Predictive maintenance process

Abstract

Predictive maintenance method for an electrical device (14) supplied by a cable (16) having at least two conductors, having first and second electrical conductors (18i) electrically insulated from each other and in which first and second electrical currents, having first and second intensities, respectively, flow, said method comprising a real-time repetition of an updated cycle comprising the following successive steps: 1) at an updated instant, measurement of the first and second intensities, and determination, by computer, of first and second updated values, for at least one attribute of the first and second intensities, respectively; 2) by computer, calculation of - at least one indicator as a function of the first updated value and a first "previous" value determined, for the first intensity, prior to the updated cycle, for said at least one attribute;and - at least one indicator based on the first and second updated values, without using said previous values, then determination of a failure risk based on the difference between the value of each indicator and a corresponding reference value; 3) based on the failure risk, generation, by computer, of an alert and, preferably, intervention on the device. Figure for the abbreviation: Fig. 2;

Description

Title of the invention: Predictive maintenance method technical field

[0001] The present invention relates to a predictive maintenance method and a monitoring device implementing such a method to monitor an industrial electrical device. Previous technique

[0002] An industrial installation typically uses a wide variety of electrical equipment of varying ages and loads. The failure of any equipment can have very damaging economic consequences. Maintenance operations are therefore scheduled to regularly check the condition of the equipment. However, these often require a production shutdown.

[0003] In particular, motors generally need to be disassembled in order to undergo vibration analysis in the laboratory. This analysis makes it possible to assess the level of aging and, consequently, to decide whether to scrap a motor or refurbish it.

[0004] There is a permanent need for a solution enabling simplified and more efficient predictive maintenance.

[0005] One object of the present invention is to meet, at least partially, this need. Summary of the invention

[0006] According to the invention, this goal is achieved by means of a predictive maintenance method for an electrical device powered by a cable comprising at least two conductors, comprising first and second electrical conductors electrically insulated from each other and in which first and second electrical currents, having first and second intensities respectively, flow, said process involves a real-time repetition of an "updated" cycle comprising the following successive steps: 1) at an updated instant, measurement of the first and second intensities, and determination, by computer, of first and second updated values, for at least one attribute of the first and second intensities, respectively, preferably for at least the amplitude or a phase shift of said intensities; 2) by computer, calculation of - at least one indicator based on the first updated value and a first "previous" value determined, for the first intensity, prior to the updated cycle, for said at least one attribute, preferably determined during a previous cycle of steps 1) to 3); and - at least one indicator based on the first and second updated values, without using said previous values, i.e. determined, for the first and / or second intensity, prior to the updated cycle, for said at least one attribute, then determination of a failure risk based on the difference between the value of each indicator and a corresponding reference value; 3) depending on the risk of failure, computer-generated alert and, preferably, intervention on the device.

[0007] An indicator can be a function of, or even equal to, the difference between the first and second updated values.

[0008] An indicator can be a function of, or even equal to, the difference between one of the first and second updated values ​​and the corresponding previous value.

[0009] As will be seen in more detail later in the description, a method according to the invention makes it possible to detect slow changes in the operation of the device, but also sudden changes in this operation.

[0010] Advantageously, the method makes it possible to prevent a failure of the device, and therefore to avoid the disruption resulting from the occurrence of such a failure. Furthermore, it can be implemented in situ, on the device while it is in operation.

[0011] A method according to the invention may further include one or more of the following optional features: - the said previous values ​​are measured more than 1 hour, more than 1 day, more than 1 week, more than 1 month, and / or less than six months before the updated time; - the first and second intensities are out of phase in normal operation of the device, preferably by 180° (two-phase current) or 120° (three-phase current), and at least one indicator measures an imbalance between said intensities, in particular a deviation from the phase shift in normal operation; - alternatively, the first and second intensities are in phase, as in a multi-conductor cable comprising several conductors insulated from each other and supplied by the same phase; - the device is a motor, a compressor or an extruder, said device preferably having a rated power greater than 1 kW; - in step 1), the first and second intensities are measured using magnetic field sensors arranged around the cable, configured to measure, for each of the conductors, at least one component of the magnetic field produced by the current flowing in the conductor; - The aforementioned first and second intensities are measured by following steps consisting of: 1. place a plurality of said magnetic field sensors around said cable, in at least one location of said cable; E2. simultaneously measure, for each of said conductors, at least one component of the magnetic field produced by the current flowing in said conductor, by means of said plurality of magnetic field sensors; E3. for each of said conductors, determine the angle between said conductor and the nearest magnetic field sensor of said plurality of magnetic field sensors, said angle being defined with respect to the center of said cable and by assimilating said conductors and said sensors to points; the current intensities in said conductors being related to the components of the magnetic field measured by the relation B = kMI where B is the matrix of said components of the magnetic field, I is the matrix of said current intensities, M is a matrix comprising a plurality of proportionality coefficients depending on said angles between said conductors and said magnetic field sensors and k is a predetermined coefficient, E4. Calculate the inverse M1 of the matrix M, so as to deduce the values ​​of the said current intensities I = (po / 2ir).M'.B, where po is an equivalent magnetic permeability which takes into account the presence of insulating materials in said cable; - each sensor is powered by a respective battery, a gauge preferably transmitting to a computer implementing steps 2) and 3), a charge level of said battery, the computer taking into consideration or not an intensity supplied by said sensor depending on whether the charge level exceeds or does not exceed a threshold charge level; - at least one of the said reference values ​​is determined, by computer, according to the operating mode of the device; - in step 2), at least one indicator is calculated by computer as a function of the difference between the first and second discounted values, and / or as a function of the difference between the first discounted value and the first previous value; - in step 2), at least one indicator chosen from the intensity and phase shift between the different currents is calculated by computer; - the first and second electric currents are alternating currents and at least one indicator is a function or consists of a difference between the phases of the intensities of said currents or a difference between the amplitudes of the intensities of said currents; - at least one indicator and / or at least one reference value is / are determined, by computer, by a statistical processing of historical data taken from historical devices identical to said device powered by the cable, the statistical processing being configured to establish a correlation between indicator values ​​and the occurrence of a device failure; - in step 2), we determine, by comparing the value of each indicator with a corresponding reference value, a predicted time interval before a failure occurs; - the reference value is stored in a memory accessible to the computer or determined, by computer, according to the operating conditions of the device at the updated time; - the process is implemented while the device is in operation; - in step 3), the first and second updated intensities are displayed in the same graph, on a computer screen, the graph showing, preferably in superposition, the temporal evolutions of the first and second intensities; - the cable comprises first, second and third electrical conductors that are electrically insulated from each other, In step 1), the first, second, and third intensities of the first, second, and third electric currents flowing in the first, second, and third conductors, respectively, are measured. In step 2), the risks of failure are determined for each pair of two intensities chosen from the first, second and third intensities, that is to say by using, for each pair of two intensities, the updated and previous values ​​of said intensities.

[0012] The invention also relates to the use of a method according to the invention for: - detect imbalances between the phases of the current intensities flowing in the different conductors of the cable, particularly in a cable supplying, in three-phase, a motor or a compressor, and / or - detect a drop in current in a cable, particularly in a cable supplying an extruder.

[0013] The invention also relates to a monitoring device for implementing a method according to the invention, said device comprising: - at least two sensors, comprising first and second sensors capable of measuring first and second intensities; - a computer configured to receive, from the first and second sensors, the measurements of the first and second intensities respectively, and including code instructions for the implementation of steps 2) and 3) and for the determination of the first and second updated values ​​in step 1).

[0014] In one embodiment, the device further includes a third sensor, measuring said third intensity of the third electric current flowing in the third conductor.

[0015] The computer preferably includes a screen capable of displaying, in the same graph, the temporal evolutions of the different intensities and / or of displaying said alert.

[0016] The invention also relates to: - a computer program comprising program code instructions for the execution of steps 2) and 3) and for determining the first and second updated values ​​in step 1) of a method according to the invention, when said program is executed by a computer, - a computer medium on which such a program is recorded, for example a memory or a CD-ROM.

[0017] The invention finally relates to an industrial installation comprising: - a monitoring device according to the invention; - an electrical appliance; - a cable supplying electrical energy to said device, said cable comprising at least two conductors, these conductors comprising first and second electrical conductors electrically insulated from each other and in which first and second electrical currents flow, respectively; the first and second sensors being coupled to the first and second conductors, respectively, so as to measure the first and second intensities of the first and second electric currents, respectively; the device's computer being configured to receive, from the first and second sensors, the measurements of the first and second intensities respectively, and having access to a program containing code instructions for the implementation of steps 2) and 3) and for the determination of the first and second updated values ​​in step 1). Definitions

[0018] A "cable" comprises a set of electrical wires, or "electrical conductors", or "conductors", electrically insulated from one another, adjacent to one another and fixed to one another, the conductors being conventionally housed in a common sheath.

[0019] A repetition of cycles "in real time" means that the time interval between two successive cycles is less than 5 s, preferably less than 3 s, preferably less than 2 s, preferably less than 1 s.

[0020] The term "computer" refers to a computing unit, which includes a set of several machines with computing capabilities. This unit may be integrated into a tablet, a mobile phone, or be a PC or a server, for example, a server located remotely from the user, such as the cloud. The computer may also be, at least partially, integrated into the sensor housing. The sensors and the computer include means of communication for exchanging information.

[0021] Typically, a computer includes, in particular, a processor, memory, A human-machine interface, typically comprising a screen, a communication module via the internet, Wi-Fi, Bluetooth®, or the telephone network. Software configured to implement a method of the invention is loaded into the computer's memory. The computer can also be connected to a printer.

[0022] Unless otherwise specified, "comprising" or "comprising" should be interpreted in a non-restrictive manner. A cable comprising two conductors may therefore comprise more than two conductors, for example, be a three-phase power supply cable. Brief description of the figures

[0023] Other features and advantages of the invention will become apparent upon reading the detailed description that follows and upon examination of the accompanying drawing in which: - [Fig.1] [Fig.1] represents, schematically, the different stages of a process according to the invention; - [[Fig.2]] [Fig.2] schematically represents an installation according to the invention, and - [[Fig.3]] [Fig.3] illustrates the operation of a set of magnetic field sensors to evaluate an electrical intensity.

[0024] An index "i" applied to a reference to an object generically designates the different occurrences of that object. For example, if a cable has first, second, and third conductors 18;, this means that it has a first conductor 18i, a second conductor 182, and a third conductor 183. Detailed description

[0025] The objective of the method according to the invention is to alert on the probable occurrence of a failure of an electrical device in an industrial installation, thus allowing a preventive intervention on the device.

[0026] As illustrated in [Fig.2], the industrial installation 10 comprises: - a monitoring device 12 according to the invention; - an electrical appliance 14; - a power cable 16 supplying electrical energy to said device, said cable comprising at least two conductors, these conductors comprising first and second electrical conductors 18; electrically insulated from each other and in which first and second electrical currents i; flow, respectively.

[0027] The electrical device 14 can be any type. Typically, its rated power is greater than 1 kW, greater than 3 kW, or greater than 5 kW, and / or less than 1000 kW, or less than 100 kW, or less than 10 kW. In particular, it can be a motor, a compressor, or an extruder, especially an extruder used to form a sheath around electrical conductors to manufacture a cable.

[0028] Such an extruder conventionally comprises one or more heating rings arranged to melt plastic granules so that the resulting molten material can be extruded around electrical conductors and then cooled to form a cable.

[0029] The electrical device is conventionally supplied with energy, generally from the public electricity network, by means of the power cable 16, generally three-phase for high power.

[0030] The monitoring device is designed so that maintenance can be carried out before failures occur.

[0031] The monitoring device 12 comprises: - at least two sensors comprising first and second sensors 20; coupled to the first and second conductors, respectively, so as to measure the first and second intensities, respectively; - a computer 22 in communication with the first and second sensors, so that the computer can receive the measurements of the first and second intensities respectively.

[0032] The computer 22 can be any type of computer. However, it is specifically programmed for the implementation of steps 2) and 3) and for determining the first and second updated values ​​in step 1). It is preferably installed in the industrial installation, but could be, at least in part, located remotely from the installation, with the means of communication used to communicate with the computer 22 being adapted accordingly.

[0033] The sensors are preferably located outside the conductors whose current intensity they measure.

[0034] In a preferred embodiment, the sensors are magnetic field sensors arranged around the cable at at least one location on the cable. Preferably, at least one component of the magnetic field produced by the current flowing in each conductor is measured simultaneously using the sensors. For each conductor, the angle between the conductor and the nearest magnetic field sensor is determined, this angle being defined with respect to the center of the cable and by considering the conductors and sensors as points. The current intensities in the conductors are related to the measured magnetic field components by the relation B = kMI, where B is the matrix of magnetic field components, I is the matrix of current intensities, M is a matrix containing a plurality of proportionality coefficients depending on the angles between the conductors and the magnetic field sensors, and k is a predetermined coefficient. We can then calculate the inverse M1 of the matrix M, in order to deduce the values ​​of the current intensities. currents I = (po / l^.M '.B, where po is an equivalent magnetic permeability which takes into account the presence of insulating materials in the cable.

[0035] Since a polyphase cable has several current densities from its different conductors, we can first calculate the magnetic field produced by the current using the Biot-Savart law, then invert all the local measurements of the components of the magnetic field by determining an inverse matrix, in order to obtain the current intensities from the different current sources.

[0036] The preferred method for measuring current intensity is described in detail below for a cable having five conductors, but the cable could have a different number of conductors. This method consists of simultaneously measuring the current intensity in the conductors of the plurality of conductors by performing a first step E1 consisting of placing a plurality of magnetic field sensors around the cable, at at least one location on the cable.

[0037] During the sensor installation operation in step 11, no movement is required of either the sensors or the cable, whether translational, rotational, or any other type of movement. Furthermore, there are no constraints regarding the knowledge or application of the current intensity values ​​flowing in the cable conductors.

[0038] Then a step E2 consists of simultaneously measuring, for each of the conductors of the cable, at least one component of the magnetic field produced by the current flowing in that conductor.

[0039] These simultaneous measurements are carried out using a plurality of magnetic field sensors.

[0040] In a particular embodiment, the plurality of magnetic field sensors can measure for each conductor only the tangential component or only the radial component of the magnetic field.

[0041] Alternatively, for greater accuracy, the plurality of magnetic field sensors can measure for each conductor both the tangential and radial components of the magnetic field.

[0042] Then, in step E3, for each conductor, the angle α between that conductor and the nearest magnetic field sensor of the plurality of magnetic field sensors is determined. The angle α is defined with respect to the center of the cable cross-section, and the conductors and sensors are considered as points. Indeed, for simplicity, it is assumed that each conductor has an infinitesimally small cross-section and that the magnetic field detected by each sensor is localized at a point corresponding to the sensor's location.

[0043] Figure 3 schematically illustrates, by way of non-limiting example, a cable of A circular section comprising five conductors evenly distributed within the cable, which are considered as five points Cl to C5 equidistant from each other on the circumference of a circle T. Only the magnetic field sensor closest to conductor Cl has been represented and is symbolized by point A. The radius of the circle T is denoted by r, the distance between conductor Cl and sensor A is denoted by db, the line segment connecting sensor A and the center of circle T is denoted by db, the angle at point A between the line segment d'i and the line segment connecting conductor Cl and point A is denoted by [3]. The magnetic field detected by sensor A is represented by the vector Bb which is orthogonal to the line segment connecting conductor Cl and point A.

[0044] For a cable comprising N conductors each producing a magnetic field B; , i = 1, ..N, the component BA of the magnetic field at point A originating from the N conductors is defined as follows:

[0045] [Math.l] BA = • cos(^) = Arcsin^^à. ) ) .

[0046] where: Bim is the component of the magnetic field coming from the ith conductor captured by the mth sensor; the angles a; and [3; are defined for the i-th conductor in a similar way to the angles a and [3 defined above, respectively; d; denotes the distance between the i-th conductor and the nearest sensor; Po is an equivalent magnetic permeability that takes into account the presence of insulating materials in the cable; r is, as defined above, the radius of circle T; and Iim is the value of the intensity of the current flowing in the th conductor and deduced from the magnetic field measurement carried out by the mth sensor.

[0047] The number of magnetic field sensors is at least equal to the total number N of conductors in the cable.

[0048] Thus, the current intensities in the conductors are linked to the measured magnetic field components by the relation B = kMI where B is the matrix of magnetic field components measured in all conductors by all sensors, I is the matrix of current intensities flowing in all conductors, M is a matrix comprising a plurality of proportionality coefficients depending on the angles between the conductors and the magnetic field sensors and k is a predetermined coefficient.

[0049] It follows that I = (po / 2ir).M *.B, where M 1 is the inverse matrix of M.

[0050] Thus, following step E3 of determining the angle α between each conductor and the nearest sensor, we carry out a step E4 consisting of calculating the inverse matrix M1, so as to deduce the values ​​of the current intensities I in all the conductors of the cable.

[0051] By way of non-limiting example, for N = 5 conductors and five magnetic field sensors placed respectively at points A, B, C, D and E, the analytical expression allowing the current intensities in the conductors to be deduced is as follows:

[0052] [Math.2]

[0053] where h, i = 1, ..., 5 denotes the intensity of the current flowing in the th conductor and BA, Bb, Bc, Bd and Be denote the components of the magnetic field respectively measured by the five sensors.

[0054] The equivalent magnetic permeability po is a macroscopic permeability, which makes the calculation simpler than if one considered the local magnetic permeability.

[0055] In a particular embodiment, the angle α between this conductor and the magnetic field sensor of the nearest plurality of magnetic field sensors is determined so as to maximize the following function F:

[0056] [Math.3]

[0057] where h denotes the intensity of the current flowing in the th conductor.

[0058] There are various methods of mathematical convergence known in themselves to maximize F, such as the Levenberg-Marquardt algorithm (also called the LM algorithm) or the Nelder-Mead method.

[0059] The computer 22 is programmed to apply the angle determination steps E3 and inverse matrix calculation steps E4 of the process described above, in order to deduce the intensity of the current flowing in each of the conductors.

[0060] This embodiment allows the current intensity in all the cable conductors to be measured in a single operation, quickly obtaining extremely precise results without having to remove the cable sheath. Furthermore, the installation of the magnetic field sensors around the cables is easy and causes no damage, marking, or deformation to the cable. Moreover, given the position of the sensors Since the magnetic field relative to the cable conductors is unknown, taking into account the angular offset between sensors and conductors allows for increased accuracy of the intensity values ​​obtained.

[0061] Preferably, each sensor 20 is powered by an "autonomous" power source, preferably a battery 24 associated with the sensor, preferably physically incorporated into the sensor. Even more preferably, a gauge 26 is associated with each battery 24 to measure the charge level of the associated battery 24.

[0062] The monitoring device is used to implement steps 1) to 3), as shown in [Fig.1].

[0063] The current in a conductor, preferably in each conductor, is preferably an alternating current, preferably sinusoidal.

[0064] In step 1), the sensors 20; measure, at an "updated instant", the intensity of the currents i; which flow in the conductors 18;.

[0065] Each sensor 20; can directly measure the intensity of the current i; which flows in a conductor 18; which is associated with it, and / or communicate to the computer 22 the measurement which it makes so that the computer calculates the updated intensity from all the measurements received from the sensors.

[0066] In a particularly advantageous embodiment, each gauge 26 communicates to the computer 22, preferably at regular intervals, the instantaneous charge level of the battery powering the associated sensor. The delay between two transmissions of the charge level is preferably less than 24 hours, preferably less than 1 hour. The charge level transmission can also be in real time.

[0067] Communication between the sensors and gauges on the one hand, and the computer 22 on the other hand, can be ensured by any means, preferably by radio or wired means, preferably by wired means.

[0068] In step 2), the computer 22 analyzes the updated intensities received from the sensors in order to search for signs identified as harbingers of a future failure of the device.

[0069] The intensity of a current in a conductor is classically characterized by a set of attribute values.

[0070] The attributes are not exhaustive.

[0071] Attributes can be distinguished as "intrinsic" or "relative" depending on whether they only take into account values ​​measured on a single conductor or whether they take into account values ​​measured on several conductors. Attributes can also be distinguished depending on whether they only take into account values ​​measured at the current instant, or "instantaneous attributes," or whether they take into account... account of the values ​​measured at the current time and the values ​​measured at a previous time, or "temporal attributes".

[0072] For example, the intensity of a sinusoidal current in a first conductor, as a function of time t, can be substantially equal - to ii(t) = iorsin(cort + 9 / in a time interval centered on an instant T, where iOi is the amplitude of the intensity, in amperes (A), 91 is the phase shift, or initial phase, expressed in radians, and coi is the angular frequency, in radians per second (rad s *), and - at i / (t) = ioi' sin(cüi't + 9 / ) in a time interval centered on an instant T', where i 01' is the amplitude of the intensity, in amperes (A), 9 / is the initial phase, expressed in radians, and 0) / is the angular frequency, in radians per second, and the intensity of a sinusoidal current in a second conductor, as a function of time t, can be approximately equal - to i2(t) = i02-sin(co2-t + 92) in a time interval centered on an instant T, where i02 is the amplitude of the intensity, in amperes (A), q>2 is the initial phase, expressed in radians, and co2 is the angular frequency, in radians per second (rad s *), and - at i2'(t) = io2' sin(co21 + 92') in a time interval centered on an instant T', where i 02' is the amplitude of the intensity, in amperes (A), q>2' is the phase at the origin, expressed in radians, and co2' is the angular frequency, in radians per second.

[0073] The origin is a "zero" instant from which an intensity is measured.

[0074] ii, iOi, o)।.q)i.i2, io2, «2 and q>2 are examples of intrinsic and instantaneous attributes, ior io2,<Oi - œ2 et 91 - q> 2 are examples of relative and instantaneous attributes, ii - i / , iOi- iOi', coi - <Oi’, <pi - 9 / , i2 - i2’, i02- i02’, <o2 - œ2’, 92 - 92’, (h - h’) / (T-T’), (ior iOi’) / (T-T’), (œi - O / fT-T’), (91 - 9i’) / (T-T’), (i2 - i2’) / (T-T’), (i02- i02 ) / (T-T’), (co2 - <o2’) / (T-T’), (92 - 92’) / (T-T’) sont des exemples d’attributs intrinsèques et temporels, et ir <2 » ior <02 » æi — <o2’, 91 — 9 / , (ior <02 ) / (T-T’), ((Oi — œ2’ / (T-T’), (91 — 92’) / (T-T’) sont des exemples d’attributs relatifs et temporels.

[0075] The combinations of attributes mentioned above are also examples of attributes.

[0076] An indicator is an attribute whose value, compared with a corresponding reference value, provides information on the probability of occurrence of a future failure of the device, i.e. on a "risk of failure".

[0077] A statistical analysis makes it possible to determine the indicators, depending on the device in question. In particular, for a given device, one can identify the attributes whose values ​​change significantly (from a statistical point of view) before the failure occurs, and deduce one or more indicators using these attributes.

[0078] One can also compare the attribute values ​​for a device that has not undergone failure, with the values ​​for these same attributes for the same device that has suffered a failure.

[0079] It is thus possible to associate a value of an indicator with a probability of occurrence of the failure in the future.

[0080] In practice, the pulsation is imposed by the electrical network. The preferred attributes, and the preferred indicators are functions, or even consist of ii - i2, i0i - io2 ("difference in amplitudes of the current intensity in different conductors"), q>i - <p2(ou « différence de phases à l’origine de l’intensité du courant dans différents conducteurs », ou « déphasage entre phases »)ii - i / , iOi - iOi , q> i - q>i', i2 - i2', io2- te , Ç>2 - qÇ, (ii - ii') / (T-T'), (iOr iOi') / (T-T') (or "rate of change of the amplitude of the current intensity in a conductor", (q>i - q>i') / (T-T') (or "rate of change of the phase at the origin of the current intensity in a conductor")), (i2 - i2') / (T-T'), (io2- io2 ) / (T-T'), (q>2 - <p2’) (t-t’). les indices « 1 » et 2 donnés ici désignent indifféremment les premier, deuxième, troisième, ou ième conducteur du câble, i étant un entier supérieur égal à 1, de sorte que par exemple, pour câble triphasé i.e. comportant trois conducteurs, attributs indicateurs susmentionnés peuvent impliquer conducteurs.

[0081] For example, in a motor, the indicator could be an imbalance between the currents of the different phases. Indeed, if the current is not equivalent across the three phases, this could mean that there is a malfunction that could cause premature wear.

[0082] The statistical analyses described above also make it possible to determine, for an indicator, a reference value defining a limit between an acceptable situation, i.e. for which the probability of occurrence of the failure is low enough to be acceptable, and an unacceptable situation, i.e. for which the probability of occurrence of the failure is high enough for a maintenance intervention to be planned.

[0083] Preferably, similar statistical analyses can also be used to evaluate the time interval between the current time and the time of failure. The computer can thus advantageously evaluate the time of failure, that is, the future time at which the failure is statistically expected to occur, based on the difference between the value of an indicator and the corresponding reference value. Determining the time of failure advantageously allows for optimal planning of the intervention on the device in step 3).

[0084] Preferably, similar statistical analyses can also be used to evaluate a threshold charge level to assess whether the energy supplied by a battery to a sensor is sufficient for the measurement made by the sensor to be considered reliable.

[0085] The reference values ​​of the various indicators and the threshold load level are accessible to the computer. In particular, they can be stored in computer-accessible memory.

[0086] In a preferred embodiment, the reference value of at least one indicator is determined by the computer based on the operating mode of the device. The computer 22 then receives information characteristic of the device's operation, for example, information relating to the instructions given to the device, and deduces the reference value according to pre-established rules to which the computer has access. This information can be entered by an operator or result from measurements on the device.

[0087] For example, if the device is an extruder, the amplitude of the current intensity will differ depending on whether the temperature supplied by the heating rings is increasing, for example at the start-up of the extruder, or whether it is stable. Between these two phases, a large change in the amplitude of the current intensity can be considered acceptable, whereas it will not be acceptable between two instants during the stable operating phase.

[0088] For example, if the device is a motor, the normal current intensity will differ depending on whether the motor is operating at full load or reduced load. The computer therefore determines a higher reference value when the motor is operating at full load.

[0089] Indicators using relative attributes are particularly well suited when a "normal" relationship (i.e., in the absence of faults) is known between the current attributes in the different conductors. In particular, for a three-phase current, the currents in the three conductors typically evolve in the same way over time, with a phase shift of 120°. They have the same amplitude. A deviation from this normal relationship can constitute an indicator that is advantageously easy to evaluate.

[0090] When the computer uses an indicator requiring values ​​at different times, i.e., uses a temporal attribute, the values ​​of the attributes determined during successive cycles are time-stamped and stored in memory accessible by the computer. The computer can then retrieve these values ​​from memory to calculate the indicator.

[0091] Preferably, the computer compares the values ​​of several indicators with corresponding reference values ​​in order to better assess the probability of a future occurrence of a failure.

[0092] Preferably, the computer 22 also analyzes the charge level received from each gauge 26. If a sensor's charge level is below a threshold charge level, the computer - issues a corresponding alert so that a battery recharge or replacement can be scheduled and / or - rejects the intensities determined from the measurements taken by the sensor, considering that their quality is insufficient.

[0093] In one embodiment, the sensors communicate with the computer 22 wirelessly, preferably by radio frequency. Preferably, the computer evaluates the quality of the received signals, using any known method, and, if the quality is insufficient, - issues a corresponding alert so that a maintenance operation can be scheduled, and / or - rejects the intensities determined from these signals.

[0094] In step 3), if the result of comparing a value of at least one indicator with the corresponding reference value is associated with an unacceptable situation, for example, if the value of said indicator is higher or lower than the reference value, the computer issues an alert. The alert can be of any kind, for example, visual and / or audible. In particular, it can be a message and / or an image displayed on the computer screen.

[0095] An operator who has received the alert can plan a maintenance operation to return the device to normal operation. The intervention may consist in particular of replacing the device or modifying it, notably to update or repair it.

[0096] The alert can also be processed electronically by a maintenance planning tool.

[0097] The cycle of steps 1) to 3) is repeated, the time interval between two steps preferably being less than 1 hour, preferably less than 1 minute, preferably less than 5 seconds.

[0098] The cycle of steps 1) to 3) is preferably repeated instantaneously, for real-time monitoring.

[0099] The measurements are preferably displayed on a computer screen, preferably in the form of a graph showing, as a function of time, the measured intensities and / or the value of one or more intensity attributes, the curves of the different intensities being preferably superimposed.

[0100] In one embodiment, the time scale represents a duration greater than 1 hour, preferably greater than 1 day or greater than 1 week and / or less than 6 months.

[0101] Preferably, the computer displays an interface allowing the operator to: - select the information they wish to view, and / or - select display options for the selected information, and / or - select one or more sensors in order to limit the display to the information provided by that sensor or those sensors, and / or - to perform operations on the information received, for example to add intensities.

[0102] Remarkably, the method according to the invention makes it possible, in particular, to detect imbalances in the currents in the cable conductors, leading to overconsumption by the device. This overconsumption can, in particular, indicate fatigue in the device, especially in a motor, and therefore a risk of future failure.

[0103] The method according to the invention also makes it possible to detect a drop in current in a conductor of a cable supplying an extruder. This drop can, in particular, indicate a deterioration of a heating collar and therefore a risk of failure, for example, poor mixing of the material to be extruded, a jam, or physical degradation of the heating collar.

[0104] As is now clear, the invention makes it possible, by analyzing the intensities of the electric currents flowing in the different conductors of a device's power cable, to detect slow changes in indicators by using current measurements taken at different times, particularly on the same conductor, but also to detect faster changes, notably by using current measurements taken at the same time on different conductors. Monitoring these indicators, preferably in real time, makes it possible to anticipate a device failure and thus to intervene preventively on the device to avoid this failure.

[0105] The use of autonomous sensors, arranged outside the conductors, facilitates the implementation of the invention.

[0106] Measuring the gauges of the batteries powering the sensors and not using measurements made with an insufficient charge level improves the reliability in assessing the risk of failure.

[0107] Of course, the invention is not limited to the embodiments described and represented above.

[0108] In particular, the preceding description refers to first and second conductors, but is not limited to a configuration comprising only two conductors. Preferably, the invention is notably applied to a power cable comprising three conductors for supplying a three-phase device, said first and second conductors being any pair of two conductors chosen from among the three conductors of the cable.

[0109] The invention is also not limited to alternating currents. The currents can also be direct.

[0110] The conductors of the cable are not necessarily identical.

Claims

Demands

1. A method for predictive maintenance of an electrical device (14) powered by a cable (16) having at least two conductors, having first and second electrical conductors (18i) electrically insulated from each other and in which first and second electrical currents, having first and second intensities, respectively, flow, said method comprising a real-time repetition of an updated cycle comprising the following successive steps: 1) at an updated time, measurement of the first and second intensities, and determination, by computer, of first and second updated values, for at least one attribute of the first and second intensities, respectively; 2) by computer, calculation of - at least one indicator as a function of the first updated value and a first "previous" value determined, for the first intensity, prior to the updated cycle, for said at least one attribute;and - at least one indicator based on the first and second updated values, without using said previous values, then determination of a risk of failure based on the difference between the value of each indicator and a corresponding reference value; 3) based on the risk of failure, generation, by computer, of an alert and, preferably, intervention on the device.

2. A method according to the immediately preceding claim, wherein the apparatus (14) is a motor, a compressor or an extruder, said apparatus having a rated power greater than 1 kW.

3. A method according to any one of the preceding claims, wherein, in step 1), the first and second intensities are measured by means of magnetic field sensors arranged around the cable, configured to measure, for each of the conductors, at least one component of the magnetic field produced by the current flowing in the conductor.

4. A method according to the immediately preceding claim, wherein said first and second intensities are measured by following steps consisting of: E1. Place a plurality of said magnetic field sensors around said cable, at at least one location on said cable; E2. Measure simultaneously, for each of said conductors, at least one component of the magnetic field produced by the current flowing in said conductor, by means of said plurality of magnetic field sensors; E3. For each of said conductors, determine the angle between said conductor and the nearest magnetic field sensor of said plurality of magnetic field sensors, said angle being defined with respect to the center of said cable and by treating said conductors and said sensors as points; the current intensities in said conductors being related to the components of the magnetic field measured by the relation B = kMLet I be the matrix of the components of the magnetic field, I the matrix of the current intensities, M a matrix comprising a plurality of proportionality coefficients depending on the angles between the conductors and the magnetic field sensors, and k a predetermined coefficient. Calculate the inverse M1 of the matrix M, so as to deduce the values ​​of the current intensities I = (po / 2ir) * M * B, where po is an equivalent magnetic permeability that takes into account the presence of insulating materials in the cable.

5. A method according to any one of the two immediately preceding claims, wherein each sensor is powered by a respective battery, a gauge preferably transmitting to a computer (22) implementing steps 2) and 3), a charge level of said battery, the computer taking into consideration or not an intensity supplied by said sensor depending on whether the charge level exceeds or does not exceed a threshold charge level.

6. A method according to any one of the preceding claims, wherein at least one said reference value is determined, by computer, according to the operating mode of the device.

7. A method according to any one of the preceding claims, wherein, in step 2), at least one indicator is calculated by computer as a function of the difference between the first and second discounted values, and / or as a function of the difference between the first discounted value and the first previous value.

8. A method according to any one of the preceding claims, in in which, in step 2), at least one indicator chosen from the intensity and phase shift between the different currents is calculated by computer.

9. A method according to any one of the preceding claims, wherein the first and second electric currents are alternating currents and at least one indicator is a function of, or consists of, a difference between the phases of the intensities of said currents or a difference between the amplitudes of the intensities of said currents.

10. A method according to any one of the preceding claims, wherein at least one indicator and / or at least one reference value is / are determined, by computer, by statistical processing of historical data taken from historical devices identical to said device (14) powered by the cable, the statistical processing being configured to establish a correlation between indicator values ​​and the occurrence of a device failure.

11. A method according to any one of the preceding claims, wherein, in step 2), a predicted time interval before a failure occurs is determined by comparing the value of each indicator with a corresponding reference value.

12. A method according to any one of the preceding claims, implemented while the apparatus (14) is in operation.

13. A method according to any one of the preceding claims, wherein, in step 3), the first and second updated intensities are displayed in the same graph on a computer screen, the graph showing, in superposition, the temporal evolutions of the first and second intensities.

14. Use of a method according to any one of the preceding claims for: - detecting imbalances between the phases of the currents flowing in the different conductors of the cable, in particular in a cable supplying, in three-phase, a motor or a compressor, and / or - detecting a drop in current in a cable, in particular in a cable supplying an extruder.

15. A monitoring device for implementing a method according to any one of claims 1 to 13, said device comprising: - at least two sensors, comprising first and second sensors (20;) capable of measuring the first and second intensities (1;); - a computer (22) configured to receive, from the first and second sensors, the measurements of the first and second in tensions respectively, and including code instructions for the implementation of steps 2) and 3) and for the determination of the first and second updated values ​​in step 1) of said process.

16. Computer program comprising program code instructions for carrying out steps 2) and 3) of a method according to any one of claims 1 to 13, and for determining the first and second updated values ​​in step 1) of said method.

Images