Thermal Error Modelling for the Parallelism Between the Gear Axis and the Z-axis of GMI

Abstract

Gear measuring instrument (GMI) is the mainstream equipment for gear measurement. As the measurement accuracy of gears improves and GMI is gradually used in production, thermal errors are becoming an important factor affecting the measurement accuracy of GMI. In the precision measurement of the gears, the parallelism errors between the gear axis and the Z-axis of the GMI are the major error source. This error is usually reduced by mechanical adjustment when a GMI is assembled. When the temperature of the GMI and its environment changes, the thermal deformation will lead to new parallelism errors between the gear axis and the Z-axis of the GMI, which will further cause errors in the GMI. In this paper, the effect of the parallelism thermal errors of the GMI is studied, and the parallelism thermal errors model under room temperature environment was established. First, the temperature field of the GMI and its environment and the parallelism thermal errors were measured; then, the TMP are grouped using the FCM, and the key TMP were selected by optimizing the group of the measurement points using Pearson correlation coefficient method; finally, the parallelism thermal errors were established by RBF (Radial Basis Function) neural network and tested by using the measured data. The experimental results show that the predictive accuracy of the established parallelism thermal error model can reach 60.89% in the X direction and 82.26% in the Y direction, indicating good generalization ability and robustness. The model can be used for subsequent thermal error compensation.