OBRABOTKAMETALLOV technology Vol. 26 No. 3 2024 a b Fig. 6. Dependence of the roughness parameter Rz on tool wear: a – distribution; b – correlation The rectangular window due to the high level of side lobes, which can lead to significant spectral leaks, and the Kaiser window due to the complexity of tuning the parameters to achieve the optimal result were not used in this work. As can be seen from Fig. 5, the Hamming window function effectively reduces the spectral leakage that occurs when applying the Fourier transform. At the same time, the Hamming window function provides a good compromise between the width of the main lobe and the level of side lobes in the spectral representation, is easy to implement and does not require significant computational resources. This makes it the preferred choice for many digital signals processing applications, providing the high quality and accuracy of analysis needed to implement an online monitoring system. To determine the influence of the degree of tool wear on the parameters of the microrelief and the frequency response of the acoustic signal, experiments were carried out, the results of which are shown in Fig. 6 and Fig. 7. As can be seen from Fig. 6, the surface roughness of the processed material directly depends on the degree of tool wear, and the following correlation dependence has been established: r = −0.9678 (strong, negative). The deviation between the tool diameter with increasing number of machined surfaces is different, at the beginning there is a small dimensional wear of 2–4 μm, which increases Rz by 20 %. Further, within 6 μm, the roughness increases by 50 % of the minimum values obtained. At the same time, the deflection of the tool increases (Fig. 7), as evidenced by the increase in the resonant frequency ω2 = 100 Hz. It is important to regularly monitor the condition of the tool and replace it in a timely manner. As can be seen from Fig. 7, the sound spectrograms of a worn tool contain higher values than the sound of a new tool, with identical cutting process parameters. This fact is also confirmed by the graph of vibration acceleration at various degrees of tool wear. It is worth noting the difference in density around 50 Hz, but as previously stated, further work is still needed to understand the discrepancies for each specific case. Thus, based on the conducted research, we can conclude that with the use of a VA complex based on the formation of DS, Online Monitoring is possible. However, special attention should be paid to the obtained frequency response data of the acoustic signal from the standpoint of its processing and filtering. A significant amount of data obtained should be optimally processed in order to analyze and correlate it with the condition of the cutting tool. To do this, the study determined the frequency ranges of the acoustic signal, within which it is possible to draw conclusions about the current state of the instrument. The next step was to select a specific window function. It depended on signal filtering requirements, such as the acceptable level of spectral leakage and the required frequency resolution. The study analyzed the results of applying these window functions to audio signals to determine the optimal acoustic signal from the point
RkJQdWJsaXNoZXIy MTk0ODM1