main work of this paper includes the following aspects:By analyzing the heat source that affects the measurement accuracy of the GMC,the temperature and thermal error detection scheme of the GMC is designed,and the temperature and thermal errors detection system are built.The parallelism thermal errors between the gear axis and the Z axis of the GMC and the spatial thermal error of the probe in the spatial thermal error model at the probe were detected.While detecting the thermal errors,the temperature of the surrounding environment of the GMC and the key position of the fuselage was also collected in real time.The collection of these data lays the foundation for the thermal error modeling of the GMC.The thermal error model of the parallelism between the gear axis and the Z axis of the GMC is established.Firstly,combined with the fuzzy clustering method and the correlation coefficient method,the initial temperature measuring points of the parallelism between the gear axis and the Z axis of the GMC are grouped and optimized to determine the temperature sensitive points,using to reduce the workload of model learning and the influence of strong correlation between data sets on model learning.Then,the thermal error model of the parallelism between the gear axis and the Z axis of the GMC is established by radial basis function(RBF)neural network and multiple linear regression method respectively.Finally,the experimental data verify that the prediction results of the model are basically consistent with the change trend of the experimental data.The fitting ability of the model is more than 75 % and the relative error of the models are small,indicating that the fitting ability of the models are good.Experiments show that the models can effectively predict new data,and prove the correctness and generalization ability of the models.The spatial thermal error model at the probe is established.Firstly,the initial temperature measuring points of the thermal error of the probe at different spatial positions are grouped and optimized by combining the fuzzy clustering method(FCM)and the correlation coefficient method to determine the temperature sensitive points in each sensitive direction of the probe.Then,the spatial thermal error model at the probe is established by combining the point-by-point multiple linear regression modeling method and the inverse distance weighting method(IDW).The experimental results show that the thermal error models are consistent with the actual experience,the fitting ability of the models are more than 90% and the relative error of the models are small,which shows that the fitting ability of the spatial thermal error models are good,and the accuracy and generalization ability of the spatial thermal error models are verified by experiments.