Method for monitoring the condition of mechanical components on a shafting, associated monitoring device and system

Abstract

The invention relates to a method for monitoring the condition of mechanical components of the bearing and gear mechanism type on a shafting of rotating mechanical equipment. The method comprises acquiring at least one absolute acceleration measurement value y c [k] of the shaft r [k] of the shaft, determining (E30) a matrix H[k] that makes it possible to define a state model described by (I), determining (E40) an estimate of the vector x[k] from the data of the state model, the set of steps further comprising, for at least one mechanical component: determining (E50) a characteristic quantity of the contribution of the components to the vector Y[k] from the estimate, comparing (E60) the quantity to a threshold, detecting (E70) a potential fault of the at least one mechanical component as a function of the result of the comparison.

Description

Background Technology

[0001] This invention belongs to the general field of monitoring and predictive maintenance of mechanical components. More specifically, the invention relates to a method for monitoring the condition of mechanical components such as bearings and gears integrated into the shaft system of rotating machinery. The invention also relates to monitoring devices and systems configured to implement this monitoring method. The invention finds particularly advantageous applications in the context of real-time monitoring of the health status of such mechanical components, but is not limited thereto, especially when the rotating machinery involved is an aircraft engine.

[0002] Traditionally, shaft systems integrated into rotating machinery are equipped with various mechanical parts or components, such as bearings and gears. It should be noted that the term "rotating machinery" generally refers to machines whose kinematics obey the physics of rotational mass, and therefore encompasses any rotating system, such as electric motors, turbine engines, electromechanical actuators, wind turbines, etc.

[0003] Monitoring the health of such mechanical components plays a fundamental role, on the one hand, in ensuring the good lifespan of the shaft system they are equipped with (and subsequently the rotating machinery integrated with the shaft system), and on the other hand, in preventing any events or even accidents that may be caused by excessive wear of the mechanical components.

[0004] Therefore, it should be understood that, in order to avoid the aforementioned inconveniences, and more generally, to ensure the early detection of any defects that may affect the mechanical components, it is of utmost importance to ensure regular monitoring of their health status.

[0005] Therefore, devices are known capable of continuously monitoring rotating machinery through vibration analysis. Such devices are schematically operated based on vibration measurements, for example, using accelerometers placed on structural elements of the machinery. These vibration measurements can also be performed using displacement sensors capable of measuring the relative displacement of a shaft relative to the structural elements. When vibration measurements are performed, frequency analysis of the recorded spectrum is then used, starting from principles that can identify faults in the mechanical components, for example, by evaluating the amplitude of spectral lines characterizing the kinematics of the mechanical components in question. In this way, the operating condition of one or more components on the shaft system can be examined over time.

[0006] The fact remains that the vibration analysis described above is still not well-suited for monitoring rotating equipment with variable revolutions per minute (RPM) (under load and at varying speeds), making it practically unsuitable for most rotating machinery. This is particularly relevant in the aerospace field, where the rotating machinery in question operates based on large variations in RPM (e.g., aircraft engines during takeoff and landing, especially turbine engines).

[0007] Subsequently, some relatively new technological developments were proposed to allow real-time monitoring of the condition of mechanical components and to accommodate variable revolutions per minute. The fact remains, however, that these developments are still far from satisfactory. In practice, their implementation can only be conceived for shaft speeds less than 60 revolutions per minute, effectively excluding any applications that do not adhere to this constraint (and thus particularly in the aerospace field). Summary of the Invention

[0008] The object of this invention is to overcome all or part of the shortcomings of the prior art, particularly the aforementioned shortcomings, by providing a solution that enables accurate, real-time monitoring of the condition of mechanical components, such as bearings and gears on shaft systems of rotating machinery. Furthermore, the solution proposed in this invention allows the monitoring to be performed at variable revolutions per minute (RPM) and is not limited to shaft speed, making it advantageously applicable to any technical field requiring the use of rotating machinery.

[0009] Therefore, according to a first aspect, the present invention relates to a method for monitoring the condition of mechanical components such as bearings and gears on a shaft system equipped with rotating machinery. The method includes acquiring at least one measured value y of the absolute acceleration of the shaft in a fixed reference frame associated with the rotating machinery. c The steps of [k], where k is an integer index, and the set of steps below:

[0010] - Obtain at least one measurement value y previously acquired. c The value of the rotational frequency f of the axis at time [k]. r [k],

[0011] - Determine a matrix H[k], which allows the definition of a state model as described below:

[0012] x[k+1]=x[k]+w[k] and

[0013] Where w[k] is random noise, and v[k] is the noise relative to the at least one measured value y. c [k] is associated noise, H[k] is determined by the value f r [k] is defined, SE[k] equals y. c [k] and y c The expected value of the product of the conjugate values ​​of [k], where x[k] is a vector, and for each mechanical component, the vector comprises subvectors whose components represent the contribution of the mechanical component to the vector Y[k].

[0014] - Determine the estimate of vector x[k] from the matrix H[k].

[0015] For at least one mechanical component, the set of steps further includes the following steps:

[0016] - Determine the quantitative characteristics of the contribution associated with the mechanical component from the estimated quantity.

[0017] - Based on the comparison of the quantity with the threshold, detect possible defects in the at least one mechanical component.

[0018] Therefore, the monitoring method proposed in this invention is based on the use of a state model that enables the representation (modeling) of the corresponding contribution of the mechanical components of the shaft system to the vibration experienced by the shaft during the operation of rotating machinery.

[0019] More specifically, the contribution associated with the gear-type mechanical component is in the component y of the vector Y[k]. c The contribution associated with bearing-type components is at the level of the component SE[k] of the vector Y[k]. In other words, the vector Y[k] forms a kinematic representation of the mechanical component whose state should be monitored.

[0020] This method of modeling the vibration behavior of the shaft, and therefore the mechanical components it is equipped with, advantageously distinguishes the present invention from the prior art. In fact, by doing so, it is no longer necessary to limit the monitoring of the mechanical components to a shaft speed of less than 60 revolutions per minute. Therefore, the method according to the invention allows for such monitoring regardless of the shaft speed. It is further independent of the operating speed of the rotating machinery, and thus can be implemented in cases of variable speeds.

[0021] The monitoring method according to the invention also provides the possibility of performing the necessary steps for detecting any defects in the mechanical components after each acquisition of the absolute acceleration measurement of the shaft. This makes it possible to establish synchronous and real-time monitoring of the state of the mechanical components, thereby helping to detect potential defects early.

[0022] Finally, an additional advantage of the monitoring method according to the invention is attributed to the robustness of the estimator of the determined vector x[k]. This robustness stems from the fact that no assumptions are made regarding the statistical properties of the noise in the state model, and that the estimator considers not only the measurement-related errors but also the errors related to the modeling of the state model.

[0023] In a specific implementation, the monitoring method may also include one or more of the following features, which may appear individually or in any combination that is technically possible.

[0024] In a specific implementation, the estimate of vector x[k] is determined by means of a minimax optimization algorithm or a least squares optimization algorithm.

[0025] In a specific implementation, the quantity represents the magnitude, phase, or energy of the contribution.

[0026] In a specific implementation, if a defect is detected, the method further includes the step of issuing an alarm.

[0027] The fact that an alert is issued when a defect is confirmed to be detected allows information about the existence of the defect to be obtained without delay.

[0028] In a specific implementation, multiple absolute acceleration measurements are acquired cyclically, and the set of steps is performed after each acquisition of an absolute acceleration measurement.

[0029] According to another aspect, the present invention relates to a computer program comprising instructions for implementing the monitoring method according to the invention when the computer program is executed by a computer.

[0030] The program can use any programming language and can take the form of source code, object code, or intermediate code between source code and object code, such as a partially compiled form, or any other desired form.

[0031] According to another aspect, the present invention relates to a computer-readable information or recording medium having a computer program according to the present invention recorded thereon.

[0032] Information or recording media can be any entity or device capable of storing programs. For example, media can include storage devices such as ROMs, such as CD ROMs or microelectronic circuit ROMs, or magnetic recording devices such as floppy disks or hard disks.

[0033] On the other hand, the information or recording medium can be a transmissible medium, such as electrical or optical signals, which can be transmitted via cables or optical fibers, via radio, or by other means. The program according to the invention can be downloaded specifically from an Internet-type network.

[0034] Alternatively, the information or recording medium may be an integrated circuit with a program that is adapted to execute or be used to execute the method in question.

[0035] According to another aspect, the present invention relates to a device for monitoring the condition of mechanical components such as bearings and gears on a shaft system of rotating machinery. The processing device includes:

[0036] - A first acquisition module, configured to acquire at least one measurement of the absolute acceleration y of the shaft in a fixed reference frame associated with the rotating machinery. c The steps are defined in [k], where k is an integer index.

[0037] - A second acquisition module, configured to acquire the at least one measurement value y during acquisition. c The value of the rotational frequency f of the axis at time [k].r [k],

[0038] - A first determining module, configured to determine a matrix H[k], which enables the definition of a state model as described below:

[0039] x[k+1]=x[k]+w[k] and

[0040] Where w[k] is random noise, and v[k] is the noise relative to the at least one measured value y. c [k] is associated noise, H[k] is based on the value f r [k] is defined, SE[k] equals y. c [k] and y c The expected value of the product of the conjugate values ​​of [k], where x[k] is a vector, and for each mechanical component, the vector comprises subvectors whose components represent the contribution of the mechanical component to the vector Y[k].

[0041] - A second determining module, configured to determine an estimate of vector x[k] from the matrix H[k].

[0042] - A third determining module, configured to determine at least one quantitative characteristic of the contribution associated with the mechanical component from the estimated quantity.

[0043] - A comparison module configured to compare the at least one quantity with a threshold to obtain a comparison result.

[0044] - A detection module configured to detect potential defects in the at least one mechanical component based on comparison results.

[0045] According to another aspect, the present invention relates to a system for monitoring the condition of mechanical components such as bearings and gears on a shaft system of rotating machinery. The monitoring system includes:

[0046] - At least one measurement y is used to obtain the absolute acceleration of the shaft in a fixed reference frame associated with the rotating machinery. c A device of [k], where k is an integer index.

[0047] -The monitoring device according to the present invention.

[0048] Finally, according to the last aspect, the present invention relates to an aircraft including a monitoring system according to the present invention. Attached Figure Description

[0049] Other features and advantages of the invention will become apparent from the following description, with reference to the accompanying drawings, which illustrate exemplary embodiments without any limitation. In the drawings:

[0050] [ Figure 1 ] Figure 1 The illustration schematically depicts a specific embodiment of the monitoring system according to the invention in its environment, the monitoring system being configured to monitor the condition of mechanical components such as bearings and gears on the shafting of an aircraft engine;

[0051] [ Figure 2 ] Figure 2 Indicatively representing belonging to Figure 1 An example of the hardware architecture of the monitoring device according to the present invention;

[0052] [ Figure 3 ] Figure 3 Represent the process in the form of a flowchart. Figure 2 The monitoring equipment implements a specific mode of the monitoring method according to the present invention. Detailed Implementation

[0053] This invention is described in the context of its applications in the aerospace field, and more specifically in the context of the monitoring and predictive maintenance of mechanical components, such as bearings and gears in shafting systems of aircraft engines.

[0054] For the remainder of this specification, the aircraft referred to, but not limited to, is an aircraft type, such as a civil aircraft capable of carrying passengers and equipped with multiple engines of the same turbine engine type. For example, the engine is a turbojet engine.

[0055] However, based on other examples not detailed herein, other types of turbine engines, such as turboprop engines, are considered, but non-turbine engines, such as piston engines, are not excluded. The invention is practically applicable to any type of aircraft engine including a shaft system on which mechanical components such as bearings and gears are arranged, and whose condition will be monitored over time, as described in detail later. Of course, consideration is not excluded for other types of aircraft, such as helicopters.

[0056] Figure 1 The illustration schematically depicts a particular embodiment of the monitoring system 10 according to the invention in its environment.

[0057] The monitoring system 10 is configured to monitor the condition of mechanical components such as bearings and gears on the shafting of an aircraft engine. For this purpose, the system 10 includes a data acquisition device 11 configured to acquire:

[0058] - At least one measurement of the absolute acceleration of the axis in a fixed reference frame associated with the aircraft engine. c [k] (k is an integer index indicating the acquisition sequence in a series of absolute acceleration measurements),

[0059] - At least one measured value of the shaft rotational speed v r [k'](k' is an integer index indicating the sequence of acquisitions in a series of rotational speed measurements).

[0060] The "absolute acceleration" referred to here typically refers to the total acceleration of the shaft, that is, the acceleration that reflects the vibration experienced by the shaft due to its fixation to the aircraft's engine. This absolute acceleration is defined in a fixed reference frame associated with the aircraft's engine, and it is understood that it remains independent of the aircraft's displacement velocity.

[0061] Traditionally, the acquisition device 11 includes an acquisition chain, which includes at least one component dedicated to measuring the absolute acceleration value y. c [k] Sensors, such as accelerometers of known design, and at least measurements of the rotational speed v of the shaft. r [k'] Sensors, such as optical encoders or even tachometers. The acquisition chain also includes other components, such as, in particular, acquisition cards, electronic amplification and / or filtering devices, analog-to-digital converters, etc. These aspects will not be described in detail here, as they are beyond the scope of this invention.

[0062] Typically, those skilled in the art know how to perform and acquire measurements of the absolute acceleration and rotational speed of a shaft, and therefore know in particular how to select appropriate sensors for each considered quantity (absolute acceleration, rotational speed), for example, from product catalogs provided by specialized manufacturers. They also know how to locate these sensors and how to select sampling frequencies to acquire the relevant measurements. As a non-limiting example, the sampling frequency is chosen to be at least ten times greater than the maximum frequency of the absolute acceleration / rotational speed signal acquired due to the sensor dedicated to acquiring this signal.

[0063] Finally, at the output of the acquisition device 11, the acquired absolute acceleration / rotational speed measurement corresponds to digital data forming a sampling signal, which can be processed by a computer device to monitor the state of mechanical components arranged on the rotating shaft, as described in more detail below.

[0064] It should be noted that in the remainder of this specification, the term "measured value y" is used interchangeably with "measured value y". c [k]” or “sampled signal y” c (Therefore, the sample of this signal corresponds to the measured value y) c [k]) was quoted indiscriminately, and the two statements have the same meaning. Of course, the measured value v of the shaft speed... r [k'] Such annotations are also valid.

[0065] The absolute acceleration / rotational speed measurements acquired by acquisition device 11 are taken over one or more time intervals, referred to as "acquisition intervals." It should be understood that each acquisition interval includes the period during which the aircraft's engines are operational. Here, "operation of the aircraft's engines" refers to the fact that the aircraft's engines are running. This configuration certainly includes the taxiing phase before and after landing, the takeoff phase, the cruise phase, and the landing phase, but also includes the phases during which the aircraft has not yet left its parking position before takeoff or has reached its parking position after landing, but whose engines are still running.

[0066] It should be noted that there are no restrictions on the duration of the acquisition interval, or even on the number of conceivable acquisition intervals. In practice, the duration of the acquisition interval is defined directly, for example, by data on the duration in seconds, or even indirectly, for example, by data on the number of rotation cycles of the rotating shaft to be acquired.

[0067] It should also be noted that the absolute acceleration measurement value y is collected. c The time of [k] may be related to the acquisition of the speed measurement value v. r The times [k'] may differ or be the same. In any case, once measurements of both quantities are acquired during the same acquisition interval, the absolute acceleration and rotational speed measurements can, of course, be correlated with a given time included in the acquisition interval, for example, through interpolation techniques.

[0068] The monitoring system 10 also includes a monitoring device 12, which is configured to perform processing operations by implementing the monitoring method according to the invention, in particular based on the measurements acquired by the acquisition device 11, so as to enable monitoring of the condition of mechanical components such as bearings and gears on the shafting of the aircraft engine (i.e., performing wear tracking).

[0069] In this embodiment, the monitoring device 12 is integrated into the aircraft's computing unit and is also referred to as FADEC (Full Authority Digital Engine Control). The FADEC unit is configured in a manner known per se to optimally drive and regulate engine operation. However, unlike prior art FADEC units, the FADEC unit described herein, due to the integration of the monitoring device 12, is also capable of monitoring the status of mechanical components such as bearings and gears on the shaft system.

[0070] However, the choice to integrate the monitoring device 12 into the FADEC unit of the aircraft constitutes only one variation of the embodiments of the present invention. Therefore, it is not excluded that the processing device 12 may be located outside the FADEC unit (i.e., not electronically integrated), but still arranged in the aircraft.

[0071] Figure 2 Indicatively representing belonging to Figure 1 An example of the hardware architecture of the monitoring system 10 according to the monitoring device 12 of the present invention.

[0072] like Figure 2 As shown, the monitoring device 12 has a computer hardware architecture. Therefore, the monitoring device 12 specifically includes a processor 1, a random access memory 2, a read-only memory 3, and a non-volatile memory 4. It also includes a communication device 5.

[0073] The read-only memory 3 of the monitoring device 12 constitutes a recording medium according to the invention, which is readable by the processor 1 and contains a computer program PROG according to the invention, including instructions for performing steps of the monitoring method according to the invention. The program PROG defines functional modules of the monitoring device 12 that depend on or control the hardware components 1 to 5 of the monitoring device 12, and specifically includes:

[0074] - The first acquisition module MOD_OBT1 is configured to acquire at least one measurement of the absolute acceleration y of the axis. c [k],

[0075] - A second acquisition module MOD_OBT2, which is configured to acquire the at least one measurement value y in the previous acquisition. c The value of the rotational frequency f of the axis is obtained at time [k]. r [k],

[0076] - The first determining module MOD_DET1 is configured as a determining matrix H[k], which enables the definition of a state model as described below:

[0077] x[k+1]=x[k]+w[k] and

[0078] Where w[k] is random noise, and v[k] is the noise relative to the at least one measured value y. c [k] is associated noise, H[k] is determined by the value f r [k] is defined, SE[k] equals y. c [k] and y c The expected value of the product of the conjugate values ​​of [k], where x[k] is a vector, and for each mechanical component, this vector comprises subvectors whose components represent the contribution of that mechanical component to vector Y[k].

[0079] - The second determining module MOD_DET2 is configured to determine an estimate of vector x[k] based on the matrix H[k].

[0080] - A third determining module, MOD_DET3, is configured to determine the quantitative characteristics of the contribution associated with the mechanical component based on the estimated quantity.

[0081] - The comparison module MOD_COMP is configured to compare the quantity with a threshold to obtain a comparison result.

[0082] - The detection module MOD_DETECT is configured to detect potential defects in the at least one mechanical component based on the comparison results.

[0083] - The alarm module MOD_ALERT is configured to issue an alarm if a defect in a mechanical part is detected.

[0084] The communication device 5 is configured to allow the monitoring device 12, in particular, to communicate with the acquisition device 11, so as to directly acquire one or more absolute acceleration and / or rotational speed measurements therefrom. For this purpose, the communication device 5 is based on a wired or wireless communication interface, capable of implementing any known protocol (Ethernet, Wi-Fi, Bluetooth, 3G, 4G, 5G, etc.) suitable for data exchange between the monitoring device 12 and the acquisition device 11. Note that in this exemplary embodiment, the communication device 5 integrates the first acquisition module MOD_OBT1. Of course, the acquisition device 11 itself has a communication device suitable for transmitting one or more absolute acceleration and / or rotational speed measurements.

[0085] In this embodiment, the second acquisition module MOD_OBT2 is also integrated into the communication device 5, and each value f of the axis rotation frequency... r [k] is determined by the acquisition device 11 itself based on the acquired rotational speed measurement value. That is, in the embodiment described here, the monitoring device 12 measures each measurement value y of the absolute acceleration of the shaft. c [k] Obtain the value f of the rotational frequency of the axis. r [k], here the value is obtained in the form of data exchange between the acquisition device 11 and the second acquisition module MOD_OBT2.

[0086] The value of the rotational frequency of the shaft can be determined using any technique known to those skilled in the art, with a dedicated speed measurement device (v) fixed on the shaft. r [k'] sensor.

[0087] For example, the rotational speed signal can initially be analyzed to determine its properties. It can be a square wave or a sine wave, or even a series of pulses. Based on this analysis, an algorithm can then be implemented in a second step that makes it possible to estimate the rotational frequency of the reference axis. Such an algorithm can be decomposed into three implementation variants: the first variant is based on the detection of the rising edge when the signal from the speed sensor is a square wave; the second variant is based on the estimation of the instantaneous frequency of the sine wave; and the third variant is based on the time position of the peak when the signal from the speed sensor is a series of pulses.

[0088] It should be noted that the acquisition device 11 obtains each measured value of the absolute acceleration y relative to the axis. c The selection of the rotation frequency value associated with [k] constitutes only one exemplary embodiment of the present invention.

[0089] However, other variations of the embodiment are conceivable. For example, once the monitoring device 12 has acquired the acquired rotational speed signal from the acquisition device 11, the monitoring device 12 can perform processing operations that enable the determination of one or more values ​​of the rotational frequency. It can then be understood that, for this variation of the embodiment, the acquisition module MOD_OBT2 has the functionality to implement the processing operations.

[0090] It is not excluded that each value of the rotational frequency f of the axis can be imagined. r [k] is determined by the shaft's rotational speed signal, rather than by one or more speed sensors, for example, directly from aircraft engine control signals. In this way, there is no need to use such speed sensors, and once the value f of the shaft's rotational frequency is determined from the control signals... r [k] These values ​​can be received using the acquisition module MOD_OBT2, or the acquisition module itself can determine the value f from the control signals transmitted to it. r [k].

[0091] Typically, there are no restrictions on how the monitoring device 12 can acquire one or more values ​​of the rotational frequency of the shaft.

[0092] In the remainder of this specification, and for the purposes of simplification only, the mechanical components present on the shaft are considered to correspond to gears formed by two meshing wheels and bearings, but are not limited thereto.

[0093] To be purely illustrative, the monitoring method for this particular pattern is implemented during aircraft flight, more specifically during the cruise phase.

[0094] Figure 3 A specific embodiment of the monitoring method according to the present invention is illustrated in flowchart form. The steps of the specific embodiment are... Figure 2The monitoring equipment 12 is executed.

[0095] In this embodiment, the measured value of the absolute acceleration y of the shaft is considered to be... c [k] is acquired cyclically (e.g., periodically) by the acquisition device 11, and each measurement value y c [k] is immediately transmitted to monitoring device 12 after collection.

[0096] like Figure 3 As shown, for each measured value y of the absolute acceleration of the shaft acquired by the acquisition device 11 c [k], The monitoring method includes acquiring the measured value y. c [k] Step E10 Step E10 is implemented by the first acquisition module MOD_OBT1 equipped with the monitoring device 12.

[0097] In this embodiment, the monitoring method acquires the measured value of absolute acceleration y each time. c [k] is followed by the step set ENS_E.

[0098] The remainder of this specification is intended to describe the steps included in the set ENS_E for measuring the absolute acceleration y previously acquired and obtained after performing step E10. c [k]. It should be noted that the steps described in the set ENS_E are intended to be performed iteratively after each measurement of the absolute acceleration of the shaft is acquired by the monitoring device 12.

[0099] like Figure 3 As shown, the set of steps ENS_E first includes Step E20 , used to obtain the previously acquired measurement value y c The value of the rotational frequency f of the axis at time [k]. r [k]. Step E20 is implemented by the second acquisition module MOD_OBT2 equipped with the monitoring device 12.

[0100] As mentioned earlier, the value f of the rotational frequency of the shaft r [k] is the value corresponding to the measured value y from the data acquisition device. c [k](The correspondence here involves the measured value y) c The acquisition time of [k] is determined, therefore, in this embodiment, the acquisition of the value, which is the subject of step E20, involves data transmission between the acquisition device 11 and the monitoring device 12.

[0101] It should be noted that in this embodiment, steps E10 and E20 are described as being performed one after another. Of course, this is merely an optional implementation, and it is not excluded that these steps E10 and E20 may be performed in parallel.

[0102] Once the rotational frequency f of the axis is obtained r The set of steps, ENS_E, further includes step E30, which determines the matrix H[k] so that the state model can be defined. Step E30 is implemented by the first determination module MOD_DET1 equipped with the monitoring device 12.

[0103] The matrix H[k] describes the changes in the health state characteristics of the mechanical components arranged on the rotation axis, which are implicitly contained in the state vector of the state model, as will be explained in detail later.

[0104] It should be noted that the state model according to the present invention involves not only the acceleration signal y c It also involves the signal SE, which corresponds to the signal y. c and the signal y c The expected value of the product of the conjugate values ​​(equivalently, the signal SE corresponds to the acquired signal y). c (The square of the envelope). Therefore, the state model can be viewed as a dynamic model of the "enhanced" signal, which corresponds to the acceleration signal y. c The collection of samples of the signal SE is represented here as a vector, denoted by Y. Considering such a state model advantageously enables the simultaneous monitoring of the health status of both gears and bearings.

[0105] Therefore, the monitoring method proposed in this invention is based on the use of a state model that enables the representation (modeling) of the corresponding contribution of the mechanical components of the shaft system to the vibration experienced by the rotating shaft during the operation of the rotating machinery.

[0106] More specifically, the contribution associated with the gear-type mechanical component is in the component y of the vector Y[k]. c The contribution associated with bearing-type components is at the level of the component SE[k] of the vector Y[k]. In other words, the vector Y[k] forms a kinematic representation of the mechanical component whose state should be monitored.

[0107] This method of modeling the vibration behavior of a shaft, and therefore, not to mention the method of modeling the vibration behavior of the mechanical components it is equipped with, advantageously makes it possible to perform the monitoring regardless of the shaft's rotational speed. It is further independent of the operating revolutions per minute (RPM) of the rotating machinery, and thus can be implemented in cases with variable RPMs.

[0108] An example of determining the matrix H[k] will now be described in detail. For this, it is important to first note that the y-values ​​provided by the accelerometer... c The vibration signal can be decomposed according to the following expression:

[0109] y c [k]=y r [k]+y g [k]+b[k]

[0110] Where y r The vibration signal generated by the bearing is called the "bearing signal". g 'b' is the signal generated by the contact between gears, called the "meshing signal," and 'b' is the measured noise representing the vibration of other engine components.

[0111] meshing signal y g It can be written as follows:

[0112]

[0113] In the expression:

[0114] -M is the number of harmonics in the meshing component (frequency). This number M is predetermined;

[0115] -a m [k] and Ф m [k] represents the amplitude and phase modulation of the m-th meshing component, respectively;

[0116] -θ m [k] is the instantaneous phase of the engagement of the m-th meshing component, which can be written as:

[0117]

[0118] Where f r Z is the rotational frequency of the shaft, Z is the cycle order of the meshing frequency (i.e., the ratio of the contact frequency between the two teeth of the gear to the contact frequency of the rotating shaft), and t e It is the sampling period (i.e., the sampling frequency f). e (the reciprocal).

[0119] Given meshing signal y g The analytical expression for the vibration signal y c It can still be written as:

[0120]

[0121] in Corresponding to carrier The complex envelope, v1, corresponds to the measurement noise that includes the bearing vibration signal and vibration signals from other parts of the system. This is a characteristic of the gear's state. Therefore, the latter's diagnosis is equivalent to estimating and analyzing this complex envelope.

[0122] A defect on one gear is manifested by the periodicity of the complex envelope. This periodicity is equal to the rotation period of the faulty gear. From an analytical perspective, the complex envelope... It can be approximated using Fourier series. By using θ r [k] is defined as equal to θ m The ratio of [k] to quantity Z xm (θ) r [k] Therefore, corresponding to the angular displacement along the rotation axis, complex envelope It can be written as follows:

[0123]

[0124] In the expression:

[0125] -N b It is the maximum number of side lines (bands) around the meshing frequency. The number N b It was predetermined;

[0126] -o r1 and o r2 These are the cycle orders of the rotational frequencies of the first and second gears, respectively.

[0127] -α n,m and β n,m These are Fourier coefficients;

[0128] -b[k] is a vector, which can be written as:

[0129]

[0130] Where b r1 [k] equals

[0131]

[0132] And b r2 [k] equals

[0133]

[0134] -pm[k] is a vector, which can be written as:

[0135]

[0136] Where, p m,r1 [k] equals p m,r2 [k] equals

[0137] Therefore, the vibration signal y c Satisfy the following expression:

[0138]

[0139] In the expression:

[0140] -c e [k] equals It is the vector of the carrier wave;

[0141] -x e [k] equals It contains the proximity engagement signal y g Complex envelope The vector of coefficients of the Fourier series.

[0142] Complex envelope The estimate is simplified to include the variable x e The estimation of the Fourier coefficients in the variable revolutions per minute (RPM) is performed by taking into account the possible variations in the Fourier coefficients. Therefore, in this exemplary embodiment, the Fourier coefficients and thus the variable x are assumed. e It varies according to the random walk given by the following expression:

[0143] x e [k+1]=x e [k]+w e [k]

[0144] In the expression, w e [k] is a random signal with any statistical properties.

[0145] Finally, the y given above c [k] and x e The relationship [k] forms a model of the meshing signal. Due to this model, real-time gear monitoring is simplified to estimating the meshing signal y. g Complex envelope Fourier coefficients.

[0146] So far, the meshing signal y has been discussed. g The analytical model is as follows. Furthermore, the bearing signal y is now proposed. r The analysis model is used to ultimately give the analysis model of the signal SE[k], and then the matrix H[k] that forms the state model according to the present invention is expressed.

[0147] The vibration generated by a bearing mounted on a shaft can be expressed by the following expression:

[0148]

[0149] In the expression:

[0150] -M[k] is the load distribution function when the inner ring of the bearing is subjected to radial load. Under static conditions, this distribution function is known to be periodic, with a period equal to the period of rotation of the reference shaft. For more details on these aspects, those skilled in the art can refer to, for example, the reference: "Cyclic spectral analysis of rolling-element bearing signals: Facts and fictions", J. Antoni, Journal of Sound and Vibration 304, 2007, 497-529;

[0151] -κ(ω[k]) is a modulation function that depends on the angular velocity ω of the axis;

[0152] -A i Let A be the amplitude of the i-th impact. It has a Gaussian distribution between 0 and 1, such that A... i =A+δA i A is the mean of the distribution, δA i It is the random part;

[0153] -I is the impulse response of the bearing structure;

[0154] - Refers to the integer part of a decimal number x;

[0155] -d represents the number of impacts that may be caused by bearing defects;

[0156] -T i It is the moment when the i-th impact occurs, hence T i =t(iθ) d +δθ i ) is the angular period of the possible bearing defect, and δθ i It is a central Gaussian distribution.

[0157] Vibration signal y c The square of the envelope of the signal, i.e., the signal SE[k], is given by the following expression:

[0158]

[0159] Among the symbols Let denote the conjugate of the complex number a, and It represents the mathematical expectation.

[0160] In this exemplary embodiment, it is assumed that the engagement signal y g Bearing signal y r It is uncorrelated with the noise b. Therefore, the square of the envelope becomes:

[0161]

[0162] In the expression, n2[k] equals It is considered noise of any statistical nature.

[0163] The presence of defects in the bearing is manifested in the bearing signal y. r It is periodic and has an autocorrelation function. Therefore, the signal SE[k] is periodic and contaminated with noise n2[k]. Based on this observation, the signal SE[k] can be approximated by a Fourier series and can be written as follows:

[0164]

[0165] In the expression:

[0166] -μ z [k] and Ф z [k] represents the amplitude and phase of the z-th component of the Fourier series, respectively;

[0167] -l indicates the number of lines associated with bearing defects (l represents the acceleration signal y used to monitor different components of the bearing). c The number of harmonics of interest in the square of the envelope. This number l is predetermined;

[0168] -v2 is noise of any statistical nature, and when l is much lower than the number of valid lines associated with bearing defects, it includes noise n2 and a portion of the possible lines associated with bearing defects.

[0169] Through trigonometric transformation, the signal SE[k] can be finally written as:

[0170]

[0171] In the expression:

[0172] -h z [k] equals

[0173] -x z [k] equals

[0174] By grouping the components of the Fourier series, the signal SE[k] can be expressed in the following form:

[0175]

[0176] In the expression:

[0177] -c r[k] equals

[0178] -x r [k] equals

[0179] With engagement signal y g The model is analyzed and modeled in the same way. In this exemplary embodiment, it is assumed that the coefficients of the Fourier series of the signal SE[k] follow a random walk, such that:

[0180] x r [k+1]=x r [k]+w r [k]

[0181] In the expression, w r [k] is finite-energy random noise.

[0182] Based on previous developments, the state model can now be expressed using the following expression:

[0183]

[0184] In the expression:

[0185] -v[k] is related to the measured value y c [k] Associated noise;

[0186] -x[k] is a vector, called the "state vector", which for each mechanical component includes a subvector (x e [k] or x r [k]), where the components of the subvector represent the contribution of the mechanical component to vector Y[k]. It is worth noting that the state vector x contains unknowns from the state model.

[0187] Based on previous developments, the state vector x can also be represented in the following way:

[0188]

[0189] In the expression, w[k] is the state noise of the model, and F represents the identity matrix.

[0190] The expressions for y[k] and x[k+1] represent the vibration signal y c An enhanced state model is derived. The parameters of this state model, and therefore the quantitative characteristics of the bearing and gear states, can be determined based on these expressions.

[0191] Finally, once the matrix H[k] is determined, the monitoring method includes determining an estimate of the vector x[k] based on the matrix H[k]. Step E40Step E40 is implemented by the second determination module MOD_DET2 equipped in the monitoring device 12.

[0192] According to a particular exemplary implementation, an estimate of vector x[k] is determined using a minimax optimization algorithm. More specifically, starting with the previously described expression for the state vector x, a minimax estimate is achieved for a linear combination of state variables, denoted as: And defined as follows:

[0193]

[0194] In the expression, It is a robust estimate of the state vector x[k]. The estimate... The following recursive equation is satisfied:

[0195]

[0196] Where g[k] corresponds to the gain of the estimator.

[0197] By paying attention to e[k], the estimation error (i.e., Then the gain of the minima estimator can be determined by minimizing the following quadratic cost function J:

[0198]

[0199] In the expression, the function J is strictly positive, and P[1], Q, and R are diagonally weighted matrices of the initialization error e[1], the state noise w[k], and the measurement noise v[k], respectively (such matrices, especially their respective settings, are known to those skilled in the art). In this exemplary embodiment, the cost function J is minimized for the worst case, which is equivalent to making J relative to... Minimize and maximize J with respect to e[1], w[k], and v[k]. This leads to the following minimax optimization formula:

[0200]

[0201] Such optimization problems can be solved using the Lagrange multiplier method, thus obtaining the following expression for the gain g[k]:

[0202] g[k]=P[k]Γ[k-1]H T [k]R -1

[0203] In the expression:

[0204] -P is a positive definite symmetric matrix that satisfies the following Riccati equation:

[0205] P[k]=P[k-1]Γ[k-1]+Q

[0206] -Г[k] is given by the following expression:

[0207]

[0208] Where γ is strictly less than Sup R -1 (“Sup” defines the supremum of the inverse of the weighted matrix R).

[0209] For further details regarding the minimax optimization and gain acquisition g[k] developed in this exemplary embodiment, those skilled in the art can refer to the reference “Discrete H-infinity filter design with application to speech enhancement”, Shen, X., ICASSP, Vol. 2, pp. 1504–1507, 1995.

[0210] It is important to note that, according to other exemplary embodiments of step E40 of the monitoring method, it is not excluded that the estimate may be determined by an algorithm other than the minimax optimization algorithm described above. For example, a least squares optimization algorithm may be used.

[0211] Once the estimator of vector x[k] is determined, potential defects in each mechanical component placed on the shaft of the aircraft engine can be detected. To this end, in Figure 3 In the implementation of the method, the set of steps ENS_E includes multiple steps for each of the mechanical components.

[0212] For the purposes of documentation, in this embodiment, the mechanical components arranged on the shaft include a bearing and a gear formed by two meshing wheels. Furthermore, for the remainder of the specification, the various steps of the set ENS_E are initially performed and described with respect to the bearing. However, it should be understood that this is purely an arbitrary choice, and the various steps could also be initially performed with respect to the gear.

[0213] like Figure 3 As shown, the multiple steps first include Step E50 The contribution Q_R of the bearing to the vector Y[k] is determined based on the estimated value of the vector x[k]. Step E50 is implemented by the third determination module MOD_DET3 equipped in the monitoring device 12.

[0214] The quantity Q_R more specifically characterizes the health state of the bearing. It is obtained from the bearing's contribution, that is, from the subvector x included in the state vector x[k] by employing the exemplary implementation described above for step E30. r [k] is obtained. In this exemplary embodiment, the contribution of the bearing corresponds to the estimation of the signal SE[k]. Note that such an estimation is accessible due to the previously obtained estimate of the state vector x[k] (step E40), not to mention providing the subvector x. r An estimate of [k] is obtained so that the quantity Q_R can be determined.

[0215] As a non-limiting example, the quantity Q_R represents the energy of the signal SE[k].

[0216] However, it is not ruled out that another type of quantity Q_R could be conceived, such as the phase or energy contribution of the bearing.

[0217] The multiple steps also include Step E60 The quantity Q_R determined for the bearing is compared with the threshold S_R to obtain a comparison result. Step E60 is implemented by the comparison module MOD_COMP equipped in the monitoring device 12.

[0218] In fact, the comparison result corresponds to information data indicating whether the indicator Q_R is greater than or less than the threshold S_R, and possible defects in the bearing can be detected from this information data.

[0219] It should be noted that the threshold S_R can be determined by an expert and recorded in the storage device of the monitoring device 12 (e.g., in non-volatile memory 4), or, according to another example, correspond to the amount Q_R determined in a previous implementation of the monitoring method according to the invention, in which no defects in the bearing were detected.

[0220] Therefore, the multiple steps of the set ENS_E also include Step E70 The bearing is then inspected for potential defects based on the comparison results obtained after step E60. Step E70 is implemented by the MOD_DETECT detection module equipped in the monitoring device 12.

[0221] Therefore, in the current implementation, if the quantity Q_R is higher (and correspondingly lower) than the threshold S_R, a defect in the bearing is detected (correspondingly, no defect exists in the bearing). This detection is based on a comparison between numerical quantities (Q_R and S_R). In other words, at this stage of the monitoring method, information about the bearing's potentially non-compliant health condition corresponds to digital information, typically encoded in bits.

[0222] Furthermore, in this embodiment, the multiple steps of the set ENS_E also include Step E80 If a defect is detected in the bearing, an alarm is issued. Step E80 is implemented by the alarm module MOD_ALERT equipped in the monitoring device 12.

[0223] For example, the alarm corresponds to an audio signal, a light signal, etc., which can be transmitted to the ground operator responsible for maintaining the aircraft and / or the aircraft's cockpit to alert the crew.

[0224] Furthermore, communication with users (e.g., ground operators or crew members) regarding the status of the monitoring system 10 can be conducted in the form of graphics projected onto a display device (e.g., a screen) dedicated to this purpose.

[0225] It's important to note that if no defect is detected, then of course no alarm will be issued. Figure 3 In Step E85 ).

[0226] Subsequently, for the gears arranged on the shaft, steps E50 to E70 and E80 or E85 are repeated. To this end, the amount Q_E characteristic of the gear's contribution to the vector Y[k] is determined (in this exemplary embodiment, this is the complex envelope). The amplitude, phase, or energy of the gear is measured, and then this quantity Q_E is compared with a threshold S_E to detect possible defects in the gear.

[0227] Although steps E50 to E70 and E80 or E85 have been described as being performed sequentially for each mechanical component arranged on the shaft, the present invention also covers the case where these steps are performed in parallel for each of the components.

[0228] In addition, if in Figure 3 While the implementation takes into account the execution of steps E50 to E70 and E80 or E85 on all mechanical components arranged on the shaft, other implementations are also conceivable, in which these steps involve only a portion of the mechanical components.

[0229] The invention has so far been described with the monitoring device 12 deployed in an aircraft in mind. However, it is not excluded that the latter may be located on the ground and communicate with the acquisition device deployed in the aircraft via any protocol known to those skilled in the art.

[0230] Furthermore, the measurement of absolute acceleration y received by monitoring device 12 in each instance is also taken into account. cThe monitoring method is described by implementing the steps in set ENS_E after [k]. However, this monitoring method is still applicable to the case where the monitoring device stores multiple absolute acceleration measurements (after receiving them from the acquisition device 11) and then implements the steps in set ENS_E for each such stored measurement.

[0231] The monitoring method is also described by considering an implementation in which an alarm is issued if a defect is detected. However, it is not excluded that the monitoring device 12 may not issue an alarm, but rather that the detection results are analyzed by experts (in real time or in a delayed manner), and the experts may therefore decide whether it is necessary to verify the defect detection performed by the monitoring device 12.

[0232] It should also be noted that if Figure 3 The implementation itself does not include acquiring measurements useful for monitoring the health status of mechanical components placed on the shaft, so it is still conceivable that one or more steps of the monitoring method (e.g., acquiring at least one measurement y) c [k] steps and / or acquiring rotational speed v r Other modes of acquiring measurements during step [k']. In this case, the monitoring method is no longer implemented solely by the monitoring device 12, but by all or part of the acquisition device 11.

[0233] Finally, it should be noted that, more generally, the implementation of this invention is not limited to the field of aviation. It is indeed still applicable to any type of rotating machinery, regardless of the technical field involved, such as electric motors, power transmission motors, electromechanical actuators, wind turbines, gear shafts, etc.

Claims

1. A method for monitoring the state of mechanical components on a shaft system of rotating machinery, the method comprising acquiring at least one measurement of the absolute acceleration y of the shaft in a fixed reference frame associated with the rotating machinery. c The steps of [k], where k is an integer index, and the set of steps (ENS_E): - Obtain at least one measurement value y previously acquired. c The value of the rotational frequency f of the axis at time [k]. r [k], - Determine a matrix H[k] such that a state model can be defined as follows: Where w[k] is random noise, and v[k] is the noise relative to the at least one measured value y. c [k] is associated noise, H[k] is determined by the value f r [k] is defined, SE[k] equals y. c [k] and y c The expected value of the product of the conjugate values ​​of [k], where x[k] is a vector, and for each mechanical component, the vector comprises subvectors whose components represent the contribution of the mechanical component to the vector Y[k]. - Determine the estimate of vector x[k] from the matrix H[k]. For at least one mechanical component, the set of steps further includes the following steps: - Determine the quantitative characteristics of the contribution associated with the mechanical component from the estimated quantity. - Based on a comparison of the amount of contribution associated with the mechanical component with a threshold, a possible defect in the at least one mechanical component is detected.

2. The method according to claim 1, wherein the estimator of vector x[k] is determined by means of a minimax optimization algorithm or a least squares optimization algorithm.

3. The method of claim 1, wherein the amount of the contribution associated with the mechanical component represents the magnitude, phase, or energy of the contribution.

4. The method according to any one of claims 1 to 3, wherein if a defect is detected, the method further includes the step of issuing an alarm.

5. The method according to any one of claims 1 to 3, wherein a plurality of absolute acceleration measurements are acquired cyclically, and the set of steps (ENS_E) is performed after each acquisition of an absolute acceleration measurement.

6. A computer program comprising, when executed by a computer, instructions for implementing a method for monitoring the state of mechanical components on a shaft system of rotating machinery according to any one of claims 1 to 5.

7. A computer-readable recording medium having a computer program as described in claim 6 recorded thereon.

8. A device (12) for monitoring the condition of mechanical components on a shaft system of rotating machinery, the device comprising: - The first acquisition module (MOD_OBT1) is configured to acquire at least one measurement of the absolute acceleration y of the axis in a fixed reference frame associated with the rotating machinery. c [k], where k is an integer index. - The second acquisition module (MOD_OBT2) is configured to acquire the at least one measurement value y during acquisition. c The value of the rotational frequency f of the axis at time [k]. r [k], - The first determining module (MOD_DET1) is configured to determine a matrix H[k], which enables the definition of a state model, described as follows: Where w[k] is random noise, and v[k] is the noise relative to the at least one measured value y. c [k] is associated noise, H[k] is based on the value f r [k] is defined, SE[k] equals y. c [k] and y c The expected value of the product of the conjugate values ​​of [k], where x[k] is a vector, and for each mechanical component, the vector comprises subvectors whose components represent the contribution of the mechanical component to the vector Y[k]. - The second determining module (MOD_DET2) is configured to determine an estimate of the vector x[k] from the matrix H[k]. - A third determining module (MOD_DET3) is configured to determine at least one quantitative characteristic of the contribution associated with the mechanical component from the estimated quantity. - A comparison module (MOD_COMP) is configured to compare at least one amount of the contribution associated with the mechanical component with a threshold to obtain a comparison result. - A detection module (MOD_DETECT) is configured to detect potential defects in the at least one mechanical component based on comparison results.

9. A system (10) for monitoring the condition of mechanical components on a shaft system of rotating machinery, the system comprising: - At least one measurement y used to obtain the absolute acceleration of the shaft in a fixed reference frame associated with the rotating machinery. c The device (11) of [k], where k is an integer index. - The device (12) for monitoring the condition of mechanical components on a shaft system equipped with rotating machinery, as claimed in claim 8.

10. An aircraft comprising a system (10) for monitoring the condition of mechanical components on a shaft system of rotating machinery, as claimed in claim 9.

Images