Method of grinding a tooth or profile of a workpiece

Abstract

The invention relates to a method for grinding teeth of a workpiece by means of a grinding tool in a grinding machine, wherein during the engagement of the grinding tool with the teeth to be ground, first and second machine parameters (PS, PW) are measured and the measured two machine parameters (PS, PW) are compared with predetermined stored values, wherein a signal is output if at least one machine parameter (PS, PW) exceeds or falls below the predetermined stored value, taking into account a tolerance range. In order to be able to draw better conclusions about the grinding process by monitoring the relevant variables, the invention proposes that at least one machine parameter (PS, PW) comprises a periodic signal component, wherein the signal component is decomposed into individual frequency components by frequency analysis and the frequency components are used to compare their frequencies and / or amplitudes.

Description

[0001] The present invention relates to a method for grinding teeth or contours of a workpiece using a grinding tool in a grinding machine, wherein a grinding tool is received on a tool spindle and the tool spindle is rotated by a first drive motor, wherein a workpiece is received on a workpiece spindle and the workpiece spindle is rotated by a second drive motor, wherein at least during engagement of the grinding tool with the teeth or contours to be ground, a first machine parameter and a second machine parameter are measured, and the two measured machine parameters or variables derived from the machine parameters are compared with predetermined stored values, wherein a signal is output if at least one of the machine parameters or variables derived therefrom exceeds or falls below the predetermined stored value, taking into account tolerance ranges.

[0002] This method is known from WO 2020 / 193228 A1. Here, the described type is monitored to detect whether grinding wheel breakage has occurred in the grinding worm, wherein, under given conditions, a warning indicator is output indicating an unacceptable process deviation.

[0003] When monitoring a grinding process, it is common practice to observe relevant process parameters, such as the current consumption of the motor driving the grinding spindle, and check whether this value remains within a predetermined tolerance range during the grinding stroke. This tolerance range naturally varies throughout the grinding stroke. If the value exceeds the tolerance range, a signal is output indicating that the process is not under suitable conditions. The defined tolerance range can be understood as a range of values ​​that the system has "learned" based on past experience. For relevant prior art, please refer to tool monitoring at Nordmann GmbH & Co. KG in Cologne, Germany (see: https: / / www.nordmann.eu / huellkurven.html).

[0004] In fact, doing so can provide an account of the grinding process, especially from the perspective of whether a suitable workpiece was ground. However, it has been found that in many cases, this is insufficient to fully explain the grinding process and the quality of the ground workpiece. In particular, without a suitable grinding process, it is difficult to determine the cause of defects.

[0005] This invention aims to further develop a general process that allows for improved conclusions about the grinding process by monitoring relevant variables. Therefore, it should be possible to better assess whether a suitable workpiece has been ground. If necessary, it should also be possible to indicate which problems prevent the grinding process from proceeding properly. Therefore, one object of this invention is process safety monitoring and, as comprehensively as possible, workpiece quality monitoring in the grinding process, which should be achievable quickly and inexpensively, and thus eliminate the need for additional downstream measurements of the ground workpiece, particularly gears.

[0006] In response to this objective, the present invention provides at least one of the machine parameters or variables derived from the machine parameters that contains a periodic signal component, wherein these signal components are decomposed into individual frequency components by frequency analysis, and these frequency components are used to compare their frequencies and / or their amplitudes.

[0007] Preferably, the machine parameters or variables derived from these machine parameters contain periodic signal components.

[0008] Preferably, the frequency components are used only to compare their amplitudes.

[0009] According to the preferred procedure, the step frequency analysis is performed by Fast Fourier Transform (FFT).

[0010] However, alternative and known methods can also be used, particularly Discrete Fourier Transform (DFT), Root Mean Square Analysis (to determine the RMS spectrum), Amplitude Spectrum Determination, Cepstral Analysis, Compensated Sine Function, or Power Spectrum Determination (PSD Analysis). The signal analysis procedures mentioned are known in themselves and therefore do not need to be discussed in detail here.

[0011] When grinding a workpiece using a grinding tool, the first and second machine parameters can be measured at predetermined time intervals.

[0012] When grinding a workpiece using a grinding tool, the measurement of the first and second machine parameters can also be performed along a predetermined feed path. Therefore, the parameters mentioned (especially the grinding power of the tool spindle) are not considered in a timely manner, but rather locally (during the grinding stroke). This allows for viewing of particularly relevant portions and comparison with already stored data.

[0013] According to a preferred embodiment of the present invention, the first machine parameter is the current consumption of the motor of the tool spindle. The second machine parameter is preferably the current consumption of the motor of the workpiece spindle.

[0014] Regarding the power or current consumption of the tool motor, according to another embodiment of the invention, a process is proposed to subtract the current consumption from the total current consumption of the motor, resulting in a grinding stroke without cutting material (or consuming relatively little energy in the case of only the final stroke). In this respect, this method does not consider the total spindle power, but only the cutting power. Therefore, the portion generated by the idle stroke (or final stroke) is subtracted from the total power consumption. During the idle stroke (or final stroke), only power losses are recorded, such as power losses caused by the bearings of the shaft or hydrodynamic losses caused by the cooling lubricant.

[0015] When we talk about the current consumption (I) of the motor here, we are of course also referring to the power absorbed by the motor (P), which can be calculated using the following relationship.

[0016]

[0017] It is given when a voltage U is applied.

[0018] One of the machine parameters being detected can also be solid-state sound from the grinding machine or a part thereof, where solid-state sound is detected by a solid-state sound sensor.

[0019] Another option is to derive the variable from the machine parameters as the volumetric cutting energy (in J / mm²) required to machine a given volume of material to be removed from the teeth or contour. 3 (Measurement). The power or current consumption of the tool spindle can be used to determine the energy consumed over a specified time period, which is necessary for machining the machining allowance (especially after deducting portions not caused by machining, see above). Taking into account the geometry of the tool (grinding worm) and the teeth or profile, as well as the process parameters (especially the tool feed to the workpiece), the corresponding machining volume can be determined. Utilizing the cutting energy per unit volume of the aforementioned parameters facilitates the transferability between different grinding processes (especially generating grinding processes) for different workpieces.

[0020] Characteristic values ​​that characterize the grinding process of the workpiece can be determined and output from the measured machine parameters and / or variables derived from the machine parameters.

[0021] The preferred grinding tool is a worm gear, and the preferred workpiece is a gear.

[0022] In this regard, the proposed concept suggests monitoring and evaluating grinding processes, particularly generating grinding processes, by combining several signals, thereby considering at least two distinct process parameters. For each workpiece being ground, a characteristic value can be defined that provides information about the grinding process that has been performed.

[0023] The described procedure effectively monitors the grinding process and detects faults or obvious defects in the early stages. In particular, it detects defects on the workpiece blank, ripples on the sides of the ground gears, and tool defects.

[0024] By retrieving and evaluating stored data (within machine control), knowledge is learned from previous grinding processes. This allows for more effective detection of process deviations (abnormalities), enabling warnings to be issued to machine operators or the grinding process to be stopped.

[0025] During the processing, several signals from within the control system (i.e., signals present in the machine control system) and signals from outside the control system (e.g., recorded by a sensor that picks up solid-borne sounds, such as those originating from the machine tool or hall floor) may be recorded and taken into account, depending on the circumstances.

[0026] In addition, internal machine data (such as setting calibration, grinding worm diameter, and generated path for workpiece and tool to guide each other) can be used to evaluate the process, and for this purpose, it can be adaptively filtered and classified as necessary (e.g., by subdividing the entire grinding process into different strokes, into workpiece and tool feed, outfeed, and full engagement).

[0027] Based on the variables affecting the process, especially the generated course (affected by screw diameter and calibration), characteristic values ​​can be calculated and output.

[0028] In addition to previously known methods (see above), the aforementioned subdivided eigenvalues ​​(statistical eigenvalues, such as standard deviation, variance, kurtosis, skewness, frequency distribution, percentile classification, slope, area content) can be determined from individually recorded signals; this also applies to determining eigenvalues ​​from signal analysis (such as frequency, amplitude, order analysis).

[0029] To ensure the general validity and transferability of measurement data across different grinding processes, the aforementioned characteristic values ​​can be standardized. For this purpose, for example, relevant metal removal rates, relevant grinding power, relevant cutting forces, and contact time between the workpiece and tool can be used.

[0030] The combination of different characteristic values ​​obtained from various sensor signals provides information about the grinding process.

[0031] These data are evaluated using algorithms derived from statistics, particularly from machine learning, to determine the quality of the editing process. Well-known algorithms from machine learning include supervised and unsupervised learning, deep learning, and reinforcement learning.

[0032] This allows for improved monitoring quality, thereby stabilizing the grinding process. It enables improved identification of whether a process is faulty, and consequently, improved identification of the fault type. This requires that the stored data (in machine control) has taught about this or similar errors.

[0033] In addition to identifying errors, it can also troubleshoot problems more quickly. Based on this knowledge, the machine can adaptively intervene in the process and optimize it.

[0034] As mentioned, the measurement signals in the process can be divided into regions that are meaningful for process evaluation based on knowledge of the grinding process. Therefore, it is not necessary to evaluate the entire machining process, but only the relevant regions.

[0035] Therefore, the proposed program monitors the grinding process and evaluates it based on indices. This means calculating an index (i.e., a numerical value) for each individual workpiece to assess the quality of the grinding process. These index values ​​can then be provided to the machine operator as a progress chart, allowing them to easily obtain an overview of the grinding process that has occurred. Several eigenvalues ​​can be used to determine the aforementioned indices. These eigenvalues ​​then evaluate the grinding process, dressing process, and workpiece alignment in the grinding machine in different ways. Thus, errors that would otherwise only be detectable in the measuring chamber or not at all can be detected at an early stage. These errors include workpiece blank errors or setup errors.

[0036] Furthermore, if irregularities are taught into the algorithm, specific information about those irregularities can be output. In this way, the process can be optimized, and errors can be eliminated in a targeted and time-efficient manner.

[0037] The data recorded during the actual grinding process is compared with the data stored in the database. In this way, for example, certain process signals can be assigned to certain areas of the grinding worm; correspondingly, for example, signals can be assigned to certain worm diameters and associated speeds of the grinding worm.

[0038] The accompanying drawings illustrate examples of embodiments of the present invention.

[0039] Figure 1 Firstly, for the first grinding process, the power consumption or current consumption of the motor driving the grinding spindle is shown in the upper figure in a general manner, and the power consumption or current consumption of the motor driving the workpiece spindle is shown in the lower figure.

[0040] Figure 2 For the second grinding process, according to an embodiment of the method of the present invention, the power consumption or current consumption process of the motor driving the workpiece spindle is shown in the left figure, and the amplitude of the frequency component of the periodic process of power or current obtained from the fast Fourier transform (FFT) is shown in the right figure.

[0041] Figure 1 The curves showing the change of power P or current I over time t are presented, representing the results for the motors driving the grinding spindle (S) and the workpiece spindle (W) in a gear grinding machine (showing the rough grinding stroke, during which the workpiece is ground to a certain quality). Figure 1At the top, a curve showing the change of power PS over time when the power PS is absorbed by the motor of the grinding spindle is plotted; the same curve for the motor's ampere value IS is obtained correspondingly by applying the motor voltage. This also applies to... Figure 1 The lower figure in the image shows the same information for the motor that drives the workpiece spindle.

[0042] The expected power and current curves for manufacturing a suitable gear are shown using dashed lines. The dashed lines above and below the dashed lines indicate the allowable tolerance ranges for the power or current curves. EL represents the region where the grinding worm enters the gear being ground, VS represents the actual full cut that occurs during grinding, and AL represents the tool leaving the workpiece.

[0043] The lines drawn represent the actual flow of power or current during the grinding of a specific workpiece.

[0044] It can be seen that the power or current consumption of the grinding spindle is normal, indicating that the grinding is appropriate. However, this is not actually the case. Due to a side profile error (i.e., due to the opposite side line deviation between the right and left tooth sides), the grinding spindle consumes the expected current, but the workpiece spindle does not consume the expected current. The power consumption of the workpiece spindle is outside the expected range, which can be seen from... Figure 1 As can be seen in the image below.

[0045] Therefore, it should be noted that the power consumption or current consumption at the grinding spindle may still be within the allowable tolerance range depending on the location or direction of the side shape error. Only when the power consumption or current consumption at the workpiece spindle is considered simultaneously does it indicate that there is a problem with the pre-machining of the gear and that there is a corresponding side shape error.

[0046] Therefore, in this embodiment, typically, at least during the engagement of the grinding tool with the teeth to be ground, a first machine parameter in the form of power or amperes for the tool spindle and a second machine parameter in the form of power or amperes for the workpiece spindle are provided. The two measured machine parameters, PS / IS and PW / IW, are compared with predetermined stored values. Figure 1 As shown in the lower part of the diagram, it is immediately apparent that the power or current intensity process has deviated from the allowable tolerance range, thus triggering a warning signal. Therefore, the machine control indicates a problem in the grinding process and that good parts may not be produced.

[0047] Figure 2 A specific procedure according to the invention is shown, which has been specially designed herein to identify problems for improvement in the grinding process and to issue corresponding warning messages.

[0048] Figure 2The graph on the left again shows the power or current consumption of the workpiece spindle (see the drawn line), which is within the allowable tolerance. However, it can also be seen that the power or current consumption varies periodically, indicating a superimposed oscillation process.

[0049] This superimposed vibration can produce harmful fine ripples on the tooth surface, which can generate noise when the gear mechanism is in use; this is particularly detrimental in the field of electric vehicles.

[0050] According to a possible embodiment of the invention, it is therefore envisioned that, in the improvement of the above procedure, variables derived from machine parameters (here: from the power consumption or current consumption of the workpiece spindle) are considered in order to evaluate or assess the grinding process.

[0051] right Figure 2 The waveform recorded in the left-hand portion of the image undergoes a Fast Fourier Transform (FFT) to decompose the periodic signal into its components (“harmonics”). This is in Figure 2 The diagram on the right is shown here. The amplitude A of the individual frequency component of the detected periodic signal is schematically plotted above the order (Or). It can be seen that the frequency component is above the limit value Gr, therefore it can be concluded that a suitable grinding process did not occur.

[0052] It should be noted that Figure 2 The illustrations are merely illustrative; of course, individual limit values ​​can be set for each individual frequency component generated by the FFT. In particular, specialized knowledge stored in the machine control system can be used here to trigger a warning signal under given circumstances.

[0053] The proposed procedure can be used to very quickly identify unsuitable process conditions, especially before an entire batch of workpieces might be produced incorrectly. This can save considerable costs.

[0054] Of course, other parameters can also be used to monitor the process, especially recording machine solid-state noise or hall floor solid-state noise. This also allows for reporting, particularly after analyzing the recorded signals (e.g., via FFT), about how the grinding process is proceeding and whether good parts were expected to be ground.

[0055] This implementation illustrates the use of FFT. Alternatively, any other known frequency analysis method can be used, such as, in particular, Discrete Fourier Transform (DFT), frequency analysis via root mean square analysis (determining the RMS spectrum), frequency analysis via determining the amplitude spectrum, frequency analysis via cepstral analysis (including variables such as power cepstral spectra), frequency analysis via compensated sine functions, or frequency analysis via determining the self-power spectrum (PSD analysis). The methods described in the measurement analysis are all known and therefore do not need to be discussed in more detail here. The only important thing is to determine the individual frequency components of the measured periodic signal component via frequency analysis, and the results obtained are used for comparison with permissible limits (especially the maximum permissible amplitude for the individual amplitude of the "harmonics").

Claims

1. A method of grinding the teeth or contours of a workpiece using grinding tools in a grinding machine. The grinding tool is received on a tool spindle, and the tool spindle is rotated by a first drive motor. The workpiece is received on the workpiece spindle, and the workpiece spindle is rotated by a second drive motor. During at least the engagement of the grinding tool with the tooth or profile to be ground, a first machine parameter (PS) and a second machine parameter (PW) are measured, and the two measured machine parameters (PS, PW) or a variable (A) derived from the machine parameters (PS, PW) are compared with predetermined stored values. If at least one of the machine parameters (PS, PW) or the variable (A) derived therefrom exceeds or falls below the predetermined stored value, taking into account tolerance ranges, a signal is output. Its features At least one of the machine parameters (PS, PW) or the variable (A) derived from these machine parameters (PS, PW) contains periodic signal components, wherein these signal components are decomposed into individual frequency components by frequency analysis, and the frequency components are used to compare their frequencies and / or their amplitudes.

2. The method according to claim 1, characterized in that, Both machine parameters (PS, PW) or the variable (A) derived from these machine parameters (PS, PW) contain periodic signal components.

3. The method according to claim 1, characterized in that, The frequency components are used only to compare their amplitudes.

4. The method according to claim 1, characterized in that, The frequency analysis was performed using Fast Fourier Transform (FFT).

5. The method according to claim 1, characterized in that, The frequency analysis was performed using the Discrete Fourier Transform (DFT).

6. The method according to claim 1, characterized in that, The frequency analysis is performed by root mean square analysis, by determining the amplitude spectrum, by cepstral analysis, by compensating for a sine function, or by determining the self-power spectrum.

7. The method according to claim 1, characterized in that, During the grinding of the workpiece using the grinding tool, the first machine parameter (PS) and the second machine parameter (PW) are measured at predetermined time intervals.

8. The method according to claim 1, characterized in that, During the grinding of the workpiece using the grinding tool, the first machine parameter (PS) and the second machine parameter (PW) are measured along a predetermined feed path.

9. The method according to claim 1, characterized in that, The first machine parameter (PS) is the power consumption or current consumption of the motor of the tool spindle.

10. The method according to claim 1, characterized in that, The second machine parameter (PW) is the power consumption or current consumption of the motor of the workpiece spindle.

11. The method according to claim 9, characterized in that, The process of subtracting the power consumption or current consumption from the total power consumption or current consumption of the motor results in a grinding stroke with no material cutting or only low cutting at the end of the grinding stroke.

12. The method according to claim 1, characterized in that, One of the machine parameters is the solid-state sound of the grinding machine or a part thereof, which is detected by a solid-state sound sensor.

13. The method according to claim 1, characterized in that, The variables derived from the machine parameters (PS, PW) are the unit volume cutting energy required to process the predetermined volume of material to be removed from the tooth or the profile.

14. The method according to claim 1, characterized in that, Characteristic values ​​that characterize the grinding process of the workpiece are determined and output from the measured machine parameters (PS, PW) and / or the variables (A) derived from the machine parameters (PS, PW).

15. The method according to claim 1, characterized in that, The grinding tool is a grinding worm, and the workpiece is a gear.

Images