Automatic process control in gear machining equipment

Abstract

The present invention relates to a method for monitoring a machining process in which the tooth flank of a previously machined toothed workpiece 23 is machined using a precision machining apparatus 1. In the method, a plurality of measured values ​​are collected while a precision machining tool 16 is in machining engagement with the workpiece. These measured values ​​include the value of a power indicator indicative of the current power consumption of the tool spindle during machining of the tooth flank of the workpiece. A normalization procedure is applied to at least some of the measured values ​​or to values ​​of variables derived from the measured values ​​to obtain normalized values. The normalization procedure depends on at least one of the following parameters: geometric parameters of the precision machining tool, in particular its outer diameter; geometric parameters of the workpiece; and setting parameters of the precision machining apparatus, in particular the radial infeed and the axial feed.

Description

Technical Field

[0001] The present invention relates to a machine for finishing gears, in particular to a method for monitoring a machine that performs a generating process. The present invention further relates to a finishing machine configured to execute such a method, a computer program for executing such a method, and a computer-readable medium including such a computer program.

Background Art

[0002] Hard precision machining (hard finishing) of pre-machined gears is a very demanding method where even a slight deviation from the process specifications can cause the machined workpiece to be considered scrap ( "NIO parts", where NIO means "not in order"). This problem can be particularly well illustrated by the example of continuous generating grinding, but it equally applies to other generating finishing methods such as single flank generating grinding, gear honing, hard skiving, etc. Although not to the same extent, similar problems also occur in the case of non-generating methods such as discontinuous contour grinding or continuous contour grinding.

[0003] In continuous generating grinding, a pre-machined gear blank is machined by rolling engagement with a grinding wheel (grinding worm) having a worm-shaped contour. Generating grinding is a very demanding generating machining method based on a large number of synchronized high-precision individual movements and affected by many boundary conditions. Information on the basis of continuous generating grinding can be found, for example, in Chapter 2.3 ( "Basic Methods of Generating Grinding") on pages 121 to 129 of the book "Continuous Generating Gear Grinding" (ISBN 978-3-033-02535-6) written by H. Schriefer et al. and published by Reishauer AG in Wallisellen in 2010.

[0004] Theoretically, the tooth surface shape in continuous generation grinding is determined solely by the dressed contour shape of the grinding worm and the machine's setting data. However, in practice, deviations from the ideal state occur in automated production, and these deviations can have a decisive impact on the grinding results.

[0005] Traditionally, the quality of gears produced by continuous generating gear grinding methods has only been evaluated after the completion of the machining process, through gear measurements performed outside the machine ("offline") using numerous measured variables. Various standards exist that specify how gears are measured and how to check whether the measurement results are within or outside the dimensional tolerance specifications. A summary of such gear measurements can be found, for example, in Chapter 3 ("Quality Assurance in Continuous Generating Gear Grinding"), pages 155-200, of the aforementioned book by Schriefer et al.

[0006] It is known from current technology that machines can be corrected based on gear measurements to eliminate detected machining errors. This is discussed in Chapter 6.10 ("Analyzing and Eliminating Gear Tooth Deviations"), pages 542-551, of the aforementioned book by Schriefer et al.

[0007] For time and cost reasons, only random checks are usually performed during gear inspection, so machining errors are often detected very late. This can, in some cases, result in a significant number of parts in a production lot having to be discarded as NIO parts. Therefore, it is desirable to detect machining errors as early as possible "online" during machining, ideally before the machining error reaches a degree that necessitates rejecting the workpiece as an NIO part.

[0008] To this end, it is desirable to provide an automated process monitoring system that can detect undesirable process deviations, obtain signs of potential machining errors, and modify machine settings in a targeted manner, so that these machining errors are avoided or at least reduced. Ideally, process monitoring should also allow for retrospective conclusions about process deviations, even if machining errors are only detected later, for example, during EOL (End of Line) testing.

[0009] To date, only rudimentary strategies are known for automated process monitoring in gear machining, based on the latest technologies.

[0010] For example, Patent Document 1 describes monitoring parameters in a gear machine and performing a gear check to determine whether several measured machine parameters deviate from nominal values.

[0011] Non-patent document 1 describes various means of tool monitoring relating to general metal cutting machining tools (page 3). This presentation shows application examples in various metal cutting processes, including a few examples of processes related to gear machining, particularly hobbing (pages 41 and 42), hard skiving (page 59), and honing (page 60). Continuous generation grinding is mentioned only incidentally (e.g., pages 3 and 61).

[0012] Methods for automated process monitoring in various machining processes are known from the following documents, particularly Patent Documents 2, 3, 4, and Non-Patent Document 2. However, even in these documents, the precision machining of gears is not discussed in detail.

[0013] One difficulty in process monitoring in gear machining is that the monitored measurement variables depend very complexly on numerous geometric properties of the tool (e.g., diameter, module, number of threads, helix angle, etc., in the case of grinding worms), geometric properties of the workpiece (e.g., module, number of teeth, helix angle, etc.), and machine setting parameters (e.g., radial infeed, axial feed, tool and workpiece spindle speed, etc.). Due to these diverse and complex dependencies, on the one hand, it is very difficult to draw direct conclusions from the monitored measurement variables about specific process deviations and the resulting machining errors. On the other hand, it is extremely difficult to compare measured variables from different machining processes. A further challenge arises when using dressable tools. Dressing changes the diameter of the tool during machining of a series of workpieces, and therefore the contact conditions also change. As a result, even within the same series of workpieces, with all other conditions remaining the same, monitored measurement variables from different dressing cycles cannot be directly compared to one another.

[0014] Known process monitoring methods do not yet adequately take into account these specific characteristics of gear machining. [Prior art documents] [Patent Documents]

[0015] [Patent Document 1] German Patent Application Publication No. 102014015587 [Patent Document 2] U.S. Patent No. 5,070,655 [Patent Document 3] U.S. Patent No. 3,809,970 [Patent Document 4] U.S. Patent No. 4,894,644 [Non-patent literature]

[0016] [Non-Patent Document 1] The version of the corporate presentation "NORDMANN Tool Monitoring" dated October 5, 2017, accessed on February 25, 2019, from https: / / www.nordmann.eu / pdf / praesentation / Nordmann_presentation_ENG.pdf [Non-Patent Document 2] "Process Monitoring during Grinding and Dressing" by Klaus Nordmann (Schleifen + Polieren 05 / 2004, Fachverlag Moeller, Velbert (DE), pages 52-56) [Overview of the project]

[0017] The object of the present invention is to provide a process monitoring method in gear machining that enables the detection of process deviations and the cancellation of the effects of process deviations in a targeted manner. This method should be particularly suitable for use with dressable tools.

[0018] This objective is achieved by the method of claim 1. Further embodiments are provided in the dependent claims.

[0019] A method is provided for monitoring the machining process in which the tooth surface of a pre-toothed workpiece is machined in a finishing machine (i.e., a precision machining machine). The finishing machine has a tool spindle that drives a finishing tool (i.e., a precision machining tool) to rotate around a tool axis, and a workpiece spindle that drives and rotates the pre-toothed workpiece. The method is as follows: Detecting multiple measurements while the finishing tool is engaged with the workpiece for machining, The process involves applying a normalization process to at least some of the measured values ​​or a number of values ​​derived from the measured values ​​to obtain normalized values, Includes, The normalization process depends on at least one process parameter, which is selected from the geometric parameters of the finishing tool, the geometric parameters of the workpiece, and the setting parameters of the finishing machine.

[0020] Therefore, it is proposed to record measurements in the finishing machine and to normalize (standardize) at least some of these measurements or values ​​derived from them. The normalization process takes into account the influence of one or more process parameters on the measurements, particularly the geometric parameters of the finishing tool (especially its dimensions, specifically its outer diameter), the geometric parameters of the workpiece, and / or the setting parameters of the finishing machine (especially radial infeed, axial feed, and tool and workpiece spindle speeds). The resulting normalized values ​​are therefore independent of the aforementioned process parameters or have at least much less dependence on them. The normalization process allows for the comparison of normalized values ​​between different machining operations, even if these process parameters differ.

[0021] Normalization is particularly important when the measurements include power indicator values ​​that show the current power consumption of the tool spindle during machining of the tooth surface of a workpiece. In particular, the detected power indicator can serve as a measure of the current consumption of the tool spindle. Such power indicators are especially affected by these parameters. Therefore, applying normalization to the power indicator value or the variables derived from this power indicator is particularly advantageous.

[0022] The normalization process is preferably based on a model that describes the expected dependency of the measured values ​​on the parameters described above. When the measured values ​​are power indicator values, the model preferably represents the dependency of process power (i.e., the mechanical power or electricity required for the machining process being performed) on the parameters described above. In particular, the process power model may be based on a force model that describes the expected dependency of the cutting force acting at the contact point between the finishing tool and the workpiece on the geometric parameters of the finishing tool, the geometric parameters of the workpiece, and the setting parameters of the finishing machine. The process power model may also take into account the effective lever arm length between the tool axis and the contact point between the finishing tool and the workpiece. This lever arm length can be approximated in particular by the outer diameter of the finishing tool. In addition, the process power model may also take into account the speed of the tool spindle.

[0023] Normalization can involve, for example, multiplying the acquired measurements or variables derived therefrom by a normalization coefficient. However, more complex normalization processes are also possible. If the measurements include power indicator values, the normalization coefficient can be, in particular, an inverse power variable calculated based on a process power model of the actual machining conditions or a variable derived therefrom.

[0024] The normalization process is preferably applied directly to the acquired measurements after filtering as necessary. The normalization process is advantageously performed in real time, i.e., during the machining process, particularly during the machining of the workpiece, i.e., while the finishing tool is still engaged with the workpiece for machining. This means that the normalized values ​​are immediately available during the machining process and can be used in real time to monitor the machining process.

[0025] In particular, normalized values ​​are advantageous when analyzed in real time to identify unacceptable process deviations during the machining process. This allows for the immediate identification of workpieces deemed to have unacceptable process deviations after machining, and, if necessary, the removal of those workpieces from the workpiece batch for further handling. For example, such workpieces could undergo further measurement or be directly classified as NIO parts.

[0026] In a preferred embodiment, the method includes calculating characteristic parameters of a machining process from measured values ​​or values ​​derived from measured values. Calculating characteristic parameters from measured values ​​or values ​​derived therefrom has advantages, whether or not a normalization process is performed as part of the process. In some embodiments, at least one of the characteristic parameters is a normalized characteristic parameter, i.e., a normalization process is applied at some point during the calculation of the characteristic parameters. This can be done by calculating the characteristic parameters from normalized measured values. This can also be done by applying the normalization process only to intermediate results (i.e., quantities derived from measured values) when calculating the characteristic parameters.

[0027] At least one of the characteristic parameters is preferably specific to the machining process. This at least one characteristic parameter is therefore preferably not only a statistical measure such as a mean or standard deviation that can be formed independently of the specific machining process, but also a parameter that takes into account the characteristics of the specific machining process.

[0028] Preferably, at least one of the characteristic parameters correlates with a predetermined machining error of the workpiece. This is particularly advantageous when there is a one-to-one relationship, especially a simple proportional relationship, between the characteristic parameter and the size of the machining error. This makes it possible to obtain immediate information about the occurrence of a particular machining error by monitoring the characteristic parameters of different workpieces. This allows even inexperienced operators to correctly interpret the characteristic parameters and perform corrective actions.

[0029] There are certain advantages when characteristic parameters are directly related to the gear measurement results. For this reason, the method is: Perform gear measurements on selected workpieces to obtain at least one gear measurement that characterizes a predetermined machining error for each such workpiece, To determine a correlation parameter that characterizes the correlation between at least one characteristic parameter and at least one gear measurement, It can include...

[0030] The calculation of at least one of the characteristic parameters may include spectral analysis of measured values, particularly power indicator (preferably normalized) and / or accelerometer values. Spectral components at multiples of tool speed and / or workpiece speed are particularly preferred. Thus, the calculation of the corresponding parameters is specific to the machining process. Such spectral analysis is also advantageous when normalization is not performed, as may be the case with values ​​from an accelerometer.

[0031] When a precision machining process is a generating process, particularly a generating grinding process, in which precision machining tools and workpieces engage in rolling action, it is advantageous for the characteristic parameters to include at least one of the following variables. The cumulative pitch indicator is calculated by evaluating the spectral components of the measured values, particularly the normalized power indicator, at the rotational speed of the workpiece spindle, and correlating this with the cumulative pitch error of the workpiece. A contour shape indicator calculated by evaluating the spectral components of measurements, particularly those from an acceleration sensor, at the tooth occlusion frequency, and correlating these with the contour shape deviation of the workpiece.

[0032] It is also advantageous if the characteristic parameters include a wear indicator. The wear indicator is calculated from the low-pass filtered spectral component of the measured values, particularly the normalized power indicator measurements, and correlates with the degree of wear of the precision machining tool.

[0033] It is advantageous to determine the process deviation of the machining process from the target process based on the transition of at least one of the characteristic parameters of multiple workpieces. For this purpose, a favorably selected value of a characteristic parameter is correlated with the value of another characteristic parameter or another process variable. Advantageously, the machining process is then adjusted to reduce the process deviation, or the limits used in the real-time analysis scenario described above are adjusted to identify unacceptable process deviations. Process deviations can be determined by a trained machine learning algorithm.

[0034] The method may include storing a dataset in a database, the dataset including a unique identifier for the workpiece, at least one process parameter, and at least one of the characteristic parameters. Furthermore, the method may include: Retrieving a dataset of multiple workpieces from a database, To graphically output at least one value from the characteristic parameters of multiple workpieces, or a value derived therefrom, It can include...

[0035] The computational and memory requirements for these steps are very reasonable compared to processing raw data, so these steps can be performed, for example, in a web browser.

[0036] The method preferably provides a recalculation of the normalization process whenever at least one of the process parameters changes. The recalculation of the normalization process then preferably involves applying the model with the changed process parameter.

[0037] The recalculation of the normalization process can be based particularly on the changed dimensions of the finishing tool, especially its outer diameter, and may include compensation for the changed dimensions. This is especially important when this dimension changes during the machining of a series of parts, as is typically the case with dressable tools. In this way, characteristic parameters obtained in different dressing cycles can be directly compared to one another.

[0038] The present invention also provides a finishing machine for machining the tooth surfaces of a pre-toothed workpiece. This finishing machine comprises a tool spindle that drives a finishing tool (precision machining tool) to rotate around a tool axis, a workpiece spindle that drives and rotates a pre-toothed workpiece, a control device that controls the process of machining the workpiece using the finishing tool, and a process monitoring device. The process monitoring device is configured in particular to perform the method described above.

[0039] For this purpose, the process monitoring device is preferably, A detection device that detects multiple measurements while the finishing tool is engaged with the workpiece for machining, A normalization device that obtains a normalized value by applying a normalization process to at least a portion of the measured values ​​or the values ​​of quantities derived from the measured values, Equipped with, The normalization process depends on at least one process parameter, which is selected from the geometric parameters of the finishing tool, the geometric parameters of the workpiece, and the setting parameters of the finishing machine.

[0040] Preferably, the normalization device is configured to perform the normalization process in real time while the finishing tool is engaged with the workpiece for machining.

[0041] In some embodiments, the process monitoring device includes a defect detection device configured to analyze normalized values ​​in real time in order to detect unacceptable process deviations.

[0042] The finishing machine may be equipped with a workpiece handling device configured to automatically remove workpieces that are determined to have unacceptable process deviations.

[0043] In some embodiments, the process monitoring device includes a characteristic parameter calculation device that calculates characteristic parameters of the machining process from measured values ​​or values ​​derived from measured values. The characteristic parameter calculation device can be configured to perform spectral analysis of measured values, values ​​derived from measured values ​​or normalized values ​​for at least one of the characteristic parameters, and in particular can be configured to evaluate spectral components at multiples of the speed of the tool spindle and / or workpiece spindle.

[0044] The process monitoring device may include a data communication device that sends a dataset to a database containing a unique identifier for a workpiece, at least one process parameter, and at least one of the characteristic parameters.

[0045] The process monitoring device may include a deviation detection device that detects process deviations of the machining process from a target process based on the value of at least one of the characteristic parameters of multiple workpieces. The deviation detection device may include a processor device programmed to run a trained machine learning algorithm that detects process deviations.

[0046] A process monitoring device may include a normalization calculation device that recalculates the normalization process when at least one of the process parameters changes. The normalization calculation device is preferably configured to apply a model describing the expected dependency of the measured values ​​on the process parameters, particularly a model of process force or process power, to the recalculation of the normalization process. The normalization calculation device may also be advantageously configured to provide compensation for the dimensions of the finishing tool, particularly its outer diameter.

[0047] Detection devices, normalization devices, defect detection devices, characteristic parameter calculation devices, normalization calculation devices, deviation detection devices, and data communication devices can be implemented, at least in part, by software running on one or more processors of a process monitoring device.

[0048] The present invention also provides a computer program. This computer program includes instructions to cause one or more processors of a process monitoring device in the type of finishing machine described above, in particular, to perform the method described above. This computer program can be stored in a suitable memory device, for example, a computer separate from the machine control unit.

[0049] Furthermore, the present invention provides a computer-readable medium on which computer programs are stored. This medium can be a non-volatile medium such as flash memory, a CD, or a hard disk.

[0050] Preferred embodiments of the present invention are described below with reference to the drawings. The drawings are for illustrative purposes only and should not be construed as limiting. [Brief explanation of the drawing]

[0051] [Figure 1] This is a schematic diagram of the Genesis Grinding Machine. [Figure 2] This is an enlarged view of Area II in Figure 1. [Figure 3] This is an enlarged view of Area III in Figure 1. [Figure 4] This figure shows two illustrative traces of the current consumption of a tool spindle during machining of a tooth surface. [Figure 5] This figure shows the maximum current consumption of the tool spindle for multiple workpieces as a function of the outer diameter of the grinding worm, with squares representing denormalized values ​​and cross marks representing normalized values. [Figure 6] This figure shows the spectrum of current consumption of the tool spindle during machining of a workpiece. [Figure 7] This figure shows the spectrum of measurements taken from an accelerometer during the machining of a workpiece. [Figure 8] This figure shows the cumulative pitch indicator value of multiple workpieces as a function of the grinding worm position. [Figure 9] This figure shows the values ​​of the contour shape indicators of multiple workpieces as a function of the grinding worm position, with triangles representing the first workpiece spindle and cross marks representing the second workpiece spindle. [Figure 10] This figure shows the wear indicator values ​​for multiple workpieces as a function of the grinding worm position. [Figure 11] This figure shows the vibration indicator values ​​of multiple workpieces as a function of the grinding worm position. [Figure 12] This is a flowchart of the process for monitoring the processing of workpiece batches. [Figure 13] This is a flowchart of the process for automatically detecting and correcting process deviations. [Figure 14] This is a flowchart showing the parameters and the method for graphically outputting the information obtained from those parameters. [Figure 15] This is a schematic block diagram of the functional unit of the process monitoring device. [Modes for carrying out the invention]

[0052] <Example structure of a grinding machine> Figure 1 shows a generator grinding machine 1 as an example of a finishing machine for machining the tooth surfaces of a pre-toothed workpiece. This machine comprises a machine bed 11 on which a tool carrier 12 is guided to displace along the radial infeed direction X. The tool carrier 12 carries an axial slide 13 which is guided to displace along the feed direction Z relative to the tool carrier 12. A grinding head 14 is mounted on the axial slide 13 and can be swung around a pivot axis that extends parallel to the X axis (so-called A axis) to match the helix angle of the gear being machined. The grinding head 14 further carries a shift slide which can move the tool spindle 15 along the shift axis Y relative to the grinding head 14. A worm-shaped grinding wheel (grinding worm) 16 is mounted on the tool spindle 15. The grinding worm 16 is driven by the tool spindle 15 to rotate around the tool axis B.

[0053] The machine bed 11 also carries a swivel workpiece carrier 20 in the form of a turret that can swivel around axis C3 between at least three positions. Two identical workpiece spindles are mounted on the workpiece carrier 20 opposite each other, and in Figure 1 only one workpiece spindle 21 with an associated tailstock 22 is visible. The workpiece spindle visible in Figure 1 is in a machining position where a workpiece 23 clamped to it can be machined using a grinding worm 16. The other workpiece spindle, offset by 180° and not visible in Figure 1, is in a workpiece exchange position where a finished workpiece can be removed from this spindle and a new blank can be clamped. A dressing device 30 is mounted offset by 90° relative to the workpiece spindle.

[0054] All driven axes of the Generative Grinding Machine 1 are digitally controlled by the Machine Control Unit 40. The Machine Control Unit 40 comprises several axis modules 41, a control computer 42, and a control panel 43. The control computer 42 receives operator commands from the control panel 43 and sensor signals from various sensors of the Generative Grinding Machine 1, and calculates control commands for the axis modules 41. The control computer 42 also outputs operating parameters to the control panel 43 for display. Each axis module 41 provides control signals at its output to one machine axis (i.e., at least one actuator used to drive the associated machine axis, such as a servo motor).

[0055] A process monitoring device 44 is connected to a control computer 42. This device continuously receives multiple measurements from the control computer 42 and other sensors as needed. On the one hand, the process monitoring device 44 continuously analyzes the measurements to detect machining errors at an early stage and remove affected workpieces from the machining process. On the other hand, the process monitoring device 44 uses the measurements to calculate various characteristic parameters of the final machined workpiece. These processes are described in more detail below.

[0056] The process monitoring device 44 sends a dataset for each workpiece to the database server 46. This dataset includes a unique workpiece identifier and selected process and characteristic parameters. The database server 46 stores these datasets in a database so that the corresponding dataset can then be retrieved for each workpiece. The database server 46, which has this database, can be located inside the machine or remotely from the machine. The database server 46 can connect to the process monitoring device 44 via a network, as shown by the cloud in Figure 1. In particular, the database server 46 can connect to the process monitoring device 44 via the machine or a company LAN, via a WAN, or via the internet.

[0057] Client 48 can connect to database server 46 to retrieve, receive, and evaluate data from the database server 46. This connection can also be made over a network, particularly a LAN, WAN, or the Internet. In particular, client 48 may include a web browser that can visualize the received data and its evaluation. The client does not need to meet any special requirements regarding computing power, nor does the client application require high network bandwidth.

[0058] Figure 2 shows Section II of Figure 1 at a magnified scale. The tool spindle 15 with the grinding worm 16 clamped to it can be seen. A measuring probe 17 is swivelably mounted to the fixed part of the tool spindle 15. This measuring probe 17 can swivel between the measuring position and the parking position shown in Figure 2. In the measuring position, the measuring probe 17 can be used to measure the gear of the workpiece 23 on the workpiece spindle 21 by contact. This is done "inline," i.e., while the workpiece 23 is still on the workpiece spindle 21. This makes it possible to detect machining errors at an early stage. In the parking position, the measuring probe 17 is positioned in an area protected from collision with the workpiece spindle 21, tailstock 22, workpiece 23, and other components on the workpiece carrier 20. The probe 17 remains in this parking position during workpiece machining.

[0059] The centering probe 24 is positioned on the side of the workpiece 23 opposite to the grinding worm 16. In this example, the centering probe 24 is designed and positioned in accordance with International Publication No. 2017 / 194251. The operating mode and positioning of the centering probe are explicitly mentioned in the aforementioned international publication. In particular, the centering probe 24 may include an inductive proximity sensor or a capacitive proximity sensor, as is known from the current state of the art. However, optically operating sensors may also be used for the centering operation, for example, a sensor that guides a light beam onto the gear to be measured and detects the light reflected from the gear, or a sensor that detects the obstruction of the light beam by the gear while the gear rotates around the workpiece axis C1. Furthermore, one or more additional sensors capable of directly recording process data onto the workpiece may be positioned on the centering probe 24, for example, as proposed in U.S. Patent No. 6,577,917. Such additional sensors may include, for example, a second centering sensor for a second gear, a temperature sensor, an additional structural transmission noise sensor, a pneumatic sensor, and so on.

[0060] In addition, Figure 2 symbolically illustrates the accelerometer 18. The accelerometer 18 is used to reveal vibrations of the stator of the tool spindle 15 that occur during the grinding process of the workpiece and when dressing the grinding worm. In practice, the accelerometer is not typically located in the housing (as shown in Figure 2), but rather directly on the stator of the drive motor of the tool spindle 15, for example. This type of accelerometer is known.

[0061] The coolant nozzle 19 guides the coolant jet into the machining zone. An acoustic sensor (not shown in Figure 2) can be provided to indicate the noise transmitted through this coolant jet.

[0062] Figure 3 shows Section III of Figure 1 at an enlarged scale. In this figure, the dressing device 30 is particularly easy to recognize. A dressing spindle 32, to which a disc-shaped dressing tool 33 is clamped, is positioned on a swivel drive unit 31 and can be swiveled around axis C4. Alternatively or in addition, a fixed dressing tool may also be provided, in particular a so-called head dresser intended to engage only with the head area of ​​the worm thread of a grinding worm for dressing these head areas.

[0063] <Processing workpiece batches> To machine an unmachined workpiece (blank), the workpiece is clamped by an automatic workpiece changer on a workpiece spindle at the workpiece exchange position. The workpiece is exchanged in parallel with the machining of another workpiece on a different workpiece spindle at the machining position. Once the new workpiece to be machined is clamped and machining of the other workpiece is complete, the workpiece carrier 20 is rotated 180° around the C3 axis so that the spindle with the new workpiece to be machined reaches the machining position. Before and / or during the rotation process, a centering operation is performed using the corresponding centering probe. For this purpose, the workpiece spindle 21 is rotated and the position of the tooth grooves of the workpiece 23 is measured using the centering probe 24. Based on this, the rolling angle is determined. In addition, the centering probe can be used to predict signs of excessive variation in tooth thickness and other pre-machining errors, even before machining begins.

[0064] When the workpiece spindle, which carries the workpiece 23 to be machined, reaches the machining position, the workpiece 23 is engaged with the grinding worm 16 without misalignment by moving the tool carrier 12 along the X axis. At this time, the workpiece 23 is machined by the grinding worm 16, which is engaged in rolling. During machining, the workpiece is continuously advanced along the Z axis with a constant radial infeed in the X direction. In addition, the tool spindle 15 is slowly and continuously moved along the shift axis Y to allow the use of unused areas of the grinding worm 16 during machining (so-called shift movement). Immediately after machining of the workpiece 23 is complete, the workpiece is optionally measured inline using a measuring probe 17.

[0065] Simultaneously with the machining of the workpiece, the finished workpiece is removed from another workpiece spindle, and another blank is clamped onto this spindle. Each time the workpiece carrier is rotated around the C3 axis, the selected components are monitored before or during the rotation, i.e., without affecting the cycle time, and the machining process is not resumed until all specified requirements are met.

[0066] If, after machining several workpieces, the grinding worm 16 has been used extensively and has become excessively dull and / or the geometry of the flanks is excessively inaccurate, the grinding worm is dressed. To do this, the workpiece carrier 20 is rotated ±90° so that the dressing device 30 is positioned opposite the grinding worm 16. At this time, the grinding worm 16 is dressed using the dressing tool 33.

[0067] <Data acquisition for process monitoring> The process monitoring device 44 is used to monitor the finishing process performed in the generation grinding machine 1, and, if necessary, to automatically detect and remove improperly machined workpieces and / or to intervene in the finishing process to correct improperly machined workpieces.

[0068] To this end, the process monitoring device 44 receives from the control computer 42 several different measurement data, including sensor data directly recorded by the control computer 42 and data read by the control computer 42 from the axis module 41, such as data indicating current consumption or power consumption in the tool spindle and workpiece spindle. For this purpose, the process monitoring device can be connected to the control computer 42 via a known interface, such as a known Profinet standard.

[0069] The process monitoring device 44 may also have its own analog and / or digital sensor inputs that directly receive sensor data from other sensors as measurement data. Additional sensors are typically those that are not directly required to control the actual machining process, such as acceleration sensors or temperature sensors that detect vibrations.

[0070] For the purposes of the following discussion, it is assumed that the process monitoring device 44 records at least the following measurement data. • Instantaneous angular velocity or instantaneous rotational speed (rpm) of the tool spindle 15; • Instantaneous angular velocity or instantaneous rotational speed (rpm) of the workpiece spindle 21; • Current consumption or power consumption of tool spindle 15; • Linear acceleration of the tool spindle housing 15 along three different spatial directions.

[0071] Of course, the process monitoring device 44 can also record several other measurement data.

[0072] The process monitoring device 44 continuously records measurement data during the machining of the workpiece. In particular, the current consumption or power consumption of the tool spindle 15 is recorded at a sufficiently high sampling rate such that there is at least one value, preferably multiple values, for each power consumption during machining of each tooth surface.

[0073] <Normalization Process> In the process monitoring device 44, filtering, such as low-pass filtering or band-pass filtering, is first applied to the recorded values ​​of the tool spindle's current consumption or power consumption to reduce high-frequency noise as needed. Next, a normalization process is applied to these values ​​(and possibly filtered values). The result of the normalization process is a normalized power indicator. The value of the normalized power indicator is obtained by normalizing the current consumption or power consumption by a normalization coefficient N. P It is calculated by multiplication of . The normalization coefficient takes into account the geometric parameters of the finishing tool, the geometric parameters of the workpiece, setting data of the finishing machine such as the speed of the tool spindle, radial infeed and axial feed per rotation of the workpiece, and the resulting contact condition between the tool and the workpiece.

[0074] This is based on the following considerations: The current consumption or power consumption of the work spindle depends heavily on the geometric parameters of the finishing tool, the geometric parameters of the workpiece, and the setting data of the finishing machine. For example, a grinding worm with a larger diameter requires more torque than a grinding worm with a smaller diameter, due to a longer effective lever arm, and therefore a higher current consumption is expected, given that all other machining conditions are the same. Also, for example, a higher axial feed rate or a larger radial infeed, given all other conditions being the same, will result in a higher current consumption of the tool spindle, as well as a higher speed for the tool spindle. The normalization factor takes such effects into account. As a result, the normalized power indicator generally no longer depends on such effects, or depends on them to a much lesser extent than in the case of directly measured current consumption or power consumption. Since these effects are already taken into account in the calculation of the normalized power indicator, deviations from the target process can be detected much more easily using the normalized power indicator than in the case of directly measured current consumption or power consumption.

[0075] This is explained in more detail in Figure 4. Figure 4 shows two representative curves 61 and 62 of the current consumption of the tool spindle during the generation grinding process of a single tooth surface. Curve 61 is measured for a relatively large radial infeed, and curve 62 is measured for a much smaller radial infeed, with all other machining conditions being the same. Both curves have a similar shape; that is, after the running-in phase, the current remains almost constant before decreasing again in the running-down phase. On the other hand, these curves differ significantly in their current amplitudes.

[0076] Within the scope of process monitoring, the temporal evolution of current consumption is continuously analyzed to detect unacceptable process deviations. This can be done in various ways. One possible method is to define an envelope curve that current consumption must not exceed or fall below. If the current consumption exceeds or falls below such an envelope curve, it can be concluded that an unacceptable process deviation has occurred. Such an envelope curve 63 is shown in a very simplified form as an example in Figure 4. During machining with radial infeed, for which the current curve 62 is recorded, the envelope curve 63 represents the upper limit of the current that must not be exceeded. However, this envelope curve is no longer useful when a larger radial infeed is set, as is clearly shown by the current curve 61. That is, with a larger radial infeed, the current consumption will exceed the envelope curve 63 even if the machining process is correct. Therefore, the envelope curve 63 must be recalculated each time the radial infeed is changed. This must be done based on test machining or empirical values. Both methods are time-consuming and tend to introduce errors.

[0077] For this reason, current consumption or power consumption measurements are normalized during the process monitoring. This normalization process takes radial infeed into particular consideration. This allows the normalized measurements to be directly comparable to each other, independently of the radial infeed values. Therefore, the same envelope curve can always be used for different radial infeed values. This curve only needs to be calculated once and can then be used for numerous different machining conditions.

[0078] Similar considerations apply to other methods of analyzing measurements, such as when several spectral components of the measurement are continuously monitored in the frequency domain.

[0079] The influence of the grinding worm's outer diameter on the recorded measurements is particularly important because the grinding worm's outer diameter changes with each dressing process. This is illustrated in Figure 5. Figure 5 shows the maximum current consumption i_max of the tool spindle measured during each generation grinding process on multiple workpieces. All workpieces were machined using the same grinding worm, which was dressed after a certain number of workpieces had been machined. The grinding worm's outer diameter decreases during each dressing operation. As a result, the workpieces are machined with a variable outer diameter. Figure 5 shows the grinding worm's outer diameter along the horizontal axis and the maximum current consumption of workpieces machined with this outer diameter along the vertical axis. Directly measured maximum current consumptions are marked with squares. It is easy to see that this current consumption also decreases with decreasing outer diameter. This means that the measured maximum current consumptions for different outer diameters are not directly comparable. In contrast, the normalized maximum current consumption, obtained by applying a normalization process to the measured maximum current consumption, is marked with an "X". The normalization process takes into account the variable outer diameter of the grinding worm. As a result, the value of the normalized maximum current consumption no longer depends on the outer diameter of the grinding worm.

[0080] The normalization process is preferably performed in real time while the workpiece is being machined in the finishing machine 1. On the one hand, this makes it possible to analyze the normalized measurements in real time during workpiece machining, and to detect unacceptable process deviations before or immediately after the completion of machining. Affected workpieces can be identified and rejected immediately in real time accordingly. On the other hand, it is ensured that the characteristic parameters of the machining process for each workpiece can be calculated from the normalized measurements immediately after the completion of machining of the workpiece. Thus, the calculated characteristic parameters are available immediately after the machining of the workpiece is completed. On the one hand, analysis of the calculated characteristic parameters makes it possible to detect further process deviations at an early stage. On the other hand, it is not necessary to store directly recorded measurements (i.e., raw data) for a longer period, as in the case of offline evaluation. Instead, it is sufficient to store the calculated characteristic parameters along with the identifier of each workpiece and the selected process parameters. This makes it possible to keep memory requirements very low.

[0081] <Example of calculating characteristic parameters from measured variables> The process monitoring device 44 calculates various characteristic parameters from a (preferably normalized) power indicator, as well as other measured variables that characterize the machined workpiece and its machining process. These characteristic parameters are advantageously process-specific, and therefore, they allow for direct conclusions about process deviations in the machining process. In particular, the characteristic parameters make it possible to predict several machining errors of the workpiece. Thus, the number of workpieces that undergo individual gear measurements can be reduced, while process deviations can still be reliably detected at an early stage, allowing for corrective actions to be taken in the machining process as needed.

[0082] The calculation of characteristic parameters from the measured variables is shown below using, as an example, the following three characteristic parameters. (a) Cumulative pitch indicator I fP (b) Wear indicator I Wear (c) Profile shape indicator I ffa

[0083] All three parameters are determined by spectral analysis of the time evolution of the measured variables over the machining of the workpiece.

[0084] <(a) Cumulative pitch indicator I fP > Cumulative pitch indicator I fP [[ID=2⑷To obtain the cumulative pitch indicator I, the spectral components of the power indicator (preferably normalized) at the workpiece speed n C are evaluated.

[0085] This is shown in Figure 6. This shows the spectrum of the normalized current consumption (the absolute value of the spectral component of the normalized current consumption as a function of the frequency "f"). Such a spectrum can be obtained by FFT of the time evolution of the normalized current consumption. The arrow indicates the spectral component of the normalized current consumption at the rotational speed n C of the workpiece spindle. To quantify this spectral component, the spectral intensity at this frequency can be determined, or the spectrum in a narrow range around this frequency can be integrated. The resulting quantity is the cumulative pitch indicator I fP .

[0086] The greater the cumulative circular pitch error of the pre-machined tooth portion of the workpiece and / or the worse the concentricity of the workpiece, the cumulative pitch indicator I fP generally becomes larger. Therefore, the cumulative pitch indicator I fPFrom this, it is possible to estimate the existing cumulative pitch error of the unprocessed parts from the pre-processing stage, and / or the concentricity error caused by, for example, the inaccurate alignment of the workpiece clamping device.

[0087] <(b) Wear indicator I Wear > Wear indicator I Wear To obtain this, the static portion of the normalized power indicator, i.e., the portion below a high-frequency cutoff frequency such as 2 Hz, is needed. For this purpose, for example, the time evolution of the power indicator can be filtered using a low-pass filter and integrated.

[0088] Wear indicator I Wear It can be understood as a measure of the normalized cutting energy applied to the workpiece after all geometric influences and the influence of technical data used, such as radial infeed and axial feed, have been removed through normalization processing. In short, wear indicator I Wear The higher the value of , the more material the grinding worm has removed from the workpiece with a given drive power. Therefore, a decrease in the wear indicator reflects a deterioration in the tool's removal action from the workpiece, assuming no other conditions have changed. In this respect, wear indicator I Wear A decrease in this value indicates increased tool wear.

[0089] <(c) Contour shape indicator I ffa > Normalized current consumption becomes increasingly meaningless as the frequency increases. For this reason, other measurements, such as those from the accelerometer 18, are preferably used to calculate parameters derived from high-frequency process components.

[0090] Contour shape indicator I ffaTo determine this, the spectral components of such measurements at the tooth clenching frequency are evaluated. The tooth clenching frequency corresponds to the workpiece velocity multiplied by the number of teeth z in the workpiece. That is,

number

[0091] This is shown in Figure 7. Figure 7 shows the spectrum of the signal from the accelerometer 18 during machining of the workpiece. The spectral component at the tooth clenching frequency is marked with an arrow. This can be quantified, for example, by integrating the spectrum over a narrow range near the tooth clenching frequency.

[0092] The greater the contour shape deviates from the ideal contour shape according to the specifications, the more the contour shape indicator I ffa This generally increases. Contour shape indicators can therefore be used to draw conclusions about contour shape deviations or process deviations that result in such contour shape deviations.

[0093] <Examples of calculations for other characteristic parameters> The calculation of characteristic parameters was explained above using three examples. However, it goes without saying that many other characteristic parameters can be calculated.

[0094] Another example is the vibration indicator I Vib This characteristic parameter is obtained by integrating the absolute value of the acceleration sensor measurement signal in the frequency domain.

[0095] <Detection of process deviations> By monitoring changes in the required characteristic parameters across multiple workpieces, indicators of deviations in the machining process from an idealized target process can be obtained. Based on this, the machining process can be adjusted accordingly to reduce the deviation. To detect unacceptable process deviations with greater accuracy in real time, the comparison of the required characteristic parameters across multiple workpieces can be used to more precisely define limit values, such as the envelopes mentioned above, for real-time monitoring.

[0096] This will be explained below using Figures 8 to 11.

[0097] When interpreting these figures, several characteristics of the processing methods selected here must be taken into consideration beforehand.

[0098] Firstly, it should be noted that each workpiece is machined in two stages: rough machining and finish machining. The shift strategy is adapted to this as follows: Each workpiece is first roughly machined in a specific grinding worm region. Next, the grinding worm is shifted by a specific amount (towards a larger Y value in the figure, i.e., to the left) so that an unused grinding worm region is used for finish machining. After finish machining, the grinding worm is shifted back to the end of the grinding worm region that was last used for rough machining, and the next grinding worm region is used for rough machining of the next workpiece. As a result, almost all grinding worm regions are used first to finish one workpiece and then used to rough machine subsequent workpieces. Only the grinding worm region at the far right, close to Y=0, is used solely for rough machining in this shift strategy. The machining position Y in Figures 8 to 11 indicates which worm region along the grinding worm width was used to machine each workpiece during rough machining.

[0099] On the one hand, it should be noted that the grinding worm is newly dressed each time the end of the grinding worm is reached during the shift. The outer diameter of the grinding worm decreases during dressing. On the one hand, this changes the leverage ratio that converts the drive torque into cutting force when grinding the workpiece, and on the other hand, the contact conditions of each tooth surface during machining also change. Figures 8 to 11 show the parameters over several dressing cycles as a function of the machining position Y during rough machining, respectively. Due to the normalization process, the dressing has little to no effect on each parameter. This is because the effect of the dressing operation on the geometry of the grinding worm is taken into account by the normalization coefficient. As a result, the normalized power indicator values ​​in different dressing cycles are directly comparable to each other. Correspondingly, the characteristic parameters shown for different dressing cycles are also directly comparable to each other despite the fluctuating outer diameter of the grinding worm. This is a major advantage of the proposed normalization process.

[0100] Figure 8 shows the cumulative pitch indicator I of the rough machining process of multiple workpieces over several dressing cycles as a function of the machining position Y along the grinding worm width. fP This is shown. This figure shows whether the workpiece was machined on the first workpiece spindle or on the second workpiece spindle (I fP (C1) or I fP (C2)) is distinguished. Using a triangle, the cumulative pitch indicator I of the machined workpiece on the first workpiece spindle fP (C1) is shown, while the cross mark indicates the cumulative pitch indicator I of the machined workpiece on the second workpiece spindle. fP This indicates (C2).

[0101] It is immediately apparent that the cumulative pitch indicator of the first workpiece spindle is, on average, considerably higher than that of the second workpiece spindle. For similarly pre-machined workpieces, this indicates a concentricity error of the workpiece in the first workpiece spindle, resulting from inaccurate alignment of the clamping device. Such a concentricity error can introduce unwanted noise when using gears produced in this manner. At the same time, normalization processing reveals that the dressing operation does not actually affect the desired cumulative pitch indicator value.

[0102] Thus, characteristic parameter I fP However, by correlating with another quantity in the machining process, in this case position Y along the width of the grinding worm, and displaying it visually, it becomes easier for the operator to recognize the concentricity error and its corresponding cause. Instead of position Y, other quantities can also be used here, in the simplest case being correlation with consecutive workpiece numbers.

[0103] Figure 9 shows the contour shape indicators I of multiple workpieces. ffa However, it is similarly expressed as a function of machining position Y for roughing operations across several dressing cycles. Regardless of the dressing cycle, the contour shape indicator can be seen to be considerably smaller at the right end of the grinding worm (near Y=0), where machining begins after each dressing operation, than further along the grinding worm, and increases on average toward the left end of the grinding worm (Y=40mm) with large fluctuations. This figure thus shows that while the first workpiece in a dressing cycle is always produced with the correct contour shape, the contour shape deviation increases with each subsequent workpiece and also fluctuates violently. From the perspective of the shift strategy described above, this indicates that the grinding worm is overloaded in the area where it is first used for finishing operations and then for roughing operations.

[0104] In Figure 9, parameter I ffa This is correlated with another quantity in the machining process, in this case the position Y along the grinding worm width, and by visually displaying this other quantity together, it becomes easier for the operator to recognize contour shape errors and their causes.

[0105] Figure 10 shows wear indicators I of multiple workpieces. Wear However, it is similarly expressed as a function of the machining position Y during the roughing operation. The wear indicator has a relatively large value at the right end of the grinding worm, close to Y=0. As the machining position progresses along the width of the grinding worm, the wear indicator decreases rapidly to a significantly lower value. It should be noted that the wear indicator is not a direct measure of wear itself, but a measure of the amount of material removed from the tooth surface. This means that across the widest range of the grinding worm width, less material is removed from the tooth surface than in the range at the right end, close to Y=0. This indicates that wear is increasing in all areas except the range at the right end. The change in the wear indicator across the width of the grinding worm, therefore, supports the insights that can also be obtained from the change in the profile shape indicator across the width of the grinding worm. That is, the wear indicator indicates that the grinding worm is excessively worn everywhere except in the area at the right end, close to Y=0.

[0106] Figure 11 shows the vibration indicator I of multiple workpieces as a function of the machining position Y during rough machining. Vib This indicates that, in the wear area of ​​the grinding worm, a higher vibration load is generated despite the slow material removal rate. This, in turn, can lead to unwanted noise when using gears produced in this manner. The transition of the vibration indicator across the width of the grinding worm thus reaffirms the findings already obtained from the transitions of the contour shape indicator and the wear indicator.

[0107] To perform compensatory actions here, for example, the tool spindle speed, or the radial infeed or axial feed per rotation of the workpiece can be reduced.

[0108] <Comparison with measurements from gear measurements> The characteristic parameters obtained for the selected workpiece can be compared with the results of gear measurements using a gear measuring machine. In this way, parameters representing the correlation between the characteristic parameters and the actual shape deviation can be quantitatively determined. For example, if there is a linear correlation between the characteristic parameters and the shape deviation, the coefficient of the linear correlation can be determined by performing a linear regression. This makes it possible to directly quantify the shape deviation using the characteristic parameters for each machined workpiece. Otherwise, this would only be possible through gear measurements, which would involve a disproportionate amount of effort.

[0109] <Web-based interface> The graphical display of the required characteristic parameters and their correlation with other characteristic values ​​of the machining process can be performed via a web browser on any client computer, particularly in a platform-independent manner. Other evaluation metrics can also be implemented accordingly, in a platform-independent manner. This facilitates remote analysis.

[0110] <Automatic detection of machining errors> In the example above, the analysis of the changes in various characteristic parameters over the grinding worm width was performed visually by the operator on the machine or by a specialist on any client computer. Alternatively, such analysis can be performed entirely automatically.

[0111] For this purpose, the process monitoring device 44 can execute an algorithm that automatically recognizes patterns in the characteristic parameters obtained across several workpieces. Machine learning algorithms, as various versions of which are known, are particularly well suited to this purpose. Such algorithms are often also called "artificial intelligence." One example of this is a neural network algorithm. It is clear that such an algorithm in the above example can easily detect, for example, the difference in the cumulative pitch indicator between the first workpiece spindle and the second workpiece spindle or the wear behavior across the grinding worm width described above. For this purpose, the algorithm can be trained in the usual way using a training dataset. The training dataset may particularly take into account parameters that represent the correlation between the characteristic parameters and the actual shape deviation according to gear measurements.

[0112] At this time, necessary measures can be taken to eliminate process deviations. For example, when a concentricity error is detected, the centering of the workpiece clamping device in the corresponding workpiece spindle can be corrected manually or automatically. If excessive wear is detected, the radial infeed and / or axial feed can be reduced accordingly. These measures can also be taken manually or automatically.

[0113] <Force Model> It is preferable to calculate the normalization coefficients using a model-based approach.

[0114] Regarding generative grinding, there are models in the literature that describe the dependency of cutting force on the geometric and technical parameters of the tool and workpiece. As an example, refer to Chapter 4.7.3 "Cutting Force" (pp. 319-322) of the aforementioned book "Continuous Generating Gear Grinding" (ISBN 978-3-033-02535-6) by H. Schriefer et al., edited in 2010 by Reishauer AG in Valiselen.

[0115] The following references the force model used in the paper "Numeric Simulation of Continuous Generating Grinding taking into Account the Dynamic Response of the System Machine - Tool - Workpiece" by C. Dietz (Diss. ETH Zurich No. 24172, https: / / doi.org / 10.3929 / ethz-b-000171605). This paper discloses a method for numerically modeling the continuous generating grinding process. In particular, a model for calculating cutting forces is presented, and the procedure for how the parameters of this model can be experimentally determined by measurement is shown.

[0116] The force model for the normal force is given in equation 4.27 of C. Dietz's paper as follows:

number

[0117] Cutting force F c F is the normal force. n It is proportional to μ, and the constant of proportionality μ is called the ratio of forces.

number

[0118] The ratio μ of forces is also a quantity that can be determined empirically.

[0119] Geometric variable a e a p and l k This represents the cutting zone. These geometric variables can be calculated analytically, as shown in Chapter 4.5.1 of the paper by C. Dietz, or they can be determined numerically by transmission calculations.

[0120] In particular, the contact length l k The following analytical relationships can be derived.

number

number

[0121] Cutting width a pThe following analytical relationships can be derived regarding this.

number

number

[0122] The cutting depth corresponds to the nominal grinding tolerance, and the nominal grinding tolerance is related to the radial infeed Δx as follows.

number

[0123] cutting speed v c and feed rate v f This is derived from the kinematics of the regenerative grinding process. As described in Chapter 4.7.1 of the paper by C. Dietz, analytical formulas for these variables can also be given. For example, the following equation is given for cutting speed v c Applies to this.

number

[0124] peripheral speed v cu The following formula applies to this:

number

[0125] Axial velocity component v ca The following formula applies to this:

number

[0126] Rolling speed v cw The following formula applies to this:

number

[0127] feed rate v f The following formula applies to this:

number

[0128] Alternatively, a geometric variable a that describes the cutting zone. e a p and l k and cutting speed v c and feed rate v f This can also be determined from numerical process simulations.

[0129] The constants F0 and k, the exponents ε1 and ε2 of the powers, and the ratio μ of the forces can be determined empirically, as given, for example, in Chapter 5.3 of C. Dietz's paper.

[0130] In C. Dietz's paper, the following values ​​were empirically determined for generative grinding of gears made of hardened steel using vitrified bond tools with aluminum oxide abrasive.

number

[0131] Using these values, the process force actually measured can be reproduced with very high accuracy by the force model.

[0132] In the case of other material combinations, the values ​​of the parameters mentioned above may deviate from those shown above. However, such values ​​can be easily determined empirically by comparing the measured force value with the calculated force value.

[0133] <Modeling process power> Power P supplied by the tool spindle B This is the cutting force F of the grinding worm at the contact point between the grinding worm and the workpiece. c and peripheral velocity v cu It is obtained as the product of the following. On the other hand, this peripheral speed is the rotational speed n of the tool spindle. B It is proportional to the diameter d of the grinding worm at the contact point. pSS It is proportional to the effective lever arm, which is half of the effective lever arm.

[0134] Process power can be modeled as follows:

number

[0135] Diameter d of the grinding worm at the contact point pSS Its outer diameter d is a good approximation. aSS It can be replaced with this.

[0136] <Calculation of normalization coefficients> Based on this model of process power, the normalization coefficient can be selected, for example, as follows:

number

[0137] For the exponents E1, E2, and E3 of the power, the following formulas apply in this model.

number

[0138] However, in extended models, the exponents of these powers may also deviate from 1, and these can be determined empirically.

[0139] The first coefficient (with exponent E1) is obtained directly from the force model. This coefficient takes into account the geometry of the grinding worm and workpiece, as well as the technical specifications, particularly radial infeed and axial feed.

[0140] The second coefficient (with an exponent E2) takes into account the leverage ratio at the contact point, which varies depending on the grinding worm diameter.

[0141] The third coefficient (with exponent E3) takes into account the process power's dependency on the tool spindle speed.

[0142] As can be seen from the above considerations of the force model and the further explanation in C. Dietz's paper, the cutting force varies to some extent over the machining of the flank. However, for the purpose of process control, the cutting force can be considered constant over the machining of one flank, neglecting run-in and run-out. Therefore, after each dressing operation (changing the geometry of the grinding worm) and after each change in technical parameters (especially radial infeed and / or axial feed), the normalization coefficient N P It is sufficient to recalculate this normalization coefficient N. PIt can then be used on all workpieces in the dressing cycle.

[0143] Normalization coefficient N P Of course, it can also be calculated in ways different from those described above. More complex normalization processes are possible, for example, a normalization process that involves subtraction to remove the offset first, followed by only multiplication or division.

[0144] The above considerations apply to generative grinding. For other finishing operations, there are other models of cutting forces, and therefore, for other finishing operations, the normalization coefficient will differ from the normalization coefficient shown above.

[0145] <Flowchart of an exemplary method> Figure 12 shows a flowchart illustrating an exemplary method for monitoring the process during the generation grinding of a batch of similar workpieces using the type of finishing machine shown in Figure 1.

[0146] In step 110, the finishing machine is set up and relevant process parameters (particularly geometric parameters of the grinding worm and workpiece, as well as technical parameters such as radial infeed and axial feed) are input into the machine control unit 42 via the control panel 43. In step 111, the grinding worm 16 is dressed and the outer diameter of the dressed grinding worm is determined. In step 112, a normalization factor is calculated based on the process parameters and the outer diameter of the grinding worm.

[0147] In block 120, individual workpieces of a batch are machined. During the machining process, measured variables are continuously recorded by the process monitoring device 44 in step 121. In step 122, at least some of the measured variables, particularly those relating to the current consumption of the tool spindle, are normalized in real time. In step 123, while individual workpieces are still being machined, the partially normalized measured variables at this point are continuously analyzed in real time to directly detect possible machining errors online based on process deviations. If a possible machining error is detected, the corresponding information variable is set in the process monitoring device. Steps 121-123 are continuously repeated while the workpieces are being machined.

[0148] Immediately after the machining of the workpiece is complete, characteristic parameters are calculated in step 124 from the partially normalized measured variables. The characteristic parameters are compared to the specifications. If the parameters are found to deviate excessively from the specifications, information variables for machining errors are set.

[0149] In step 125, the workpiece handling system is instructed by an information variable to remove any workpieces in which signs of machining errors have been detected. These workpieces may undergo further inspection or be immediately rejected as NIO parts.

[0150] In step 126, a dataset for each workpiece is stored in the database. This dataset includes a unique workpiece identifier, the most important process parameters, the requested characteristic parameters, and optionally, information variables.

[0151] The machining of the workpiece is repeated in the same manner until the grinding worm is worn down to the point where a new dressing operation is required. If a dressing operation is required, step 111 is repeated. That is, the grinding worm is dressed again and its new outer diameter is determined. Therefore, the normalization factor is recalculated in step 112. The machining of the workpiece 120 is then continued using the newly dressed grinding worm and the new normalization factor.

[0152] Figure 13 illustrates how further process deviations can be automatically detected from stored datasets of several workpieces. In step 131, datasets of several workpieces are read from the database. In step 132, these datasets are analyzed by an AI algorithm (artificial intelligence) to identify process deviations that may not have been directly detectable during the machining of individual workpieces. In step 133, the results of this analysis are used to automatically initiate corrective measures for the machining process (e.g., reducing axial feed). These steps can be performed each time a certain minimum number of workpieces are machined. However, this analysis can also be continued after the machining of a batch of workpieces is complete to continue identifying workpieces affected by machining errors, for example.

[0153] This process requires only a moderate amount of computation and memory because the dataset stored is very small compared to the amount of data directly retrieved during processing.

[0154] As shown in Figure 14, independently of this, the dataset can be read from the database at any time via the network using a client computer (step 141), and can be graphically edited and output (step 142). This procedure also requires only a very moderate amount of computation and memory. This makes it possible to perform this process using a plugin in a web browser. Based on this output, the operator can perform error analysis and, for example, recalculate the envelope described above that is applied during real-time analysis.

[0155] <Block diagram of the functional blocks of the process monitoring device> Figure 15 shows an example block diagram schematically illustrating the various functional blocks of the process monitoring device 44. These functional blocks are functionally connected to each other via the command / data exchange component 401.

[0156] The normalization calculation device 410 calculates normalization coefficients as needed. The detection device 420 is used to detect the measurements. The normalization device 430 normalizes at least a portion of the measurements immediately after they are detected. The defect detection device 440 analyzes the partially normalized measurements to identify unacceptable process deviations. The handling device 441 (strictly speaking, not part of the process monitoring device) then removes the workpieces that have been processed with unacceptable process deviations. After the machining of the workpieces is complete, the characteristic parameter calculation device 450 calculates characteristic parameters from the partially normalized measurements. The data communication device 460 is used to communicate with the database server. The deviation detection device 470 is used for automatic detection of process deviations. For this purpose, the deviation detection device 470 includes a processor device 471 that executes an AI algorithm.

[0157] It goes without saying that numerous variations of the examples shown above are possible.

Claims

1. A method for monitoring a machining process in which a tooth flank of a pre-toothed workpiece (23) is machined using a finishing machine (1), said finishing machine (1) having a tool spindle (15) for driving a finishing tool (16) to rotate it about a tool axis (B) and at least one workpiece spindle (21) for driving a pre-toothed workpiece (23) to rotate it, The method comprises: detecting a plurality of measurements while said finishing tool (16) is in machining engagement with a workpiece (23); Including, The method comprises: applying a normalization process to values ​​of at least some of said measurements or quantities derived from said measurements to obtain normalized values; Including, 10. The method of claim 9, wherein the normalization process depends on at least one process parameter, the at least one process parameter being selected from the geometric parameters of the finishing tool (16), the geometric parameters of the workpiece (23), and the setting parameters of the finishing machine (1).

2. 2. The method according to claim 1, wherein the detected measurements comprise a value of a power indicator indicative of the current power consumption of the tool spindle (15), or in particular of the current value of the tool spindle (15), and the normalization process is applied to the value of the power indicator or to a variable derived from the power indicator.

3. 3. The method of claim 1 or 2, wherein the normalization process is performed in real time while the finishing tool (16) is in machining engagement with the workpiece (23).

4. analyzing the normalized values ​​in real time to detect unacceptable process deviations; The method of claim 3, comprising:

5. Removing workpieces determined to have unacceptable process deviations; The method of claim 4, comprising:

6. calculating characteristic parameters of the machining process from the measurements or values ​​derived from the measurements; The method of any one of claims 1 to 5, wherein at least one of the characteristic parameters is correlated with a predetermined machining tolerance of the workpiece (23).

7. A method for monitoring a machining process in which a tooth surface of a pre-toothed workpiece (23) is machined using a finishing machine (1), said finishing machine (1) having a tool spindle (15) for driving a finishing tool (16) to rotate it about a tool axis (B) and at least one workpiece spindle (21) for driving a pre-toothed workpiece (23) to rotate it; The method comprises: detecting a plurality of measurements while said finishing tool (16) is in machining engagement with a workpiece (23); calculating characteristic parameters of said machining process from said measurements; Including, The method, wherein at least one of the characteristic parameters is correlated to a predetermined machining error of the workpiece (23).

8. performing gear measurements on selected workpieces (23) to determine at least one gear measurement per workpiece that characterizes said predetermined machining error; determining a correlation parameter characterizing the correlation between the at least one characteristic parameter and the at least one gear measurement; 8. The method of claim 6 or 7, comprising:

9. A method according to any one of claims 6 to 8, wherein said calculation of at least one of said characteristic parameters comprises spectral analysis of measurements or values ​​derived from said measurements to obtain a plurality of spectral components.

10. A method according to claim 9, wherein spectral components at multiples of the rotational speed of the tool spindle (15) and / or the workpiece spindle (21) are evaluated.

11. The machining process is a generating process in which the finishing tool (16) and the workpiece (23) are in rolling engagement, and the characteristic parameters are: a cumulative pitch indicator (IfP) calculated from the spectral components at the rotational speed of the workpiece spindle (15) and correlating with the cumulative pitch error or concentricity error of the workpiece (23); a wear indicator (IWear) calculated from low frequency spectral components and correlated with the degree of wear of said finishing tool (16); a profile shape indicator (Iffa) calculated from the spectral components at the tooth meshing frequency and correlating with the profile shape deviation of the workpiece (23); 11. The method of claim 9 or 10, comprising at least one of:

12. storing a data set in a database (46) including a unique identifier of the workpiece, at least one process parameter, and at least one of the characteristic parameters; The method according to any one of claims 6 to 11, comprising:

13. performing an analysis of values ​​of at least one of said characteristic parameters for a plurality of workpieces (23) to determine process deviations; modifying the machining process to reduce the process deviation; 8. The method of claim 6 or 7, comprising:

14. The method of claim 13 , wherein the analysis is performed by a trained machine learning algorithm.

15. The analysis correlating values ​​of at least one of said characteristic parameters of a plurality of workpieces (23) with another parameter of said machining process; 15. The method of claim 13 or 14, comprising:

16. - graphically outputting, or in particular to a web browser, the values ​​of at least one of said characteristic parameters of a plurality of workpieces or a value derived from at least one of said characteristic parameters; The method according to any one of claims 6 to 15, comprising:

17. Varying at least one of the process parameters; recalculating the normalization of the process parameters after the modification; The method according to any one of claims 1 to 6, comprising:

18. 18. The method of claim 17, wherein the recalculation of the normalization process comprises using a model representing the expected dependency of the measurement on the process parameters, or in particular using a model of process forces or process powers.

19. 19. The method according to claim 17 or 18, wherein the recalculation of the normalization process includes compensation for variable dimensions of the finishing tool (16), or in particular compensation for its outer diameter (daSS).

20. a tool spindle (15) for driving a finishing tool (16) to rotate about a tool axis (B); At least one workpiece spindle (21) for driving and rotating a pre-toothed workpiece (23); a control device (40) for controlling the process of machining the workpiece (23) with the finishing tool (16); a process monitoring device (44) configured to perform the method for monitoring a machining process according to any one of claims 1 to 18; a finishing machine for said machining of tooth surfaces of a pre-toothed workpiece, comprising:

21. The process monitoring device (44) a sensing device (420) for sensing a plurality of measurements while the finishing tool (16) is in machining engagement with a workpiece (23); a normalization device (430) for applying a normalization process to at least some of the measurements or values ​​of quantities derived from the measurements to obtain normalized values; Equipped with 21. The finishing machine according to claim 20, wherein the normalization process depends on at least one process parameter, the at least one process parameter being selected from geometric parameters of the finishing tool (16), geometric parameters of the workpiece (23), and setting parameters of the finishing machine (1).

22. 22. A finishing machine according to claim 21, wherein the detected measurements comprise the current power consumption of the tool spindle (15), or in particular the value of a power indicator indicative of the current value of the tool spindle (15), and the normalization process is applied to the value of the power indicator or to a variable derived from the power indicator.

23. 23. The finisher of claim 21 or 22, wherein the normalizing device (430) is configured to perform the normalizing process in real time while the finishing tool (16) is in machining engagement with the workpiece (23).

24. 24. The finishing machine of claim 23, wherein the process monitoring device comprises a fault detection device (440) configured to analyze the normalized value in real time to detect unacceptable process deviations.

25. 25. The finishing machine of claim 24, comprising a workpiece handling device (441) configured to automatically remove workpieces determined to have unacceptable process deviations.

26. The process monitoring device a sensing device (420) for sensing a plurality of measurements while the finishing tool (16) is in machining engagement with a workpiece (23); and a characteristic parameter calculation device (450) for calculating a characteristic parameter of the machining process from the measurements or values ​​derived from the measurements.

27. 27. The finishing machine of claim 26, wherein the characteristic parameter calculation device (450) is configured to perform a spectral analysis of the measurements or values ​​derived from the measurements to obtain a plurality of spectral components.

28. A finishing machine as described in Claim 27, wherein the characteristic parameter calculation device (450) is configured to evaluate spectral components at multiples of the rotational speed of the tool spindle (15) and / or the workpiece spindle (21).

29. The finishing machine is configured to perform a generating process in which the finishing tool (16) and the workpiece (23) are in rolling engagement, and the characteristic parameters are: a cumulative pitch indicator (IfP) calculated from the spectral components at the rotational speed of the workpiece spindle (15) and correlating with the cumulative pitch error or concentricity error of the workpiece (23); a wear indicator (IWear) calculated from low frequency spectral components and correlated with the degree of wear of said finishing tool (16); a profile shape indicator (Iffa) calculated from the spectral components at the tooth meshing frequency and correlating with the profile shape deviation of the workpiece (23); 28. The finisher of claim 27, including at least one of:

30. The process monitoring device a data communication device (460) for transmitting a data set including the unique identifier of the workpiece, at least one process parameter, and at least one of the characteristic parameters to a database (46); 30. A finishing machine according to any one of claims 26 to 29, comprising:

31. The process monitoring device a deviation detection device (470) for detecting a process deviation of the machining process from a target process based on values ​​of at least one of the characteristic parameters of a plurality of workpieces (23); Finishing machine according to any one of claims 26 to 30, comprising:

32. 32. The finishing machine of claim 31, wherein the deviation detection device (460) comprises a processor device (461) programmed to execute a trained machine learning algorithm to detect the process deviation.

33. The process monitoring device (44) a normalization calculation device (410) for recalculating the normalization process when at least one of the process parameters changes; 26. A finishing machine according to any one of claims 21 to 25, comprising:

34. 34. The finishing machine of claim 33, wherein the normalization calculation device (410) is configured to apply a model representing an expected dependency of the measurements on the process parameters, or in particular a model of process forces or process powers, when recalculating the normalization process.

35. Finishing machine according to claim 33 or 34, wherein the normalization calculation device (410) is configured to compensate for the dimensions of the finishing tool (1), or in particular for its outer diameter (daSS).

36. A computer program comprising instructions for causing a process monitoring device in a finishing machine (1) according to any one of claims 20 to 35 to carry out the method according to any one of claims 1 to 19.

37. 37. A computer readable medium having stored thereon the computer program of claim 36.

Images