METHOD AND ARRANGEMENT FOR TESTING THE FUNCTIONALITY OF A COORDINATE MEASURING DEVICE

Description

[0001] The invention relates to a method and an arrangement for testing the functionality of a coordinate measuring machine.

[0002] The term coordinate measuring machine covers all types of devices that can be used to determine the coordinates of workpieces. In one class of coordinate measuring machines, the coordinates are surface coordinates, i.e., the coordinates of surface points of workpieces are determined. Another class of coordinate measuring machines is alternatively or additionally capable of determining coordinates inside workpieces. These include coordinate measuring machines that use invasive radiation that penetrates the material of the workpiece and, in particular, measure the intensity of the radiation passing through the workpiece. Typically, the workpiece is irradiated from different directions, and based on the results of the irradiation, a reconstruction of the scanned workpiece, particularly a computer-aided one, is carried out. Such processes are also known as computed tomography (CT).The term coordinate measuring machine also includes classic coordinate measuring machines, such as portal or gantry-style devices, horizontal arm machines, and articulated arm machines. The term coordinate measuring machine also covers machines that are not primarily designed as coordinate measuring machines, but are set up to operate like a coordinate measuring machine. In particular, these machines have at least one measuring sensor used to determine the coordinates. Known examples include robots, such as robot arms with swivel joints, to which a sensor for detecting the workpiece surface (e.g., a fringe projection sensor) is attached instead of a tool, or machine tools to which a measuring sensor (e.g., a tactile sensor) is attached instead of a machining tool or in addition to a machining tool.Hexapod mechanisms are also known, for example, to which a sensor for detecting the workpiece surface (for example a tactile sensor) is attached instead of a machining tool.

[0003] The invention is also not limited with regard to the types of sensors used by a coordinate measuring machine to determine the coordinates. Tactile sensors have already been mentioned as an example, which can be, for example, switching or measuring types. Tactile sensors can, in particular, be passive or active sensors. Active sensors can be designed to generate a contact force with which a tactile probe contacts the surface of a workpiece to be measured. Optical sensors are often used alternatively or in addition. There are also, for example, capacitive sensors and inductive sensors.

[0004] Coordinate measuring machines can change their properties, particularly due to wear and tear or external influences, but also due to changes in environmental conditions (such as ambient temperature). This can reduce the measurement accuracy, especially when measuring the coordinates of workpieces.

[0005] It is therefore beneficial to periodically check the functionality of a coordinate measuring machine. In the field, such a functional test is often referred to as a health check. If the device is fully functional and, in particular, complies with the underlying specifications regarding measurement accuracy, it can be considered "healthy" or fully functional.

[0006] It is possible for an experienced service technician to perform the functional test on-site and / or evaluate data sets obtained during the functional test, thus determining a functional test result. The service technician can not only determine whether the coordinate measuring machine is fully functional with regard to the desired measurement accuracy and / or the intended measuring task. In many cases, they can also determine whether a deterioration in functionality is to be expected in the near future. If necessary, the coordinate measuring machine can then be serviced or parts of the machine can be replaced.

[0007] The data sets obtained during the functional test are, in particular, data sets relating to the hardware of the coordinate measuring machine and can be represented, in particular, as two-dimensional or three-dimensional diagrams. For example, a suitable physical quantity, such as the motor current of a drive or a position determined by the position measuring system of the coordinate measuring machine, is measured repeatedly and / or continuously. In particular, the physical quantity can alternatively or additionally be a deviation (also called lag) of a target position from an actual position of a movable part of the coordinate measuring machine driven by a drive, an electrical voltage, in particular of a drive for moving a movable part of the coordinate measuring machine, a speed of a movable part and an acceleration of a movable part.When this description refers to the position measuring system of the coordinate measuring machine, this may refer to the only position measuring system. Alternatively, it may also refer to one of several position measuring systems, the position measurement values ​​of which are then received and further processed, for example, by a control device of the coordinate measuring machine. A second position measuring system may, in particular, be the measuring system of a measuring head, for example a measuring head on which at least one probe for tactile probing of a workpiece is arranged or can be arranged. Such measuring heads may, in particular, be active measuring heads and thus have at least one force generator that exerts a probing force on the surface of the workpiece being contacted via the tactile probe. Such measuring heads can also be subject to functional testing.

[0008] In particular, temporal profiles of the physical quantity and / or respective measured values ​​of the physical quantity can be recorded during the repeated execution of a movement sequence by the coordinate measuring machine and / or by an additional measuring device. It is also possible to subject the coordinate measuring machine to a disturbance that is undesirable or does not occur during normal operation of the coordinate measuring machine for functional testing, such as applying a mechanical shock or applying a voltage pulse to a drive motor. During this time and / or in the time after the disturbance has occurred, the profile of at least one physical quantity can be recorded to capture the coordinate measuring machine's response. This is referred to in the field as a "step response" or "jump test."

[0009] The at least two-dimensional diagram then shows the values ​​of the respective physical quantity as a function of an order quantity, which, for example, is again a physical quantity such as the time during a single functional test and / or across repeated functional tests and such as the position(s) determined by the position measuring system, or which is a non-physical quantity, for example an index quantity for distinguishing the repeatedly acquired values ​​of the physical quantity. In the present invention, the functional test can also be carried out as described above, the aforementioned dependencies can occur, and / or the acquired data sets can be or are represented in an at least two-dimensional diagram as described.

[0010] The service technician mentioned above or other experts can evaluate the data records obtained during the functional test and obtain a statement regarding the functionality of the coordinate measuring machine. These data records, each of which is assigned such a statement, can be stored for comparison, for example in paper form or as computer-readable digital data, for example in a digital data storage device or distributed data storage device (i.e. distributed across multiple storage media). Computer-generated reports on the functional test can, for example, including the diagrams and / or the associated statements regarding functionality, be saved as electronic documents, for example in the JavaScript Object Notation (JSON) data format.

[0011] While expert knowledge is highly valued, each expert makes the aforementioned statement regarding functional capability based on their own experience and skills. In rare cases, an expert may overlook or mistakenly identify indications of nonexistent or declining functional capability.

[0012] Furthermore, the evaluation of the obtained data sets by experts takes a considerable amount of time.

[0013] For certain data, numerical methods can be used for evaluation, such as mean value filters and subsequent limit value analysis, in order to arrive at clear statements.

[0014] However, this usually does not apply to data where a specific, usually temporal, data pattern is expected within the data sets, which is intended to provide a statement about the quality of the machine settings. One possibility is to evaluate the correlation of several data sets, including their temporal patterns, in order to be able to make a statement about the quality. In many cases, this is no longer possible using simple numerical methods. The evaluation is therefore often based on the experience of the evaluator.

[0015] DE 10 219 217 740 B3 describes a method and system for checking the condition of a coordinate measuring machine. This can include checking its functionality. A movement of the machine is initiated, and its functionality is tested.

[0016] ARENHART RODRIGO SCHONS ET AL: "Devices for Interim Check of Coordinate Measuring Machines: A Systematic Review", MAPAN, SPRINGER INDIA, Vol. 36, No. 1, 18 January 2021, Pages 157-173, XP037458000, ISSN: 0970-3950, DOI: 10.1007 / S12647-020-00406-0 deals with the reliability of measurements from coordinate measuring machines.

[0017] It is an object of the present invention to provide a method and a device for testing the functionality of a coordinate measuring machine, which enable an improved evaluation, in particular assessment, of a test of the functionality of a coordinate measuring machine. This object is achieved by a method and an arrangement according to the independent claims.

[0018] It is proposed to use artificial intelligence (AI) to evaluate the data sets obtained during functional testing. In particular, the AI ​​can be trained using training data sets, each of which is assigned additional information regarding the functionality of a coordinate measuring machine. For example, the training data sets are the data sets already mentioned above. The additional information assigned to the training data sets therefore each contains at least one piece of information about the functionality of a coordinate measuring machine that can be determined from the respective training data set and optionally also contains information about a possible cause for impaired functionality.

[0019] For AI to evaluate a test dataset, the coordinate measuring machine for which the test dataset was obtained must also be the coordinate measuring machine for which the training datasets were obtained, and / or at least one technically comparable coordinate measuring machine, for example, of the same or similar type, for which the training datasets were obtained. The training datasets can therefore, for example, have been obtained both for the same coordinate measuring machine as the test dataset and for at least one technically comparable coordinate measuring machine. However, they can also have been obtained exclusively for the same coordinate measuring machine or exclusively for at least one technically comparable coordinate measuring machine.

[0020] In order to generate at least one test data set for evaluation, the coordinate measuring machine is operated and operating signals are automatically recorded. The operating signals are generated by at least one first generating device, which is part of the coordinate measuring machine or is an additional generating device. The operating signals correspond to values ​​of at least one physical quantity that is characteristic of the operation of the first generating device and / or the coordinate measuring machine. Examples of operating signals are the signals generated by a position measuring system of the coordinate measuring machine or generated and / or received by a drive device with a drive control or a drive control of a drive of the coordinate measuring machine, such as a motor current or a motor speed.An example of an additional generating device is a laser interferometer, which records the position of a moving part of the coordinate measuring machine over time. In this case, the corresponding operating signals are generated within the laser interferometer, for example, interference signals corresponding to the interference of two reflected laser beams.

[0021] The test data set to be evaluated is automatically generated by a second generation device from the acquired operating signals, for example by at least one computer connected to the at least one first generation device. The at least one computer can be a computer that is part of the coordinate measuring machine and / or part of a control system of the coordinate measuring machine and, for example, controls operation during normal operation of the coordinate measuring machine. Alternatively or additionally, it can be, for example, a computer that is connected to the at least one first generation device, or it can be, for example, several computers that are connected to one another and to the at least one first generation device via connections for transmitting computer-readable signals.The first generating device itself can also be designed to generate the test data set from the recorded operating signals (and in this case also be the second generating device) or at least to contribute to this.

[0022] Furthermore, the at least one test data set to be evaluated is automatically transferred to an evaluation device. This evaluation device incorporates artificial intelligence (AI). Although this is possible, the evaluation device does not have to be implemented by a single computer, but can, in particular, be a distributed and / or networked computer system.

[0023] Connections for transmitting computer-readable signals between multiple computers, possibly also from and to the at least one generating device for generating the operating signals, and / or the parts of the computer system, can be local connections and / or remote connections (for example, connections between parts of data transmission networks such as the Internet). In any case, it is preferred that at least a part of the evaluation device, through which the AI ​​is at least partially implemented, is connected via a remote connection to the at least one generating device that generates the operating signals. This connection can be implemented indirectly via a device or system that generates the test data set from the acquired operating signals.

[0024] According to the invention, it is therefore proposed: A method for testing the functionality of a coordinate measuring machine, comprising the following steps: Operating the coordinate measuring machine and automatically recording operating signals which are generated by at least one device of the coordinate measuring machine and / or an additional device during operation and which correspond to the values ​​of at least one physical quantity which is characteristic of the operation of the device and / or the coordinate measuring machine, automatically generating a test data set from the recorded operating signals and automatically transmitting the test data set to an evaluation device which has an artificial intelligence, evaluating the test data set by means of the artificial intelligence by the evaluation device, wherein the artificial intelligence is based on training using training data sets with associated additional information,which each contain at least one piece of information about the functionality of the coordinate measuring machine or a comparable coordinate measuring machine that can be determined from the respective training data set, is in an operational state and wherein the artificial intelligence derives a statement regarding the functionality of the coordinate measuring machine from the test data set, outputting the statement regarding the functionality of the coordinate measuring machine. ,

[0025] According to the invention, it is further proposed: An arrangement for testing the functionality of a coordinate measuring machine, comprising: at least one first generating device which is designed to automatically generate operating signals during operation of the coordinate measuring machine, wherein the operating signals correspond to values ​​of at least one physical quantity which is characteristic of the operation of the first generating device and / or the coordinate measuring machine, a second generating device which is the first generating device or another device and which is designed to automatically generate a test data set from the detected operating signals and to automatically transmit the test data set to an evaluation device, the evaluation device which has an artificial intelligence, wherein the evaluation device is designed to evaluate the test data set by means of the artificial intelligence, wherein the artificial intelligence is in an operational state on the basis of training using training data sets with associated additional information, each of which contains at least one piece of information about a functionality of the coordinate measuring machine or a comparable coordinate measuring machine that can be determined from the respective training data set, and wherein the artificial intelligence is designed to derive a statement regarding the functionality of the coordinate measuring machine from the test data set.

[0026] Embodiments of the arrangement emerge from the description of embodiments of the method. For example, the arrangement comprises a training device configured to train the artificial intelligence using the training data sets with the associated additional information. This speaks of an embodiment of the method in which the training of the artificial intelligence using the training data sets with the associated additional information is part of the process.

[0027] The training device is, for example, a computer or a set of computers with computer software that controls the execution of the training. The computer, the set of computers, or one of the computers in the set can also incorporate artificial intelligence, which evaluates the test data and derives a statement regarding the functionality of the coordinate measuring machine from the test data.

[0028] The artificial intelligence derives a statement regarding the functionality of the coordinate measuring machine from the test data set. The AI ​​uses the results of its training to do this. Therefore, the training data sets, along with the associated additional information, are used to assign the statement regarding functionality to the test data set.

[0029] The use of artificial intelligence trained on training data sets has the advantage of eliminating subjective results based on the personal experience of human experts when evaluating test data sets. Furthermore, evaluation time is saved, allowing experts to focus on other tasks. This does not preclude at least one expert from reviewing the AI's results and / or monitoring its operation. Furthermore, since the availability of test data sets for evaluation is greater when using AI, functionality testing can be performed more frequently and, for example, scheduled in advance.If the evaluated test data sets are stored, preferably with the statement regarding functionality derived from their evaluation, particularly in a computer-readable format, a comparison can also be made with the condition of the coordinate measuring machine at the time of the previously acquired test data sets when evaluating at least one test data set obtained through a new functional test. This, in turn, makes it possible to make predictions about future changes in the coordinate measuring machine's functionality. For example, the previous temporal progression of functionality can be extrapolated into the future.

[0030] The following describes examples of a "first" device of a coordinate measuring machine that generates operating signals during operation of the coordinate measuring machine. The corresponding physical quantities that are characteristic of the device and / or coordinate measuring machine's operation are also mentioned.

[0031] A first example concerns coordinate measuring machines with a tactile sensor or an optical sensor that optically detects a point-like or near-point-like area of ​​the workpiece to be measured. Such optical sensors include confocal sensors, which measure the distance to the surface area and radiate a spectrum of measuring radiation onto the surface area. Due to optical dispersion, the sensor's focal distance depends on the wavelength or frequency of the measuring radiation. Therefore, the spectral component of the measuring radiation with the greatest intensity is reflected back to the sensor, whose focal distance is equal to the current distance of the sensor from the irradiated surface area.

[0032] With the aforementioned tactile or optical sensors, it is important that the position of the point-like or near-point-like surface area detected by the sensor is accurately recorded. This is generally not the case with coordinate measuring machines, which are not capable of reproducibly probing the surface area with the sensor. It is also often desired to be able to repeat an optical or tactile probing of a specific surface area or to precisely probe a measuring location on the surface of a workpiece. In this case, too, the reproducibility and accuracy of the probing is crucial.

[0033] It is therefore proposed, in particular, to determine the probing reproducibility (often referred to as probing error) of the coordinate measuring machine under test. In this case, the position measuring system of the coordinate measuring machine is used as the device that generates position values. The physical quantity is therefore the position of a movable part of the coordinate measuring machine, the movement of which moves the sensor or which is the sensor, or the physical quantity is the relative position of the sensor or of a part of the coordinate measuring machine connected to the sensor to another object, such as the workpiece to be measured. The quantity to be evaluated can then be the aforementioned probing error, which is the deviation between a target value and an actual value.The actual value is obtained from the position measuring system, while the target value is obtained, for example, from the motion control of the coordinate measuring machine, or more generally from a specification for the movement of the respective part to be moved.

[0034] The probing deviation can be determined, for example, by probing a plurality of surface points of a workpiece, and in particular of a test standard, at predetermined distances from one another, for example, along a straight line. The distances between the probing points (i.e., the target values) can be predetermined. It is also possible, after probing a plurality of probing points with an initial distance between each two points, to increase the respective distance between subsequent points.

[0035] Alternatively or additionally, the same measuring point on a workpiece can be repeatedly contacted, in particular by having the sensor reach the detection location via different paths. In the case of a tactile sensor, the detection location is, in particular, the location of the measuring point. In the case of an optical sensor, the detection location is the location from which the sensor detects the surface of the workpiece with the measuring point.

[0036] As a result of the functional test, it can be determined, in particular, whether at least one of the probing deviations is too large. Alternatively, it can be determined whether a specified maximum probing deviation has been reached and / or exceeded for more than a specified number or for more than a specified percentage of surface points probed. For example, a corresponding limit value can be specified in each case. Furthermore, alternatively or additionally, it can be determined whether a trend in the probing deviations of the successively probed surface points indicates that the maximum probing deviation will be reached in the near future when the functional test is repeated.

[0037] The probing error can be represented as a two-dimensional diagram by plotting the probing error of the individual probed surface points against the ordinal number of the probed surface points. For example, the first probed surface point has the ordinal number one, etc. Alternatively, the probing error can be plotted against the position of the surface points along a spatial direction. For example, with constant target distances between any two consecutive surface points in this spatial direction, this results in probing errors plotted at constant intervals. To determine the probing behavior when using an active measuring head with a tactile probe, the following procedure can be used: A surface point of a workpiece, in particular a test standard, is probed, i.e.The tactile contact between the probe and the surface point is established by moving the probe towards the surface of the workpiece. During the process, and in particular from the time tactile contact is established, the measured values ​​of the position measuring system of the measuring head, with which the deflection of the probe relative to a reference point of the measuring head is measured, are continuously recorded. Such measuring heads with tactile probes have oscillating systems, in particular due to mechanical suspension and coils for generating the measuring force, but also due to the elasticity of the probe itself. Therefore, from the time tactile contact is established, a deflection of the measured values ​​of the position measuring system is observed, which decays over time in the manner of an oscillation and approaches a constant value. The result of the functional test can be the magnitude of the deflection and / or the time until the oscillation decays, i.e.until a constant value is reached. In particular, it can be checked whether the deflection is too high and / or the time until the oscillation decays is too long. For example, corresponding limit values ​​are specified that must not be exceeded to pass the functional test.

[0038] As a two-dimensional diagram, the probing behavior, in particular as described, can be represented for each measured value of the position measuring system of the measuring head as a function of time during the described probing process.

[0039] For varying probing distances (e.g., the sensor travels a linear path before reaching the contact point), the deviation of each contact point can be plotted in the diagram as a function of the respective path. In addition to the quality of the probing behavior—especially the scatter—the resulting diagram also shows other characteristics of the system, such as natural frequencies that are excited at certain distances, or the effect of possibly poorly adjusted damping of the active sensor in the form of increased scatter at certain probing distances.

[0040] According to the claims, the test data set and the training data sets comprise the values ​​of the at least one physical quantity as a function of an order quantity (or in the case of a more than two-dimensional dependency on more than one order quantity). An order quantity is understood in particular to mean that the values ​​of the physical

[0041] Quantity(s) can be ordered according to the order of the assigned values ​​of the ordinal quantity and, in particular, can be represented in a diagram according to this order. An example of a diagram is a two-dimensional diagram in which the values ​​of the ordinal quantity are plotted along a first axis (e.g., the X-axis) and the values ​​of the respective physical quantity are plotted along a second axis (e.g., the Y-axis).

[0042] Such a dependency generally makes it possible to represent the values ​​of the physical quantity(s) in a corresponding diagram. Such diagram creation can be optional and can lead to the diagram being output, for example, on a screen or via a printer.

[0043] In any case, it is provided according to the invention that the artificial intelligence, when evaluating the test data set on the basis of its training, takes into account diagram features of at least one diagram of the dependence of the physical quantity on an order quantity of the test data set a) determines a correspondence or similarity between the test data set and an examination situation recorded in the training data sets or b) determines that there is no correspondence or similarity between the test data set and an examination situation recorded in the training data sets.

[0044] In case a), the AI ​​derives the statement regarding the functionality of the coordinate measuring machine according to the corresponding or similar test situation recorded in the training data sets.

[0045] Although the creation of diagrams is purely optional, it is always possible to use the diagram features to determine at least one training dataset that corresponds to or is similar to the test dataset. Diagram features can also be determined directly from the values ​​of the physical quantity as a function of the ordinal quantity (such as time, location, or position). The diagram features can, in particular, be any features that describe the diagram as a multidimensional, at least two-dimensional, diagram and take into account the values ​​of the physical quantity in their dependence on the ordinal quantity.Examples of features are: positions of the values ​​in the plane of the diagram (in the case of a two-dimensional diagram) or the space of the diagram (in the case of a three-dimensional diagram; whereby more than three-dimensional diagrams and dependencies are also possible and, for example, in the case of a four-dimensional dependency, i.e. the dependence of three physical quantities on an ordinal quantity, one can also speak of a space); parameters derived from these positions of the values; features that describe the course of the values ​​as a function of the ordinal quantity; maxima and / or minima of the values; statistical quantities such as mean values ​​and standard deviation of the values ​​of the physical quantity.

[0046] As mentioned, the evaluation device has artificial intelligence (AI), for example based on a neural network, which is implemented by the evaluation device. When training the evaluation device, a machine learning method can be applied, for example by training the neural network using learning data (which the training data sets comprise). In particular, the training data sets can have the same data format as the test data sets that are fed to the evaluation device as part of the functionality test of the coordinate measuring machine. In contrast to the functionality test, however, during training the evaluation device also receives at least one piece of additional information for each training data set that is linked to the training data set.The additional information relates to the functionality of a coordinate measuring machine through whose operation the data set was obtained, whereby at least one training data set can also be obtained by simulating the operation of a coordinate measuring machine as an alternative to actual operation. In particular, the additional information can have been obtained by an expert through evaluation of the data set. The expert can be a person or a group of people, at least one computer, for example with artificial intelligence, or an interaction between at least one person and at least one such computer. In any case, expert knowledge is used to generate the additional information. The original goal may not have been to generate additional information for training the evaluation device, but rather to test the functionality of a coordinate measuring machine.However, such a test result, which contains the data set and the additional information regarding functionality, is suitable for training the evaluation device.

[0047] The training of the artificial intelligence can be part of the method using the training data sets with the associated additional information and / or the arrangement for testing the functionality of the coordinate measuring machine can be designed to also carry out the training of the artificial intelligence.

[0048] If the artificial intelligence of the evaluation device has already been trained using a plurality of training data sets and associated additional information, it can be further trained using at least one additional training data set and associated additional information. This can achieve an improved training state. This also makes it possible to use the evaluation device to repeat previously performed evaluations of test data sets based on the improved training state. The at least one additional training data set can, in particular, be a previously evaluated test data set.

[0049] Artificial intelligence can be implemented, in particular, by at least one neural network. Training a neural network specifically involves adjusting and / or optimizing the neural parameters and / or the connection structure of the network's neurons. These parameters include, in particular, weights for signal transmission between the neurons of the neural network, but also any threshold values ​​for the neurons' activity. The connection structure of the neurons can be modified, in particular, by creating new connections between neurons or deleting existing connections. This includes, for example, the possibility of providing and removing bias units.

[0050] A ready-to-use, trained network can therefore be defined as a neural network whose neural parameters and / or structure have been optimized such that data sets from the inspection of a coordinate measuring machine can be correctly evaluated with sufficient probability. Training and operational readiness each relate specifically to a specific type of inspection of the coordinate measuring machine and thus to a specific type of data set. Examples of different types of inspection have already been given, such as testing for probing reproducibility and testing the sensor's behavior when probing a surface of a workpiece to be measured.

[0051] The neural network can, for example, be a convolutional neural network (CNN). In particular, different types of data sets can optionally be evaluated using different types of neural networks as artificial intelligence.

[0052] Embodiments and optional features of the invention will now be described with reference to the accompanying figures. The individual figures of the drawing show: Fig. 1 a coordinate measuring machine in gantry design as an example of a type of coordinate measuring machine on which a functional test can be carried out, Fig. 2 schematically steps of an embodiment of the method for testing the functionality of a coordinate measuring machine, Fig. 3 a first two-dimensional diagram showing fifty values ​​of the probing deviation with respect to a linear axis of a coordinate measuring machine as a function of the characteristic number of the respective probing of a surface of a workpiece with a tactile probe of the coordinate measuring machine and corresponding to a first test data set or training data set, Fig. 4 a second two-dimensional diagram which, in comparison to Fig. 3 fifty other values ​​of the probing deviation with respect to the linear axis of the coordinate measuring machine as a function of the characteristic number of the respective probing and corresponds to a second test data set or training data set, Fig. 5 a diagram which represents deviations from the ideal circular movement for a circular movement of a movable part of a coordinate measuring machine within an xz plane of a coordinate system of the coordinate measuring machine, namely based on read values ​​of the position measuring system of the coordinate measuring machine, and Fig. 6 schematically shows a coordinate measuring machine which is connected via a data transmission network with a connected, in particular distributed, data storage (data cloud) to an evaluation device for evaluating test data sets.

[0053] The Fig. 1 The coordinate measuring machine (CMM) 211 shown in gantry design has a measuring table 201, above which columns 202, 203 are arranged to be movable in the Y direction of a Cartesian coordinate system XYZ. The columns 202, 203, together with a cross member 204, form a gantry of the CMM 211. The cross member 204 is connected at its opposite ends to the columns 202 and 203, respectively. Electric motors (not shown) act as drive motors to cause the linear movement of the columns 202, 203 and the parts carried by the columns 202, 203 along the movement axis, which runs in the Y direction. For example, an electric motor is assigned to only one of the two columns 202, 203 or to each of the two columns 202, 203. The cross member 204 is combined with a cross slide 207, which, for example, B. is air-supported and movable along the cross member 204 in the X direction of the Cartesian coordinate system.The current position of the cross slide 207 relative to the cross member 204 can be determined using a scale division 206. The movement of the cross slide 207 along the movement axis in the X direction is driven by at least one further electric motor as a drive (not shown). A vertically movable sleeve 208 is mounted on the cross slide 207 and is connected at its lower end to a measuring head 205 via a mounting device 210. An angled single-axis rotary joint 215 is coupled to the measuring head 205 via an exchange interface 209. A stylus 111 with a stylus ball 121 is connected to the rotary joint 215. The rotary joint 215 can be rotated by a further electric motor (not shown) about an axis of rotation of the Cartesian coordinate system that runs parallel to the Z direction, so that the stylus, for example, B. is aligned in the direction of a measuring object 217 standing on the measuring table 201.Unlike in . Fig. 1 Instead of the swivel joint 215, for example, an active measuring head can be arranged at the changeover interface 209, to which the stylus is in turn arranged. The active measuring head can have its own position measuring system, the position measurement values ​​of which are transmitted to the control system of the CMM 211.

[0054] By a plurality of short vertical lines at the lower end of the cross member 204 in Fig. 1 A scale of the position measuring system of the CMM 211 is shown. Additional scales can extend, in particular, in the vertical direction (Z direction) along the sleeve 208 and in the horizontal direction along the longitudinal edges of the measuring table 201 (in the Y direction). Using associated position measuring sensors, each arranged on a part of the CMM 211 relative to which the scale moves, the position measuring system measures (directly with the involvement of the sleeve or indirectly without the involvement of the sleeve) the position of the sleeve 208, in particular in the Y direction, X direction, and Z direction.

[0055] Shown in Fig. 1 Furthermore, there is an evaluation device 220, which receives the measurement signals from the measuring head via a schematically illustrated connection 230. A schematically illustrated controller 222 of the CMM 211 controls the drives (e.g., the aforementioned electric motors). In particular, the controller 222 is capable of moving the stylus 111 to a desired position and aligning the stylus 111 in a desired measuring direction by controlling the drives.

[0056] The controller 222 is further combined with or comprises a test device 221, by means of which a functional test of the CMM 211 can be carried out.

[0057] In Fig. 1 Further, additional elements 216 and 218 on the CMM 211 are schematically shown, which can optionally be used to perform the functional test and record the values ​​of the aforementioned physical quantities. These additional elements 216, 218 can, for example, be removed again after the functional test has been performed. These elements 216, 218 can be, for example, position sensors or elements of an additional position measuring system.

[0058] The test device 221 is connected to the Internet, for example, via a radio connection or cable connection (not shown) and is thus connected, for example, to a data cloud.

[0059] Not only in relation to the Fig. 1 With the coordinate measuring machine shown, the data obtained during functional testing can be converted into a human- and computer-readable data format, for example, before being transmitted remotely (e.g., via the internet or another data transmission network) and, for example, to the data cloud. Even if the data format is only computer-readable, the use of a uniform data format for the training data sets and the test data sets is preferred. In any case, the transmission of the acquired data can be carried out using suitable software and at least one computer.

[0060] Fig. 6 shows schematically a coordinate measuring machine 211, which is only, for example, the one shown in Fig. 1 The coordinate measuring machine 211 is connected to an evaluation device 7 via a data transmission network (indicated by double arrows), to which a distributed data memory 6 is connected. The evaluation device 7 has the artificial intelligence 8.

[0061] When a functional test of the coordinate measuring machine 211 is performed, data is generated which, optionally after preprocessing at the location of the coordinate measuring machine 211, is transmitted via the data transmission network. The data is preferably stored in the data storage 6. In particular, if the test data set that can be evaluated by the artificial intelligence 8 is already stored in the data storage 6, the artificial intelligence 8 can access the data storage 6 and evaluate the test data set in the manner described in this description.

[0062] To generate the test data set, the coordinate measuring machine 211, for example, has a generation device 5. Alternatively, this generation device can be located at a different location and generate the test data set there from the raw data of the functional test. Fig. 6 Depicted is a generating device, which in many cases is part of the coordinate measuring machine and is designed to automatically generate operating signals during operation of the coordinate measuring machine, wherein the operating signals correspond to values ​​of at least one physical quantity that is characteristic of the operation of the first generating device and / or the coordinate measuring machine. Examples of such a generating device have already been discussed.

[0063] At the Fig. 2 shown embodiment of a method for testing the functionality of a CMM, for example the CMM 211 from Fig. 1 , the test is started in step S1, in the following step S2 the test data set or a plurality of test data sets are generated, in the following optional step S3 the test data set or the test data sets are transferred to a data storage device (such as via the Internet to a data cloud), in the following step S4 the test data set is evaluated using the artificial intelligence of an evaluation device and in the following step S5 the result of the evaluation is output with a statement regarding the functionality of the coordinate measuring machine.

[0064] In principle, the evaluation device can be located remotely from the coordinate measuring machine, in the immediate vicinity of the coordinate measuring machine, or configured as part of the coordinate measuring machine. In particular, if the evaluation device is located remotely from the coordinate measuring machine, the evaluation device can access the at least one test data set via a data transmission connection. It does not have to retrieve the entire test data set, although this is possible depending on the design of the method. If the entire test data set is not retrieved, the evaluation device can continuously access the data from the test data set required for the evaluation.However, it is preferred that the evaluation device is not located at the location of the coordinate measuring machine, but rather remotely, and only accesses the data of the at least one test data set already stored at a storage location remote from the coordinate measuring machine. In this case, for example, the above-mentioned optional step S3 from . Fig. 2 The evaluation device accesses the data stored in the cloud during the evaluation. Alternatively or additionally, the evaluation device can retrieve the data of at least one test data set in whole or in part from the cloud or from the storage location remote from the coordinate measuring machine and only then begin the evaluation.

[0065] However, it is also possible for the data obtained during the functional test to be transmitted remotely by other means, for example, by saving it on a data storage device and manually transporting the device. In any case, it is possible to create the test data set to be evaluated from the data obtained during the functional test in a format that can be evaluated by the AI, at the location of the coordinate measuring machine and / or remotely.

[0066] The advantages of transferring at least one test data set to a storage location remote from the coordinate measuring machine are the possibility of secure storage against unauthorized access and / or data loss. In particular, the data can be archived there for a long time and made available for evaluation. This makes it possible, in particular, to collect test data sets from coordinate measuring machines located at different locations at the specified storage location. They are thus available for comparing the test data sets of coordinate measuring machines of the same type or otherwise comparable.In particular, if the additional information is available, which in particular for each test data set and / or each tested coordinate measuring machine contains at least one piece of information about a statement on the functionality of the coordinate measuring machine determined from the respective test data set, a comparison of the additional information and / or a training of an artificial intelligence in the manner according to the invention is also possible.

[0067] If the description of the figures or the description of exemplary embodiments and configurations refers to test data sets, this also implies, unless otherwise stated in the description, the presence of additional information associated with the test data set, provided that the respective test data set has already been evaluated and, in this way, the result of the evaluation with the statement about the functionality of the coordinate measuring machine is available as additional information. In other words, the result of the evaluation is stored together with the test data set or is linked to it in data format. For example, the statement is a statement about a change in the tested coordinate measuring machine over time, for example due to wear and tear and other stress on the mechanical and / or electrical devices of the coordinate measuring machine.

[0068] If the artificial intelligence detects such a change based on its training for one or more test data sets, it preferably not only detects the change itself, but also uses its training to make and output a statement about a type of error (e.g., worn brushes on a drive motor, worn bearings on a drive, worn suspension on a measuring head, adjustment or calibration of a coordinate measuring machine device that has not been performed or needs to be repeated, temporal changes to mechanical parts of the coordinate measuring machine - e.g., drive belts or air bearings - and the resulting necessary readjustment of parameters, contamination of the measuring systems). Therefore, during training in this case, it is necessary that the training data sets also contain statements about error types in the associated additional information.

[0069] As already described, the test data sets can each contain data that contain a dependency of at least one physical quantity on at least one order quantity and can therefore be represented as two- or multi-dimensional diagrams. Furthermore, it has already been described that when evaluating the respective test data set, the artificial intelligence, taking into account diagram features of at least one diagram of the dependency of the physical quantity on an order quantity of the test data set, determines, as far as possible, a correspondence or similarity of the test data set to a test situation recorded by the training data sets and derives the statement regarding the functionality of the coordinate measuring machine based on the corresponding or similar test situation recorded by the training data sets.In short, artificial intelligence derives the statement regarding functionality from what it has learned, taking into account the diagram features. Because it is an artificial intelligence, its capabilities go beyond identifying a similar or most similar training dataset. Rather, the artificial intelligence no longer needs access to the training datasets when evaluating the test dataset.

[0070] Preferably, however, not only the at least one test data set, but also the training data sets are and / or are designed such that they can be evaluated during training of the artificial intelligence with regard to diagram features of the dependence of the at least one physical quantity on at least one ordinal quantity. For example, all data sets therefore contain values ​​of the same physical quantity as a function of the same ordinal quantity, and the dependencies of the physical quantity on the ordinal quantity contained in the respective data sets are formatted identically and / or the quantities are scaled identically, or the data sets each contain information about the possibly different scaling. However, it is also possible to initially normalize the differently scaled training data sets, in particular using numerical methods.For example, data series can be provided with equidistant support points through interpolation in order to ensure comparability.

[0071] If no comparable data set exists yet, the test data set can become the first comparison data set for comparing further test data sets. In this case, it is preferable for a human expert to evaluate the test data set and assign it a statement regarding the functionality of the coordinate measuring machine being tested. If a sufficient number of such comparison data sets with additional information are available, they can be used as training data sets to train the artificial intelligence.

[0072] Fig. 3 and Fig. 4 For the physical quantity of probing deviation explained above, each diagram represents a test data set, but can also represent a training data set. Fifty probing deviation values ​​are shown, plotted against their characteristic number. The characteristic number can, in particular, correspond to the sequence of probing the surface of a workpiece (especially a test feature) during which the coordinate measurement values ​​for determining the probing deviation were measured. Ideally, all probing deviations are zero, i.e., there is no discernible difference between the target value and the actual value of the probing.

[0073] In the case of the diagram of the Fig. 3 The probing errors are close to zero. Two horizontal lines in each of the diagrams indicate Fig. 3 and Fig. 4 Limit values ​​for maximum permissible probing deviations are shown. In the example, these limits are 1 micrometer and -1 micrometer.

[0074] In the case of the diagram of the Fig. 4 The probing deviations (i.e., their values) also do not fall below the lower limit. However, they exhibit a downward trend. In this exemplary embodiment, the key figures for the probing deviations are arranged in the order in which the workpiece was probed. The downward trend therefore indicates that the overall probing deviation (in terms of magnitude) increases as the number of probing operations progresses.

[0075] Artificial intelligence can also recognize this downward trend after training with corresponding training data sets that also exhibit such tendencies and to which corresponding additional information with statements about such tendencies is assigned.

[0076] In particular, artificial intelligence can be used to evaluate the test data set, which contains the information about the diagram from Fig. 3 or Fig. 4 The data of the test data set, which contains the information about the dependence of the probing error on the characteristic value, is converted into image data. This creates data that corresponds to the Fig. 3 or Fig. 4 Each pixel is defined, for example, by an x-coordinate and a y-coordinate. The x-coordinate corresponds to the key figure, or the probing distance, and the y-coordinate to the value of the probing deviation.

[0077] In the next step, the AI, based on its training and the image data, checks whether it recognizes an inspection situation for which it already knows an associated statement about the functionality of the coordinate measuring machine. Possible statements are associated with the dependency of this specific physical quantity (here, the probing error) contained in the inspection data set as a function of the order quantity (here, the key figure corresponding to the sequence of probing operations). If additional inspection data sets with other dependencies on physical quantities are available, then these dependencies are also associated with possible statements about functionality, such as, in the case of the following error, "following error OK" or "following error too large" or "following error tends to increase over time."

[0078] If the AI ​​has recognized a test situation with a known associated statement about the functionality of the coordinate measuring machine, it outputs this statement, for example by assigning the statement to the test data set and saving it.

[0079] Fig. 5 shows a diagram corresponding to another test data set containing data from the execution of another functional test, namely the test to what extent a coordinate measuring machine is capable of moving a moving part, such as a stylus, on an ideal circular path. In the exemplary embodiment, therefore, the extent to which an ideal circular path can be executed in an xz-plane of the coordinate measuring machine's coordinate system is to be tested. The axis designations of such a coordinate system were determined based on Fig. 1 for an example of a coordinate measuring machine. For example, the probe ball 121 of the coordinate measuring machine 211 can be made of Fig. 1 To perform the test, the workpiece must be moved along such a circular path. The corresponding drives for movement in the x- and z-directions are operated, and the corresponding values ​​from the position measuring system of the coordinate measuring machine 211 are read out.

[0080] The diagram in Fig. 5 shows the deviation from the ideal circular path with a radius of "25." The deviations can be represented at the selected diagram scale between a radius of "24.98" and a radius of "25.02." These deviations are plotted as a function of the angle in the range between 0 degrees and 360 degrees (i.e., over one revolution of the moving part around the center of the circle). The line shown corresponds to a movement of the moving part along the ideal circle in a first direction, namely clockwise. In practice, the circular movement is also performed in the opposite direction, and the test is also based on the corresponding second movement curve.

[0081] It can be seen that there are larger deviations toward a larger radius than toward a smaller radius than "25." The largest deviations occur at approximately 0 degrees and 180 degrees. Larger deviations also occur at angles of 90 degrees and 270 degrees. This will be discussed in more detail below.

[0082] Again, the artificial intelligence can be designed to generate image data from the test data set, which can be used by the Fig. 5 described dependency. For example, the pixels that are on the Fig. 5 The coordinates "radius" (i.e. distance from the center of the circle) and "angle in the direction of orbit" are described. It is possible that the angular coordinate in the generated image is a coordinate along a straight coordinate axis, or (similar to Fig. 5 ) is alternatively represented as an angular coordinate.

[0083] If the AI ​​has recognized a test situation with a known associated statement about the functionality of the coordinate measuring machine, it gives this statement as in the case of Fig. 3 or Fig. 4 for example the statement "maximum deviation of the radius from the target value is OK" or "maximum deviation of the radius from the target value is at the angular position alpha_0" or "drive of the x-axis needs to be readjusted".

[0084] Additionally, the deviations ("noses") from the ideal circular path at the reversal points of the movement with respect to one of the linear axes of movement, i.e., at 0°, 90°, 180°, and 270°, are of interest and can be considered separately by the AI, for example. These points correspond to a zero crossing in the speed of the respective axis with reversal of the movement. Shortly before these reversal points, the speed of an axis decreases progressively, reaches zero for a short time, and then changes direction. At the point of the zero crossing, (adhesive) friction effects and hysteresis effects become more pronounced, which manifests itself in the larger deviations of the movement from the ideal circular movement in the area of ​​these points mentioned above. This can be seen from the Fig. 5 The curve shown indicates that it is a clockwise circular motion, since the clockwise "noses" occur after the zero crossings of the speed at 0°, 90°, 180° and 270°.

[0085] Now it is the task of the AI ​​to optimize the control parameters for controlling the CMM drives in such a way that these effects are reduced and thus the deviations from the ideal circular path are minimized.

[0086] Those skilled in the art are familiar with other functional tests for testing the functionality of a coordinate measuring machine. Also known is the respective physical quantity measured and recorded during the test, as a function of an order quantity depending on the type of test. For all of these dependencies of a physical quantity, a diagram can be displayed, and image data can be generated through which the values ​​of the physical quantity are defined, in particular by means of the coordinates in relation to a two-dimensional or multi-dimensional diagram. The AI ​​can calculate the coordinates, or the coordinates are already contained in the test data set. Instead of the AI, the coordinates can also be calculated by a separate computing device, for example, implemented by software on a computer. For example, the test data set can therefore be preprocessed before the AI ​​evaluates it.

[0087] After all intended functional tests have been performed and evaluated by the AI, an overall statement about the health status of the coordinate measuring machine can be generated. The procedure for generating the overall statement can be specified. For example, it can be specified that the statement "the coordinate measuring machine is healthy" (i.e., it is fully functional) is only generated and output if all individual functional tests have been passed. Optionally, an informative statement can be added that maintenance and / or adjustment of the coordinate measuring machine is recommended within a specific period of time from the performance of the functional test.

[0088] The described use of artificial intelligence makes it possible to largely dispense with the use of a human expert with extensive experience in testing the functionality of coordinate measuring machines on-site. This applies in particular if the operator of the coordinate measuring machine initiates the functional test themselves or permits the functional test to be carried out remotely, and at least one test data set generated during the functional test is transmitted remotely from the location of the coordinate measuring machine (e.g., via the internet) so that the AI ​​can access it. In particular, the functional test can therefore be initiated remotely and, if necessary, its execution can be controlled. The AI ​​can then also remotely or remotely evaluate the at least one generated test data set and automatically generate the desired statement about the functionality of the coordinate measuring machine.In this way, for example, it can be decided whether and to what extent maintenance and / or adjustment of the coordinate measuring machine is necessary.

[0089] The training of artificial intelligence can also be automated, for example, controlled by software, and repeated once evaluated test data sets with the aforementioned additional information have been added to the training data sets already used to train the AI. Nevertheless, it is possible for at least one expert to review the evaluation performed by the AI, at least on a sample basis, and to correct it in justified cases. It is advantageous in this case if the data contained in the test data sets on the dependencies of the respective physical quantity(s) can be presented as a diagram and, in particular, image data has been generated that can be easily and quickly displayed, for example, on a computer screen. The AI ​​and the human expert therefore assess the test data sets based on the same information.Furthermore, the analysis based on diagram features, especially image data, is simple and clear for both the AI ​​and the human expert. AI, too, is particularly good at capturing differences and similarities with comparable data using image data.

[0090] Overall, the evaluation of test data sets from the functional test of a coordinate measuring machine is therefore objectified, in contrast to the subjective evaluation of a human service technician or operator. The evaluation can be standardized and carried out automatically. Furthermore, standardized recommendations for maintenance, adjustment, and operation of the tested coordinate measuring machine can be automatically generated. The time spent by service technicians can be reduced, both in terms of conducting and evaluating the functional test and in terms of travel to the coordinate measuring machine's location.

[0091] Furthermore, test data sets can be collected with the aforementioned additional information on the functionality or non-functionality of a coordinate measuring machine. This not only allows for the training of artificial intelligence but also for other comparisons. In particular, the test data sets obtained at different times from repeated functional tests of the same coordinate measuring machine can be evaluated with regard to their chronological sequence. If these evaluation results or additional evaluations are made available to the operator of the coordinate measuring machine, the operator is better informed about the condition of their coordinate measuring machine than with conventional procedures. They can automatically receive information about the wear and tear on the hardware, as well as suggestions for recommended measures.

[0092] Because the software solution allows operators to collect the data themselves, and the data can be sent and analyzed automatically, an on-site technician is no longer required, and the entire process can be completed in a shorter time. This allows operators to receive a diagnosis and recommended action within minutes of data collection.

Claims

1. Method for testing a functionality of a coordinate measuring machine (211), having the following steps of: - operating the coordinate measuring machine (211) and automatically capturing operating signals which are generated during operation by at least one first generating device, which is part of the coordinate measuring machine (211), or an additional generating device and correspond to values of at least one physical variable characteristic of the operation of the coordinate measuring machine (211), - automatically generating a test data set from the captured operating signals and automatically transmitting the test data set to an evaluation device (7) which has artificial intelligence (8), - evaluating the test data set by means of the evaluation device (7) using the artificial intelligence (8), wherein the artificial intelligence (8) is in a ready-to-use state on the basis of training using training data sets with associated additional information, which each contain at least one item of information about a functionality of the coordinate measuring machine (211) or a comparable coordinate measuring machine (211) that can be determined from the respective training data set, and wherein the artificial intelligence (8) derives a statement regarding the functionality of the coordinate measuring machine (211) from the test data set, - outputting the statement regarding the functionality of the coordinate measuring machine (211), wherein the test data set and the training data sets have the values of the at least one physical variable in dependence on an ordinal variable, and wherein the artificial intelligence (8), when evaluating the test data set on the basis of its training taking into account diagram features of at least one diagram of the dependence of the physical variable on an ordinal variable of the test data set, a) determines a correspondence or a similarity of the test data set to a test situation recorded by the training data sets, or b) determines that there is no correspondence or similarity between the test data set and a test situation recorded by the training sets, and in case a), derives the statement regarding the functionality of the coordinate measuring machine (211) according to the corresponding or similar test situation recorded by the training data sets, characterized in that the physical variable or at least one of the physical variables - is a repeatedly and / or continuously measured motor current of a drive of the coordinate measuring machine (211), and the ordinal variable is a time or an index variable, - is a probing deviation between an actual position of a moving part of the coordinate measuring machine (211) determined by a position measuring system of the coordinate measuring machine (211) and a target position, wherein the actual position corresponds in each case to optical or tactile probing of a surface of a workpiece by means of a sensor of the coordinate measuring machine, and wherein the ordinal variable relates to a sequence of the probing operations or is an index variable for distinguishing the repeatedly obtained values of the deviation, - is a deviation of a movement of a moving part of the coordinate measuring machine (211) from an ideal circular path, wherein the deviation of the movement depends on an angle in the range between 0 degrees and 360 degrees, i.e. over one rotation of the moving part around the centre of the circle, or - describes a probing behaviour of an active measuring head (205) with a tactile probe (111), wherein the probing behaviour relates to a movement of the probe to a surface of a workpiece up to tactile contact of the probe with the surface, and wherein a deflection of the probe relative to a reference point of the measuring head is measured, during the movement and also from the production of the tactile contact, for describing the probing behaviour, by continuously recording measured values of a position measuring system of the measuring head.

2. Method according to Claim 1, wherein the training of the artificial intelligence (8) using the training data sets with the associated additional information is part of the method.

3. Arrangement for testing a functionality of a coordinate measuring machine (211), comprising: - at least one first generating device which is configured to automatically generate operating signals during operation of the coordinate measuring machine (211), wherein the operating signals correspond to values of at least one physical variable characteristic of the operation of the coordinate measuring machine (211), - a second generating device (5) which is the first generating device or another device and is configured to automatically generate a test data set from the captured operating signals and to automatically transmit the test data set to an evaluation device (7), - the evaluation device (7) which has artificial intelligence (8), wherein the evaluation device (7) is configured to evaluate the test data set using the artificial intelligence (8), wherein the artificial intelligence (8) is in a ready-to-use state on the basis of training using training data sets with associated additional information, which each contain at least one item of information about a functionality of the coordinate measuring machine (211) or a comparable coordinate measuring machine (211) that can be determined from the respective training data set, and wherein the artificial intelligence (8) is configured to derive a statement regarding the functionality of the coordinate measuring machine (211) from the test data set, wherein the test data set and the training data sets have the values of the at least one physical variable in dependence on an ordinal variable, and wherein the artificial intelligence (8) is configured, when evaluating the test data set on the basis of its training taking into account diagram features of at least one diagram of the dependence of the physical variable on an ordinal variable of the test data set, a) to determine a correspondence or a similarity of the test data set to a test situation recorded by the training data sets, or b) to determine that there is no correspondence or similarity between the test data set and a test situation recorded by the training sets, and in case a), to derive the statement regarding the functionality of the coordinate measuring machine (211) according to the corresponding or similar test situation recorded by the training data sets, characterized in that the physical variable or at least one of the physical variables - is a repeatedly and / or continuously measured motor current of a drive of the coordinate measuring machine (211), and the ordinal variable is a time or an index variable, - is a probing deviation between an actual position of a moving part of the coordinate measuring machine (211) determined by a position measuring system of the coordinate measuring machine (211) and a target position, wherein the actual position corresponds in each case to optical or tactile probing of a surface of a workpiece by means of a sensor of the coordinate measuring machine, and wherein the ordinal variable relates to a sequence of the probing operations or is an index variable for distinguishing the repeatedly obtained values of the deviation, - is a deviation of a movement of a moving part of the coordinate measuring machine (211) from an ideal circular path, wherein the deviation of the movement depends on an angle in the range between 0 degrees and 360 degrees, i.e. over one rotation of the moving part around the centre of the circle, or - describes a probing behaviour of an active measuring head (205) with a tactile probe (111), wherein the probing behaviour relates to a movement of the probe to a surface of a workpiece up to tactile contact of the probe with the surface, and wherein a deflection of the probe relative to a reference point of the measuring head is measured, during the movement and also from the production of the tactile contact, for describing the probing behaviour, by continuously recording measured values of a position measuring system of the measuring head.

4. Arrangement according to Claim 3, wherein the arrangement has a training device which is configured to train the artificial intelligence (8) using the training data sets with the associated additional information.