Double-sided or single-sided machine tool and method for operating a double-sided or single-sided machine tool

Abstract

This invention relates to a double-sided or single-sided machine tool, comprising a first working disc, preferably annular, and a reverse bearing element, wherein the first working disc and the reverse bearing element are rotatably driven relative to each other by a rotary drive, and wherein a preferably annular working clearance for double-sided or single-sided machining of a flat workpiece, preferably a wafer, is formed between the first working disc and the reverse bearing element. The double-sided or single-sided machine tool includes a plurality of sensors that record measurement data related to tool and / or machining parameters of the double-sided or single-sided machine tool during operation. A control device is provided that acquires the measurement data recorded by the sensors during operation of the double-sided or single-sided machine tool, and the control device includes an artificial neural network designed to create a state vector of the double-sided or single-sided machine tool from the measurement data and compare the state vector with at least one target state vector. This invention also relates to a method for operating a double-sided or single-sided machine tool.

Description

Technical Field

[0001] The present invention relates to a double-sided or single-sided machine tool comprising a first, preferably annular, working disk and a preferably annular counter-bearing element, wherein the first working disk and the counter-bearing element can be rotationally driven relative to one another by means of a rotary drive, and wherein a preferably annular working gap for double-sided or single-sided processing of flat workpieces, preferably wafers, is formed between the first working disk and the counter-bearing element. The double-sided or single-sided machine tool comprises a plurality of sensors for recording measurement data relating to the tool and / or processing parameters of the double-sided or single-sided machine tool during operation of the double-sided or single-sided machine tool. The present invention also relates to a method for operating such a double-sided or single-sided machine tool.

[0002] For example, in a double-side polishing machine, flat workpieces, such as wafers, are polished between preferably annular work disks. A preferably annular working gap is provided between the work disks, where the flat workpiece, such as a wafer, is held during processing. For this purpose, a so-called rotor disk is typically positioned in the working gap, where the workpiece can be mounted in a floating manner in a groove. During processing, the work disks are driven relative to each other by a rotary drive, and the rotor disks rotate in the working gap due to their external teeth, which mesh with corresponding teeth on a pin ring. Thus, the workpiece is transported through the working gap along a cycloidal path during processing. Furthermore, during double-side polishing, a polishing agent, known as a slurry, is introduced into the working gap and ensures the abrasive process. Furthermore, in a double-side polishing machine, the work disks have polishing cloths, known as polishing pads, regularly spaced on their surfaces, which delimit the working gap. Prior Art

[0003] The goal of machining is to achieve a fully machined workpiece shape that is as plane-parallel as possible. The working gap geometry is crucial for this. DE 10 2006 037 490 B4 discloses a double-sided machine tool having a component for producing a global deformation of one of the working disks. In particular, the upper working disk can be deformed between a global concave and a global convex shape. With this global deformation, the concave or convex shape of the working disk is initially produced over its entire diameter, viewed in the radial direction. The annular surface of the preferably annular working disk, which delimits the working gap, remains flat; however, the opposing annular portions of the annular surface are deformed relative to one another so that the overall concave or convex shape is produced.

[0004] DE 10 2016 102 223 A1 also discloses a double-sided machine tool having a component for producing a localized deformation of one of the working disks, particularly between a localized convex shape and a localized concave shape. With this localized deformation, a convex or concave shape is produced in the radial direction between the inner and outer edges of, for example, an annular working disk. Unlike global deformation, with localized deformation, the annular portion itself is deformed concavely or convexly.

[0005] The two aforementioned embodiments can be combined in a double-sided machine tool. This allows for a wide range of working gap geometries to be produced. Thus, for example, in the event of partial wear of the polishing cloth or temperature fluctuations in the components defining the working gap, it is always possible to ensure that the workpiece is processed as plane-parallel as possible, or to ensure that the working gap is set in a manner that is optimal for workpiece quality, regardless of whether it is parallel or not.

[0006] The geometry of the working gap has a decisive influence on the shape and uniformity of the machined workpiece. In addition to the working gap geometry, the machining result is also influenced by a number of additional tool and / or machining parameters, such as the temperature of various machine components, the thickness and possible wear of the working bushing (e.g., polishing cloth), the speed of the working discs and / or counter-bearing elements rotating relative to each other and the rotor disc rotatably mounted in the working gap, or the load between, for example, the first working disc and the counter-bearing element.

[0007] It is known to monitor the tools and / or processing parameters of double-sided or single-sided machine tools with the aid of sensors during their operation. It is also known to use corresponding sensors to detect the shape and thickness of flat workpieces (e.g., wafers) being processed in the work gap. A suitable parameter window for the operation of the double-sided or single-sided machine tool must be found from a large number of these tool and processing parameters, initially as part of the double-sided or single-sided machine tool setup before commencing processing of flat workpieces. The double-sided or single-sided machine tool must be adjusted to the prevailing conditions at the relevant location of use, such as the type of workpiece insert (such as a polishing cloth), the type of polishing compound, if applicable, and other operator specifications. During the subsequent production operation of the double-sided or single-sided machine tool, the process must be monitored with the aid of sensors. Deviations from specified target values ​​(e.g., GBIR or SFQR values ​​of the processed wafers) must be identified early in this process and, if applicable, corrected during the processing sequence.

[0008] Particularly due to the large number of different machining processes, sensor measurement results must be interpreted by specialists in order to extract the correct conclusions for production operations. Such specialists are not available at every location where double-sided or single-sided machine tools are used. This can negatively impact production processes. Furthermore, production operations are often only adjusted after any unfavorable parameter deviations have occurred, with considerable delay. One reason for this is that, due to the large number of tools and machining parameters that influence the production process, it is difficult for specialists to identify relevant deviations in the measured parameters in advance. This typically occurs only after measuring the finished workpiece. If undesirable deviations are subsequently discovered during the production process, a large number of rejects are also produced. Summary of the Invention

[0009] Continuing from the explained prior art, the object of the present invention is to provide a double-sided or single-sided machine tool and a method for operating a double-sided or single-sided machine tool, with the help of which the production process of the double-sided or single-sided machine tool can be monitored more quickly and reliably while minimizing rejects.

[0010] The invention solves this object via independent claims 1, 11 and 12. Advantageous embodiments can be found in the dependent claims, description and drawings.

[0011] With regard to a double-sided or single-sided machine tool of the type mentioned at the outset, the present invention achieves the following object: to provide a control device which receives measurement data recorded by sensors during operation of the double-sided or single-sided machine tool, wherein the control device comprises an artificial neural network which is designed to create a state vector of the double-sided or single-sided machine tool from the measurement data and to compare this state vector with at least one target state vector.

[0012] With regard to a method of the type mentioned at the outset, the invention achieves the following object: an artificial neural network is trained by inputting a large number of target state vectors which lead to acceptable machining results for flat workpieces.

[0013] The double-sided or single-sided machine tool according to the present invention can particularly be a double-sided or single-sided polishing machine. However, the double-sided or single-sided machine tool can also be a double-sided or single-sided lapping machine or a double-sided or single-sided grinding machine. The double-sided or single-sided machine tool has a first, preferably annular, working disk and a preferably annular counter-bearing element. In a single-sided machine tool, the counter-bearing element can be designed, for example, as a simple weight or pressure cylinder. The counter-bearing element can preferably be a second, preferably annular, working disk. The first working disk and the counter-bearing element can be driven rotationally relative to one another, and a preferably annular working gap for machining flat workpieces such as wafers is formed between the first working disk and the counter-bearing element. In particular, when it is a double-sided or single-sided polishing machine, at least the first working disk, preferably also the counter-bearing element, or respectively the second working disk, can have a polishing bushing (polishing pad) on its surface delimiting the working gap. Furthermore, during machining, a polishing medium, such as a polishing agent, in particular a polishing liquid (abrasive slurry), can be introduced into the working gap in a manner known per se. The working disk may also be provided with tempering channels, through which a tempering liquid, for example cooling water, is conducted in order to temper the working disk during operation.

[0014] Double-sided or single-sided machine tools are particularly used for plane-parallel machining of flat workpieces. For machining, the workpiece can be accommodated in a floating manner in a known manner in a recess in a rotor disk disposed in the working gap. During operation, the first working disk and the counter-bearing element are driven in rotation relative to one another, for example, by corresponding drive shafts and at least one drive motor. It is possible that only one of the first working disk and the counter-bearing element is driven in rotation. However, both the first working disk and the counter-bearing element can also be driven in rotation, typically in opposite directions. For example, in the case of a double-sided machine tool, the rotor disk can also be moved in rotation through the working gap by a suitable kinematic system during relative rotation between the first working disk and the counter-bearing element, so that the workpiece disposed in the recess of the rotor disk describes a cycloidal path in the working gap. For example, the rotor disk can have teeth on its outer edge that mesh with corresponding teeth on a pin ring. Such machines form a so-called orbital kinematic system.

[0015] The first working disc and / or the counter-bearing element can each be held by a supporting disc. Similar to the first working disc and the counter-bearing element, the supporting disc can also be annular or have at least an annular support portion.

[0016] In a manner known per se, sensors, in particular suitable measuring devices, record measurement data related to the tool and / or processing parameters of a double-sided or single-sided machine tool during operation. These parameters may be, in particular, the tool and / or processing parameters mentioned at the outset. The sensors record the measurement data, in particular at defined intervals or continuously. The measurement data characterize the operating and tool parameters of the double-sided or single-sided machine tool and, therefore, the production process. The measurement data recorded by the sensors are also supplied, in particular, at defined intervals or continuously, to the control unit of the double-sided or single-sided machine tool. The tool and / or processing parameters can be recorded in real time. This also applies to the forwarding of the measurement data to the control unit and to the processing of the measurement data described below. The recorded measurement data can also be stored in a data memory and, for example, forwarded from the data memory to the control unit in real time or with a delay.

[0017] For processing measurement data, the control device according to the present invention includes an artificial neural network that creates a state vector for a double-sided or single-sided machine tool from the received measurement data. The state vector is formed from the current measurement data of the sensors, or is formed from the current measurement data. The state vector thus characterizes the double-sided or single-sided machine tool, in particular the current production process. The artificial neural network compares this state vector with at least one, preferably multiple, and in particular a set of target state vectors. During training, the method according to the present invention assigns a target state vector to the artificial neural network as a state vector for acceptable machining results during machining of workpieces on the double-sided or single-sided machine tool. The target state vector can be defined based on various objectives, for example, taking into account specific quality parameters (e.g., GBIR and / or SFQR) and / or production throughput or other parameters. By comparing the state vector created from the current measurement data with the target state vectors available to the artificial neural network, it can be determined whether the current state vector matches one of the target state vectors that have been trained as acceptable. If it is determined that the recorded state vector does not match any of the acceptable target state vectors, then preventive measures can be implemented, for example, if there are relevant deviations from the acceptable target state vectors. For example, it is possible to intervene in the production process by adjusting tool and / or processing parameters. Of course, defined tolerances can be specified for the comparison, within which identified minor deviations from the target state vectors are classified as acceptable. By adjusting tool and / or processing parameters based on the comparison, the production process can be influenced in such a way that the currently created state vector (again) matches at least one target state vector to a sufficient degree.

[0018] Unlike an operator, an artificial neural network can very quickly create a state vector and, therefore, tool and / or process parameters from a large amount of measurement data, and can also very quickly compare this state vector with at least one, preferably a large number, of target state vectors. Consequently, inadmissible deviations of a production process from an acceptable process can be quickly and reliably identified, particularly when fully trained or experienced personnel are not available at the production location of a double-sided or single-sided machine tool. The present invention exploits the fact that in an optimal production process, measurable tool and / or process parameters have a fixed relationship to one another. An artificial neural network trained using the tool and / or process parameters of the optimal process can therefore quickly and reliably identify deviations from the current process. The artificial neural network forms an anomaly detector that identifies inadmissible deviations (anomalies) in the production process. Consequently, even when a production process is triggered after an initial setup procedure, process optimization can be performed more quickly and with a significantly smaller number of test production processes. In the optimal scenario, only a single production test is required, which does not require external downstream measurement of the processed workpiece. The production process can be monitored more simply and quickly, while rejects are minimized even at the start of production. In particular, the production of workpieces that exceed the required tolerances in double-sided or single-sided machine tools can be reduced or, in the best case scenario, completely prevented.

[0019] According to one embodiment, the control device can be designed to issue a warning message if the generated state vector deviates from at least one target state vector. This warning message can be issued to the operator of the double-sided or single-sided machine tool, for example, via the machine tool's user interface. In the simplest case, the operator receives a warning message indicating that tool and / or machining parameters have deviated to an unacceptable degree from acceptable values ​​for an optimal production process. Based on this, the operator can manually intervene in the process, in particular by adjusting the tool and / or machining parameters in a targeted manner, so that the state vector generated from the current measurement data (again) corresponds to the at least one target state vector.

[0020] In another variation, the warning message may include an adjustment recommendation for adjusting specific tools and / or processing parameters. This adjustment recommendation may be issued by the control device based on adjustment rules stored in the control device. Such adjustment rules may have been previously created by an operator of the double-sided or single-sided machine tool. The operator can then evaluate the adjustment recommendation and, if necessary, implement it. Thus, by combining the determined deviation between the state vector value and at least one target state vector with the formalized causal relationship between the tool and / or processing parameters of the double-sided or single-sided machine tool, the control device can automatically create a recommendation for changing the tool and / or processing parameters.

[0021] According to another embodiment, the control device may further include a regulating device designed to control a double-sided or single-sided machine tool, in particular a tool and / or operating parameters of the double-sided or single-sided machine tool, in the event of a deviation between a created state vector and at least one target state vector, determined by comparison, so that the created state vector matches the at least one target state vector. In particular, the regulating device may control actuators for influencing the tool and / or machining parameters. Additional automation is achieved by means of the regulating device because the device autonomously controls the double-sided or single-sided machine tool based on the comparison performed so that the state vector created from the current measurement data (again) corresponds to at least one of the target state vectors. The regulating device may be integrated into the control device.

[0022] A control device can be designed to control tools and / or operating parameters of a double-sided or single-sided machine tool based on control rules stored in the control device. In particular, the control rules can specify specific control rules for the control device that are associated with specific determined deviations in the state vector. Furthermore, the control rules can be pre-created by an operator, for example. This allows for automated control based on control specifications pre-stored in the form of control rules, particularly without operator intervention.

[0023] According to another embodiment, an additional artificial neural network may be provided. The network is designed to evaluate measurement data related to tools and / or processing parameters using machine learning, and based on the evaluation, controls a double-sided or single-sided machine tool, in particular the tools and / or operating parameters of the double-sided or single-sided machine tool, and / or creates and / or modifies adjustment rules stored in the control device. This additional artificial neural network may be in addition to the first artificial neural network mentioned above that forms the anomaly detector. However, it is also conceivable that the additional artificial neural network is designed to be integrated with the artificial neural network mentioned above that forms the anomaly detector. The control device may also be integrated into the additional artificial neural network.

[0024] In particular, the additional artificial neural network provided may include a so-called learning classifier system (LCS), also known as an artificial intelligence system. This type of system is based on established if-then relationships and can modify tools and / or machining parameters of a double-sided or single-sided machine tool based on outliers (i.e., deviations between a current state vector and at least one target state vector detected by the (first) artificial neural network). The LCS generates output data from input data and rules. The control device may also include a memory storing previously acquired tool and / or machining parameters, including data related to the workpiece to be machined. This stored data can be made available to the artificial neural network, in particular the LCS, which takes this data into account when evaluating measurement data and resulting control data for a double-sided, double-sided or single-sided machine tool. An artificial neural network, preferably designed as an LCS, can identify the probability of deviations of workpiece machining results from specified target values, for example, characteristic values ​​such as GBIR and SFQR, early in the production process. On this basis, the artificial neural network can intervene early in the production process, or at the latest later in the production process, by controlling actuators such as specific tools and / or processing parameters, to prevent defective products. The artificial neural network can also be used to improve initially created adjustment rules, for example, based on the operator's further experience with the production process through machine learning. To this end, the artificial neural network can modify the adjustment rules stored in the control device. It is also conceivable that the adjustment rules are created using the artificial neural network and then optimized, if necessary, based on additional process data. With the above-mentioned embodiments, maximum automation of the production process is possible without requiring any operator intervention.

[0025] According to another embodiment, the limiting condition may be that the sensor includes a measuring device for measuring the working gap, in particular the shape and / or width of the working gap, more particularly the distance between the first working disk and the counter-bearing element; and / or for measuring the temperature of the first working disk and / or the counter-bearing element and / or other machine components of the double-sided or single-sided machine tool; and / or for measuring the temperature and / or flow rate of the processing agent supplied to the working gap for processing the workpiece; and / or for measuring the rotational speed of the first working disk and / or the counter-bearing element and / or a rotor disk rotatably mounted in the working gap; and / or for measuring the load between the first working disk and the counter-bearing element; and / or for measuring the rotational speed and / or torque and / or temperature of the rotary drive machine; and / or for measuring the pressure and / or force of a component for generating a deformation of the first working disk and / or the counter-bearing element; and / or for measuring the thickness of the working bushing of the first working disk and / or the counter-bearing element; and / or for measuring the thickness and / or shape of the workpiece processed in the double-sided or single-sided machine tool. The aforementioned metrology devices may be present together or in any desired combination. The processing agent may be, for example, a polishing agent, particularly a polishing liquid such as an abrasive slurry. The aforementioned metrology devices record tool and process parameters relevant to the production process of double-sided or single-sided machine tools, including, for example, environmental data.

[0026] If a double-sided or single-sided machine tool (in particular, a tool and / or processing parameters of a double-sided or single-sided machine tool) is controlled based on a deviation between a created state vector and at least one target state vector determined by means of a comparison, this may in particular include controlling actuators for influencing the working gap, in particular the shape and / or width of the working gap, more particularly the distance between the first working disk and the counter-bearing element; and / or for influencing the temperature of the first working disk and / or the counter-bearing element and / or other machine components of the double-sided or single-sided machine tool; and / or for influencing the temperature of a processing agent supplied to the working gap for processing the workpiece. The invention also provides a method for controlling the speed and / or flow rate of a first working disk and / or a counter-bearing element and / or a rotor disk rotatably mounted in a working gap; and / or for influencing the load between the first working disk and the counter-bearing element; and / or for influencing the speed and / or torque and / or temperature of a rotary drive; and / or for measuring the pressure and / or force of a component used to deform the first working disk and / or the counter-bearing element; and / or for influencing the thickness of a working bushing of the first working disk and / or the counter-bearing element; and / or for influencing the thickness and / or shape of a workpiece machined in a double-sided or single-sided machine tool. The aforementioned examples of influencing or separately controlling the actuators can occur together or in any desired combination. The processing agent can, for example, be a polishing agent, in particular a polishing liquid such as an abrasive slurry. The actuator to be controlled thus influences tool and processing parameters of the double-sided or single-sided machine tool that are relevant to the production process, including, for example, environmental data.

[0027] According to a further embodiment, it can be provided that the counter-bearing element is formed by a second, preferably annular, working disk, wherein the first and second working disks are arranged coaxially with respect to one another and can be driven in rotation relative to one another, wherein a working gap for double-sided or single-sided machining of flat workpieces is formed between the working disks.

[0028] The present invention also relates to a system comprising at least two double-sided or single-sided machine tools according to the present invention, wherein a higher-order artificial neural network connected to the artificial neural networks of the at least two double-sided or single-sided machine tools is provided, wherein the higher-order artificial neural network is designed to train at least one artificial neural network of the at least two double-sided or single-sided machine tools based on data obtained from the artificial neural networks of the at least two double-sided or single-sided machine tools by inputting a state vector that leads to an acceptable machining result for a flat workpiece.

[0029] In this embodiment, a system is provided that comprises at least two, and particularly more than two, double-sided or single-sided machine tools according to the present invention. Furthermore, a higher-order artificial neural network is provided that is connected to the artificial neural networks of the at least two double-sided or single-sided machine tools. The higher-order artificial neural network is designed to train the artificial neural networks of the at least two double-sided or single-sided machine tools based on data obtained from the artificial neural networks of the at least two double-sided or single-sided machine tools. The higher-order neural network thus forms a higher-order structure into which similar double-sided or single-sided machine tools can be integrated. Furthermore, a cross-plant memory that obtains data on similar double-sided or single-sided machine tools in the system and also provides this data to the higher-order artificial neural network can be provided. In this way, individual double-sided or single-sided machine tools in the system can be optimized by using their individual data, taking into account the data stored in the memory, as applicable. The aforementioned embodiments can achieve advantageous effects, for example, in production planning, fleet management, or predictive maintenance.

[0030] The double-sided or single-sided machine tool according to the present invention can be designed to carry out the method according to the present invention. Therefore, the method according to the present invention can be carried out using the double-sided or single-sided machine tool according to the present invention.

[0031] As explained, in the method according to the present invention, an artificial neural network is trained by inputting a large number of state vectors that result in acceptable machining results for flat workpieces. Training can occur during a production process performed by an operator using different tools and / or machining parameters on a double-sided or single-sided machine tool. Depending on the machining results, the artificial neural network is informed as to whether the production process has resulted in acceptable machining results for each tool and / or machining parameter. In this case, the associated tools and / or machining parameters are stored in the artificial neural network as target state vectors. This initial training typically occurs before regular machining of flat workpieces on a double-sided or single-sided machine tool begins.

[0032] Furthermore, the trained artificial neural network can be further trained during operation of the double-sided or single-sided machine tool by inputting additional target state vectors that result in acceptable machining results for a flat workpiece. As a result of this additional training performed during the production process using the double-sided or single-sided machine tool, tool and / or machining parameters are further optimized.

[0033] According to another embodiment, a trained artificial neural network is used to train an additional artificial neural network by inputting a large number of target state vectors that result in acceptable machining results for flat workpieces during operation of a double-sided or single-sided machine tool. This additional artificial neural network can be untrained or pre-trained. For example, the additional artificial neural network can be a copy of the trained artificial neural network and further trained based on it. This can be useful, for example, if the trained artificial neural network is a general neural network that was trained for a specific type of double-sided or single-sided machine tool but has not been specialized for a specific double-sided or single-sided machine tool, particularly with respect to the individual machining parameters at the site. Thus, specialized versions of the trained artificial neural network can be generated, which can ultimately replace the trained artificial neural network. One possible application scenario is that a double-sided or single-sided machine tool can be delivered using a trained artificial neural network. This training is based on test or laboratory data from the manufacturer of the double-sided or single-sided machine tool, respectively. This is followed by further specialization for the individual manufacturing processes at the user's end using an additional state-of-the-art neural network. This requires less understanding of the manufacturing processes at the installation site of the double-sided or single-sided machine tool. Simple diagram description

[0034] The following is a more detailed explanation of exemplary embodiments of the present invention using schematic diagrams, wherein: FIG1 shows a section of a double-sided or single-sided machine tool according to the invention in a first operating state. FIG2 shows the view from FIG1 in the second operating state. FIG3 shows the view from FIG1 in the third operating state. FIG4 is a schematic representation of the functions of a double-sided machine tool according to the present invention according to a first exemplary embodiment. FIG5 is a schematic diagram showing the functions of a double-sided machine tool according to another exemplary embodiment of the present invention. FIG6 is a schematic diagram showing the functions of a double-sided machine tool according to another exemplary embodiment of the present invention. FIG. 7 is a schematic diagram showing the functions of a double-sided machine tool according to another exemplary embodiment of the present invention. FIG8 is a schematic diagram showing the functions of a double-sided machine tool according to another exemplary embodiment of the present invention. FIG9 is a schematic representation of the functions of a double-sided machine tool according to another exemplary embodiment of the present invention, and [Fig. 10] A schematic diagram showing a system according to the present invention. Unless otherwise specified, like reference numerals refer to like objects in the drawings. Implementation Method

[0035] The double-sided machine tool, depicted merely as an example in Figures 1 to 3, has an annular upper support plate 10 and also an annular lower support plate 12. A first annular upper working plate 14 is fastened to the upper support plate 10, and a second annular lower working plate 16 is likewise fastened to the lower support plate 12. An annular working gap 18 is formed between the annular working plates 14, 16, in which a flat workpiece, such as a wafer, is processed on both sides during operation. The double-sided machine tool can be, for example, a polisher, grinder, or lapper.

[0036] The upper support disk 10 and upper working disk 14 and / or the lower support disk 12 and lower working disk 16 can be driven in rotation relative to one another by a suitable drive system comprising, for example, an upper drive shaft and / or a lower drive shaft and at least one drive motor. The drive system is known per se and will not be described in detail for the sake of clarity. In a manner also known per se, the workpiece to be machined can be held in a floating manner in the working gap 18 between the rotor disks. Suitable kinematics (e.g., planetary kinematics) can ensure that the rotor disks also rotate through the working gap 18 during the relative rotation of the support disks 10, 12 or working disks 14, 16, respectively. Temperature control channels can be provided in the upper working disk 14 or upper support disk 10, and possibly also in the lower working disk 16 or lower support disk 12, through which a temperature-controlled fluid, for example, a temperature-controlled liquid such as cooling water, can be conducted during operation. This is also known per se and not illustrated in greater detail.

[0037] The double-sided machine tool shown in Figures 1 to 3 further includes a distance measuring device, which is a sensor known per se. For example, the sensor can operate optically or electromagnetically (e.g., an eddy current sensor). In the illustrated example, three distance measuring devices 20, 22, and 24 are provided, which measure the distance between the upper work disk 14 and the lower work disk 16 at three radially spaced locations on the work disk 18, as indicated by the arrows in Figure 1. As can be seen, the measuring device 20 measures the distance between the upper work disk 14 and the lower work disk 16 in the region of the radially outer edge of the working gap 18. The distance measuring device 24 measures the distance between the upper work disk 14 and the lower work disk 16 in the region of the radially inner edge of the working gap 18. The distance measuring device 22 measures the distance between the upper work disk 14 and the lower work disk 16 in the middle of the working gap 18.

[0038] For reasons of clarity, the distance measuring devices 20, 22, 24 are not shown in FIGS. 2 and 3. The measurement data of the distance measuring devices 20, 22, 24 is applied to the control device 34.

[0039] In the current situation, the lower working disk 16 is only fastened to the lower support disk 12 in the area of its outer edge and in the area of its inner edge, for example, screwed along a partial circle in each case, as shown by reference symbols 26 and 28 in FIG. 1. In contrast, the lower working disk 16 is not fastened to the lower support disk 12 between these fastening positions 26 and 28. In fact, between these fastening positions 26 and 28, an annular pressure volume 30 is located between the lower support disk 12 and the lower working disk 16. The pressure volume 30 is connected via a dynamic pressure line 32 to a pressure fluid reservoir not shown in more detail in the figure, such as a liquid reservoir, especially a water reservoir. In the dynamic pressure line 32, a pump and a control valve can be configured, which can be controlled by the control device 34. In this way, the desired pressure acting on the lower working disk 16 can be accumulated in the pressure volume 30 by introducing a fluid into the pressure volume 30. The pressure existing in the pressure volume 30 can be measured by a pressure measuring device not shown in more detail. The measurement data of the pressure measuring device can also be applied to the control device 34, so that the control device 34 can set the specified pressure in the pressure volume 30.

[0040] Due to its freedom of movement between the fastening positions 26 and 28, the lower working disk 16 can be locally made into a convex shape by setting a sufficiently high pressure in the pressure volume 30, as indicated by the dotted line at reference symbol 36 in FIG. 2. If it is assumed that the pressure p0 in the pressure volume 30 is in the operating state where the lower working disk 16 in FIG. 1 has a flat shape, then the convex deformation of the lower working disk 16 shown at 36 in FIG. 2 can be achieved by setting a pressure of p1 > p0. On the other hand, the local concave deformation of the lower working disk 16 can be achieved by setting a pressure of p2 < p0 in the pressure volume 30, as illustrated by the dotted line at reference symbol 38 in FIG. 3.

[0041] In this situation, it can be seen that, when observed in the radial direction, the lower working disk 16 can be locally convex (FIG. 2) or respectively locally concave (FIG. 3) between its inner edges, in the area of the fastening position 26, and at its outer edge, in the area of the fastening position 28.

[0042] In addition to this localized radial deformation of the lower working disk 16, means for overall deformation of the upper working disk 14 can also be provided. These means can be designed as described above or, respectively, as described in DE 10 2006 037 490 B4. The upper support disk 10 and the upper working disk 14 fastened thereto are deformed as a whole so that an overall concave or convex shape of the working surface of the upper working disk 14 is produced over the entire cross section of the upper working disk 14. In contrast, the upper working disk 14 can remain flat between its radially inner edge and its radially outer edge or be locally deformed in the manner described above by means of the pressure volume 30. The means for adjusting the shape of the upper working disk 14 can also be controlled by means of the control device 34.

[0043] The distance measuring devices 20, 22, and 24 form sensors that record measurement data related to the tool and / or processing parameters of the double-sided machine tool. In the present case, they particularly record the thickness and geometry of the working gap 18 during operation of the double-sided machine tool. Preferably, the double-sided machine tool includes a plurality of additional sensors with corresponding additional measuring devices. These measuring devices can particularly be measuring devices of the type described above. These measuring devices record additional tool and / or processing parameters during operation of the double-sided machine tool.

[0044] The measurement data recorded by the sensors are fed to the control device 34. From these measurement data, the control device 34 uses an artificial neural network 34 integrated in the device to create a state vector of the double-sided machine tool and compares this state vector with at least one target state vector, preferably a set of target state vectors assigned to acceptable production processes within a training range.

[0045] The training of the artificial neural network 34 will be explained in more detail based on FIG. A double-sided machine tool according to the present invention is shown in FIG. 4 by reference numeral 40. An unprocessed workpiece 42, such as an unprocessed wafer, is supplied to the machine tool for processing, and a processed workpiece 44, in particular a processed wafer 44, is output by the double-sided machine tool 40. A data memory 46 is provided, for example, to which measurement data related to tool and processing parameters recorded by sensors is supplied, where this data is made available to an operator 48, as shown at 50 in FIG. Furthermore, measurement data related to the geometry of the processed workpiece, for example, is fed into the data memory 46 as additional tool and / or operating parameters, where this data is also fed to the operator 48, as shown at 52 in FIG. Finally, external environmental data may also be made available to the data memory, as shown at 54. This external environmental data can also be fed to the operator 48. Based on this, the operator 48 evaluates the production process based on the corresponding data to determine whether the processing results are acceptable. The operator 48 makes this assessment available to the artificial neural network 34 that controls the device 34, as shown at 56 in Figure 4. The corresponding state vector is stored by the artificial neural network 34 as a target state vector.

[0046] FIG5 illustrates how a double-sided machine tool can be operated on this basis. In this case, program data related to the tool and / or processing parameters can be directly supplied to the artificial neural network 34 of the control device 34, as indicated by 58 in FIG5 . The artificial neural network 34 creates a state vector from the acquired measurement data related to the tool and / or processing parameters and compares this state vector with a stored target state vector. If an inadmissible deviation or discrepancy is detected, the control device 34 issues a corresponding warning message to the operator 48, as indicated by 60 in FIG5 . Based on this, and if applicable, taking into account the measurement data of the processed workpiece 44 available via 52 , the operator 48 can control the double-sided machine tool 40, in particular the actuators for influencing the tool and / or operating parameters, as indicated by 62 in FIG5 , so that the continuously monitored and created state vector is consistent with at least one target state vector. In this case, the operator 48 thus determines the outcome of the evaluation of the received data. The operator 48 is supported by the control device 34 as an anomaly detector.

[0047] FIG6 shows another automated variant of the procedure shown in FIG5 . In this embodiment, the control device 34 , in particular its artificial neural network 34 , further includes a connected regulating device 64 , as indicated by 66 in FIG6 . If a deviation between a generated state vector and at least one target state vector is detected by comparison, regulating intervention by the regulating device 64 in the actuators of the double-sided machine tool 40 occurs, particularly without intervention by the operator 48 . This results in adjustments to the recorded tool and / or processing parameters of the double-sided machine tool 40 , as indicated by 68 in FIG6 . All relevant data can be stored in the data memory 46 . For example, the integrated control and regulating devices 34 , 36 can control the tool and / or operating parameters of the double-sided machine tool 40 based on regulation rules, for example, stored in the regulating device 64 . The regulation rules can, for example, contain specific control instructions for specific established deviations from the tool and / or processing parameters generated by the operator 48 , according to which the control and regulating devices 34 , 64 control the actuators.

[0048] FIG7 shows another embodiment of the process explained with reference to FIG6 . Operator 48 also participates in this embodiment. The operator also receives process data related to the recorded tool and processing parameters, as indicated at 70 in FIG7 , and process data related to the processed workpiece, as indicated at 52 . Finally, operator 48 also receives control commands executed by control device 34 , as indicated at 72 in FIG7 . Based on this, operator 48 can monitor the respective adjustments being performed and, if applicable, adjust the adjustments of control device 64 in an appropriate manner, as indicated at 74 in FIG7 .

[0049] FIG8 presents another example of a possible training of an artificial neural network as an anomaly detector. Here, training begins with the control device 34, utilizing a pre-trained artificial neural network 34, as explained above with reference to FIG4 . The control device 34 transmits any deviation or anomaly data to the operator 48, as shown at 60 in FIG8 and explained above with reference to FIG5 . Based on this, the operator 48 can train an additional artificial neural network 76, which is fed (additional) target state vectors for acceptable tool and / or machining parameters for the double-sided machine tool 40 during operation of the double-sided machine tool 40, as shown at 78 in FIG8 . The additional artificial neural network 76 can be an untrained artificial neural network 76. However, it can also be a pre-trained artificial neural network 76, for example, a copy of the neural network 34 of the control device 34. On this basis, specialized training of the additional artificial neural network 76 can be performed at the start of production operation for the respective individual program parameters of the application scenario of the double-sided machine tool 40, based on, for example, the artificial neural network 34 of the control device 34 (pre-)trained for a generic type of double-sided machine tool 40. After this training, the additional artificial neural network 76 can replace the previously trained artificial neural network 34 of the control device 34.

[0050] Another embodiment of the present invention will be explained based on Figures 9 and 10 . It specifically includes an additional artificial neural network 86 designed for machine learning. This may be a learning classification system (LCS), or an artificial intelligence system. In the embodiment shown in Figure 9 , sensor measurement data related to tool and / or process parameters is supplied to a data memory 46 and, as shown at 80 in Figure 9 , to a control device 34 . Workpiece data, particularly measurement data related to the geometry of the processed workpiece, is also supplied to both the data memory 46 and the control device 34 , as shown at 82 in Figure 9 . The control device 34 also exchanges data with the data memory 46 , as shown at 84 in Figure 9 . Figure 9 shows an additional artificial neural network 86 associated with the control device 34 . This additional artificial neural network 86 may also be combined with the artificial neural network 34 of the control device 34 . This additional artificial neural network 86 is designed for machine learning and, in particular, forms a learning classification system (LCS).

[0051] Measurement data related to the geometry of the machined workpiece 44 is also fed to the LCS 86 via 82. If an inadmissible deviation between the currently recorded state vector and the acceptable values ​​of the tool and / or processing parameters stored as target state vectors is detected by the control device 34, in particular its artificial neural network 34, during operation of the double-sided machine tool 40, a corresponding anomaly signal is output to the LCS 86, as shown at 88 in FIG. 9 . Past measurement data is also available to the LCS 86 from the data memory 46. Based on this, the LCS 86 can autonomously make decisions regarding changes to specific process parameters, in particular for controlling actuators that influence the recorded tool and / or processing parameters, and thus control the actuators, as shown at 90 in FIG. 9 . In this way, maximum automation and autonomy can be achieved.

[0052] FIG10 shows another embodiment of a variation of the one shown in FIG9 . Specifically, FIG10 shows a system according to the present invention comprising at least two double-sided machine tools 40 . Of course, the system may also comprise more than two double-sided machine tools 40 , as illustrated by three dots in FIG10 . In FIG10 , for illustrative purposes, two devices 92 i are shown as dashed blocks, which may correspond in design and functionality to the embodiment shown in FIG9 . Devices 92 i may also be designed differently, for example, to achieve different goals, such as optimal wafer quality or maximum output. These devices are connected to a common data memory 46 via 80 and 84 . Furthermore, a higher-level artificial neural network 94 , which also includes an LCS and can be linked to an operator 48 , is provided in the system shown in FIG10 . The higher-level artificial neural network 94 is also connected to the data memory 46 , as indicated by 96 . Furthermore, the higher-level artificial neural network 96 receives control commands, which are executed by the LCS 86 in each case, as indicated by 98 in FIG10 . On this basis, the higher-level LCS 94 may further optimize or specialize the LCS 86 of the device 92i, for example by specifying collective or individual control rules and / or target state vectors for the individual device 92i.

[0053] 10: Upper support plate 12: Lower support plate 14: Upper working plate 16: Upper working plate 18: Work break 20: Distance measuring equipment 22: Distance measuring equipment 24: Distance measuring equipment 26: Fastening position 28: Fastening position 30: Pressure Volume 32: Dynamic pressure line 34: Control Equipment / Artificial Neural Network 36: Arrow 38: Arrow 40: Double-sided machine tool 42: Unprocessed workpiece 44: Processed workpiece 46:Data Memory 48: Operator 50: Arrow 52: Arrow 54: Arrow 56: Arrow 58: Arrow 60: Arrow 62: Arrow 64: Adjustment equipment 66: Arrow 68: Arrow 70: Arrow 72: Arrow 74: Arrow 76: Artificial Neural Network 78: Arrow 80: Arrow 82: Arrow 84: Arrow 86: Artificial Neural Network 88: Arrow 90: Arrow 92i:Equipment 94: Artificial Neural Network 96: Arrow 98: Arrow

Claims

1. A double-sided or single-sided machine tool comprising a first working disc, preferably annular, and a reverse bearing element, wherein the first working disc and the reverse bearing element are rotatably driven relative to each other by means of a rotary drive, and wherein a preferred annular working clearance (18) for double-sided or single-sided machining of a flat workpiece (42, 44) is formed between the first working disc and the reverse bearing element, wherein the double-sided or single-sided machine tool includes recording data during operation of the double-sided or single-sided machine tool. A plurality of sensors for measuring data relating to tools and / or machining parameters of a double-sided or single-sided machine tool, characterized by providing a control device (34) that acquires the measurement data recorded by the sensors, the sensors including a measuring device for measuring the thickness and / or shape of a workpiece (44) machined in the double-sided or single-sided machine tool, and providing an artificial neural network that creates a state vector of the double-sided or single-sided machine tool from the measurement data and compares the state vector with at least one target state vector, wherein the control device (34) further includes an adjustment device (64) designed to control the double-sided or single-sided machine tool, in particular the tools and / or operating parameters of the double-sided or single-sided machine tool, based on an adjustment rule stored in the adjustment device (64), in the presence of a deviation between the created state vector and the at least one target state vector by means of the comparison determination, such that the created state vector matches the at least one target state vector.

2. A double-sided or single-sided machine tool as requested in item 1, wherein the control device (34) is designed to issue a warning message in the event of a deviation between the created state vector and one of the at least one target state vector.

3. A double-sided or single-sided machine tool as requested in item 2, wherein the adjustment device (64) is integrated into the control device (34).

4. A double-sided or single-sided machine tool as requested in item 1 or 2, wherein an additional artificial neural network (86) is provided, the additional artificial neural network being designed to evaluate the measurement data relating to the tools and / or processing parameters by means of machine learning, and to control the double-sided or single-sided machine tool, in particular the tool and / or operating parameters of the double-sided or single-sided machine tool, and / or to create and / or modify an adjustment rule stored in an adjustment device (64) based on the evaluation.

5. A double-sided or single-sided machine tool as requested in item 4, wherein the adjustment device (64) is integrated into the additional artificial neural network (86).

6. A double-sided or single-sided machine tool as claimed in item 1 or 2, wherein the measuring equipment is used to measure the working clearance (18), in particular a distance between the first working disc and the reverse bearing element; and / or to measure a temperature of the first working disc and / or the reverse bearing element and / or other machine components of the double-sided or single-sided machine tool; and / or to measure a temperature and / or a flow rate of a processing fluid supplied to the working clearance (18) for machining the workpieces (42, 44); and / or to measure the temperature of the first working disc and / or the reverse bearing element and / or other machine components of the double-sided or single-sided machine tool. The rotational speed of the disk and / or the reverse bearing element and / or the rotor disk rotatably mounted in the working gap (18); and / or for measuring a load between the first working disk and the reverse bearing element; and / or for measuring a rotational speed and / or a torque and / or a temperature of the rotary drive; and / or for measuring a pressure and / or a force of a component that causes deformation of the first working disk and / or the reverse bearing element; and / or for measuring the thickness of a working bushing of the first working disk and / or the reverse bearing element.

7. A double-sided or single-sided machine tool as claimed in claim 1 or 2, wherein the reverse bearing element is formed by a second working disc, preferably annular, wherein the first working disc and the second working disc are coaxially arranged and rotatably driven relative to each other, wherein the working gap (18) for double-sided or single-sided machining of flat workpieces (42, 44) is formed between the working discs.

8. A double-sided or single-sided machine tool as requested in item 1 or 2, wherein it is designed to perform a method in which the artificial neural network is trained by inputting a large number of target state vectors that result in an acceptable machining outcome of a flat workpiece (42, 44).

9. A double-sided or single-sided machine tool system comprising at least two double-sided or single-sided machine tools as claimed in any one of claims 1 to 8, wherein a higher-order artificial neural network (94) is provided connected to the artificial neural networks of the at least two double-sided or single-sided machine tools, wherein the higher-order artificial neural network (94) is designed to train at least one artificial neural network of the at least two double-sided or single-sided machine tools based on data obtained from the artificial neural networks of the at least two double-sided or single-sided machine tools by means of a state vector resulting in an acceptable machining outcome of a flat workpiece (42, 44).

10. A method for operating a double-sided or single-sided machine tool as claimed in any of items 1 to 7, wherein the artificial neural network is trained by inputting a large number of target state vectors that result in an acceptable machining outcome of one of the flat workpieces (42, 44).

11. The method of claim 10, wherein the trained artificial neural network is further trained during operation of the double-sided or single-sided machine tool by inputting an additional target state vector that results in an acceptable machining outcome of a flat workpiece (42, 44).

12. The method of claim 10 or 11, wherein an additional artificial neural network (76) is trained during operation of the double-sided or single-sided machine tool by means of a large number of target state vectors that result in an acceptable machining outcome of a flat workpiece (42, 44).

Images