METHOD FOR SUPPORTING THE MAINTENANCE OF AN INDUSTRIAL TOOL, CORRESPONDING TOOL AND SYSTEM, AS WELL AS PROGRAM FOR EXECUTING THE METHOD

Description

Scope of the invention

[0001] The field of the invention is that of industrial tooling, and in particular tools designed to perform screwing or drilling with one or more determined torques.

[0002] The invention relates more specifically to the maintenance of such tools, and in particular preventive maintenance, for example to identify a defect or a state of wear of the tool, and where appropriate to ensure an effective and simplified intervention. Prior art and its drawbacks

[0003] In industrial production, such as in the automotive or aircraft sectors, screwdriving and / or drilling tools are widely used. These tools, which can be stationary or portable (and in the latter case, battery-powered), incorporate motors, typically electric or pneumatic, depending on the intended application. These tools are generally connected (wirelessly or via a cable) to a controller or concentrator (often housed in a box) which allows for the management of various operating cycles.

[0004] For example, in the case of a screwdriver, the screwing process is controlled by a servo system, and the tool measures the torque applied to the screw. This measurement is transmitted to the controller, which verifies that its value falls within the limits specified by the screwing procedure. In this way, the controller can trigger a work stoppage when the torque measurement reaches a threshold value. The screwing results can be recorded in quality databases for later processing and / or used by the operator to verify whether the tightening was correct. The controller also ensures traceability of the operations performed by the tool, for example, by recording results such as the final tightening torque, the screwing speed, the final tightening angle, the date and time of the operations, and tables representing the quality (good or bad, based on predetermined parameters) of the screwing performed.

[0005] The measurement taken by the sensor on the tool is a representation of the torque applied to the screw. This measurement is affected by various transmission elements that add noise to the measured signal. The screwdriver includes at least the following components: a motor; one or more epicyclic gear trains designed to increase the torque produced by the motor; and a torque sensor.

[0006] It may also include other elements, notably an angle return.

[0007] All these factors can disrupt the sensor signal, which then no longer perfectly reflects the torque applied to the screw. This noise can, for example, cause the tool to stop at a point other than the predefined stop point, resulting in incorrect screw fastening even though the screwdriver is reporting a success. Furthermore, degradation of the transmission components can increase noise in the measured signal. This increased noise leads to a decrease in tool performance (reduced accuracy) and / or a reduction in the quality of the screw fastening.

[0008] It is therefore desirable to check the accuracy of these screwing means and the proper functioning of the tools, whether for the control and / or monitoring of screwing parameters or for the proper execution of these parameters.

[0009] In other words, the reliability of the tools must be verifiable regularly throughout their service life to allow for maintenance, preferably preventive. The aim is to avoid, or at least reduce, wear and tear that leads to poor screw quality and / or tool breakage on the production line, requiring line shutdowns for tool replacement. This, of course, negatively impacts the overall efficiency of the production line in question.

[0010] Inspections can be carried out in various ways. For example, document FR2882287 describes a screw-driving tool comprising a rotating element mounted on a body and a torque sensor. The torque measurement provides information for determining the wear of the components. More specifically, this document explains how to process a frequency spectrum to extract at least one vibration frequency associated with a rotating element. This frequency is then compared with a reference frequency to determine the wear of the rotating element in question.

[0011] This allows for the identification of any defects in the screwdriver, but not the level of tightening precision.

[0012] Another well-known example is found in document WO2018177669, which shows the preamble to claim 1, where wear characteristics of a drill are used for tool maintenance purposes.

[0013] Preventive maintenance involves tracking the number of times a tool is used (number of screws driven, or cumulative usage time) and comparing this count with recommendations from tool suppliers. These recommendations dictate a maintenance schedule. This method uses an empirical approach that provides a statistical estimate of a tool's wear but doesn't accurately reflect its current condition.

[0014] In other words, the operator, in order to perform maintenance on a tool, proceeds as follows: Consulting the instructions and reading the manual, dismantling the tool, visually inspecting the components, identifying the defective component, consulting the supplier to find out the availability of the component, ordering... receiving and replacing the component.

[0015] This process is lengthy and therefore expensive.

[0016] Furthermore, it can lead to unnecessary interventions (replacement of a part that is not worn) or interventions that are too late (failure to detect a worn part).

[0017] It is also known to perform regular checks on a test bench. These checks also require a production stoppage, as the tool is moved to the test bench, usually at a location separate from the production area. This results in a slowdown in production and / or the implementation of replacement tooling.

[0018] The ISO5393 standard therefore requires that at least 25 test screwing operations be carried out to check a screwing tool.

[0019] Therefore, depending on the applications and / or tools: a need to implement a more efficient and effective tool maintenance technique, particularly in detecting defective components, a need for rapid and local processing (defect analysis and intervention) (and for example without information being disseminated outside the company or workshop), a need for preventive failure detection, allowing anticipation of a necessary component replacement, a need for maintenance rationalization, by changing only truly defective components, a need for maintenance assistance, facilitating the task of the operator in charge of maintenance, and / or a need for monitoring the maintenance and quality of the tool.

[0020] The present invention aims in particular to provide a simple and effective solution to this requirement. Description of the invention

[0021] According to a first aspect of the invention, a method is proposed to assist in the maintenance of an industrial tool such as a screwdriver or a drill, implementing several rotating moving components according to claim 1.

[0022] Thus, only the necessary maintenance interventions are carried out, at the most opportune time, and they are carried out easily, thanks to the support information.

[0023] Depending on the implementation, the memory can be associated with the tool or with a hub controlling it. Thus, the term "associated with the tool" means in particular "integrated into the tool" or "carried by the tool" (for example in the form of an optional module) or "remote, but paired with the tool" for example in the hub.

[0024] Depending on the circumstances, and in particular the available processing power, the analysis step can be carried out by the tool and / or by the concentrator.

[0025] Similarly, depending on the case, the step of identifying a faulty component can be done by the hub, which then transmits this information to the tool with the signature, the tool itself and / or the terminal.

[0026] The operator has, directly on his terminal, a description of a dismantling, locating, assembly operation... The three-dimensional information on said tool, combined with images of the tool taken with a camera of said terminal, makes it possible to provide the operator with an augmented reality representation on the screen of said terminal, for example to identify a defective component and / or maintenance operations to be carried out.

[0027] The intervention is thus facilitated, as the operator is guided visually, directly on a view of the tool itself. The three-dimensional information allows elements to be superimposed on the images to guide the operator, for example in the form of colored zones, arrows or other identifying features for a zone, a section or a component, a manipulation to be performed (screwing, unscrewing, insertion and / or relocation point, etc.), written instructions, etc.

[0028] According to a particular embodiment, said step of obtaining support information includes a step of connecting to a remote maintenance server, containing a set of information relating to said tool, referred to as the digital twin of the tool, and including at least one of the pieces of information belonging to the group comprising: a three-dimensional representation of the tool, an exploded view of the tool, a data sheet for the tool, a nomenclature for the tool, a calibration report for the tool, a maintenance history for the tool, a theoretical signature for the tool, an initial signature for the tool, at least one previous signature for the tool.

[0029] The implementation of such a digital twin makes it possible to have a large amount of data, potentially updated with each intervention, and to facilitate the detection and resolution of problems for each type of tool, and for the tool in question in particular. A method for assisting in the maintenance of a tool according to claim 1, characterized in that said analysis step is implemented in a concentrator connected to said tool, receiving said measurement data from said tool, and performing said analysis.

[0030] This approach is particularly useful when the tool's processing capacity is insufficient to perform such an analysis, or when equipping the tool with such processing capacity is not desirable. In this case, the tool performs the measurements and transmits them to the concentrator. If the memory is stored within the tool, a transmission from the concentrator to the tool, and specifically to this memory, is then implemented.

[0031] According to a particular implementation, said signature comprises a plurality of frequency lines.

[0032] In this case, the identification step may include comparing the amplitude of each line with a predetermined threshold value (for example, in the form of a reference signature).

[0033] Alternatively or in addition, the said identification step may also implement an analysis of the evolution of the amplitude of each line between two signatures, in particular the last two signatures.

[0034] According to a particular embodiment, the analysis step takes into account an aggregation of measurement data corresponding to at least two screw tightenings. Indeed, to obtain a reliable signature, it may be desirable to have data recorded over a sufficient angular range (for example, at least 720°). If a single tightening does not cover such a range, several tightenings may be taken into account, preferably according to an optimized aggregation. The invention includes a step of guiding the operator, by the terminal, to perform an intervention on a defective component.

[0035] According to a particular embodiment, the process may also include a step of predicting wear or a defect of a component, by analyzing a series of at least two signatures of said tool and / or a batch of similar tools and / or by comparison to predetermined threshold values.

[0036] The invention also relates to a system for assisting in the maintenance of an industrial tool such as a screwdriver or a drill according to claim 8.

[0037] According to another specific aspect, said remote maintenance server contains a set of information relating to said tool, called the digital twin of the tool, and including at least one piece of information belonging to the group comprising: a three-dimensional representation of the tool, an exploded view of the tool, a data sheet for the tool, a nomenclature for the tool, a calibration report for the tool, a maintenance history for the tool, a theoretical signature for the tool, an initial signature for the tool, at least one previous signature for the tool.

[0038] The invention also relates to computer programs comprising program code instructions for implementing the control method described above (according to any one of the various embodiments mentioned above), when executed on a computer and / or by a microprocessor.

[0039] These programs can be implemented respectively in the tool and / or in the terminal and / or in a remote device capable of communicating with the terminal, to perform all or part of the steps of the control process. List of figures

[0040] Other objects, features and advantages of the invention will become more apparent upon reading the following description, given by way of simple illustration and not limitation, in relation to the figures, among which: [ Fig. 1 ] is a cross-sectional view of a tool integrated into a tooling set according to the invention; [ Fig. 2 ] is a functional diagram of a tool according to the invention; [ Fig. 3 ] presents on an example diagram the variations in tightening torque depending on the number of tightening operations performed by the same tool; [ Fig. 4 ] presents a flowchart of the main steps for the implementation of a quality level control process according to an initial implementation; [ Fig. 5a ] presents an example of the linear characteristic of a given stiffness, according to the first implementation;[ Fig. 5b ] presents an example of a curve representing the first relationship, according to the first implementation; [ Fig. 5c ] presents an example of a curve representing the second relationship, according to the first implementation; [ Fig. 5d ] presents an example of a curve representing the third relationship, according to the first implementation; [ Fig. 6a ] presents an example of a representative curve of the first table, according to a second implementation; [ Fig. 6b ] presents an example of a representative curve of the second table, according to the second implementation; [ Fig. 6c ] presents an example of a representative curve of an intermediate table, illustrating the fact that the tool always stops at the level of a maximum, according to the second implementation; [ Fig. 6d ]presents an example of a representative curve of the third table, according to the second implementation; [ Fig. 7 ] presents a flowchart of the main steps for implementing the quality control process according to a first embodiment of the invention; [ Fig. 8a ] presents an example of two curves representing the first two tables obtained for two screw tightenings by the screwdriver of the Fig. 1 , according to the first embodiment; [ Fig. 8b ] presents an example of two curves representing two third tables corresponding to the two curves of the Fig. 8a , according to the first embodiment; [ Fig. 8c ] presents an example of representative curves of a set of intermediate aggregate tables obtained from the two curves of the Fig. 8b , according to the first embodiment; [ Fig. 9 ] presents a flowchart of the main steps for implementing the quality control process according to a second embodiment; [ Fig. 10a ] presents an example of a truncated version of the curve representing the third table obtained for the first screw tightening by the screwdriver of the Fig. 1 , according to the second embodiment; [ Fig. 10b ] presents an example of a shifted version of the representative curve of the third table obtained for a second screw tightening by the screwdriver of the Fig. 1 , according to the second embodiment; [ Fig. 10c ] presents an example of a representative curve of the third candidate aggregate table obtained by optimized concatenation of the curves of the Fig. 10a and of the Fig. 10b , according to the second embodiment; [ Fig. 11 ]illustrates an example of the result of measurement analysis, for a screwdriver, allowing the definition of a tool signature; [ Fig. 12 ] presents an example of the implementation of the process of the invention; [ Fig. 13 ] illustrates an example of a signature record of a tool by an operator; [ Fig. 14 ] illustrates an example of maintenance assistance, from an operator's portable terminal. [ Fig. 15 ] is a simplified flowchart of one embodiment of the invention. Detailed description of embodiments of the invention

[0041] The general principle of the described technique relies in particular on storing, in each tool's memory, a signature of that tool, containing data representative of the quality of the work, for example, of a screw tightening, and specifically of the quality of each component in a predetermined set of components. It also relies on implementing augmented reality operator assistance to guide the operator during maintenance, the need for which is detected by analyzing the signature. The analysis of the quality, or condition, of these components can be carried out, in particular, according to the technique described in patent document FR2882287. Other solutions are described, by way of example, in the appendix, which forms an integral part of this description.

[0042] Other methods of obtaining information on the quality of a tool's work can of course be used, or adapted, depending on the type of tool.

[0043] An example of measurement results for the tightening angle and torque is illustrated on the Fig. 11 . The illustrated histogram shows the frequency as a function of 6 sigma (detailed calculations in the appendix). A series of peaks appears at specific frequencies. Each of these peaks corresponds to one of the components (operating frequency or harmonics), and it is therefore possible to detect a defect by considering the peak amplitude and / or its frequency shift relative to a reference frequency. For example, analyzing the line amplitude allows us to determine if the dispersion (overall and specific) is too high (for example, as a percentage of the line dispersion relative to the overall dispersion).

[0044] This allows us to build a signature of the tool, for example in the form of associating each component with a frequency value and / or an amplitude, or a simpler quality information (for example 1 if the value(s) considered are in a range considered acceptable, and 0 if these values ​​exceed a predetermined threshold, set by the manufacturer).

[0045] The signature may also contain information such as the tool's identifier, its commissioning date, the date of the last intervention, usage time, etc.

[0046] This signature is stored in a dedicated memory of the tool, for example an RFID chip, which can be read by a short-range contactless link, for example according to the NFC standard, so that reading is possible even when the tool is not powered.

[0047] The operator has a mobile terminal, for example a smartphone, a tablet or a dedicated terminal, capable of reading the contents of the RFID chip, and recording the signature of the tool.

[0048] If necessary, several signatures can be stored in the tool, such as a reference signature corresponding to an ideal theoretical signature, an initial signature from when the tool was originally built, and / or one or more recent signatures to analyze changes in component wear. Alternatively, these signatures can be stored on the operator's terminal, in the concentrator, or on a remote server to which the terminal connects.

[0049] As detailed in the appendix, the signature is obtained by analyzing one or more series of measurements performed by the tool. If the tool has sufficient computing power, it can perform the calculations itself, determine the current signature, and store it in the RFID memory.

[0050] However, it is common for tools to lack such internal computing power. In this case, the calculations can be performed by the concentrator. The tool transmits the angle and torque measurements to the concentrator, either periodically or continuously. The concentrator then performs the necessary processing, for example, according to the approaches described in the appendix, determines the signature, and transmits it to the tool for storage.

[0051] This approach is schematically illustrated in Fig. 12 .Tool 121 continuously transmits the results of torque and angle measurements for each tightening operation, illustrated here by curves 122, to a concentrator 123. The concentrator 123 processes the measurement data, for example by concatenating it and applying an FFT, to produce a signature, illustrated by curve 124 (corresponding to the figure 11 ).

[0052] This signature is transmitted to tool 121, which stores it in its RFID memory.

[0053] The maintenance operator can thus retrieve the tool's signature at any time, whether it is powered or not, using a terminal equipped with an NFC / RFID reader, for example. The operator can perform periodic checks of all the tools under their responsibility simply by reading the contents of the RFID memory. They can also be alerted by an alarm. For example, if a predetermined threshold is reached, the controller (or the tool, if applicable via the controller) can issue an alarm to signal a need for tool maintenance or even block the tool to prevent production with a risk of non-quality. These alarms can be issued on the controller's or tool's human-machine interface, but also sent via conventional means such as Ethernet networks or fieldbuses to the plant's supervisory systems.

[0054] Obtaining the signature is schematically illustrated by the Fig. 13 .

[0055] The tool 131 receives (F1) from the hub 123, as explained above, the signature 124, for example via a Wi-Fi connection (or any other means of communication, wired or wireless, depending on the means implemented to communicate between the hub and the tool). The Wi-Fi signal processing means 1311 transmit (F2) the signature to the tool's microprocessor (CPU) 1312, which writes it (F3) into the RFID memory, for example, of the EEPROM type, 1313. This tool synchronization operation can be performed at regular intervals so that the tool maintains an up-to-date signature, representative of its current dispersion.

[0056] Terminal 133 accesses (F4) this memory 1313, to retrieve the signature and apply any necessary maintenance accordingly.

[0057] Alternatively, the 1312 microprocessor can preprocess the data received from the concentrator. In this case, the signature includes, either as a replacement for or in addition to the processing performed, for example, using the techniques described in the appendix, data about a component to be checked or replaced. This processing can also be performed in advance by the concentrator. Otherwise, this processing is performed by the terminal after loading (F4) the signature and / or by a remote server to which the terminal connects.

[0058] It is indeed easy for this terminal, especially if it is a phone, to connect with a server, for example via a 4G connection.

[0059] This is schematically illustrated by the Fig. 14 .As previously mentioned, tool 131 receives (F5) its signature from the hub 123, for example via Wi-Fi. Terminal 133 reads (NFC scan of the tool's RFID memory) (F6) this signature. It can then connect (F7) to a remote server 141, and / or contact (F8) a remote support service 142.

[0060] The remote server 141 can contain a wealth of information about the tool and its evolution over time, for example, in the form of a "digital twin," to facilitate monitoring and maintenance. This digital twin can contain, and make available to the operator on their terminal: a 3D representation of the tool, its exploded view, its datasheet, its parts list, its calibration report, its maintenance history, and more. It can also contain the successive signatures transmitted by the terminal, and, if applicable, an analysis of these signatures, for example, to identify a variation or abnormal evolution. Thus, the digital twin is enriched with the current state of dispersion.

[0061] It is then possible to use the terminal's peripherals to optimize the delivery of information to the maintenance operator. In particular, by combining the use of the camera, the screen, and having access to information about the tool, including a 3D model of the tool (or more generally, information enabling the construction of three-dimensional data), it is possible, using software hosted on the smartphone, to show the user in augmented reality the components causing the dispersion and potentially requiring replacement, an intervention, or a check to be performed...

[0062] Three-dimensional information allows elements to be superimposed onto images to guide the operator in augmented reality, for example in the form of colored areas, arrows or other elements to identify an area, a portion or a component, a manipulation or a control to be performed (screwing, unscrewing, insertion and / or movement location to be carried out...), a measurement point, written indications...

[0063] The control and maintenance approach implemented can therefore be as follows: The maintenance operator reads the RFID chip using their terminal to acquire the tool's signature; the terminal connects to a server containing the digital twin of the tool and downloads the 3D model of the tool; the terminal combines the information from the tool's signature and the 3D model to highlight the components contributing most to the tool's dispersion; the maintenance operator uses the terminal's camera to film the tool, and processing software overlays the downloaded 3D model onto the tool and presents the defective components in augmented reality; the operator is guided through the operation to be carried out, disassembly, part replacement, reassembly, etc., by enhancements to the augmented reality images, and where necessary, additional information (audio, text, etc.).

[0064] Once the maintenance operation has been carried out, it is recorded in the digital twin, along with the serial numbers of any new components that may have been fitted into the tool.

[0065] This information, stored on a server, can also enable statistical analysis of several maintenance operations carried out, to detect fragile components, identify causes of defects, adapt instructions, issue preventive maintenance recommendations...

[0066] This can also facilitate stock management, anticipate component orders, and develop products...

[0067] The process of the invention according to one embodiment is summarized in the Fig. 15 .

[0068] The tool periodically (151) or continuously performs measurements M of the torque C and / or angle A, which are then analyzed (152) to determine a signature S. If the tool has sufficient processing capacity, this analysis is performed internally. Otherwise, the measurement data M is transmitted to an external device, for example, a concentrator 123, which determines this signature S and returns it to the tool, which stores it (153) in its internal memory. It is, of course, also possible to distribute the analysis processing between the tool and the concentrator.

[0069] Thus, the tool permanently has information in its memory that can be read remotely, for example via RFID. The maintenance operator can at any time, without interrupting the use of the tool or having to move it to a maintenance area, read (154) the contents of the memory, and obtain the information useful for maintenance, and in particular the S signature.

[0070] The remote server 157 contains reference information, for example a digital twin JN of the tool and / or a reference signature, which allows identification (155) of a possible fault D (defective component, need for adjustment or intervention...), by comparison with the signature S. Depending on the available processing capacity, this fault identification can be carried out by the terminal and / or the server.

[0071] To assist the operator in resolving this defect D, an augmented reality representation (156) is displayed on the terminal screen, combining images I of the tool obtained using a camera 157 mounted on the terminal and additional AR information provided by the server 158, including 3D information and, where applicable, instructions, animations, illustrations, etc., guiding the operator in their maintenance operations. If necessary, information can also be projected directly onto the tool from the terminal, provided it has projection capabilities. APPENDIX 1. Example of a tool implementing the technique of the invention

[0072] With reference to the Fig. 1 ,A screw-driving tool according to this embodiment comprises a motor 1 mounted in the body 10 of the tool, the motor output being coupled to a gear set, or reducer, 2, formed of epicyclic gear trains, itself coupled to a right-angle gear 3 (the screw-driving axis being here perpendicular to the motor axis; the right-angle gear 3 may be absent in the case of another conceivable embodiment in which the screw-driving axis and the motor axis are coaxial) intended to drive in rotation a screw-driving head having a tip 4 provided for receiving a screw-driving socket.

[0073] As is known, a torque sensor 51 (for example, a strain gauge bridge) provides information relating to the tightening torque exerted by the tool. An angle sensor 52 is also provided at the rear of the motor. It may, for example, comprise a magnet rotating in front of a Hall effect sensor mounted on an electronic board.

[0074] On the functional diagram of the Fig. 2 , The mechanical components are this time represented schematically so as to show the motor 1, the reducer 2 and the angle drive 3. As illustrated, the torque sensor 51 is connected to a measuring microcontroller 54 which transmits the data to a control unit 55 of the tool.

[0075] Based on the data provided by the torque sensor 51, a control unit 55 controls the operation of the motor 1 via a control unit 53.

[0076] The control unit 55 further incorporates means for processing the signal provided by the torque sensor 51 to deliver at least one piece of information representative of a dispersion and / or deviation from the screwing target, resulting from disturbances generated by the screwdriver. According to this embodiment, the control unit 55 and the control unit 53 are integrated into a unit 6, designated as the "screwing controller" on the figure 2 .

[0077] The screw-driving controller 6 comprises, or may be comprised of, a microprocessor or microcontroller implementing a program stored in internal or external memory, enabling, in particular, the execution of the steps of the process of the invention, for example, according to the embodiments described below. It may further integrate or control: a communication module 61 allowing the controller 6 to be connected to an information exchange network, for example of the Ethernet type; a display 62.

[0078] Thus, when the controller 6 detects that dispersion or deviation values ​​from the setpoint no longer meet production requirements, a signal and / or an alert message is displayed on the display 62, and possibly sent to a remote station via the communication module.

[0079] In the case of a battery-powered tool, the tool control functions 55 and motor control functions 53 can be integrated into the tool.

[0080] According to one embodiment, such a message can indicate the faulty component concerned, for example by comparing the individual dispersion with respect to threshold values ​​or a percentage of the dispersion, and can also specify the type of maintenance and / or servicing to be carried out.

[0081] It is recalled that the tightening torque is determined from a voltage transmitted by the torque sensor 5. 2. Quality control of the screwing process

[0082] After detailing by examples the main devices for the implementation of the invention, we will now explain how these cooperate within the framework of a process for controlling the quality level of screwing of a tool.

[0083] The curve of the Fig. 3 This illustrates the variations in tightening torque as a function of the number of tightening operations measured. This curve is developed from several measurements (for example, from 25 to 100) allowing for statistical studies on the behavior of a tool. Analysis of the results yields, in particular, the following two data points: The dispersion of tool tightening torques, characterized by a standard deviation (σ), indicates the tool's ability to accurately reproduce a torque. Dispersion is generally expressed as 6 times the standard deviation divided by the mean as a percentage; the deviation from the target is calculated by subtracting the target from the mean. The purpose of pre-calibrating the screwdriver against the target is to achieve the smallest possible deviation.

[0084] The measurement curve typically has the shape of a Gaussian, with almost all tightenings (99.73% in the tests carried out) located in the 6σ zone.

[0085] To accurately assess the tool's precision during a single tightening operation (or over a limited number of tightening operations), all defects must be present on the torque curve. However, a defect that occurs once per revolution of the screwdriver's output shaft will not necessarily appear if the screw requires a 30° rotation to be tightened. A minimum rotation angle of 720° (at least two revolutions to analyze low frequencies) of the output shaft is desirable (obtained over one or, according to the present invention, several tightening operations).

[0086] The disturbances causing variations in torque from one screwing to another have various origins such as the meshing of the teeth or electrical disturbances of the signals by the magnetic field of the motor which generate discrepancies between the torque measurement and the torque actually applied to the screw.

[0087] These disturbances are characterized by an oscillation in the measurement of the tool's torque around what the actual value applied to the screw would be if it were measured in real time by a sensor placed between the screw and the screwdriver.

[0088] This oscillation occurs at varying frequencies and amplitudes depending on the origin of the disturbances.

[0089] In this description, it is assumed that the magnitude of the disturbances is proportional to the instantaneous torque supplied by the screwdriver. Consequently, the dispersions and deviations are of the same level regardless of the tightening torque.

[0090] According to the estimation results, the process, which is the subject of the invention, includes a step of emitting an alert signal when the controller 6 detects that the dispersion or deviation from the setpoint of the tightenings no longer meets production requirements. This alert addresses both quality and safety constraints. An alert can also be generated upon detection of an abnormal disturbance amplitude in a component, for example, to perform a diagnostic. This can notably be presented in the form of a table showing the dispersions and deviations for each disturbance, as explained in more detail later, in relation to step 4.7 of the process. Fig. 4 .

[0091] It is also possible, during a maintenance test, to estimate the dispersion and deviation from the setpoint of the screwdriver tightenings for usual test stiffnesses, thus allowing a quick check of the tool.

[0092] It should be noted that this estimate does not take into account certain effects of the screwdriver, such as insufficient motor braking when the tightening target is reached. Indeed, it is difficult, and of little use, to determine what happens after the motor stops.

[0093] One aspect of the invention is to calculate the dispersion of a screw-driving tool and its average deviation from the target tightening torque based on disturbances detected in the signal produced by its torque sensor 51. This evaluation can be performed during a single tightening operation on a production line, for example. This evaluation is therefore much faster than performing a diagnostic, which generally requires several dozen tightening operations on a test bench.

[0094] Examples of the method of the invention, which can be implemented on a microprocessor and / or in a computer, in particular in the tool, the terminal and / or the server, are described below. 3. Examples of screw control procedures 3.0 Glossary

[0095] Within the scope of this description and the claims: "First table" is a table of doublets, each containing an angle value and a torque value. Such a first table represents the torque ramp-up during screw tightening; "Second table" is a table containing a series of values ​​representing the true characteristic of the screw as a function of the angular pitch; "Third table" is a table of values ​​showing the torque as a function of the angle. Such a third table represents the disturbances induced by the screwdriver during the torque ramp-up; "Third unit screwing table" is a third table obtained from a given screw tightening operation. At least two third unit screwing tables are considered, called the third table of a first screw tightening operation and the third table of a second screw tightening operation, according to the method of the invention;"Truncated third screw-driving table" is a unit screw-driving third table from which data relating to one or more measurements have been removed; "intermediate third table" is a third table containing data from at least two unit screw-driving third tables and / or truncated third screw-driving tables; "aggregated third table" is a third table containing data from at least two unit screw-driving third tables and / or truncated third screw-driving tables comprising a number of values ​​considered sufficient to perform an analysis of the disturbances induced by the screwdriver; "candidate aggregated third table" or "selected aggregated third table" is a third table selected from at least two "aggregated third tables" according to an optimization criterion. 3.1 First implementation of a screw control process

[0096] In relation to the Fig. 4 ,A first implementation will now be described. Fig. 4 presents a flowchart of the main steps for the implementation of a process for controlling the quality level of screwing of a screwdriver, in relation to a predetermined screwing objective, according to a first example of implementation.

[0097] In step 4.1, the tool, for example a screwdriver, is switched on and performs a task according to a screw-driving procedure. During operation, the sensors measure the torque (sensor 51) and angle (sensor 52), referencing these measurements with respect to time. Measurements are taken every millisecond, for example (step 4.2). The values ​​from sensors 51 and 52 are transmitted to the controller 6.

[0098] Controller 6 stores the measured values ​​in its memory and processes them to produce a table of doublets of torque and angle values ​​as a function of time, for example at predetermined time intervals. This table is called the "raw table".

[0099] In step 4.3, the controller determines a table representing torque as a function of the tightening angle, for constant pitch angle values, to develop an initial table of pairs representing the torque increase during tightening of at least one screw, each pair including an angle value and a torque value. This step consists of: determine (I) an angular step, which can be chosen arbitrarily, and for example correspond to a mean step located between two values. In the latter case, it corresponds to the difference between the final angle and the initial angle, divided by the number of points between the two: Δθ = θ n − θ 0 n

[0100] Thus, each new angle is calculated as follows: θ ′ i = i × Δθ ; Calculate (II) the torque samples for each newly defined angle. To perform this second calculation (II), according to a first approach, this calculation can implement a linear interpolation between two torque values ​​from the first series: C ′ i = C i + C i + 1 − C i × θ ′ i − θ i θ i + 1 − θ i

[0101] Other approaches, notably polynomial interpolation of the first set of measurements, can also be used.

[0102] At the end of step 4.3, controller 6 establishes a series of S1 values ​​which represents the value of the torque as a function of the angular step (each torque value being calculated for constant angular steps).

[0103] This gives us the first table, representing the relationship: C capteur = f α obtained from the representative pairs of values ​​representing the torque increase during the tightening of at least one screw, recorded in the first table. This first table eliminates the influence of the tool's rotational speed, which can vary during tightening.

[0104] In step 4.4, the theoretical characteristic of the assembly is estimated. This step allows us to determine an image of the true characteristic of the screw by calculating a theoretical characteristic.

[0105] Several digital processing (filtering) methods are possible, such as: Linear regression applied to the torque table as a function of angle. Polynomial regression applied to the torque table as a function of angle. Low-pass filter with a cutoff frequency equal to the value of the fault with the lowest frequency. This method preserves the useful portion of the disturbances by eliminating, in particular, any potential screw defects, which may not be linear.

[0106] At the end of this step, controller 6 updates a second table containing a series of S2 values ​​representing the true characteristic of the screw as a function of the angular pitch (at constant angular pitch). This series of values ​​can be expressed by the formula: C caractéristique vraie = g α

[0107] In step 4.5, the portion of the signal resulting from disturbances generated by the tool can then be isolated and quantified. Depending on one implementation, this step is broken down into several sub-steps: 1°] Step 4.5.1 consists of retaining from the first table only the information representative of the disturbances generated by the tool.

[0108] For the same angle, the torque values ​​in the second table are subtracted from the corresponding torque values ​​in the first table. The result of these subtractions is divided by the corresponding torque values ​​in the second table, expressed as a percentage. ΔC % = f α − g α / g α

[0109] This determines the values ​​in the third table.

[0110] In a second step, at stage 4.5.2, the discrete Fourier transform is calculated from this table to perform a frequency analysis of the signal and highlight the various disturbances that appear as a spectral line characterized by a specific frequency. The table below presents, as an example, representative values ​​of the frequency and amplitude of each detected disturbance. Disturbance Frequency Amplitude 1 f1 A1 2 f2 A2 ------ ------ ------ n fn Year

[0111] n varies, for example, from 1 to 1000.

[0112] It can be noted that according to this method of calculation, errors are considered to be independent of the torque, or to have little influence on its value.

[0113] 2°] In substep 4.5.3, a linear characteristic with a defined stiffness is selected. This stiffness is a normatively defined input parameter, for example: sharp angle (30°) - elastic angle (360°), or somewhere in between (in particular, this characteristic can be the true characteristic of the actual assembly, defined by the second table). The clamping angle to be simulated, α screw, is selected.

[0114] The dispersion is evaluated for each frequency fi present in the table above. The table of torque rise as a function of angle associated with this stiffness is expressed as follows: T R α = α α vis . C consigne Or : TR is the torque of the linear characteristic, α screw is the total screw angle (from 0% to 100% of the torque, in degrees), C setpoint is the setpoint torque (in Nm).

[0115] In this implementation, the actual torque is considered perfect, meaning that the torque increases proportionally with the angle.( Fig. 5a ). Curve C11, illustrating this linear characteristic, is a straight line segment whose slope is a function of the clamping angle. 3°] During substep 4.5.4, controller 6 determines a first mathematical relationship Tc obtained by summing curve C11 (the linear characteristic) and the sinusoidal curve whose amplitude and frequency are those of the disturbance under consideration. The calculation of this sinusoid addition relationship is implemented for each disturbance.

[0116] This expresses a first relationship: T c α = T R α + C consigne . A . sin 2 . π . f . α Or : T c is the torque measured by sensor 5 (in Nm), A is the relative amplitude of the fault with respect to the setpoint torque C setpoint, from the FFT (%) f is the frequency of the disturbance (deg-1).

[0117] An example of a C12 curve produced by this first mathematical relationship is presented at the Fig. 5b .

[0118] 4°] In sub-step 4.5.5, controller 6 determines a second mathematical relationship, which expresses the fact that the tool's shutdown does not take into account torque decreases. This second relationship, Ts, is derived from the first relationship and is expressed as follows: T S α = max 0 ≤ x ≤ α T c x

[0119] Torque values ​​are maximized to eliminate decays, thus producing a second relationship expressing torque as a function of an angle. In other words, Ts is a hypothetical representation of a continuously increasing torque, meaning that the tool cannot be stopped at a torque value lower than a value previously reached during operation. According to this hypothetical representation, the tool does not stop when the torque decreases, but rather when a "maximum" value is reached.

[0120] An example of a C13 curve illustrating this second mathematical relationship is presented at the Fig. 5c .

[0121] 5°] During substep 4.5.6, controller 6 deduces from this second relation a third relation which corresponds to a subtraction, from the values ​​obtained using the second relation, of the corresponding values ​​of the linear characteristic. This third relation is expressed mathematically as follows: T S α − T R α

[0122] The table thus obtained is illustrated by curve C14 of the Fig. 5d .

[0123] Thus, at the end of the five steps 4.5.1 to 4.5.6, which are described below according to an example of implementation, the controller 6 can determine the dispersion and / or deviation from the target resulting from each disturbance generated by the tool (step 4.6).

[0124] Initially, in step 4.6.1, controller 6 evaluates the individual influence of disturbances on the dispersion and / or screw deviation relative to the target.

[0125] In one particular case, when the linear characteristic is selected in step 4.5.3, the calculation of the dispersion and deviation from the target is performed by considering a joint stiffness that can be chosen independently of the stiffness of the joint on which the values ​​of the first series were collected. In another embodiment, several dispersion and deviation calculations are performed using several stiffnesses, for example, the standard stiffnesses used to define a rigid joint and an elastic joint, or a stiffness specific to the application.

[0126] During the calculation of dispersion and deviation from the objective, the disturbances generated by the tool are considered one by one in order to evaluate, for each disturbance, its contribution to the overall disturbance.

[0127] According to one implementation, the evaluation of the individual influence of disturbances on dispersion and of the screw deviation from the target is carried out following these steps: Calculating the average of this difference over a period. This average represents the difference between the torque generated by the tool and the target tightening torque. It is expressed as follows: x ¯ = C consigne − f ∫ 0 1 f T S α − T R α dα Calculation of the standard deviation of this third relationship. This standard deviation represents the dispersion introduced by the spectral line on the torque generated by the tool. It is expressed as follows: σ = 1 x ¯ × ∫ 0 1 f T S α − T R α − C consigne + x ¯ 2 dα

[0128] These calculations are repeated for each value of n (most are close to 0, and not significant. Higher values ​​correspond to possible defects. Knowing the characteristic frequencies of each element of the tool, it is possible to determine the faulty element(s).

[0129] In a second step, at step 4.6.2, the controller 6 evaluates the influence of all disturbances on the dispersion and the screw deviation relative to the objective, for the assembly stiffness(es) previously chosen.

[0130] According to an example implementation, the calculation is performed by carrying out the following calculations: The averages are added together, and the value thus calculated represents the difference between the torque generated by the tool and the target tightening torque for all the grooves, and therefore the disturbances induced by the tool. It is expressed as follows: x moy ≥ C consigne − ∑ i = 1 n x ¯ i − C consigne

[0131] With i varying from 1 to n, i representing each of the perturbations. The standard deviations are aggregated to give a representative value of the dispersion induced by all the spectral lines. It is expressed as follows (formula 1): 6 σ ≤ ∑ i = 1 n 6 σ i 2

[0132] Indeed : σ X + Y = σ X 2 + σ Y 2 + 2 σ X σ Y ρ X Y

[0133] Gold : − 1 ≤ ρ X Y ≤ 1

[0134] SO : σ X 2 + σ Y 2 + 2 σ X σ Y ρ X Y ≤ σ X 2 + σ Y 2 + 2 σ X σ Y

[0135] Knowing that : σ X 2 + σ Y 2 + 2 σ X σ Y = σ X + σ Y 2

[0136] We obtain: σ X + Y ≤ σ X 2 + σ Y 2

[0137] This calculation (formula 1) therefore allows us to verify the presence of an increase in dispersion by the sum of the squares of the dispersions calculated for each defect.

[0138] According to a specific implementation and for safety reasons, the value obtained is increased compared to the actual value.

[0139] In step 4.7, tests are performed to determine if the dispersion and deviation fall within an acceptable range; if not, an alert is issued. The results can be presented in a table such as: Corner Dispersion Gap Assessment Threshold Assessment Threshold 30° σ 30 x 30 360° σ 360 x 360 Special angle σ special special x

[0140] Using such a table, it is possible to extract the dispersion and deviation for a given angle (in the case of a diagnosis, for a desired application of the client, we analyze for a given angle. This makes it possible to identify a possible faulty component).

[0141] The table below shows the dispersions and deviations for each disturbance: Disturbance Dispersion Gap Assessment Threshold Assessment Threshold 1 σ 1 x 1 2 σ 2 x 2 ------ ------ ------ n σ n xn

[0142] This table allows the identification of components generating abnormal inaccuracy, each of the disturbances 1 to n being associated with one of these components.

[0143] The example of implementing the process for controlling the quality level of work of a tool, which has just been described in the previous pages, is considered the most accurate.

[0144] Another implementation will now be described in the form of another example. This implementation describes a simpler but significantly less precise method. 3.2 Second implementation of a screw control method

[0145] This alternative implementation does not include FFT calculation and therefore does not impose any special conditions on the recording of the torque table.

[0146] Initially, and identically to the first implementation, the torque table expressed as a function of the angle is calculated and recorded. This step results in a series of values, forming the first table, and expressing: C capteur = f α

[0147] Figure 6.1 illustrates an example of a C21 curve representing this first table.

[0148] In a second step, and identically to the first implementation, controller 6 determines the theoretical characteristic of the screw, which is a reflection of the true characteristic. This step results in a series of values, forming the second table, and expressing: C caractéristique vraie = g α

[0149] Figure 6.2 illustrates an example of curve C22 representing this second table, superimposed on curve C21.

[0150] In a third step, and differently from the first implementation, controller 6 isolates the portion of the signal resulting from disturbances generated by the tool. This third step is broken down into several stages: 1] Determination, from the torque table transmitted by the tool as a function of angle, of a first relationship expressing the fact that the tool's stopping does not take into account torque decreases. The table thus determined is expressed as follows: h α = max 0 ≤ x ≤ α f x

[0151] With h(α) the torque (Nm) calculated by controller 6, which indicates that a tool stop cannot be activated at a torque value lower than a value previously reached during tightening. As with the previous method, the tool always stops at a maximum value.

[0152] Figure 6.3 illustrates an example of curve C23 representing this table, superimposed on curve C21. II] Determination of a second relationship that expresses the differences between the first relationship and the theoretical characteristic. In other words, this second relationship provides, as a function of the angle, the difference between the torque table calculated in the previous step and the theoretical characteristic of the screw (second table, S2). This results in a series of values ​​representing: ΔC = h α − g α III] Determining a third relationship obtained by normalizing the second relationship with respect to the theoretical characteristic, this step consists of calculating the ratio between the difference and the theoretical characteristic of the screw. This results in a series of values, constituting the third table and representing: ΔC % α = h α − g α / g α

[0153] There figure 6d illustrates an example of a C24 curve representing this third table.

[0154] Fourthly, the dispersion and deviation from the target resulting from the portion of the signal itself caused by the disturbances generated by the tool are calculated. This step allows for the calculation of the dispersion and deviation from the clamping target induced by all the disturbances.

[0155] The average of this difference across the entire signal is first calculated. This average represents the difference between the torque generated by the tool and the target tightening torque. It can be expressed by the following equation: x ¯ = C consigne − 1 n ∑ y = 1 n ΔC % y

[0156] The standard deviation of this third relationship is then calculated. This standard deviation represents the dispersion introduced into the torque measured by the tool. It can be expressed by the following equation: σ = 1 x ¯ × 1 n ∑ y = 1 n ΔC % y − C consigne + x ¯ 2 where n here represents all the measurement points taken during the operation of the tool.

[0157] Unlike the first implementation, which only considers a single period because they are all considered identical, the second implementation takes each oscillation into account. This second method has the advantage of considering any disparities between oscillations.

[0158] In a fifth step and in a manner similar to the first implementation, the evaluation results and a possible alert is / are issued.

[0159] This second implementation is, however, significantly less precise because it does not allow for results to be obtained for each disturbance frequency, and therefore for each component to be tested individually.

[0160] The present invention thus makes it possible, in particular, to determine whether a tool is capable of performing the task required of it, in real time and on the production line. The invention can also be used to determine, based on measurement results, whether the task, such as screwing, has been performed correctly. By using the values ​​characterizing the disturbances detected and calculated during operation with the produced part, quality control can be performed. in retrospectparts produced and thus call into question the quality of certain parts if it turns out that the magnitude of the disturbances was too great.

[0161] The process of the invention makes it possible, in particular, to provide, depending on the needs and applications, at least one of the following elements: an alert if the dispersion or deviation from the setpoint of the tightenings no longer meets the production requirements, possibly during the use of the screwdriver in production (quality / safety aspect); an estimation of the dispersion and deviation from the setpoint of the tightenings of the screwdriver for the usual test stiffnesses, possibly during a maintenance test (quick control aspect); the detection of an abnormal disturbance amplitude of a component and the generation of an alert (diagnostic aspect).

[0162] The invention thus allows, in particular: to warn the user in real time of the screwdriver's inability to perform the job correctly, for example when using the tool on an assembly line; to quickly provide, in real time, an estimate of the dispersion and deviation from a tightening target; to generate an alert if a component of the screwdriver degrades abnormally, in particular to allow the tool assembler or maintenance technician to identify the faulty component(s).

[0163] Although described through a number of detailed embodiments, the proposed method and the corresponding devices include various variants, modifications, and improvements that will be obvious to those skilled in the art, it being understood that these various variants, modifications, and improvements form part of the scope of the invention, as defined by the following claims. Furthermore, the various aspects and features described above may be implemented together, separately, or substituted for one another, and all the different combinations and sub-combinations of these aspects and features form part of the scope of the invention. In addition, some of the devices described above may not incorporate all the modules and functions intended for the described implementations. 4. Embodiments of the process of the invention

[0164] As previously mentioned, a minimum rotation angle of 720° (at least two turns to analyze low frequencies) of the output shaft is generally desirable. Therefore, when a single tightening does not cover this minimum angular range, it is advisable to consider measurements taken from two or more tightenings.

[0165] The data obtained from these different screw measurements must then be combined, or concatenated, to implement the technique described above, and in particular to construct the tables described above. This concatenation, however, cannot be implemented without prior processing, as the technique described above takes into account the periodicity of the signal represented by the measurements and processing.

[0166] Thus, the invention proposes a method for controlling the quality level of screwdriving of a screwdriver against a predetermined screwdriving target, taking into account a series of data representative of the torque increase of at least two screw tightenings at a predetermined angular frequency. It comprises the following steps: obtaining a sub-series of data for each of said screwings, corresponding to a sub-series of measurements; optimized aggregation of said sub-series, to form said data series, including a data deletion step corresponding to a number of measurements determined according to a periodicity optimization criterion; and analysis of said data series, delivering said at least one information representative of a dispersion and / or a deviation from said screwing objective, resulting from disturbances induced by said screwdriver.

[0167] In other words, concatenating, or aggregating, data from two (or more) subsets of measurements does not retain all available data, even though the primary objective is to have a sufficient angular range. Instead, some data is removed, so that the signal resulting from processing the retained data exhibits periodicity characteristics that are effective for determining dispersion and / or deviation information.

[0168] The purpose of this suppression is, in essence, to provide a final signal that is as periodic as possible or, in other words, that the link between the two signal portions, corresponding to the two screws taken into account, is as linear as possible (that is to say that the slopes of the two signal portions, at the level of their junction, are as close as possible to each other, so that this junction is as "smooth" as possible, without introducing a sudden transition that would disrupt the analysis).

[0169] Two embodiments are described below. 4.1 First embodiment of the invention

[0170] In relation to the Fig. 7 , A first embodiment will now be described. Fig. 7presents a flowchart of the main steps for implementing a process to control the quality level of screwdriving of a screwdriver, relative to a predetermined screwdriving target, according to a third embodiment. Some steps of this first embodiment are further illustrated by the curves shown on the Fig. 8a , Fig. 8b and Fig. 8c .

[0171] More specifically, during the implementation of step 4.3 (according to any of the implementations mentioned in paragraph 5.3), a series of doublets representative of the torque increase during the tightening of a first screw by the screwdriver of the Fig. 1is obtained (for example, after implementing steps 4.1 and 4.2 described above). Each pair includes an angle value and a torque value. In this way, a first table of values ​​associated with the first tightening is created. The first table thus obtained is illustrated by curve A1 of the Fig. 8a .

[0172] During the implementation of step 4.4 (according to any of the aforementioned implementations), a second table of values, associated with the first tightening, is determined from the first table of values. The second table of values ​​presents the torque as a function of the angle and is representative of the true characteristic of the first screw.

[0173] During the implementation of step 4.5.1 (according to any of the aforementioned implementations), a third table of values, showing the torque as a function of the angle and representing the disturbances induced by the screwdriver during the torque ramp-up of the first screw, is determined from the first and second tables. The third table thus obtained, associated with the first screw tightening, is illustrated by curve B1 of the Fig. 8b .

[0174] However, unlike the first two implementations described above, steps 4.3 (for example, after implementation of steps 4.1 and 4.2 described above), 4.4, and 4.5.1 are implemented for at least one second screw tightening operation by the screwdriver. This results in a first table of values ​​associated with the second tightening operation (illustrated by curve A2 of the Fig. 8a ),a second table of values ​​associated with the second screwing, as well as a third table associated with the second screwing (illustrated by curve B2 of the Fig. 8b ).

[0175] Indeed, the defects sought in the measurements are related to the screwdriver. Thus, even if different screws are used, the defect is present in the series of measurements taken during the different screw tightening operations. Therefore, in order to perform a detailed analysis of the measurements taken, for example, based on a sufficient number of measurement points, during step 7.1, an optimized aggregation of the third table associated with the first tightening operation and the third table associated with the second tightening operation is implemented. This optimized aggregation involves removing a number of values ​​from the third table associated with the second tightening operation, determined according to a periodicity optimization criterion. A candidate third aggregated table is thus produced after the implementation of step 7.1.

[0176] More specifically, in substep 7.1.1, the third table associated with the first screwing operation and a truncated version of the third table associated with the second screwing operation, called the truncated third table, are concatenated to form an intermediate aggregate table. The truncated third table results from removing a given number of successive values ​​corresponding to angles of minimum amplitude from among the values ​​in the third table associated with the second screwing operation. In other words, it is the first values ​​of the third table associated with the second screwing operation that are removed here.

[0177] Furthermore, the deletion in the third table associated with the second screw operation, and the concatenation with the third table associated with the first screw operation, are repeated for different values ​​of the given number of deleted values. In this way, a set of intermediate aggregated tables is obtained.

[0178] Reconsidering the example of the curves (and associated value vectors) B1 and B2 of the Fig. 8b representing the third tables associated respectively with the first and second screwing, the following formula is used for example to obtain the values ​​of each intermediate aggregate table (the different intermediate aggregate tables are indexed by x here). ∀ x entier ∈ 0 longueurB 2 2 , ∀ a entier ∈ 1 ; longueurB 1 + longueurB 2 − x , C x a = B 1 a si a ≤ longueur B 1 B 2 a − longueur B 1 + x si longueur B 1 < a < longueur B 1 + longueur B 2 − x

[0179] Curves C0 to C5, corresponding to the different intermediate aggregate tables of the resulting set of intermediate aggregate tables, are shown on the Fig. 8c .

[0180] In terms of numerical values, the values ​​corresponding to curves B1 and B2 are given in the following table: Sample B1 B2 1 0 -0,5 2 1 -2 3 3 -4 4 3 -1,5 5 1 0 6 0 0,5 7 -1 2 8 -3 4 9 -3 1,5 10 -1 0

[0181] The numerical values ​​corresponding to curves C0 to C5 are then given in the following table: Sample C0 C1 C2 C3 C4 C5 1 0 0 0 0 0 0 2 1 1 1 1 1 1 3 3 3 3 3 3 3 4 3 3 3 3 3 3 5 1 1 1 1 1 1 6 0 0 0 0 0 0 7 -1 -1 -1 -1 -1 -1 8 -3 -3 -3 -3 -3 -3 9 -3 -3 -3 -3 -3 -3 10 -1 -1 -1 -1 -1 -1 11 -0,5 -2 -4 -1,5 0 0 12 -2 -4 -1,5 0 0,5 2 13 -4 -1,5 0 0,5 2 4 14 -1,5 0 0,5 2 4 1,5 15 0 0,5 2 4 1,5 0 16 0,5 2 4 1,5 0 17 2 4 1,5 0 18 4 1,5 0 19 1,5 0 20 0

[0182] Alternatively, in substep 7.1.1, the third table associated with the second screwing operation and a truncated version of the third table associated with the first screwing operation, called the truncated third table, are concatenated to form the intermediate aggregate table. More specifically, the truncated third table results from removing a given number of successive values ​​corresponding to angles of maximum amplitude from among the values ​​in the third table associated with the first screwing operation. In other words, it is the last values ​​of the third table associated with the first screwing operation that are removed here.

[0183] Furthermore, the deletion, in the third table associated with the first screwing, and the concatenation with the third table associated with the second screwing are repeated for different values ​​of the given number of deleted values, thus delivering the set of intermediate aggregated tables in this alternative.

[0184] Back to the Fig. 7 , In substep 7.1.2, autocorrelation is implemented for each intermediate aggregate table in the previously obtained set of intermediate aggregate tables. This yields a corresponding set of autocorrelated intermediate aggregate tables.

[0185] Reconsidering the example of the curves (and associated value vectors) C0 to C5 of the Fig. 8c , The following formula is used, for example, to calculate the autocorrelation function g(y,x) of each intermediate aggregate table (y is the argument of each autocorrelation function here, the different intermediate aggregate tables C0 to C5 being indexed by x ) : ∀ whole ∈ [0; length Cx ], g y x = ∑ i = 0 longueur Cx − y − 1 C x i + y × C x i longueur Cx − y

[0186] In practice, this operation consists of summing the term-by-term multiplication of the vector C x ( i ) with a shifted version of the y value, C x ( i + y ).

[0187] In substep 7.1.3, each intermediate autocorrelated aggregate table is averaged. This produces a corresponding set of averaged values.

[0188] Reconsidering the example of the autocorrelation functions g(y,x) above, the following formula is used, for example, to obtain the averaged values ​​in question: h x = ∑ j = 1 G x g j x G x with G x the number of values ​​in the vector to be averaged. In terms of numerical values ​​associated with the curves (and associated value vectors) C0 to C5 of the Fig. 8c , This gives us: x 0 1 2 3 4 5 h ( x ): -4,778 2,236 2,417 3,147 3,953 3,95

[0189] Thus, an intermediate aggregate table whose corresponding mean value is the maximum among the mean values ​​is selected as a third candidate aggregate table according to the periodicity optimization criterion. Therefore, the third candidate aggregate table corresponds to the intermediate aggregate table exhibiting the greatest regularity (in the sense of autocorrelation) among the different tables in the set of intermediate aggregate tables. In this way, the results of an analysis based, for example, on the implementation of a Fourier transform are improved, as the discontinuities of the analyzed table are minimized. In the example above, the third candidate aggregate table is thus the C4 curve corresponding to x= 4 .

[0190] Furthermore, when several intermediate aggregate tables have the same maximum averaged value among all averaged values, an intermediate aggregate table corresponding to the deletion of a minimum number of successive values ​​(when implementing value deletion in the third table associated with the second screw operation) is selected from among the intermediate aggregate tables in question as the third candidate aggregate table, or selected according to the periodicity optimization criterion. Thus, a maximum number of values ​​is obtained in the third candidate aggregate table, thereby allowing for a better analysis resolution of the table in question.

[0191] In step 7.2, the total number of values ​​in the third candidate aggregate table is tested, for example, by comparison to a predetermined threshold. For instance, the third candidate aggregate table is determined to be the third aggregate table when the total number of values ​​in the third candidate aggregate table exceeds the predetermined threshold. Thus, the number of values ​​in the third aggregate table is considered sufficient to achieve good resolution for the screw gun's fault analysis.

[0192] Alternatively, when the total number of values ​​in the third candidate aggregate table is less than the predetermined threshold, steps 4.3 (for example, after implementing steps 4.1 and 4.2 described above), 4.4, and 4.5.1 (according to any of the aforementioned implementations) are implemented again for a new screw tightening operation by the screwdriver. A new corresponding third table is thus generated. Based on this, step 7.1 of optimized aggregation (according to any of the aforementioned embodiments) is again applied to the third candidate aggregate table and the new third table. A new third candidate aggregate table is thus generated.Thus, when the number of values ​​in the third aggregated table is not sufficient to perform a detailed analysis of the screw gun's defects, a new optimized aggregation is implemented iteratively in order to obtain a sufficient number of values ​​for a detailed analysis of the results.

[0193] In some embodiments, the new third candidate aggregate table is subjected to a new test on the number of values ​​it contains according to a new implementation of step 7.2 as described above.

[0194] In some embodiments, step 7.2 is not implemented, and the analysis is systematically performed on the third candidate aggregate table obtained from the optimized concatenation of values ​​measured during a predetermined number of tightenings (e.g., two tightenings, three tightenings, etc.). In these embodiments, the third candidate aggregate table obtained after implementing step 7.1 a number of times corresponding to the predetermined number is systematically the third aggregate table.

[0195] Back to the Fig. 7The third aggregated table is analyzed in such a way as to provide at least one piece of information representative of a dispersion and / or a deviation from the screwing target, resulting from disturbances induced by the screwdriver. For example, such an analysis is implemented according to the technique described above in relation to the first (see steps 4.5.2, 4.5.3, 4.5.4, 4.5.5, 4.5.6, 4.6, according to any of the aforementioned implementations) and second implementations of the process in paragraph 5.3. 4.2 Second embodiment of the invention

[0196] In relation to the Fig. 9 , A second embodiment will now be described. Fig. 9presents a flowchart of the main steps for implementing a process to control the quality level of screwdriving of a screwdriver, relative to a predetermined screwdriving target, according to a fourth embodiment. Some steps of this second embodiment are further illustrated by the curves shown on the Fig. 10a , Fig. 10b and Fig. 10c .

[0197] The first embodiment described above gives very good results. However, calculating the autocorrelation is quite computationally intensive and memory-intensive. The second embodiment reduces the computational load at the cost of slightly less satisfactory results. Such results are nevertheless sufficient in many practical cases.

[0198] As in the first embodiment described above, in the second embodiment, steps 4.3 (for example, after implementation of steps 4.1 and 4.2), 4.4, and 4.5.1 are implemented for at least one first and one second screw tightening operation by the screwdriver. This results in two initial value tables associated with the first and second tightening operations, two second value tables associated with the first and second tightening operations, and two third value tables associated with the first and second tightening operations.

[0199] In order to perform a detailed analysis of the measurements taken, for example based on a sufficient number of measurement points, during step 9.1, an optimized aggregation of the third table associated with the first tightening and the third table associated with the second tightening is implemented. This optimized aggregation involves removing a number of values ​​from the third table associated with the second tightening, determined according to a periodicity optimization criterion. A candidate third aggregated table is generated after the implementation of step 9.1.

[0200] To do this, in substep 9.1.1, a correlation is calculated between, on the one hand, the third table associated with the first screwing and, on the other hand, the third table associated with the second screwing. A correlation function is thus produced.

[0201] Reconsidering the example of the curves (and associated value vectors) B1 and B2 of the Fig. 8bDescribed above and representing the third tables associated respectively with the first and second screwing, the following formula is used, for example, to obtain the correlation function in question (y is here the argument of the correlation function): g y = ∑ i = 0 longueur B 1 − y − 1 B 1 i + y × B 2 i longueur B 1 − y

[0202] According to this formula, the length of B1 must be less than or equal to the length of B2. In practice, it is always possible to swap the roles of B1 and B2 if necessary to satisfy this condition.

[0203] In substep 9.1.2, an argument value ymax (corresponding in practice to a rotation angle of the screwdriver) maximizing the correlation function g(y) is determined. The periodicity optimization criterion in this case corresponds to the removal, in the third table associated with the first screw operation, of a number of successive values, called the optimized number, which is a function of the argument value maximizing the correlation function g(y).

[0204] Furthermore, when several argument values ​​maximize the correlation function g(y), the optimized number depends on the maximum argument value among those that maximize the correlation function g(y). Thus, a maximum number of values ​​is obtained in the third aggregated table, allowing for a better resolution of the analysis of that table.

[0205] In some variants, the search for the argument value(s) that maximize the correlation function g(y) is limited to the interval y ∈ longueur B 1 2 ; longueur B 1 so as to remove at most only half of the values ​​from the third table associated with the first screwing.

[0206] In substep 9.1.3, a truncated version of the third table associated with the first screw and the third table associated with the second screw are concatenated to produce a candidate aggregate third table. The truncated third table results from removing, from the third table associated with the first screw, the optimized number of consecutive values ​​corresponding to maximum-amplitude arguments from among the values ​​in the third table associated with the first screw. In other words, it is the last values ​​of the third table associated with the first screw that are removed here.

[0207] For example, reconsidering the example of the curves (and associated value vectors) B1 and B2 of the Fig. 8b , The third candidate aggregate table corresponds to the curve (and associated value vector) Cymax defined by: C y max a = B 1 a si a ≤ longueur B 1 − y max B 2 a − longueur B 1 − y max sinon

[0208] The different corresponding curves, i.e. B1(a), B2(a-(length B1)-Ymax) and Cymax, are represented respectively on the Fig. 10a , Fig. 10b and Fig. 10c .

[0209] The present second embodiment also includes step 7.2 (according to any one of the aforementioned embodiments) of testing the total number of values ​​of the third candidate aggregate table and / or of analysis (according to any one of the aforementioned embodiments) as described above in relation to the first embodiment of the method according to the invention.

[0210] In some embodiments, step 7.2 is not implemented, and the analysis is systematically performed on the third candidate aggregate table obtained from the optimized concatenation of values ​​measured during a predetermined number of tightenings (e.g., two tightenings, three tightenings, etc.). In these embodiments, the third candidate aggregate table obtained after implementing step 9.1 a number of times corresponding to the predetermined number in question is systematically the third aggregate table.

Claims

1. A method for assisting in the maintenance of an industrial tool such as a screwdriver or a drill, implementing several rotatably movable components, the method comprising the following steps of: - obtaining (151) measurement data representative of an angle and / or a torque upon using said tool; - analysing (152) said measurement data, so as to determine at least one piece of quality data representative of possible disturbances induced for each of the components of a set of controlled components, delivering a signature of said tool comprising said quality data; - storing (153) said signature in a memory associated with said tool, and contactlessly readable at a short distance; - remotely reading (154) said signature in said memory, using a maintenance assist terminal; - identifying (155) a component requiring an intervention, from said signature; said method being characterised in that it comprises the following steps of: - obtaining, by said terminal, assistance information on said intervention to be performed, comprising three-dimensional information on said tool; - guiding the operator, by said terminal, for performing an intervention on a defective component, comprising: - shooting at least one image of said tool, using a camera mounted to said terminal; - displaying (156) an augmented reality representation, using said image(s) and said three-dimensional information, identifying said defective component and / or maintenance operations to be performed.

2. The method for assisting in the maintenance of a tool according to claim 1, characterised in that said step of obtaining assistance information comprises a step of connecting to a remote maintenance server, containing a set of information relating to said tool, referred to as the digital twin of the tool, and comprising at least one of the information belonging to the group comprising: - a three-dimensional representation of the tool, - an exploded view of the tool, - a data sheet of the tool, - a nomenclature of the tool, - a calibration report of the tool, - a maintenance history of the tool, - a theoretical signature of the tool, - an initial signature of the tool, - at least one previous signature of the tool.

3. The method for assisting in the maintenance of a tool according to claim 1, characterised in that said analysis step is implemented in a hub connected to said tool, receiving said measurement data from said tool, performing said analysis.

4. The method for assisting in the maintenance of a tool according to claim 1, characterised in that said signature comprises a plurality of frequency lines, and in that said identification step performs a comparison of the amplitude of each line with a predetermined threshold value.

5. The method for assisting in the maintenance of a tool according to claim 1, characterised in that said signature comprises a plurality of frequency lines, and in that said identification step implements an analysis of the course of the amplitude of each line between two signatures.

6. The method for assisting in the maintenance of a tool according to claim 1, characterised in that said analysis step takes an aggregation of measurement data corresponding to at least two screwing operations into account.

7. The method for assisting in the maintenance of a tool according to claim 1, characterised in that it comprises a step of predicting wear or defect of a component, by analysing a series of at least two signatures of said tool and / or a batch of similar tools and / or by comparing with predetermined threshold values.

8. A system for assisting in the maintenance of an industrial tool such as a screwdriver or a drill, said tool implementing several rotatably movable components, the system comprising at least one remote maintenance server and at least one maintenance terminal capable of communicating with said tool and with said server, said tool comprising an associated memory contactlessly readable at a short distance, containing at least one signature comprising quality data representative of possible disturbances induced by each of the components of a set of controlled components, determined from an analysis of said measurement data representative of an angle and / or a torque upon using said tool, said terminal comprising contactless means for reading said signature and means for connecting to said remote server, so as to obtain assistance information on an intervention to be performed on a defective component, comprising three-dimensional information on said tool, based on an analysis of said signature, and means for guiding the operator, for performing said intervention on a defective component, comprising means for shooting at least one image of said tool, using a camera mounted to said terminal and means for displaying an augmented reality representation, using said image(s) and said three-dimensional information, identifying said defective component and / or maintenance operations to be performed.

9. The system according to claim 8, characterised in that said remote maintenance server contains a set of information relating to said tool, referred to as the digital twin of the tool, and comprising at least one of the information belonging to the group comprising: - a three-dimensional representation of the tool, - an exploded view of the tool, - a data sheet of the tool, - a nomenclature of the tool, - a calibration report of the tool, - a maintenance history of the tool, - a theoretical signature of the tool, - an initial signature of the tool, - at least one previous signature of the tool.

10. A computer program product comprising program code instructions for implementing the reading, identifying, obtaining and guiding steps of the method according to any one of claims 1 to 7, when said program is executed on a maintenance assist terminal comprising a microprocessor and / or on a computer.