Method for controlling an automatic screwing machine

Abstract

The invention relates to a method for controlling an automatic screwing machine, which comprises: a screwing unit for automatic screwing of a screw into a component; and a control unit for controlling the screwing unit using a screwing program stored in the screwing unit, which defines the temporal progression of a target rotational speed at which the screw is driven during the screwing process and, optionally also a feed force exerted on the screw during a screwing process; wherein the control unit records a dataset relating to the screwing process during a screwing process, which dataset comprises the temporal progressions of the following screwing parameters: an actual rotational speed of the screw, a torque exerted on the screw, a feed position of the screw, and a feed speed of the screw. The control unit transmits the recorded dataset to a remote data processing unit, in which the screwing process is evaluated by means of artificial intelligence based on the transmitted dataset.

Description

[0001] Method for controlling a screwing machine

[0002] The invention relates to a method for controlling an automatic screwing machine, which comprises: a screwing unit for the automated screwing of a screw into a component; and a control unit for controlling the screwing unit using a screwing program stored in the control unit, which defines the temporal progression of a target rotational speed at which the screw is to be driven during the screwing process, and optionally also a feed force to be exerted on the screw during a screwing process; wherein the control unit records a data set relating to the screwing process during a screwing process, which data set comprises the temporal progressions of the following screwing parameters: an actual rotational speed of the screw, a torque exerted on the screw, a feed position of the screw and a feed rate of the screw.

[0003] Such a procedure is generally known. It allows an experienced screw expert to determine from the recorded data set what type of screwing process was performed, for example, whether it was a metric screw element (with a standard thread, fine thread, and / or screw locking device) or a self-tapping screw element, whether the latter was screwed into sheet metal or a blind hole, and what type of material was screwed into it (e.g., plastic, aluminum, steel). If the screw expert has access to a sample data set resulting from a faultless screwing process, they can use a deviating data set to determine possible reasons that have led to, or will lead to, a faulty screw connection in the future.It goes without saying that the quality of such data set analysis and error diagnosis depends not only on the experience of the screwdriving expert, but also requires that a screwdriving expert is available at all. If an end customer has purchased a screwdriving machine but lacks the financial and / or personnel resources to hire their own screwdriving expert, the end customer can, for example, conclude a service contract with the manufacturer of the screwdriving machine, which includes regular and / or as-needed quality control and error diagnosis. In the event of a regular service appointment and / or an error occurring, a service technician from the manufacturer would then travel to the end customer to service the screwdriving machine and, if necessary, attempt to correct any existing errors.It is understood that such a service appointment entails greater financial and time expenditure the further the end customer is from the manufacturer, especially if the manufacturer and the end customer are located in different countries or even on different continents.

[0004] The invention is based on the object of creating a more reliable, more economical and more user-friendly method for controlling an automatic screwing machine.

[0005] The object is achieved by a method having the features of claim 1 and in particular by the control unit transmitting the recorded data set to a remote data processing unit in which the screwing process is evaluated by means of artificial intelligence on the basis of the transmitted data set.

[0006] In this context, a remote data processing unit is understood to be a data processing unit that is not located at the same location as the screwdriving machine, in particular not on the same company premises. The data processing unit may be located in a different city, a different country, or even on a different continent than the screwdriving machine. For example, the data processing unit may be operated by the manufacturer of the screwdriving machine, while the screwdriving machine is used by an end customer. Due to the physical distance between the screwdriving machine and the data processing unit, communication between them preferably takes place via a remote data connection, which may, for example, include the Internet.

[0007] By evaluating the data recorded in the automatic screwdriving system and transmitted to the data processing unit using artificial intelligence, screwdriving processes performed with the automatic screwdriving system can be analyzed independently of the individual experience and expertise of a service technician, thus tending to be analyzed with greater accuracy and reliability. Furthermore, it is no longer necessary to send a service technician to the end customer, for example, to correct a faulty screwdriving process, optimize a normally functioning screwdriving process, or implement a new type of screwdriving process. Instead, these tasks can be solved efficiently and reliably remotely using artificial intelligence, saving time and money in a customer-friendly manner.

[0008] Advantageous further developments of the invention can be found in the dependent claims, the description and the drawing.

[0009] According to one embodiment, the artificial intelligence is formed by a machine learning algorithm, for example, a decision tree algorithm, random forest algorithm, support vector machine (SVM) algorithm, or k-nearest neighbor (kNN) algorithm. According to another embodiment, the artificial intelligence was trained using a large number of heterogeneous data sets. In this context, heterogeneous data sets are understood to mean data sets that result from a wide variety of screwing processes, such as screwing processes in which metric screws with different threads and with or without locking are screwed into threaded holes, or in which self-tapping screws are screwed into blind holes or sheets made of different materials.Heterogeneous data sets must therefore be distinguished from homogeneous data sets, which are obtained using a single type of screwing process, in which identical screws are always screwed into identical components.

[0010] In order to be able to supply the artificial intelligence with a large amount of data on the one hand and to be able to analyze completed screw processes at a later point in time on the other, it is advantageous if the recorded data set is stored for a predetermined period of time, for example several days, weeks or months, or even permanently in a memory of the control unit.

[0011] In principle, it is conceivable for the control unit to transmit the corresponding data set to the data processing unit after each screwing process has been completed, or for multiple data sets to be transmitted automatically at predefined intervals. However, this requires permanent or at least temporary, particularly automatically established, communication between the screwing machine and the data processing unit, which is not necessarily desirable from the perspective of the user of the screwing machine. In many cases, it is therefore advantageous for the control unit to transmit the recorded data set to the remote data processing unit upon instruction from a user of the screwing machine.To improve the quality of the evaluation by the artificial intelligence, the control unit can, upon instruction from a user of the automatic screwdriving machine, transmit at least one first data set associated with an error-free screwdriving process and at least one second data set associated with a faulty screwdriving process to the remote data processing unit. Preferably, the artificial intelligence identifies an error that occurred during the faulty screwdriving process by evaluating the first and second data sets. In a sense, the artificial intelligence thus performs a target / actual comparison, on the basis of which an error that occurred during the faulty screwdriving process can be more reliably identified.In cases of uncertainty, for example, if the artificial intelligence is less than 80% certain of the error, an interaction with the user can occur, giving the user the opportunity to provide the artificial intelligence with additional data. The evaluation result can then be made available to the user of the screwdriving machine by the remote data processing unit transmitting feedback to the screwdriving machine indicating the identified error. In particular, the feedback can include a suggestion for troubleshooting, allowing the user to correct or adapt the screwdriving process accordingly.

[0012] According to a further embodiment, the control unit can, upon instruction from a user of the automatic screwdriving machine, transmit a plurality of data sets, in particular a plurality of data sets associated with error-free screwdriving processes, as well as a user-defined specification for optimizing the screwdriving process to the remote data processing unit. A specification for optimizing the screwdriving process can, for example, consist of an increased cycle time and / or a higher screwdriving quality. Based on the plurality of transmitted data sets and taking into account the transmitted optimization specification, the artificial intelligence can then create a suggestion for at least one modified machine setting and / or for a modified screwdriving program.In cases of uncertainty, for example, if the artificial intelligence is less than 80% certain about a suitable change, an interaction with the user can also occur, giving the user the opportunity to provide the artificial intelligence with additional data. The suggestion for at least one changed machine setting and / or for the changed screw program can then be transmitted from the remote data processing unit to the control unit of the screwdriving machine so that the user can implement the suggested changes and execute a correspondingly optimized screwdriving process.

[0013] According to yet another embodiment, a user of the automatic screwdriving machine can define requirements for a novel screwdriving process and enter them into the control unit. A novel screwdriving process may, for example, involve using a different screw or a different material to be screwed into. The control unit transmits the user-defined requirements to the remote data processing unit, and the artificial intelligence determines a screwdriving program that matches the user-defined requirements, which is then transmitted from the remote data processing unit to the control unit of the automatic screwdriving machine to execute the novel screwdriving process.During the implementation of the novel screwing process, as already explained above, a data set relating to the screwing process can be recorded, which can be transmitted to the data processing unit after the screwing process has been completed in order to be evaluated by the artificial intelligence so that it can, if necessary, determine a suggestion for optimising the screwing program and / or a machine setting, which is then transmitted back to the control unit of the automatic screwing machine.The invention further relates to a system for controlling an automatic screwing machine, which comprises a screwing unit for automatically screwing a screw into a component and a control unit for controlling the screwing unit based on a screwing program stored in the control unit, which program defines the temporal progression of a target speed at which the screw is to be driven during the screwing process, and optionally also a feed force to be exerted on the screw during a screwing process. The control unit is designed to record a data set relating to the screwing process during a screwing process, which data set includes the temporal progression of the following screwing parameters: an actual speed of the screw, a torque exerted on the screw, a feed position of the screw, and a feed rate of the screw.The system further comprises a remote data processing unit having artificial intelligence, to which the recorded data set can be transmitted by the control unit and in which the screwing process can be evaluated by means of the artificial intelligence on the basis of the transmitted data set.

[0014] The invention will now be described in more detail by way of example only, using various embodiments with reference to the accompanying drawings. They show:

[0015] Fig. 1 A is a sectional view of a first embodiment of an automatic screwing machine for automatically carrying out screwing processes;

[0016] Fig. 1 B is a sectional view of a second embodiment of an automatic screwing machine for automatically carrying out screwing processes;

[0017] Fig. 2 shows a data set comprising four screw curves resulting from a grooved screw connection; Fig. 3 shows a data set comprising four screw curves resulting from a metric screw connection;

[0018] Fig. 4 schematically shows the screwing machine of Fig. 1 in communication connection with a remote data processing unit.

[0019] Fig. 1 shows a first embodiment of an automatic screwing machine 10 for automatically carrying out screwing processes.

[0020] The automatic screwdriving machine 10 comprises a base unit 12, which can be mounted on a handling device (not shown), for example, a robot arm. Therefore, the automatic screwdriving machine 10 of Fig. 1A is also referred to below as a robot-mounted automatic screwdriving machine 10.

[0021] The base unit 12 carries a screw unit 14, which can be moved in a feed direction relative to the base unit 12 by means of a pneumatic or hydraulic feed cylinder 16 of the base unit 12. It is understood that the screw unit 14 can alternatively also be advanced by means of an electric drive.

[0022] The screwing unit 14 comprises a drive motor 20 which, via a drive shaft 22 extending in the feed direction, rotatably drives a screwing tool 24 which is mounted in a positioning head 26 so as to be displaceable in the feed direction.

[0023] To carry out a screwing process, the automatic screwing machine 10 is first positioned by the handling device at a short distance, e.g., a distance of approximately 20 mm, in front of two components (not shown here) to be screwed together. A screw 28 intended for screwing is fed to the positioning head 26 by means of a feed device 30. To align and hold the screw 28 in a defined screw position, the positioning head 26 has a holding mechanism 32 comprising two jaw arms 34 which are mounted on opposite sides of the positioning head 26, each rotatable about an axis of rotation oriented at right angles to the feed direction, and which can be pressed apart against the restoring force of a spring to release the screw 28.

[0024] After positioning the automatic screwing machine 10 in front of the components, the screwing unit 14 and thus the screw 28 are advanced in the direction of the components by actuating the feed cylinder 16 until the screw 28 rests on one of the components.

[0025] Next, the screw unit 14 is advanced by the feed cylinder 16 to engage the screw tool 24 with the screw 28. During this process, the screw tool 24 rotates at a low speed at most.

[0026] As soon as the screwing tool 24 engages with the screw 28, the speed of the screwing tool 24 and thus of the screw 28 is increased to a target speed in order to screw the screw 28 into the components while exerting a feed force predetermined by the feed cylinder 16.

[0027] It is understood that the automatic screwdriving machine 10 does not necessarily have to be robot-connected, but can also be hand-guided and for this purpose can have a handle 36. One embodiment of such a hand-guided automatic screwdriving machine 10 is shown in Fig. 1B and, similar to the robot-connected automatic screwdriving machine 10 of Fig. 1A, has a screw unit 14 with a drive motor 20 which, via a drive shaft 22, rotatably drives a screwdriving tool 24 which is mounted in a positioning head 26 so as to be displaceable in the feed direction. Furthermore, the hand-guided automatic screwdriving machine 10 of Fig. 1B is also provided with a feed device 30 for feeding a screw 28 to a positioning head 26 of the automatic screwdriving machine 10. In contrast to the robot-based screwing machine 10 of Fig. 1 A, the hand-guided screwing machine 10 of Fig. 1 B does not have a feed cylinder 16, since the required feed force is exerted manually here.

[0028] In both embodiments, the screwing unit 14 is controlled by a control unit 80 (Fig. 4) of the automatic screwing machine 10, in which at least one screwing program is stored, which defines the temporal progression of a target speed at which the screwing tool 24 and thus the screw 28 are to be driven during a screwing process. In the case of the robot-based automatic screwing machine 10 of Fig. 1A, the screwing program also defines the temporal progression of a feed force to be exerted by the feed cylinder 16 on the screwing tool 24 and thus on the screw 28 during the screwing process.

[0029] Furthermore, during a screwing process, the control unit 80 records a data set relating to the screwing process, which includes the temporal progression of the following screwing parameters: the actual speed of the screwing tool 24 and thus of the screw 28, detected by a suitable sensor; the torque exerted on the screw 28, detected by a suitable sensor; the feed position of the screwing tool 24 and thus of the screw 28, detected by a suitable sensor; and the time derivative of the detected feed position, i.e., the feed rate of the screw 28, also referred to in this context as the depth gradient. Each recorded data set is stored for analysis purposes for a predetermined period of time, for example, several days, weeks or months, or permanently in a memory 82 of the control unit 80. In Fig.2 shows an example of such a data set, which was recorded during a self-tapping screwing process in which a self-tapping screw 28 is screwed into at least one component. The data set comprises four screw curves which - from top to bottom - represent the actual speed of the screwing tool 24 and thus of the screw 28 in rpm, the torque exerted on the screwing tool 24 and thus on the screw 28 in Nm, the feed position of the screwing tool 24 and thus of the screw 28, referred to here as the analog depth, in mm and the time derivative of the feed position, i.e. the feed rate or the depth gradient, in mm / s.

[0030] Based on the recorded screwing curves, the screwing process can be divided into four stages, namely a stage 1 lasting 0.9 seconds, during which the screw 28 is moved towards the components to be screwed at a low speed of 200 rpm until the screw tip hits the surface of one of the components, a stage 2, during which the screw 28 penetrates the component comparatively quickly over a period of approximately 0.2 seconds at a still low speed of 200 rpm, a stage 3 lasting approximately 0.7 seconds, during which the screw 28 penetrates deeper into the component comparatively slowly at an increased speed of 500 rpm, and a stage 4 lasting approximately 0.6 seconds, during which the screw 28 is slowly screwed further into the component at a reduced speed of 200 rpm until the recorded torque reaches a maximum value of 0.6 Nm is reached and the screwing process is stopped by reducing the speed to 0 rpm. The screwing process takes approximately 2.4 seconds in total, and screw 28 has penetrated approximately 33 mm into the component after the screwing process has been completed.

[0031] For comparison, Fig. 3 shows a data set which was recorded during a

[0032] A screwing process was recorded in which a metric screw 28 was screwed through a pilot hole in a facing component into a corresponding threaded hole provided in a remote component. This data set also comprises four screw curves which - from top to bottom - represent the actual speed of the screwing tool 24 and thus of the screw 28 in rpm, the torque exerted on the screwing tool 24 and thus on the screw 28 in Nm, the feed position of the screwing tool 24 and thus of the screw 28, referred to here as the analog depth, in mm and the time derivative of the feed position, i.e. the feed rate or the depth gradient, in mm / s.

[0033] By comparing the screwing curves shown in Fig. 2 and 3, it is easy to see that the metric screwing process shown in Fig. 3 differs significantly in some aspects from the self-tapping screwing process shown in Fig. 2. For example, the metric screwing process in Fig. 3 takes a total of almost 4 seconds, significantly longer than the self-tapping screwing process in Fig. 2. This is due, on the one hand, to the fact that the metric screw 28 here has a length of 50 mm and must therefore be driven 50 mm deep into the component, and, on the other hand, to the fact that the increased speed during stage 3 is only 280 revolutions / min. Another significant difference is that the metric screw 28 can be tightened at the end of the screwing process with a maximum torque of 6 Nm, i.e., approximately ten times as tight as the self-tapping screw 28 in Fig. 2.

[0034] In other words, by comparing recorded data sets, it is possible to determine whether they are based on similar or different screwdriving processes. Based on the characteristic features of the individual screwdriving curves in a data set, it is even possible to determine whether the underlying screwdriving process is a self-tapping or metric screw connection, the length of the screw used, the thread pitch and shank diameter of the screw, and – in the case of a self-tapping screw – the type of material it was screwed into.

[0035] As shown in Fig. 4, the automatic screwing machine 10 can be temporarily or permanently connected to a remote data processing unit 110 via a remote data connection 90, which may, for example, include the Internet 100. For example, the data processing unit 110 can be located at the manufacturer of the automatic screwing machine 10 and be in a different city, a different country, or even on a different continent than the automatic screwing machine 10. Data records of the type described above and shown by way of example in Figs. 2 and 3, recorded by the control unit 14 of the automatic screwing machine 10, can be transmitted to the data processing unit 110 for analysis purposes via the remote data connection 90.Conversely, after evaluating transmitted data sets, the data processing unit 110 can send recommendations for correcting faulty screwing processes, faulty machine settings, optimized screwing programs or even new screwing programs to the control unit 14 of the automatic screwing machine 10.

[0036] Specifically, to evaluate a transmitted data set, the data processing unit 110 uses an artificial intelligence 120 which can be formed by a machine learning algorithm, for example by a decision tree algorithm, random forest algorithm, support vector machine (SVM) algorithm or k-nearest neighbor (kNN) algorithm, and which has been trained using a large number of heterogeneous data sets, i.e. using a large number of data sets that are based on a wide variety of screwing processes, which include both metric and self-tapping screw connections, which have each been implemented using screws of different lengths with different thread pitches and shaft diameters as well as in different materials. In this way, the artificial intelligence 120 is not only trained orspecialized, but it is able to use the above-mentioned characteristic features of the screwing curves of a recorded data set not only to correctly classify the underlying screwing process with a high degree of probability, but in the case of an incorrectly performed screwing process it is also able to correctly identify the error(s) with a high degree of probability, in particular if, in addition to a data set assigned to a faulty screwing process, a data set assigned to a corresponding error-free screwing process is available.

[0037] If, for example, a user of the automatic screwing machine 10 discovers a faulty screw connection and is unwilling or unable to determine the cause of the faulty screw connection, the user can transmit the data set associated with the faulty screw connection together with a data set associated with a fault-free screw connection to the data processing unit 110. The artificial intelligence 120 of the data processing unit 110 identifies an error that occurred during the faulty screwing process by comparing the transmitted data sets, and the data processing unit 110 transmits feedback indicating the identified error to the control unit 14 of the automatic screwing machine 10. This feedback may, in particular, include a suggestion for correcting the error.Such a suggestion for correcting the error could, for example, consist of a recommendation to use a different type of screw or to increase or reduce the maximum torque at the end of the screwing process. If the artificial intelligence 120 is unsure about identifying the error, for example, less than 80% certain, an interaction with the user can occur, during which the user has the opportunity to provide the artificial intelligence 120 with additional data. A situation is also conceivable in which the user of the automatic screwing machine 10 desires an optimization of an existing screwing process, for example in the form of an increased cycle time and / or higher screwing quality. To solve this optimization task, the user can transmit a large number of data sets, ideally resulting from error-free screwing processes, to the data processing unit 110.The artificial intelligence 120 can evaluate the transmitted data sets, taking into account the also transmitted optimization specification, and create a suggestion for at least one modified machine setting or for a modified screwing program. In the event of uncertainty, for example, if the artificial intelligence 120 is less than 80% certain of the required changes, an interaction with the user can also occur here, during which the user has the opportunity to provide the artificial intelligence with additional data. The suggestion for the at least one modified machine setting or for the modified screwing program can then be transmitted from the remote data processing unit 110 to the control unit 14 of the automatic screwing machine 10 so that the user can implement the suggested changes and carry out a correspondingly optimized screwing process.

[0038] A situation is also conceivable in which a user of the automatic screwing machine 10 defines requirements for a novel screwing process and enters them into the control unit 14. A novel screwing process can, for example, involve the use of a different screw or the use of components to be screwed together made of a different material. The control unit 14 can send the user-defined requirements to the remote data processing unit 110, and the artificial intelligence 120 determines a screwing program matching the user-defined requirements, which is then transmitted from the remote data processing unit 110 to the control unit 14 of the automatic screwing machine 10 to carry out the novel screwing process.During the implementation of the novel screwing process, as explained above, a data set relating to the screwing process can be recorded, which is transmitted to the data processing unit 110 after completion of the screwing process in order to be evaluated by the artificial intelligence 120 so that the artificial intelligence can, if necessary, determine a suggestion for optimising the screwing program and / or a machine setting, which is then transmitted back to the control unit 14 of the automatic screwing machine 10.

[0039] Reference symbol list

[0040] 10 screwing machines

[0041] 12 basic units

[0042] 14 Screw unit

[0043] 16 feed cylinders

[0044] 20 drive motor

[0045] 22 Drive shaft

[0046] 24 screwing tools

[0047] 26 Positioning head

[0048] 28 screw

[0049] 30 Feeding device

[0050] 32 Holding mechanism

[0051] 34 jaw arms

[0052] 36 Handle

[0053] 80 Control unit

[0054] 82 storage

[0055] 90 remote data connection

[0056] 100 Internet

[0057] 110 Data processing unit

[0058] 120 artificial intelligence

Claims

Claims 1. Method for controlling a screwdriving machine (10), comprising: a screwdriving unit (14) for the automated screwdriving of a screw (28) into a component; and a control unit (80) for controlling the screwdriving unit (14) on the basis of a screwdriving program stored in the control unit (80), which defines the time profile of a target speed at which the screw (28) is to be driven during the screwdriving process, and optionally also a feed force to be exerted on the screw during a screwdriving process;wherein the control unit (80) records a data set relating to the screwing process during a screwing process, which includes the time profiles of the following screwing parameters: an actual rotational speed of the screw (28), a torque exerted on the screw (28), a feed position of the screw (28) and a feed rate of the screw (28), characterized in that the control unit (80) transmits the recorded data set to a remote data processing unit (110), in which the screwing process is evaluated by means of artificial intelligence (120) on the basis of the transmitted data set.

2. Method according to claim 1, characterized in that The artificial intelligence is formed by a machine learning algorithm, for example, a decision tree algorithm, random forest algorithm, support vector machine (SVM) algorithm, or k-nearest neighbor (kNN) algorithm. A method according to claim 1 or 2, characterized in that the artificial intelligence (120) was trained on a plurality of heterogeneous datasets. A method according to at least one of the preceding claims, characterized in that the recorded dataset is stored in a memory (82) of the control unit (80) for a predetermined period or permanently. A method according to at least one of the preceding claims, characterized in that the control unit (80) transmits the recorded dataset to the remote data processing unit (110) upon instruction from a user of the screwdriving machine (10).Method according to at least one of the preceding claims, characterized in that the control unit (80) transmits at least one first data record assigned to a fault-free screwing process and at least one second data record assigned to a faulty screwing process to the remote data processing unit (110) upon instruction from a user of the screwing machine (10). A method according to claim 6, characterized in that the artificial intelligence (120) identifies an error that occurred during the faulty screwing process by evaluating the first and second data sets. A method according to claim 7, characterized in that the remote data processing unit (110) transmits feedback to the screwing machine (10) indicating the identified error, wherein the feedback particularly includes a suggestion for error correction. A method according to at least one of the preceding claims, characterized in that the control unit (80), upon instruction from a user of the screwing machine (10), transmits a plurality of data sets, in particular a plurality of data sets associated with error-free screwing processes, as well as a user-defined setting for optimizing the screwing process, to the remote data processing unit (110).Method according to claim 9, characterized in that the artificial intelligence (120) uses the plurality of transmitted data sets and takes into account the transmitted optimization specification to create a proposal for at least one modified machine setting and / or for a modified screwing program, which is transmitted from the remote data processing unit (1 10) to the control unit (80) of the screwing machine (10) for the purpose of carrying out an optimized screwing process. Method according to at least one of the preceding claims, characterized in that a user of the screwdriving machine (10) defines requirements for a novel screwdriving process and enters them into the control unit (80), the control unit (80) transmits the user-defined requirements to the remote data processing unit (110), the artificial intelligence (120) determines a screwdriving program that matches the user-defined requirements, and the determined screwdriving program is transmitted from the remote data processing unit (110) to the control unit (80) of the screwdriving machine (10) for the execution of the novel screwdriving process.Method according to claim 11, characterized in that during the execution of the novel screwing process using the determined screwing program, a data record relating to the screwing process is recorded, the recorded data record is transmitted to the data processing unit (110) and evaluated by the artificial intelligence (120), the artificial intelligence (120) determines a proposal for an optimized screwing program and / or an optimized machine setting on the basis of the evaluation, and the optimized screwing program and / or the optimized machine setting is transmitted from the data processing unit (110) back to the control unit (80) of the screwing machine (10). System for controlling a screwdriving machine (10), comprising: a screwdriving unit (14) for the automated screwdriving of a screw (28) into a component; and a control unit (80) for controlling the screwdriving unit (14) on the basis of a screwdriving program stored in the control unit (80), which defines the time course of a target speed at which the screw (28) is to be driven during the screwdriving process, and optionally also a feed force to be exerted on the screw during a screwdriving process;wherein the control unit (80) is configured to record a data set relating to the screwing process during a screwing process, which includes the temporal profiles of the following screwing parameters: an actual rotational speed of the screw (28), a torque exerted on the screw (28), a feed position of the screw (28) and a feed rate of the screw (28), characterized by a remote data processing unit (1 10) having an artificial intelligence (120), to which the recorded data set can be transmitted by the control unit (80) and in which the screwing process can be evaluated by means of the artificial intelligence (120) on the basis of the transmitted data set.

Images