The author verifies the proposed prediction method of strong tooth scraping on demaji nt5400dcg turning and milling machine. Among them, the cutting force of cylindrical gear scraping is measured by spike wireless sensor tool collet in Germany. The collet uses a plurality of strain gauges in different positions and directions for strain analysis, and fuses the sensors to generate axial force, torque and bending moment data; The sampling frequency is 2.5 kHz and the passing frequency (1 / TP) of the blade is 322.38 Hz.
The physical diagram of the measuring experimental device is shown in Figure 1.
In the measuring experimental device, the tool is AlCrN coated high-speed steel scraper and the workpiece material is aisi4340 steel.
In the simulation process, the Kienzle cylindrical gear scraping cutting force model suitable for aisi4340 in cutpro machining simulation software is used, and according to the effective local rake angle α N and inclination I, and the coefficients KTC, KFC and KRC of each cutting node are calculated.
The author uses the Chebyshev low-pass filter with passband frequency FP = 1 / TP + 1 / TG + 5 Hz and stopband frequency FS = 1 / TP + 1 / TG + 10 Hz to filter the data of spike tool collet on X and Y axes, and uses the moving average filter with sampling window size of 10tp to filter the data of Z axis.
The prediction results obtained by this method are compared with the measurement results in TCS, and the comparison results are shown in Figure 2.
It can be seen from Fig. 2 that the prediction results of the cutting force of cylindrical gear scraping on X and Y axes are almost consistent with the measured results, and the difference between the results on Z axis is small, which can accurately predict the cutting force of cylindrical gear scraping; This method does not need to recalibrate the cutting coefficient, and simplifies the complexity of cutting force prediction in cylindrical gear scraping.
In addition, under the same experimental conditions, the author compares the proposed strong scraping prediction method with the strong scraping prediction method based on CAD simulation model proposed by tapoglou et al.
Root mean square (RMS) error is the result of normalization of the peak cutting force on each axis (X: 544.3 n; Y: 443.5 n; Z: 605.9 n).
It can be seen that compared with the prediction method based on CAD simulation model, the RMS error of the prediction method proposed by the author is reduced by 7 n on average, and the prediction accuracy of cylindrical gear scraping cutting force is improved by about 9.7%.