Method and device for diagnosing wear in an electrical switching unit, and an electrical unit equipped with such a device

Abstract

To provide a method and a device for diagnosing an electric unit which can be easily equipped to an electronic unit and have a negligible sensitivity to temperature fluctuations.SOLUTION: The present invention relates to a method for diagnosing the state of wear of an electric switching unit, the method having a phase for monitoring an electric unit, the monitoring phase being loaded in advance, and the method using learning data showing the type of the electric unit and initialization data (CRP) corresponding to the unit to be monitored and stored in a memory during the initialization phase. The monitoring phase includes: measurement and acquisition 40 of a measurement curve (CM) of the electronic unit in an open state; determination 41 of the value of a local descriptor with a negligible sensitivity to the temperature of the measurement curve as the function of the value of the measurement curve, the initialization data, and learning data (CRU); and determinations 43 to 45 of a global status class as the function of the positioning.SELECTED DRAWING: Figure 16

Description

[Technical field]

[0001] The present invention relates to a method for diagnosing the wear state of an electrical switching unit, comprising a phase for monitoring the electrical unit.

[0002] The invention also relates to diagnostic devices and units for carrying out the method. [Background technology]

[0003] Methods and devices for diagnosing the condition of an electrical unit typically monitor one or more electrical characteristics by comparing them to a range of values. These characteristics are typically the voltage, current or time reference of a point of interest of an electrical signal. A diagnostic method of this kind is disclosed in patent application EP2584575.

[0004] Other methods determine the wear of the contacts of an electric unit, such as a contactor, by evaluating the time between two events or the duration of an event, for example the time between a command for opening or closing the contacts and the effective moment of said opening or closing. Patent application WO03054895 describes a particular way of detecting the wear of the contacts of an electric unit as a function of the moment of appearance of a primary current after a control command.

[0005] Known diagnostic methods and devices provide satisfactory results, but are not easy to deploy on existing or already installed units. Generally speaking, interventions in the electrical unit are required. Moreover, there is a need to have a higher level of accuracy and certainty in order to monitor certain electrical units, such as contactors.

[0006] Another drawback of known devices is their sensitivity to temperature fluctuations: such devices also require temperature measurements to complement the assessment of wear of the electrical contacts. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Patent application EP2584575 [Patent Document 2] Patent application WO03054895 Summary of the Invention [Problem to be solved by the invention]

[0008] The object of the invention is a method and a device for diagnosing an electric unit, which provides results with improved certainty, is easy to implement on the electric unit, and has little sensitivity to temperature variations. [Means for solving the problem]

[0009] A method for diagnosing the wear state of an electrical switching unit according to the invention, comprising a phase for monitoring said electrical unit, - The above monitoring phase is - preloaded learning data corresponding to the product type representing said electric unit; - initialization data corresponding to the unit to be monitored and stored during the initialization phase; Using And, - The above monitoring phase is - measuring and obtaining a measurement curve when said electrical unit is opened; - determining values ​​of local descriptors from the measurement curve as a function of values ​​of the measurement curve, the recorded initialization data and the first loaded learning data, at least one of the local descriptors being a descriptor that is substantially insensitive to differences in temperature and that is dependent on a point of interest of the measurement curve; - determining, for the second training data, a positioning of values ​​of a local descriptor of said measurement curve that has little sensitivity to temperature; - determining a global status class for said second training data as a function of the values ​​of the local descriptors and of said positioning of the measurement curve; Equipped with.

[0010] According to one preferred embodiment of the invention, the first descriptor corresponds to a curved surface or to an integral starting at a first point of the curve, the measurement depends on a first noteworthy point of said curve, the measurement corresponds to a first change in the direction of the variation of the signal or in the direction of change of the sign of the derivative of said measured signal, said first noteworthy point representing the start of a movement of a movable part of an electric unit.

[0011] Preferably, the integral of the signal starts at a time t1 in the range 0-10 ms before the occurrence of said first point of interest.

[0012] Advantageously, the signal integral ends when the signal of the coil voltage curve is at or close to zero volts, for example less than 2 volts in absolute value. Preferably, the signal integral lasts for 30ms to 40ms.

[0013] According to a preferred embodiment of the invention, the second descriptor corresponds to a positioning or difference of the second point of interest relative to an initial reference curve or line for a second point of interest of the new unit defined over a temperature range, the second point of interest corresponding to a variation in value such as a change in sign of the variation of a signal on a measurement curve, the signal decreasing and then increasing in relative value around the second point of interest.

[0014] Advantageously, at least one descriptor that has little sensitivity to temperature has a value difference of less than 5% between -20°C and 80°C.

[0015] Preferably, the loaded learning data represents a descriptor behavior curve which represents a predefined characteristic of said electrical signal during the life of said electrical unit.

[0016] Preferably, said descriptor behavior curves are loaded in a straight line segment definition format in order to reduce the memory space required by said training data.

[0017] Advantageously, the values ​​of said pre-loaded descriptor behaviour curves are associated with a condition or wear class of the unit to be monitored.

[0018] According to a preferred embodiment, the method comprises the steps of: - a pre-learning phase for determining, during a number of predefined operations, training data representative of at least two descriptor behavior curves representative of predefined characteristics of said electrical signal during the life of said electrical unit; - storing the learning data in a memory; - Loading the above learning data, - using the learning data from said phase for monitoring said electrical unit; Equipped with.

[0019] A device according to the invention for diagnosing the wear state of an electrical switching unit connected to an electromagnetic coil for actuating the contacts of said electrical switching unit comprises a processing circuit for implementing a diagnostic method such as that defined herein above.

[0020] Preferably, the processing circuitry comprises a local module near said unit which is connected to an external processing module remote from said unit.

[0021] An electric unit according to the invention, comprising high-power electrical contacts actuated by a control electromagnetic coil, comprises a device for diagnosing the wear state of an electrical switching unit connected to said electromagnetic coil for actuating said contacts, and for carrying out a diagnostic method such as that defined herein above.

[0022] Other advantages and features will become more clearly apparent from the following description of particular embodiments of the invention, given by way of non-limiting example and illustrated in the accompanying drawings, in which: [Brief description of the drawings]

[0023] [Figure 1] FIG. 1 is a general view of an electric unit equipped with a monitoring device according to one embodiment of the present invention. [Diagram 2] FIG. 2 illustrates a learning phase of a method according to one embodiment of the invention. [Diagram 3] FIG. 2 illustrates a learning phase of a device and method according to one embodiment of the present invention. [Figure 4] FIG. 2 illustrates an initialization and processing phase of a device and method according to one embodiment of the present invention. [Diagram 5] FIG. 2 shows measurement curves of the signal during opening of an electric unit to be monitored for different temperatures of operation. [Figure 6] FIG. 2 illustrates a first descriptor for a device and method according to one embodiment of the present invention. [Figure 7] FIG. 2 shows a curve of the variation of a first descriptor as a function of temperature and a curve of points of interest for a device and method according to one embodiment of the present invention. [Figure 8] FIG. 2 shows a curve of behavior of a first descriptor generated during a learning phase and used in a device and method according to one embodiment of the invention. [Figure 9] FIG. 2 illustrates modeling of a second descriptor for a device and method according to one embodiment of the present invention. [Figure 10] FIG. 2 illustrates modeling of a second descriptor for a device and method according to one embodiment of the present invention. [Figure 11] FIG. 2 shows a curve of the variation of a second descriptor as a function of temperature for a device and method according to one embodiment of the present invention, as well as a curve of points of interest. [Figure 12]FIG. 4 shows a curve of behavior of a second descriptor generated during a learning phase and used in a device and method according to one embodiment of the present invention. [Figure 13] FIG. 2 shows signal measurement curves for a unit in good or new condition and for a rubbed unit. [Figure 14] FIG. 2 illustrates steps of the initialization phase of a method according to one embodiment of the invention. [Figure 15] FIG. 2 illustrates a monitoring phase of a method according to one embodiment of the invention. [Figure 16] FIG. 2 illustrates steps of the monitoring phase of a method according to one embodiment of the invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0024] In FIG. 1, a contactor-type electric unit 1 comprises one or more high-power electric contacts 2 for supplying power or for interfering with the electrical supply to a load 3. The electric contacts are controlled by an electromagnetic coil 4. A control circuit 5 controls the electromagnetic coil 4 to close or open the contacts 2. The circuit 5 also controls the current flowing in the coil during the surge phase at the time of closing the contacts and during the phase to maintain the contacts closed with reduced energy and control current.

[0025] A device 10 for diagnosing the wear of a unit 1 is associated with or forms part of an electric unit. This device 10 receives a signal representative of an electrical quantity, such as the voltage or current on a coil 4 for controlling a contact. Advantageously, the device 10 is connected in parallel with the coil to receive the voltage signal generated by the coil when the contact is opened. A signal from the current flowing in the coil may also be used. However, in a preferred embodiment, the voltage signal is advantageously more stable and easier to use. The device 10 thus comprises an analog-to-digital converter 6 connected to the coil 4, a diagnostic module 7 connected to the converter for receiving the signal and for processing the diagnosis of the electric unit, as well as a device 8 for signaling the state of the electric unit, in particular the level of wear of the contacts. The diagnostic module 7 may also be connected to another external or separate processing module 9 and process the data with it. In this case, the diagnostic device is in two or more parts. It is also possible that the external part is common to several devices or is centralized. The processing module 9 may provide remote signaling of the state of the unit. It will be appreciated that communication between the module and the device may take the form of wired or wireless communication.

[0026] Thus, monitoring and diagnosis of wear of an electric unit such as a contact is preferably carried out by analysis of the voltage on the coil when said contact is opened, this voltage being representative of the speed of movement of the moving part of the contact, in which case wear of the contacts causes a reduction in the speed of movement of the moving part of the electric unit.

[0027] The diagnosis of the wear condition of an electric unit comprises a pre-learning phase carried out on a large number of products during the life cycle of said unit. Advantageously, the learning can also be carried out by numerical simulation in order to determine the relationship between the behavior of the descriptors over time, in particular D1, D2, and the quantitative characteristics of the state of the unit, such as the wear of the contacts. This learning phase allows learning data to be obtained, which are first stored in a memory, modelled and then loaded into each unit to be monitored in order to monitor its respective wear. Figure 2 illustrates the learning phase 11 and the storage 12 of the learning data.

[0028] During the learning phase, measurement curves are acquired during the opening of the unit in a wear cycle. The measurement curves represent electrical signals lasting several tens of milliseconds, for example 30 ms to 50 ms for an average size electrical unit. These durations can vary greatly depending on the size and type of unit. The signals are typically sampled using several hundred samples, for example 80 to 500 samples, although different numbers may also be used, depending on the computing power of the processor used in the diagnostic device.

[0029] During the learning phase, the electric unit undergoes a quantity of operations that allows the life of the unit to be observed. In the case of contacts, the number of operations reaches, for example, 800000, but other values ​​may also be used. Since the amount of data may be enormous, the behavior over time of the characteristics of the electric unit is defined by descriptors D1, D2 that have little sensitivity to temperature variations, these descriptors D1, D2 being associated with the electrical characteristics and with curves of the behavior CD1, CD2 of said descriptors during the life of the electric unit. In order to further reduce the size of the learning data that is subsequently loaded into the device of the unit to be diagnosed, the curves of the behavior CD1, CD2 of the descriptors are recorded in the form of data representing straight line segments of the a+bx type. The curves of the behavior over time may have several successive segments of a straight line that define the life of the unit. The set of curves of the behavior over time of the descriptors CD1, CD2 forms a space of several dimensions over the life of the product. Parts, some of the curves or values ​​of the descriptors are associated with the state or wear class of the electric unit. In terms of the number of operations of the unit, the measurement curve is not necessarily acquired at each opening. The acquisition of the measurement curve can be regularly spaced by intervals of the number of operations, or in a more relevant manner according to the variation of the unit over time. For example, the number of operations can be more frequent towards the end of the life and may be widely spaced at the beginning of the life.

[0030] The learning phase may also provide a wear reference curve CRU, which is taken from the measurement curve at the end of the life of the unit used for learning. The wear reference curve CRU may be a curve averaged over several measurement curves and / or over several worn units.

[0031] Thus, at the end of the learning phase, learning data such as the curves of the behavior of the descriptors CD1, CD2 and possibly the wear reference curve CRU are first stored in memory, modelled and then loaded into each unit for the respective wear monitoring. This learning data is common to all monitoring devices associated with the same type of unit. Figure 3 illustrates phase 13 for loading the pre-stored learning data into the unit to be diagnosed.

[0032] In a first preferred embodiment, the diagnostic method according to the invention comprises: - a pre-learning phase 11 for collecting, during a number of predefined operations, learning data representative of at least two descriptor behavior curves representative of predefined characteristics of said electrical signal over the life of said electrical unit; - storing the learning data CD1, CD2 in a memory 12; - Loading 13 of the above learning data CD1 and CD2, - using the learning data from phase 15 to monitor said electrical units; Equipped with.

[0033] In a preferred embodiment, the diagnostic method according to the invention also comprises an initialization phase 14 for determining initialization data comprising a specific reference curve CRP for the product to be controlled, and the use of said initialization data in a monitoring phase 15 of said electric unit.

[0034] During the monitoring phase, local descriptors DL1, DL2, specific to each unit to be monitored and with little sensitivity to temperature variations, are determined during the acquisition of the measurement curves CM. These descriptors are of the same type as those used to generate the global curves CD1, CD2 of behavior of the descriptors with little sensitivity to temperature variations. In a preferred embodiment with at least two descriptors D1, D2 with little sensitivity to temperature variations, these descriptors each result in a global curve CD1, CD2 of behavior previously generated, recorded and loaded, and two local descriptors DL1, DL2, specific to each unit, each resulting from a measurement curve during the opening of the electrical unit.

[0035] FIG. 5 shows measurement curves CM, namely CM-25, CM0, CM25, CM50, CM75, CM100, which represent the electric voltage signal on the control coil at the opening of the electric unit at various operating temperatures for -25, 0, 25, 50, 75 and 100 degrees Celsius, respectively. Such curves can be measurement curves CM during the monitoring phase of the electric unit as well as during the learning phase for generating curves of the behavior of the descriptor. On these curves, the first noted point 20 corresponds to a first change in the direction of variation of the voltage of the curve CM with a relative increase and then a decrease in value. The point 20 is also identified as the first reciprocal of the derivative of the signal. Subsequently, the second noted point 21 corresponds to a second change in the direction of variation of the voltage of the curve CM with a relative decrease and then an increase in value. At the second noted point 21 there is a second reciprocal of the derivative of the signal of the measurement curve CM.

[0036] FIG. 6 shows a first descriptor D1 for a device and method according to an embodiment of the invention. This first descriptor D1 corresponds to an integral or integral starting on the surface 22 of the curve CM or on a first point 23 of the measurement curve CM. Said first point 23 depends on a first noted point 20 of the measurement curve CM, which corresponds to a first change in the direction of the signal variation or in the direction of the change of the sign of the derivative of said measurement signal. The first noted point 20 represents the start of the movement of the mobile part of the electric unit. The first descriptor D1 also represents the magnetic flux at the moment of the release, in other words the moment when the spring force from the pressure of the contacts is greater than the magnetic force, which also corresponds to the start of the opening of the mobile part of the unit. This descriptor D1, whose start is determined by an integral depending on the first noted point 20, is therefore very sensitive to the wear of the contact pads and very insensitive to temperature variations. The descriptor D1 is essentially based on the behavior of the magnetic flux depending on the wear of the contacts. Preferably, the integral 22 of the signal starts at a time t1, which is in the range of 10 ms to 0 ms before the occurrence of the first point of interest 20 at a time t2 depending on the dynamic behavior of the electric unit. In FIG. 6, the integral of the signal ends at the end tf of the acquisition of the signal of the curve CM. In a preferred form of operation, the integral 22 ends when the coil voltage signal is substantially at or close to zero volts, for example less than 2 volts in absolute value. It may, however, be limited to a fixed duration, for example the integration of the signal may last 30 ms to 40 ms.

[0037] 7 shows curves of variation as a function of temperature of the first descriptor D1 for a device and method according to one embodiment of the invention, as well as the curves of only the second point of interest 21. The comparison shows that the descriptor D1 determined by the integral 22 has an accuracy of less than ±2% over the temperature range from -20°C to 100°C, while only the point 21 can reach ±6%. Overall, from 0 to 100°C, the descriptor D1 is ten times less sensitive to temperature variations than the descriptor based only on the point 21.

[0038] FIG. 8 shows a curve of the behavior CD1 of a first descriptor D1 generated during a learning phase and used in a device and method according to one embodiment of the invention.

[0039] Figures 9 and 10 show modeling of a second descriptor D2 for a device and method according to one embodiment of the invention. The second descriptor D2 corresponds to a positioning or difference 24 of the second point of interest 21 of the new unit, P-25, P25, P85, relative to an initial reference curve or line 25 defined for a second point of interest 21 of the new unit, P-25, P25, P85, defined over a temperature range, e.g. -25 to 85°C. Said second point of interest 21 corresponds to a variation in value, such as a change in the sign of the variation of the signal on the measurement curve CM, said signal decreasing and then increasing in relative value around the second point of interest. In Figures 9 and 10, the second points of interest P-25, P25, P85 of the new unit are defined by three values ​​of temperature, respectively -25°C, 25°C, and 85°C. By using a new unit-dependent baseline 25 over the temperature range, and by verifying the positioning of the point of interest 21 of the unit to be monitored on this baseline, the descriptor D2 representing this difference 24 becomes relatively insensitive to temperature since the effects of temperature fluctuations are automatically compensated for.

[0040] In FIG. 9, an overview of the acquisition of the measurements, and in FIG. 10, a detailed view, the curves CM-25, CM25 and CM85 correspond to the curves during the learning for the various temperatures of -25° C., 25° C. and 85° C., respectively. The points of interest 21 of the new units P-25, P25, P85 in the learning phase for the various temperatures of -25° C., 25° C. and 85° C., respectively, allow a reference line 25 to be defined. This line is preferably linear, but may also take various, more complex forms. The difference 24 results in a second descriptor D2 by comparing the value of a second point of interest 21 of a unit to be monitored that may be worn out with the reference line 25. In FIG. 10, at time t3, the second point of interest 21 of the worn unit is detected, which is then compared with the value 26 of the reference curve 25 in the vicinity of the same time t3. At time t3, the value of the difference 24 allows the value of the descriptor D2 to be determined.

[0041] 11 shows a curve of the variation of the second descriptor D2 as a function of temperature for a device and method according to one embodiment of the invention, as well as the curve of only the second point of interest 21. This comparison shows that the descriptor D1 determined by the difference of the point 21 of the unit to be monitored against a baseline of the point 21 of the new unit previously recorded at various temperatures has an accuracy of less than ±1% over the temperature range from -20°C to 100°C, while only the point 21 can reach ±8%. This form of determination of the descriptor D2 provides a self-compensation of the temperature.

[0042] The descriptors D1 and D2 that are little sensitive to temperature, to temperature differences or to temperature variations are selected or chosen to have a value difference of less than 5% between -20°C and 80°C. Thus, according to one embodiment of the invention, at least one of the descriptors D1, DL1, D2, DL2 that are little sensitive to temperature has a value difference of less than 5% between -20°C and 80°C.

[0043] FIG. 12 shows the curve of behavior of a second descriptor D2 generated during the learning phase and used in a device and method according to one embodiment of the invention.

[0044] The curves CD1 and CD2 represent the variation over time of the descriptors D1 and D2, which originates from a learning curve established during a learning phase by a large number of measurements. These numerous values ​​are then modelled as straight line segments, to result in curves of the behaviour of the descriptors CD1 and CD2, which are loaded into the unit to be monitored. These curves, shown in figures 8 and 12, comprise, for example, several segments. The number of segments is not limited and is advantageously in the range of 1 to 20. However, other forms of modelling are possible, for example using stepped or polynomial techniques. The loaded learning data thus represent curves of the behaviour of the descriptors CD1, CD2, which represent predefined characteristics of said electrical signal over the life of the electrical unit.

[0045] At the beginning of the curves CD1, CD2 of the behavior of the descriptors the descriptors are associated with the global class CL1 of units that are new or in the process of being run in, while at the end of the curves the descriptors are associated with the class CL3 of worn out units. Between classes CL1 and CL3 the descriptors are associated with the class CL2 of units in normal operation.

[0046] The use of at least two or more descriptors D1, D2 allows for a greater accuracy or certainty of detection of the wear state of the electric unit to be monitored. The descriptors described herein above depend on electrical quantities. However, other descriptors may also be used by combining electrical quantities with environmental quantities such as temperature or the horizontal or vertical position of the unit. Such quantities may be used to select a descriptor according to the environment. The curves of the descriptors may also be selected according to the type of use of the electric unit. Parameter settings that allow the type of product or electric unit and the use of said unit to be defined are also loaded with the corresponding learning data during the loading phase.

[0047] When learning data is loaded into a device ready to carry out diagnostic monitoring of the wear of an electric unit such as a contactor, an initialization phase allows the storage in memory of initial data specific to the unit to be monitored. Among these initial characteristics there is advantageously the acquisition of a specific reference curve CRP. This reference curve is preferably stored in memory after a "break-in" of the unit, corresponding to a predefined number of first openings, and / or when the variations of the measurement curve CM have stabilized.

[0048] Figure 13 shows two examples of curves: a first measurement curve CM of a new unit, which may also be a specific reference curve CRP, and a measurement curve CM of a worn unit, which may also be a reference curve for a worn unit CRU that may be preloaded. Other measurement curves CM over the life of the unit will fall between these two reference curves CRP and CRU.

[0049] Fig. 14 shows a flow diagram representing an initialization phase 14 for determining said initialization data comprising said specific reference curve CRP of a product to be monitored that is installed on or associated with the monitoring device. Step 30 indicates the start of the initialization phase. The process 31 for determining a specific reference curve CRP comprises a processing step 32 in which successive measurement curves are obtained, and a step 33 for monitoring the stability of the curve. A curve CM is considered stable after a predefined number of measurements Nstable and / or when the measurement curve CM changes very little between two successive measurements CMn and CMn-1.

[0050] In step 34, the measurement curve CM is recorded as a specific reference curve CRP after the minimum number of operations where the difference between the curves after each opening is small. This curve represents the electrical characteristics of voltage or current at the beginning of the product's life after "break-in". It remains valid until the end of the product's life. In another embodiment, to further improve the stability of the curve to be recorded, the specific reference curve CRP can also be the average of several stable curves CM. Step 35 represents the monitoring phase of the product.

[0051] FIG. 15 shows that after a loading phase 13 in which learning data corresponding to the type of product representing said electric unit is pre-loaded, and an initialization phase 14 in which initialization data corresponding to said unit to be monitored are stored in memory, a monitoring phase 15 performs monitoring when the electric unit is opened.

[0052] In this particular embodiment, step 36 detects the opening of the electric unit. Once the opening is detected, step 37 performs a monitoring of the unit. This detection can be performed by a signal external to the device or by analysis of the measurement signal or its variations. The monitoring phase is preferably performed at each opening of the electric unit. However, it may also be performed in a more spaced manner, for example after a predefined number of openings. The monitoring may also depend on the life of the unit, for example the monitoring may be less frequent at the beginning of the life after a running-in phase and more frequent at the end of the life for higher accuracy.

[0053] FIG. 16 shows a flow diagram of the monitoring phase 15 along with details of the monitoring step 37 according to one embodiment of the present invention.

[0054] The monitoring phase is as follows: - in step 40, measuring and obtaining a measurement curve CM when said electrical unit is opened, - in step 41, the determination of values ​​of local descriptors DL1, DL2 of said measurement curve CM as a function of the values ​​of said measurement curve CM, the recorded initialization data CRP and the first loaded learning data, where at least one of said local descriptors DL1, DL2 is a descriptor that is substantially insensitive to temperature differences and depends on the points of interest 20 and / or 21 of said measurement curve CM; - in a step 42, determining, for the second training data CD1, CD2, a positioning 42 of the values ​​of the local descriptors DL1, DL2 that are less sensitive to differences in the temperature of the measurement curve CM; - in steps 43 to 45, determining a global status class for the second training data CD1, CD2 as a function of the above-mentioned positioning values ​​of the descriptors DL1, DL2 of the measurement curve CM.

[0055] Preferably, the determination of the global status class is - in step 43, determination of the point or the closest difference between the values ​​of the local descriptors DL1, DL2 of the measurement curve CM and the corresponding descriptor behavior curves CD1, CD2; - in step 44, the determination of an intermediate status class CLI as a function of the closest positioning of the descriptor on said descriptor behavior curve; - in step 45, the selection of said global class as a function of the results of said intermediate classes; - in step 46, signalling or communication of the global wear class of the unit.

[0056] In step 43, the determination of the closest difference can preferably be performed in a global manner, for example with a point defined by the values ​​of all the descriptors, but it may also be performed for each descriptor individually, or in a mixed or semi-global manner, where one part of the descriptors is grouped and another part is individual.

[0057] In step 46, the device 8 for signaling the state of the electric units may indicate the global wear class by means of a visual indication or a separate channel for the class value. However, the class values ​​may also be grouped, in particular the class of units being run-in and the class of cleaning operation may be signaled together in a grouped manner or on a single signal.

[0058] According to one embodiment of the present invention, a device for diagnosing the wear state of an electrical switching unit is connected to an electromagnetic coil for actuating the contacts of the electrical switching unit, which comprises a processing circuit for implementing the diagnostic method described herein above.

[0059] An electrical unit with high power electrical contacts actuated by a control electromagnetic coil according to the present invention comprises a device for diagnosing the wear condition of an electrical switching unit connected to said electromagnetic coil for actuating said contacts, for carrying out the diagnostic method described herein above.

[0060] The device and the method can be implemented permanently or temporarily in the unit to be monitored. They can also be implemented on already installed electrical units. In addition, certain steps of the method can be performed in a local manner close to the unit, while other steps may be performed remotely or in a centralized manner. For example, the acquisition of the curve CM may be local and the remaining processing may be performed remotely for more elaborate calculations. In this case, the processing tasks are shared between the local module 7 and the remote module 9 as shown in FIG. 1.

[0061] Several types of load or usage categories are identified during the learning and monitoring phase. The processing of the load types is preferably performed globally during the learning phase in a manner that takes into account different situations. For example, some units are used over their lifetime for different types of load. The behavior curves of the descriptors represent the curves over units that had different load or usage conditions.

[0062] The categories of use of electrical units such as electrical contacts are, inter alia: - Load type: non-inductive, induction, slip ring motor, squirrel cage motor; - Type of control command: power on, power off, start, brake, inch; - Type of application: distribution, heating, compressors, ventilation, lifts, pumps, various machines; and / or - A combination of several criteria Depends on.

[0063] It is also possible to identify the type of load or the particular use and characterize the method accordingly, e.g., capacitive load, etc.

[0064] The classes of wear for each use category are defined by descriptors. They allow the condition of the product to be characterized and provide an indication of the product's lifespan.

[0065] A non-exhaustive list of the main classes is: - Class 1: new or break-in products; - Class 2: Products in use, - Class 3: worn products, - Class 4: Products at the end of their life It could be.

[0066] In the preferred embodiment described herein above, the assessment of the wear of the contacts is advantageously carried out on the basis of a measurement of the voltage on the coil of the contactor, however, other signals may also be used, in particular signals representative of the current flowing in the coil. [Explanation of symbols]

[0067] 1 Electrical unit 2 (high power electrical) contacts 3. Load 4. Electromagnetic Coil 5 (Control) Circuit 6 Analog-to-Digital Converter 7 Diagnostic Module, Local Module 8 Device (for signaling the level of contact wear) 9 Processing Module, Remote Module 10. Device for diagnosing wear of unit 1 13 Loading Phase 14 Initialization Phase 15 Monitoring Phase 20 First Point of Interest 21 Second Point of Interest 22 Integral 23 First Location 24 difference 25 Reference curve 26 Value

Claims

1. A method for diagnosing a wear state of an electrical switching unit (1), comprising a monitoring phase (15) for monitoring said electrical switching unit (1), said monitoring phase (15) - preloaded learning data (CRU, CD1, CD2) corresponding to the type of product representing said electrical switching unit (1); using initialization data (CRP) corresponding to said electrical switching unit (1) to be monitored and stored in a memory during an initialization phase (14); And, said monitoring phase (15) - measuring and obtaining (40) a measurement curve (CM) when said electrical switching unit (1) is opened; a step (41) of determining values ​​of local descriptors (DL1, DL2) from said measurement curve (CM) as a function of values ​​of said measurement curve (CM), of stored initialization data (CRP) and of first learning data (CRU), at least one of said local descriptors (DL1, DL2) being a descriptor that is substantially insensitive to temperature differences and that is dependent on points of interest (20, 21) of said measurement curve (CM); a step (42) of determining, relative to second training data (CD1, CD2), the positioning of said values ​​of local descriptors (DL1, DL2) that are less sensitive to the temperature of said measurement curve (CM); - a step (43-45) of determining a global status class as a function of the orientation of the determined values ​​of local descriptors (DL1, DL2) relative to the second training data (CD1, CD2) and as a function of the values ​​of the measurement curve (CM), 2. A method according to claim 1, wherein said at least one descriptor with little sensitivity to temperature (D1, D2) and said local descriptors (DL1, DL2) have a value difference of less than 5% between -20°C and 80°C.

2. 2. The diagnostic method according to claim 1, characterized in that the first descriptor (D1) and the first local descriptor (DL1) correspond to a curved surface or to an integral (22) starting at a first point of the measurement curve (CM), the measurement curve (CM) being dependent on a first point of interest (20) of the measurement curve (CM), the measurement curve (CM) corresponding to a first change in the direction of variation of an electrical signal or in the direction of change of the sign of the derivative of the electrical signal, the first point of interest (20) representing the start of a movement of a moving part of the electrical switching unit (1).

3. 3. A diagnostic method according to claim 2, characterized in that the integral (22) of the electrical signal starts at a time t1 lying in the range 0-10 ms before the occurrence of the first point of interest (20).

4. 4. The diagnostic method of claim 3, wherein the integral (22) of the electrical signal ends when the electrical signal of a coil voltage curve (CM) is at or close to zero volts.

5. 5. The diagnostic method according to claim 4, characterized in that the integral (22) of the electrical signal has a duration of 30 ms to 40 ms from start to finish.

6. 6. The diagnostic method according to claim 1, characterized in that the second descriptor (D2) and the second local descriptor (DL2) correspond to a difference (24) between the second point of interest (21) of the measurement curve (CM) at time t3 and a value (26) of an initial reference curve or line (25) at time t3, the initial reference curve or line (25) being for the second point of interest (21) of a new unit defined over a temperature range, the second point of interest (21) corresponding to a variation in value, such as a change in the sign of the variation of an electrical signal on the measurement curve (CM), the electrical signal decreasing and then increasing in relative value around the second point of interest (21).

7. 7. Diagnostic method according to any one of claims 1 to 6, characterized in that the learning data (CRU, CD1, CD2) represent curves of behavior of descriptors (D1, D2) representing predefined characteristics of the electrical signal during the life of the electrical switching unit (1).

8. 8. Diagnostic method according to claim 1, characterized in that the learning data (CRU, CD1, CD2) representing the curves of behavior of the descriptors (D1, D2) are loaded in the form of data representing straight line segments of the a+bx type in order to reduce the memory space required by the learning data (CRU, CD1, CD2).

9. 9. Diagnostic method according to any one of claims 1 to 8, characterized in that the values ​​of the learning data (CRU, CD1, CD2) representative of the curves of behavior of the descriptors (D1, D2) are associated with a condition or wear class of the electrical switching unit (1) to be monitored.

10. a pre-learning phase (11) for collecting, during several predefined operations, learning data (CRU, CD1, CD2) representative of at least two curves of the behavior of descriptors (D1, D2) representative of predefined characteristics of said electrical signal over the life of said electrical switching unit (1); storing said learning data (CRU, CD1, CD2) in a memory (12); - loading (13) of said training data (CRU, CD1, CD2), - a step (42, 43) of using said learning data (CRU, CD1, CD2) in said monitoring phase (15) for monitoring said electrical switching unit (1).

11. A device (10) for diagnosing the wear state of an electrical switching unit (1) connected to an electromagnetic coil (4) for actuating contacts (2) of said electrical switching unit (1), said diagnostic device comprising a processing circuit (10) for implementing the diagnostic method according to claims 1 to 10.

12. 12. The device according to claim 11, wherein the processing circuit (10) comprises a local diagnostic processing module (7) near the electrical switching unit (1) which is connected to an external processing module (9) remote from the electrical switching unit (1).

13. 1. An electrical switching unit (1) comprising a high-power electrical contact (2) actuated by a control electromagnetic coil (4), characterized in that it comprises a device (10) for diagnosing a wear state of the electrical switching unit (1) connected to the electromagnetic coil (4) for actuating the contact (2) and for implementing the diagnostic method according to any one of claims 1 to 10.

Images