Method, device and program for measuring erosive cavitation in a hydraulic machine

Abstract

The invention relates to a method for measuring erosive cavitation of a hydraulic machine (1) in operation, characterized by the emission of a broadband pulsed ultrasonic signal (e) by an ultrasonic transmitter transducer (20), directed through the liquid water passage conduit to another side of the wall or towards the moving part, the reception, by an ultrasonic receiver transducer (30), of an ultrasonic signal (x) in response to the signal (e), calculation, by a calculation unit (40), from the signals (e, x), of an erosive cavitation indicator, from the calculation of at least one arrival time (TOAn) of the received signal (x) relative to the emitted signal (e) and / or from an instantaneous spectral band (BWk) defined by a frequency range, for which a representative value of the power spectral density of the signal (x) is greater than a percentage (A) of its average value. Figure for the summary: Figure 5

Description

Title of the invention: Method, device and program for measuring erosive cavitation of a hydraulic machine

[0001] The invention relates to a method, a device and a computer program for measuring erosive cavitation of a hydraulic machine.

[0002] The field of the invention relates to devices for monitoring hydraulic machines.

[0003] The phenomenon of cavitation is well known and is found in components driving fluids such as gas or liquid, or driven by these fluids. Homogeneous liquids can be put under tension by reducing the pressure at which they are located or by modifying the temperature.

[0004] Cavitation means the transformation of a fluid from the liquid state to the two-phase liquid and gaseous state, caused by the liquid being put under tension by reducing the pressure.

[0005] In the context of intermittent use of hydraulic energy, operations in sub-optimal regimes are almost present, which increases the risk of cavitation phenomena. The immediate effect of these phenomena is the loss of efficiency but, under certain conditions impossible to predict, they can lead to the erosion of the machine components (for example rotating blades and guide vanes in the case of a hydraulic turbine) causing irreversible breakdowns with serious economic consequences.

[0006] When the tension in the liquid reaches a certain value, called critical value, there is a rupture in the liquid and it becomes two-phase (liquid and vapor). A liquid put under tension by the drop in pressure or heated is in a phase called metastable which can lead either to a cavitation phase (presence of fluid in the liquid state and in the vapor state), or to a boiling phase.

[0007] In the phase diagram of [Fig. 17], one can easily understand the passage of a liquid into its metastable phase and, subsequently, into the cavitation phase (liquid and vapor). The RC line (starting from the dew point R for increasing temperatures) is the curve of variation of the vaporization pressure with temperature and is the coexistence curve or limit of the phases where the liquid can exist in equilibrium with the vapors.

[0008] Generally speaking, a hydraulic structure is composed of numerous pieces of equipment, such as pumps, turbines, valves, etc., which, under certain conditions, can, by the rapid movement of their flow adjustment members, such as blades and guides, lattices, or by the flow along the hydraulic profile, cause variations in fluid pressure which may be accompanied by a passage from the fluid zone, initially in liquid state, to a mixed state: liquid-vapor.

[0009] In the context of a flow in a hydraulic machine, the appearance and development of cavitation can produce mechanical damage by implosion of the cavitation bubbles against the walls or mechanical zones. Cavitation is therefore manifested by the appearance of clouds of bubbles forming vapor particles created by the local and transient change in the pressure of the fluid. The conditions which contribute to their creation are operating regimes for which the moving parts of the installation, such as for example the blades and the guide vanes in the case of a hydraulic turbine, change configuration (for example, in a situation of change of regime (change of rotation speed of the blades or others) or in the event of variation of the fluid flow rate).Cavitation can also occur during steady state at a certain machine load - due to the fact that the hydraulic profile of the blades can generate a type of cavitation.

[0010] The consequences for the components undergoing the phenomenon of cavitations are linked to operation in a variable pressure regime. The phenomenon of cavitation, when it is not identified and especially not controlled, leads to premature mechanical wear of the parts subjected to cavitation or to irreversible destruction of these parts of the machine. This is particularly true when the cavitation is erosive.

[0011] Cavitation will be said to be of the erosive type when the cloud of bubbles caused by cavitation implodes on the organs (which may be metallic, for example) surrounding the cavitation zone.

[0012] In particular, hydroelectric turbines can be heavily affected by blade wear / erosion caused by sediment and cavitation. Wear due to erosive cavitation can significantly reduce the mean time between failures (MTBF), leading to a higher number of shutdowns for corrective maintenance.

[0013] Therefore, it is desirable to monitor this wear / erosion accurately, and above all, it is recommended to avoid the cavitation phenomenon upstream. It is recommended to avoid producing the conditions that accelerate this damage by detecting cavitation or erosion by sediments in their early stages and, ideally, continuously without interrupting the operation of the machine.

[0014] For a hydraulic group the estimated costs of repairing the effects of cavitation are very high, not to mention the consequent losses linked to shutdowns for corrective maintenance. These repair works could be avoided if the erosive cavitation were monitored during operation.

[0015] An objective of the invention is to obtain a method, a device and a computer program for measuring erosive cavitation of a hydraulic machine, which allow monitoring the onset and development of erosive cavitation and which solve the problem mentioned above.

[0016] To this end, a first object of the invention is a method for measuring erosive cavitation of a hydraulic machine in operation, the hydraulic machine comprising walls delimiting a liquid water passage conduit and at least one movable member, which is located in the liquid water passage conduit and which is set in motion during operation of the hydraulic machine, characterized in that the method comprises the following steps: during a first step, emission of at least one first broadband pulsed ultrasonic signal by at least one emitting ultrasonic transducer, which is positioned on a first side of at least one first of the walls and which is oriented through the liquid water passage conduit towards another side of the first wall (17) or towards the at least one moving member, during a second step, reception, by at least one receiving ultrasonic transducer, which is positioned on the first side of the first wall, of at least one second ultrasonic signal received in response to the first ultrasonic signal having been emitted, during a third step, calculation, by a calculation unit connected to the at least one transmitting ultrasonic transducer and to the at least one receiving ultrasonic transducer, from the first ultrasonic signal having been transmitted and from the second ultrasonic signal having been received, of at least one indicator of erosive cavitation of the hydraulic machine, from the calculation by the calculation unit of at least one arrival time of the second ultrasonic signal having been received relative to the first ultrasonic signal having been transmitted and / or from the calculation by the calculation unit of an instantaneous spectral band defined by a frequency range, for which a value representative of the power spectral density of the second ultrasonic signal having been received is greater than a first given non-zero percentage of the calculated average value of the value representative of the power spectral density over time.

[0017] Thanks to the invention, the erosive cavitation phenomena that appear on certain components in hydraulic machines are monitored in a non-intrusive manner and without the need to stop the machine, which may be, for example, in the case of a hydraulic turbine: the rotating blades, and therefore the turbine wheel, as well as the turbine guide vanes. The invention makes it possible to detect the bubbles created by the cavitation phenomena. The invention makes it possible to calculate an erosive cavitation indicator. Monitoring this erosive cavitation indicator ensures effective and low-cost control (without the need to stop the machine) over time of the cavitation phenomena and their impact on the wear of the machine.

[0018] Furthermore, the method and the device according to the invention are less expensive, more faster and more accurate than known numerical simulation techniques, more accurate than known techniques for measuring pressure fluctuations and torque variation in relation to power demands, more accurate than known techniques for measuring vibration using accelerometers and require less data volume and are more accurate than a known technique for multidimensional analysis of variables in the system, such as the position of the guide vanes, the axial deflection of the wheel, the water temperature in the case of a hydraulic turbine.

[0019] According to one embodiment of the invention, the first broadband pulsed ultrasonic signal comprises a plurality of pulses emitted with a prescribed, non-zero pulse repetition period.

[0020] According to an embodiment of the invention, each pulse has, during a modulation duration, less than the prescribed pulse repetition period, a frequency modulation greater than or equal to a low frequency modulation value, non-zero, and less than or equal to a high frequency modulation value, non-zero.

[0021] According to one embodiment of the invention, the frequency modulation is linear or cubic and goes from the low frequency modulation value to the high frequency modulation value over the modulation duration or from the high frequency modulation value to the low frequency modulation value over the modulation duration.

[0022] According to an embodiment of the invention, during the third step, the calculation unit calculates for each pulse of the first ultrasonic signal e(t) having been emitted the arrival time TOAn of the second ultrasonic signal x(t) having been received relative to the first ultrasonic signal e(t) having been emitted according to the following equations: TOA„ = argmax t [ G ( t ) ] 2 [°O231

[0024] According to an embodiment of the invention, during the third step, the calculation unit calculates for each pulse of the first ultrasonic signal e(t) having been emitted the arrival time TOAn of the second ultrasonic signal x(t) having been received relative to the first ultrasonic signal e(t) having been emitted according to the following equations: TOA„ = argmax t [ G ( t ) ] 2

[0025] — FFT^[FFT{x} n is the index of each pulse, FFT{x] is the fast Fourier transform of the second ultrasonic signal x(t) received, FFT{e] is the complex conjugate of the fast Fourier transform of the first where n is 1 index of each pulse. ultrasonic signal e(t), FFF} is the inverse fast Fourier transform.

[0026] According to an embodiment of the invention, during the third step, the calculation unit calculates the at least one indicator of erosive cavitation of the hydraulic machine as comprising the average of the arrival times TOAn for the pulses of the first ultrasonic signal e(t) having been emitted and / or the variance of the arrival times TOAn for the pulses of the first ultrasonic signal e(t) having been emitted and / or the standard deviation of the arrival times TOAn for the pulses of the first ultrasonic signal e(t) having been emitted.

[0027] According to an embodiment of the invention, during the third step, the calculation unit determines that cavitation is present, when the average of the arrival times TOAn is less than a first determined threshold.

[0028] According to an embodiment of the invention, during the third step, the calculation unit determines that cavitation is present, when the variance of the arrival times TOAn is greater than a determined threshold.

[0029] According to an embodiment of the invention, during the third step, when the calculation unit has determined that cavitation is present, the calculation unit determines a position of the cavitation in the region defined by the product between the propagation speed of the ultrasonic signals and the average of the arrival times TOAn, divided by 2.

[0030] According to an embodiment of the invention, during the third step, when the calculation unit has determined that cavitation is present, the calculation unit determines a lifetime of the cavitation from the variance of the arrival times TOAn and / or from the average of the arrival times TOAn.

[0031] According to one embodiment of the invention, during the third step, the calculation unit calculates: the power spectral density DSP;(k)(f) of the second ultrasonic signal xk(t) having been received for each of several reception time windows F; having a prescribed duration W covering the pulse n, the reception time windows F; being temporally shifted one after the other for i ranging from 1 to M, according to the following equation: ,.. , । p ty 12where M is a number DSP^(f) = IJyw+fyÇt) ■ exp( -jlnft) • dt^ prescribed time windows F; reception, the representative value DSP^^f) of the power spectral density DSP^ff) of the second ultrasonic signal xk(t) having been received being equal to the average power spectral density DSP^^f) for the power spectral density DSP;(k)(f) over the M reception time windows F;, according to the following equation: DSP(k\f) = 1 DgpW the calculated average value DSPmean of the value representative DSP(k)(f) of the power spectral density DSP;(k) over time according to the following equation: । f-^2 zx , the instantaneous BWk spectral band defined by the frequency range suppf, for which the representative value DSP(k)(f) of the density power spectral DSP;(k)(f) of the second ultrasonic signal xk(t) having been received is greater than the first given non-zero percentage A of the DSPmean mean value calculated from the representative value DSP^^f) of the DSP power spectral density; (k)(f) over time according to the following equation: BWt = supp^DSP^if ) a A-DSP,„„„] where fi is the low modulation value

[0032] of frequency, f2 is the high frequency modulation value, f is the frequency, t is time. According to one embodiment of the invention, the reception time windows F; overlap one after the other.

[0033] According to one embodiment of the invention, the prescribed duration W of the reception time window F; is greater than or equal to 1 time or 2 times or 3 times the prescribed pulse repetition period.

[0034] According to an embodiment of the invention, during the third step, when the calculation unit has determined that cavitation is present, the calculation unit determines a direction of movement of the cavitation from the instantaneous spectral band.

[0035] According to one embodiment of the invention, the first non-zero percentage A given is greater than or equal to 50% or 70% or 90% or 95%.

[0036] According to an embodiment of the invention, the at least one transmitting ultrasonic transducer comprises an active mode, which can be controlled by the calculation unit and in which the first step, the second step and the third step are carried out, and a passive mode, which can be controlled by the calculation unit and in which the first step, the second step and the third step are not carried out and the at least one transmitting ultrasonic transducer does not transmit, the calculation unit controlling by default during the third step the at least one transmitting ultrasonic transducer in the passive mode, in which the at least one receiving ultrasonic transducer receives a third signal, the calculation unit calculating during the third step the power spectral density of the third signal having been received in each of several successive time slots and each having an identical time slot duration, the computing unit controlling during the third step the at least one emitting ultrasonic transducer in the active mode, when the power spectral density of the third signal having been received in a current time slot differs by more than a second given non-zero percentage compared to the power spectral density of the third signal having been received in a time slot prior to the current time slot.

[0037] According to one embodiment of the invention, the time slot prior to the current time slot is the time slot preceding the current time slot.

[0038] A second object of the invention is a device for measuring erosive cavitation of a hydraulic machine in operation, the hydraulic machine comprising walls delimiting a liquid water passage conduit and at least one movable member, which is located in the liquid water passage conduit and which is set in motion during operation of the hydraulic machine, characterized in that the measuring device comprises: at least one emitting ultrasonic transducer, which is positioned on a first side of at least one first of the walls and which is oriented through the liquid water passage conduit towards another side of the first wall or towards the at least one moving member, the at least one emitting ultrasonic transducer being capable of emitting at least one first broadband pulsed ultrasonic signal, at least one receiving ultrasonic transducer, which is positioned on the first side of the first wall, and which is capable of receiving at least a second ultrasonic signal received in response to the first ultrasonic signal having been emitted, a calculation unit connected to the at least one transmitting ultrasonic transducer and to the at least one receiving ultrasonic transducer, the calculation unit being capable of calculating, from the first ultrasonic signal having been transmitted and from the second ultrasonic signal having been received, at least one indicator of erosive cavitation of the hydraulic machine, from the calculation by the calculation unit of at least one arrival time of the second ultrasonic signal having been received relative to the first ultrasonic signal having been transmitted and / or from the calculation by the calculation unit of an instantaneous spectral band defined by a frequency range, for which a value representative of the power spectral density of the second ultrasonic signal having been received is greater than a first given non-zero percentage of the calculated average value of the value representative of the power spectral density over time.

[0039] A second object of the invention is a computer program, comprising code instructions for implementing the third step of the method for measuring erosive cavitation of a hydraulic machine as described above, when implemented on a computer.

[0040] The invention will be better understood upon reading the following description, given solely by way of non-limiting example with reference to the figures below of the attached drawings.

[0041] [Fig. 1] represents a schematic view in open perspective of an example of a hydraulic machine on which the method, the device and the computer program for measuring erosive cavitation according to the invention can be implemented.

[0042] [Fig.2] represents a schematic view in open perspective of an example of hydraulic machine on which the method, device and computer program for measuring erosive cavitation according to the invention can be implemented.

[0043] [Fig.3] represents a schematic cross-sectional view of an example of hydraulic machine on which the method, device and computer program for measuring erosive cavitation according to the invention can be implemented.

[0044] [Fig.4] represents a schematic view in open perspective of an example of hydraulic machine on which the device for measuring erosive cavitation is arranged according to one embodiment of the invention.

[0045] [Fig.5] represents a modular block diagram of the cavitation measuring device erosive according to an embodiment of the invention.

[0046] [Fig.6] represents timing diagrams of signals that can be used by the method, device and computer program for measuring erosive cavitation according to the invention.

[0047] [Fig.7] represents timing diagrams of signals that can be used by the method, device and computer program for measuring erosive cavitation according to the invention.

[0048] [Fig.8] represents timing diagrams of signals that can be used by the method, device and computer program for measuring erosive cavitation according to the invention.

[0049] [Fig.9] represents timing diagrams of signals that can be used by the method, device and computer program for measuring erosive cavitation according to the invention.

[0050] [Fig.10A] represents signal paths which may appear in the hydraulic machine when the method, device and computer program for measuring erosive cavitation according to the invention are used in the absence of cavitation.

[0051] [Fig.10B] represents signal paths which may appear in the hydraulic machine when the method, device and computer program for measuring erosive cavitation according to the invention are used in the absence of cavitation.

[0052] [Fig. 10C] represents signals which may appear in the hydraulic machine when the method, device and computer program for measuring erosive cavitation according to the invention are used in the absence of cavitation.

[0053] [Fig.lOD] represents signal values ​​that may appear in the machine hy draulic when the method, device and computer program for measuring erosive cavitation according to the invention are used in the absence of cavitation.

[0054] [Fig.10E] represents signal values ​​which may appear in the hydraulic machine when the method, device and computer program for measuring erosive cavitation according to the invention are used in the absence of cavitation.

[0055] [Fig. 11 A] represents signal paths that may appear in the hy machine draulic when the method, device and computer program for measuring erosive cavitation according to the invention are used in the presence of cavitation.

[0056] [Fig. 1 IB] represents signal paths that may appear in the hydraulic machine when the method, device and computer program for measuring erosive cavitation according to the invention are used in the presence of cavitation.

[0057] [Fig. 1 IC] represents signals that may appear in the hydraulic machine when the method, device and computer program for measuring erosive cavitation according to the invention are used in the presence of cavitation.

[0058] [Fig. 11D] represents signal values ​​that may appear in the hydraulic machine when the method, device and computer program for measuring erosive cavitation according to the invention are used in the presence of cavitation.

[0059] [Fig. 11E] represents signal values ​​that may appear in the hydraulic machine when the method, device and computer program for measuring erosive cavitation according to the invention are used in the presence of cavitation.

[0060] [Fig. 12] represents examples of arrival times, having been calculated by the method, the device and the computer program for measuring erosive cavitation according to the invention.

[0061] [Fig. 13] shows other examples of arrival times, having been calculated by the method, the device and the computer program for measuring erosive cavitation according to the invention.

[0062] [Fig. 14] shows other examples of arrival times, having been calculated by the method, the device and the computer program for measuring erosive cavitation according to the invention.

[0063] [Fig. 15] shows other examples of arrival times, having been calculated by the method, the device and the computer program for measuring erosive cavitation according to the invention.

[0064] [Fig. 16] represents a flowchart of the method for measuring erosive cavitation according to the invention.

[0065] [Fig. 17] represents a diagram of the phases of water.

[0066] An exemplary embodiment of a method for measuring erosive cavitation of a hydraulic machine 1, as well as a device 100 for measuring erosive cavitation of the hy- machine, is described in more detail below with reference to FIGS. 1 to 16. draulique 1 implementing this process.

[0067] The hydraulic machine 1 comprises walls delimiting a conduit 12 for the passage of liquid water. The hydraulic machine 1 comprises one or more movable members 11, which are set in motion in the conduit 12 for the passage of liquid water during operation of the hydraulic machine.

[0068] The hydraulic machine 1 can be any machine using water and can be for example a hydraulic turbine 1, or others, such as for example a hydraulic pump or a hydraulic valve or others, or any hydraulic equipment in a hydroelectric plant for producing electricity.

[0069] A hydraulic machine 1 formed by a hydraulic turbine 1 is described below with reference to Figures 1 to 4. The hydraulic turbine 1 is used for the production of electricity by the passage of water and can be used, for example, in a hydroelectric plant. The hydraulic turbine 1 can be a turbo-alternator turbine for generating electricity or a turbine of an electricity generator, or the like.

[0070] The turbine 1 comprises a rotation shaft 2, which is mounted to rotate in a frame 10 about a first axis 8 of rotation. Blades 3 located in the liquid water passage conduit 12 are integral in rotation with the rotation shaft 2, to be able to rotate about the first axis 8 of rotation, in order to drive the shaft 2 in rotation about the first axis 8. The shaft 2 comprises a hub 4, on which the blades 3 are fixed and which ends in a warhead 5. The hub 4, the blades 3 and the warhead 5 form a wheel 2a, which is located in the liquid water passage conduit 12 and which is mounted to rotate relative to the frame 10 about the first axis 8 of rotation. The wheel 2a is fixed by a coupling to the shaft 2. The turbine 1 can be for example of the type with vertical shaft 2, horizontal shaft 2 or other.

[0071] The frame 10 comprises a wall 16 delimiting a water inlet path 6, in which liquid water is sent to the blades 3. The frame 10 comprises a wall 17 delimiting a water outlet path 7, in which water is sent from the blades 3. The blades 3 are located between the water inlet path 6 and the water outlet path 7. A wall 13 (or impeller mantle 13) of the frame 10 surrounds the blades 3 at a distance from them. The walls delimiting the liquid water passage conduit 12 comprise the walls 13, 16 and 17.

[0072] The passage of water against the blades 3 from the inlet channel 6 to the outlet channel 7 causes these blades 3 and the rotation shaft 2 to rotate around the first axis 8 of rotation. The blades 3 which are mobile in rotation around the first axis 8 of rotation may be provided as movable member 11. Flow adjustment members, such as guide vanes 9, or lattices, or the like, may also be provided as movable member 11.

[0073] Each blade 3 can be mounted to rotate relative to the hub 4 around a second axis of rotation 14 perpendicular to the first axis 8 of rotation and this in a manner controlled by means of regulating members. The second axes of rotation 14 of the blades 3 are spaced from each other around the first axis 8 of rotation. For example, the blades 3 are identical to each other.

[0074] Between the fluid inlet path 6 and the blades 3 are provided guide vanes 9 for controlling the fluid, which are distributed around the rotation shaft 2. The function of the guide vanes 9 for controlling the fluid is to predetermine a fluid passage section between the fluid inlet path 6 and the fluid outlet path 7.

[0075] Each sluice gate guide vane 9 is mounted to rotate relative to the frame 10 about a third axis 15 of rotation, which may for example be parallel to the first axis 8 of rotation, and this in a manner controlled by means of regulating members. The third axes of rotation 15 of the sluice gate guide vanes 9 are spaced apart from each other about the first axis 8 of rotation. For example, the sluice gate guide vanes 9 are identical to each other.

[0076] The water in the liquid state enters through the water inlet channel 6 1 to be forcibly directed by front-directing vanes 18 towards the guide vanes 9 which direct the water towards the blades 3. The blades 3, and therefore the wheel, are driven by the resulting speed vectors. It is during this path that the water travels that it can gradually transform, to pass, under certain conditions, from the liquid state to a two-phase liquid-vapor state constituting said cavitation. At the outlet, the water passes into the outlet channel 7 and is sent to the downstream reservoir of the machine 1.

[0077] According to the invention, the method for measuring erosive cavitation of the hydraulic machine 1 comprises the following steps, as shown by way of example in [Fig.16].

[0078] The device 100 for measuring erosive cavitation of the hydraulic machine 1 according to the invention comprises one (or more) emitting ultrasonic transducer 20, which is positioned on a first side of one of the walls, called the first wall. The emitting ultrasonic transducer(s) 20 is oriented through the conduit 12 for the passage of liquid water towards the moving member(s) or towards another side of the first wall. For example, a pair of emitting ultrasonic transducers 20 may be provided, as shown in [Fig. 4], or more than two emitting ultrasonic transducers 20.

[0079] In a first case, shown in Figures 3 and 4 in the example described above of the hydraulic turbine 1, the emitting ultrasonic transducer(s) 20 is positioned on the first side 131 of the wall 13 surrounding the blades 3 (this wall 13 forming in this case the first wall 13). The emitting ultrasonic transducer(s) 20 is oriented through the liquid water passage conduit 12 towards the blade(s) 3 (the the blades 3 forming in this case the moving organ(s) 11).

[0080] In a second case, shown in [Fig. 3] in the example described above of the hydraulic turbine 1, the emitting ultrasonic transducer(s) 20 is positioned on the first side 171 of the wall 17 delimiting the water outlet channel 7 (the wall 17 forming in this case the first wall 17). The emitting ultrasonic transducer(s) 20 is oriented through the liquid water passage conduit 12 towards another side 172 of the wall 17 delimiting the water outlet channel 7. This makes it possible to adapt to machines that do not have access around the moving parts 11 (for example the blades 3).

[0081] Of course, the first wall could be other than the walls 13 and 17. Of course, the moving member 11 can be other than the blades 3 and the guide vanes 9.

[0082] During a first step E1, the emitting ultrasonic transducer(s) 20 emits one (or more) first wide-band pulsed ultrasonic signal e or e(t) towards the moving member(s) 11 (which may be the blade(s) 3 in the first case described above) or towards the other side of the first wall (which may be the other side 172 of the wall 17 delimiting the water outlet channel 7 in the second case described above) (during the operation of the hydraulic machine 1, namely while the moving member(s) 11 are set in motion, i.e. while the blades 3 rotate around the first axis 8 of rotation in the case of the hydraulic turbine 1.

[0083] The device 100 for measuring erosive cavitation of the hydraulic machine 1 according to the invention comprises one (or more) receiving ultrasonic transducer 30, which is positioned on the first side of the first wall, on which the transmitting ultrasonic transducer(s) 20 is positioned (this first side of the first wall possibly being the first side 131 of the wall 13 surrounding the blades 3 in the first case described above, this first side of the first wall possibly being the first side 171 of the wall 17 delimiting the water outlet channel 7 in the second case described above). During a second step E2, the receiving ultrasonic transducer(s) 30 receives one (or more) second ultrasonic signal x or x(t) received in response to the first ultrasonic signal e having been transmitted. For example, a pair of receiving ultrasonic transducers 30 may be provided, as shown in [Fig.4], or more than two receiving ultrasonic transducers 30. .

[0084] The measuring device 102 according to the invention comprises a calculation unit 40, connected to the ultrasonic transducer(s) 20 transmitter(s) and to the ultrasonic transducer(s) 30 receiver(s), as shown by way of example in [Fig.5]. The calculation unit 40 may be or comprise one (or more) calculator, one (or more) computer, one (or more) processor, or others. The calculation unit 40 may comprise one (or more) permanent memory, one (or more) random access memory.

[0085] During a third step E3, the calculation unit 40 calculates, from the first ultrasonic signal e having been emitted and from the second ultrasonic signal x having been received, one (or more) indicators of erosive cavitation of the hydraulic machine 1. A computer program, comprising code instructions for implementing the third step E3 of the method for measuring an erosive cavitation of a hydraulic machine 1 can be recorded in the permanent memory of the calculation unit 40 (or of the elements mentioned above thereof) so that it can be executed thereon.

[0086] According to a first embodiment of the invention, the erosive cavitation indicator(s) of the hydraulic machine 1 is performed from the calculation, performed by the calculation unit 40, of one (or more) arrival times TOAn of the second ultrasonic signal x having been received relative to the first ultrasonic signal e having been emitted. The first embodiment of the invention provides a time indicator of erosive cavitation of the hydraulic machine 1.

[0087] According to a second embodiment of the invention, the erosive cavitation indicator(s) of the hydraulic machine 1 is performed from the calculation, performed by the calculation unit 40, of one (or more) instantaneous spectral band(s) BWk. The second embodiment of the invention provides a spectral indicator of erosive cavitation of the hydraulic machine 1. The instantaneous spectral band(s) BWk is defined by the frequency range, for which a representative value DSP^^f) of the power spectral density DSP;(k)(f) of the second ultrasonic signal x having been received is greater than a first given non-zero percentage A of the calculated average value DSPmean of the representative value DSP(k)(f) of the power spectral density DSP^f) over time, f denotes the frequency, t denotes the time. The symbol k represents a time slot k of emission of the first wideband pulsed ultrasonic signal e having been emitted.As shown by way of example in [Fig.6], the first broadband pulsed ultrasonic signal e having been emitted may comprise several successive time slots k-1, k, k+1, k+ 2, ... The time slots k may each have the same duration D. .

[0088] According to an embodiment of the invention, the first broadband pulsed ultrasonic signal e comprises pulses n emitted with a prescribed pulse repetition period Tr, this prescribed period Tr being non-zero. The number S is the number of pulses n in each time slot k of the first broadband pulsed ultrasonic signal e(t) having been emitted. In each time slot k, the pulses n of the first signal e(t) emitted range from n=1 to n=S. We therefore have SD.

[0089] According to an embodiment of the invention, shown as an example in [Fig.6], a frequency modulation greater than or equal to a low frequency modulation value fi, non-zero, and less than or equal to a high frequency modulation value f2, non-zero, is provided in a modulation duration Lp included in each pulse n. The modulation duration Lp is therefore less than the prescribed pulse repetition period Tr. The high frequency modulation value f2 is greater than the low frequency modulation value fi. The low frequency modulation value fi is prescribed and may be, for example, constant. The high frequency modulation value f2 is prescribed and may be, for example, constant. The modulation duration Lp may be provided, for example, at the start of the prescribed pulse repetition period Tr, or otherwise.

[0090] According to an embodiment of the invention, shown as an example in [Fig. 6], the frequency modulation is linear and goes from the low frequency modulation value fi to the high frequency modulation value f2 over the modulation duration Lp and is zero over the remainder of the prescribed pulse repetition period Tr. Alternatively, the frequency modulation is linear and goes from the high frequency modulation value f2 to the low frequency modulation value fi over the modulation duration Lp and is zero over the remainder of the prescribed pulse repetition period Tr.Of course, the frequency modulation could, over the modulation duration Lp, be non-linear from the low frequency modulation value fi to the high frequency modulation value f2 or from the high frequency modulation value f2 to the low frequency modulation value fi and be zero over the rest of the prescribed pulse repetition period Tr.

[0091] According to another embodiment of the invention, not shown, the frequency modulation is cubic and goes from the low frequency modulation value fi to the high frequency modulation value f2 over the modulation duration Lp and is zero over the remainder of the prescribed pulse repetition period Tr, or goes from the high frequency modulation value f2 to the low frequency modulation value fi over the modulation duration Lp and is zero over the remainder of the prescribed pulse repetition period Tr.

[0092] In general, the parameters of the modulation duration Lp of the emission pulses n and also of the repetition period Tr of the pulses n may have been calculated according to the actual parameters of the machine, and include the shortest distance d between the sensors and the nearest point which will interfere with the emission signal e, and the longest distance L between the sensors 20, 30 and the most distant point.

[0093] According to an embodiment of the invention, in the first case described above, the modulation duration Lp is chosen to have Lp < ^where d is the distance between the blades 3 and the wall 13 (wheel mantle) surrounding the blades 3 and c is the propagation speed of the first and second signals ultrasonic waves e and x in water, namely 1480 m / s. This allows the emission to be made before the echo (second signal received x) of each blade 3. For example, the modulation duration Lp is equal to 50 microseconds. Of course, other values ​​of the modulation duration Lp, different from this example, can be chosen.

[0094] According to an embodiment of the invention, in the first case described above, the period Tr of repetition of the pulses n is chosen to have Tr < ^where L is the distance between the wall 13 (wheel mantle) surrounding the blades 3 and the axis of rotation 8 of the blades 3. This allows the echoes (second received signal x) to arrive unambiguously at the receiving ultrasonic transducer(s) 30, if the reflection of the first emitted signal e occurs at the level of the rotation shaft 2. For example, the period Tr of repetition of the pulses n is equal to 20 milliseconds. Of course, other values ​​of the period Tr of repetition of the pulses n, different from this example, can be chosen.

[0095] The duration D of each time slot k and the duration Tr can be chosen so that the number S of pulses n is an integer. Thus in the present example, the duration D of each time slot k can be equal to 1 second for the period Tr of repetition of the pulses n equal to 20 milliseconds. In this case, in each time slot k, the pulses n of the first signal e(t) emitted go from n=1 to S=y = 50.

[0096] According to an exemplary embodiment of the invention, the low frequency modulation value fi is equal to 800 kHz. Of course, the low frequency modulation value fi may be different from this example.

[0097] According to an exemplary embodiment of the invention, the high frequency modulation value f2 is equal to 1200 kHz. Of course, the high frequency modulation value f2 may be different from this example.

[0098] [Fig.7] shows at the top the above example of the first ultrasonic signal e (amplitude in Volts on the ordinate, time t on the abscissa), having been emitted by the first ultrasonic transducer(s) 20, and at the bottom the above example of the second ultrasonic signal x (amplitude in Volts on the ordinate, time t on the abscissa), having been received by the second ultrasonic transducer(s) 30 in response to this first ultrasonic signal e.

[0099] [Fig.7] shows that the second reflected ultrasonic signal x comprises in the first time interval T1 the first emitted ultrasonic signal e, which is propagated in the metal of the wall 13 and which is a vibration signal. This vibration signal can be, according to an embodiment of the invention, extracted by the calculation unit 40, which can estimate from this vibration signal the vibration of the wall 13 (wheel mantle) surrounding the blades 3, which makes it possible to obtain an accelerometer function. This signal represents the part of the first emitted ultrasonic signal e, which propagates between the transmitting ultrasonic transducer 20 and the receiving ultrasonic transducer 30 through the metal wall 13 over a distance of approximately 10 cm. As the propagation speed in steel is approximately 4 times greater than the speed in water (5800 m / s in steel vs. 1480 m / s in water), this signal propagated in the steel arrives first at the receiving transducer 30.

[0100] [Fig.7] shows that the second reflected ultrasonic signal x includes in the second time interval T2 the second ultrasonic signal x reflected by the cavitation pocket but which is superimposed on the vibration signal mentioned above. Indeed, as the cavitation cloud is positioned between the wall 13 of the machine and the blades 3, the second ultrasonic signal x reflected by the cavitation cloud always arrives before the ultrasonic signal reflected by the blades 3, but after the vibration signal (mentioned above) propagated in the steel, because the propagation takes place in the water, at a speed approximately four times lower than that in the steel.

[0101] [Fig.7] shows that the second reflected ultrasonic signal x includes in the third time interval T3 the ultrasonic signals e propagated in the water and which are reflected by the blades 3.

[0102] According to embodiments of the invention, described below, the processing by adapted filtering, carried out by the calculation unit 40, will be correlated with a windowing that is both temporal and spectral which takes into account the measurement configuration and which will make it possible to separate the signals propagated in the metal from those reflected by the cavitation pocket.

[0103] According to a first example of the first embodiment of the invention, the calculation unit 40 calculates for each pulse n of the first ultrasonic signal e(t) having been emitted the arrival time TOAn of the second ultrasonic signal x(t) having been received relative to the first ultrasonic signal e(t) having been emitted according to the following equations: TOA n = argmax t [ G ( t ) ] 2

[0104] ^0 _ J °° r)dT°ù n is 1 each pulse and argmax is the maximum argument of G(t).

[0105] According to a second example of the first embodiment of the invention, the calculation unit 40 calculates for each pulse n of the first ultrasonic signal e(t) having been emitted the arrival time TOAn of the second ultrasonic signal x(t) having been received relative to the first ultrasonic signal e(t) having been emitted according to the following equations: TOA n = argmax t [ G ( t ) p

[0106] Gp) = FFT l[FFT{x} *FÂr{e}]]°hn is the index of each pulse, argmax is the maximum argument of G(t), FFT{x] is the fast Fourier transform of the second ultrasonic signal x(t) received, FFT[e) is the complex conjugate of the fast Fourier transform of the first ultrasonic signal e(t), FF P1 is the inverse fast Fourier transform.

[0107] In these first embodiments of the invention, the temporal processing of the ultrasonic signals e and x aims to determine the propagation or arrival times TOAn (in English Time of Arrivai) of the components described in [Fig.7]. Thus, the arrival times TOAn are calculated by the matched filtering of the received signal x with the transmitted signal e. The most efficient way to carry out this matched filtering is to calculate the fast Fourier transform (in English "Fast Fourier Transform") of the two signals e and x, then to multiply the FFT of the received signal x (in the frequency modulation band of the transducer 20 and 30, going from the low value fi of frequency modulation to the high value f2 of frequency modulation), by the complex conjugate of the FFT of the transmitted signal e to calculate the signal G(t) mentioned above.

[0108] The signal (70 — FFT^fFFT{x} *FFT{e}]] represents the spectral implementation of the matched filtering principle. Thus, given a broadband signal e(t), the filter matched temporally to this signal (i.e. with respect to the shifted versions of the signal) is constructed with the following expression:

[0109] h(t) = a ■ e\tG - t)where a is a normalization factor, t0 represents a time reference and "*" represents the Hermitian operator. As indicated by this expression, the impulse response of the filter, h, is obtained by the time reversal of the original signal. In order to show the interest of the concept of the matched filter in the context of the estimation of hydraulic parameters, we consider the simplest model for the signal received by the measurement system, knowing that the transmitted signal is e(t):

[0110] X0 - A - dt-T OA) + n(t)

[0111] where A and TOA represent, respectively, the attenuation and the propagation time of the emitted signal e(t) and n(t) represents the measurement noise. The processing by matched filtering makes it possible to provide an estimate of the arrival time TOAn which, as shown previously, helps to estimate the hydraulic parameters. The application of the matched filter on the signal x, expressed in the form of convolution between x and h, leads to the signal G(t).

[0112] The spectral form of G(t) makes it possible to determine the arrival time TOAn by: TQA — aromax [ G ( t ) ]2 According to an example of the first embodiment of the invention, the at least one erosive cavitation indicator of the hydraulic machine 1 is the average mroA of the arrival times TOAn for the pulses n of the first ultrasonic signal e(t) having been emitted, according to the following equation:

[0113] m TOA mTOA~ S^n^ï1

[0114] According to another example of the first embodiment of the invention, the at least one erosive cavitation indicator of the hydraulic machine 1 is the variance ¢7^4 of the arrival times TOAn for the pulses n of the first ultrasonic signal e(t) having been emitted, according to the following equation: [°1151 ^toa = | And 1 (TOA* ' m™A ) 2

[0116] According to another example of the first embodiment of the invention, the at least one erosive cavitation indicator of the hydraulic machine 1 is the standard deviation °toa of the arrival times TOAn for the pulses n of the first ultrasonic signal e(t) having been emitted, according to the following equation: [0117^ (JTOA - | (TO A,, - mT0A) ~

[0118] These examples can be combined with each other.

[0119] Figure 8 shows as an example at the top the successive time slots k of the first emitted ultrasonic signal e(t) as a function of time t on the abscissa, and at the bottom the timing diagram of the arrival times TOAn for these successive time slots k as a function of time t on the abscissa. The appearance of cavitation in one kc of the time slots k, shown in Figure 8, causes the appearance of a dispersion of the value of the arrival time TOAn, that is to say of different values ​​of TOAn, much lower than those corresponding to the reflection at the end of the blades 3. When cavitation appears, the reflections will have a random character, because from one emission to another the geometry of the clouds of cavitation bubbles is totally chaotic and variable in time. This therefore causes an increase in the variance CjOA which constitutes an example of a temporal indicator of the appearance of cavitation.Since this indicator is estimated for each duration D of each time slot k continuously, this allows to obtain a complete vision of this phenomenon. The duration during which the variance of the arrival time TOAn remains high indicates the lifetime DVC of the cavitation cloud C.

[0120] According to a first example of the second embodiment of the invention, shown as an example in [Fig. 6], the calculation unit 40 calculates M reception time windows F; time-shifted one after the other for i ranging from 1 to M, where M is a prescribed number of reception time windows F;. The calculation unit 40 calculates the power spectral density DSP;(k)(f) of the second ultrasonic signal xk(t) having been received in each reception time window F; having a prescribed duration W covering the pulse n, according to the following equation: DSP^Çf) -jlnft) -dt\

[0122] This power spectral density DSP;(k)(f) of the second ultrasonic signal xk(t) having been received could also constitute the indicator of erosive cavitation of the hydraulic machine 1, in addition to or instead of the instantaneous spectral band BWk. The respective prescribed durations W of the reception time windows F; may be identical to each other.

[0123] According to an embodiment of the invention, the M reception time windows F; can cover the time slot k of duration D. For example, M = D / W. The duration D of each time slot k and the duration W can be chosen so that the number M is an integer.

[0124] According to a second example of the second embodiment of the invention, as illustrated in FIG. 9, the calculation unit 40 calculates the representative value DSP(k)(f) of the power spectral density DSP^Xf) of the second ultrasonic signal xk(t) having been received, which is equal to the average power spectral density DSP(k)(f) for the power spectral density DSP;(k)(f) over the M reception time windows F;, according to the following equation: D^P^ (f) — — DSP^ (spectral density

[0125] average power DSP^^f) could also constitute the indicator of erosive cavitation of the hydraulic machine 1, in addition to or instead of the instantaneous spectral band BWk. The average power spectral density DSP(k)(f) is calculated in the reception time windows F; over the duration D of each time slot k. [Fig.9] indicates the number of samples on the abscissa. According to a third example of the second embodiment of the invention, the calculation unit 40 calculates the calculated average value DSPmean of the representative value DSP(k)(f) of the power spectral density DSP;(k) over time according to the following equation: 1 f^2 ox According to one embodiment of the invention, DSPlne(m (f'2-fï) f\DSP \f J the calculation unit 40 calculates the instantaneous spectral band BWk defined by the frequency range suppf, for which the representative value DSP(k)(f) of the power spectral density DSP;(k)(f) of the second ultrasonic signal xk(t) having been received is greater than the first given non-zero percentage A of the calculated average value DSPmean of the representative value DSP^^f) of the power spectral density DSP ,k '(f) over time according to the following equation: BWt. supp^DSP^f) 1 A ■

[0126] This instantaneous BWk spectral band can constitute the erosive cavitation indicator of the hydraulic machine 1.

[0127] According to an embodiment of the invention, the prescribed duration W of the reception time window F; is greater than or equal to the prescribed repetition period Tr pulses and may be greater than or equal to 2 times the prescribed pulse repetition period Tr. According to one embodiment of the invention, the prescribed duration W of the reception time window F; is greater than or equal to 2 times the prescribed pulse repetition period Tr and less than or equal to 3 times the prescribed pulse repetition period Tr. According to one embodiment of the invention, the prescribed duration W of the reception time window F; is greater than or equal to 3 times the prescribed pulse repetition period Tr. According to one embodiment of the invention, the prescribed duration W of the reception time window F; is equal to 2.5 times (2 and a half times) the prescribed pulse repetition period Tr. In the examples described above, the prescribed duration W of the reception time window F; may be, for example, 50 milliseconds, as shown in [Fig.6], with M = D / W = 20 in this example for D = 1 second and Tr = 20 milliseconds. .

[0128] The number N of time sampling points of each reception time window F; is equal to the sampling frequency Fe used for each reception channel, multiplied by the duration W of each reception time window F;, i.e.

[0129] N = Fe.W.

[0130] In the numerical example above, the sampling frequency Fe can be for example equal to 5 MHz.

[0131] According to one embodiment of the invention, the reception time windows F; overlap one after the other. This makes it possible to improve the quality of the estimation. This overlap may be, for example, at least 50% of the duration W of the reception time windows F; one after the other. The time offset A of the reception time windows F; one after the other may be less than or equal to W / 2. For example, in [Fig.6], the time offset A of the reception time windows F; one after the other is equal to W / 2 and the overlap between the reception time windows F; is equal to 50%.

[0132] According to one embodiment of the invention, the first non-zero percentage A given is greater than or equal to 50% or 70% or 90% or 95%. For example, the first percentage A is equal to 90%.

[0133] The spectral indicator (instantaneous band BWk) is a measure of the broadband Doppler effect and makes it possible to determine the speed and direction of movement of the cloud of cavitation bubbles. More precisely, if the instantaneous band BWk increases, from one emission time slot k to the next, the cloud of cavitation bubbles approaches the transducers 20, 30 whereas if the instantaneous band BWk decreases, the cloud of cavitation bubbles moves away. In this way, the orientation of the cloud of cavitation bubbles can be estimated. Furthermore, thanks to the lifetime DVC of the cloud C of cavitation bubbles in relation to the rotational speed of the shaft 2 (wheel), the erosive character of cavitation can be determined and estimated.

[0134] An advantage of the invention compared to existing techniques is that the invention does away with an extensive network of ultrasonic and / or vibration transducers. Indeed, existing techniques providing an extensive network of ultrasonic and / or vibration transducers are complex and do not guarantee the detection or qualification of erosive cavitation.

[0135] The invention is based on the exploitation of the temporal and / or spectral properties of the received signals x(t), so as to be able to define characteristics of the cavitation - lifetime, volume, position, movement vector - thus making it possible to evaluate the erosive nature of the phenomenon. In addition, since the monitoring can be continuous and over long periods, it is possible to monitor the cavitation phenomenon permanently and continuously, which constitutes an advance in the monitoring of hydraulic machines in service.

[0136] Figures 10A, 10B, 10C, 10D, 10E, 11A, 11B, 11C, 11D and 11E explain the principle of the invention. The innovation is based on the temporal and spectral information obtained from the second ultrasonic signals received x(t) after propagation and reflection in the environment composed by the blades 3 and / or the hub 4, and / or the warhead 5, and / or the other side 172 of the wall 17 delimiting the water outlet channel 7, and / or the wall 16 delimiting the water inlet channel 6 of the turbine 1 or machine 1 or the guard valve of the turbine and the cavitation clouds.

[0137] In the absence of cavitation, the pulses n of the first emitted signal e(t) are reflected in the second signal x(t) by the blades 3 in Figure 10A and / or by the hub 4 and / or the warhead 5 and / or the other side 172 of the wall 17 delimiting the water outlet channel 7 and / or the wall 16 delimiting the water inlet channel 6 of the turbine 1 in Figure 10B, that is to say always at the same positions. The variations of the indicators TOAn and BWk are very small because this is a stationary case: the reflections occur at the level of the blades 3 and the hub 4 and therefore the values ​​of these indicators are almost constant from one pulse n to the next. The TOAn indicators are almost identical because for each signal emitted, the reflection occurs at the same point, on the blades 3 or the hub 4 of the turbine 1. As a result, the average mT0A of the TOAn arrival times remains identical to a first average mT0A(3) of the TOAn arrival times, illustrated in figure 10D.As a result, the variance of the arrival times TOAn remains identical to a first variance V3 of the arrival times TOAn, illustrated in [Fig.lOD]. As a result, the instantaneous band BWk remains identical to a first instantaneous band BW3 of absence of cavitation, illustrated in [Fig.lOE].

[0138] [Fig.10] represents pulses n of the first emitted ultrasonic signal e(t) (amplitude in Volts on the ordinate, time t on the abscissa), and the second ultrasonic signal x(t) (amplitude in Volts on the ordinate, time t on the abscissa) reflected by the blades 3 and / or by the shaft 2 or the warhead 5 in response to these pulses n of the first emitted ultrasonic signal e(t), in the absence of cavitation.

[0139] Figure 10D represents on the ordinate the variation of the average mTOA of the TOAn arrival times and the variation of the variance of the TOAn arrival times as a function of time t on the abscissa, in the absence of cavitation.

[0140] [Fig.10E] represents the variation of the spectral indicator (instantaneous band BWk) on the ordinate as a function of time t on the abscissa, in the absence of cavitation.

[0141] When, as illustrated in Figures 1 1A and 1 1B, cavitation C is present, a portion of the first emitted ultrasonic signal e(t) is partly reflected by the clouds C of cavitation bubbles, while reflections from the blades 3 ([Fig. 11A]) and the hub 4 ([Fig. 1 1B]) and / or the warhead 5 and / or the other side 172 of the wall 17 delimiting the water outlet channel 7 and / or the wall 16 delimiting the water inlet channel 6 of the turbine 1 are still present, although to a lesser extent than in the absence of cavitation. Indeed, the cloud C of cavitation bubbles is not totally opaque to the acoustic waves of the first emitted ultrasonic signal e(t) but only a portion of these signals is reflected. The reflections occur at random positions depending on the position and the evolution of the cavitation trajectory C.

[0142] [Fig. 1 IC] represents pulses n of the first emitted ultrasonic signal e(t) (amplitude in Volts on the ordinate, time t on the abscissa), and the second ultrasonic signal x(t) (amplitude in Volts on the ordinate, time t on the abscissa) reflected by the blades 3 and / or the hub 4 and / or the warhead 5 and / or the other side 172 of the wall 17 delimiting the water outlet channel 7 and / or the wall 16 delimiting the water inlet channel 6 of the turbine 1 in response to these pulses n of the first emitted ultrasonic signal e(t), in the presence of the cavitation C.

[0143] Figure 11D represents on the ordinate the variation of the average mTOA of the arrival times TOAn and the variation of the variance of the arrival times TOAn as a function of time t on the abscissa, in the presence of cavitation C.

[0144] [Fig. 11E] represents the variation of the spectral indicator (instantaneous band BWk) on the ordinate as a function of time t on the abscissa, in the presence of cavitation C.

[0145] Figures 1 IC and 11D show that two different series of arrival TOAn times are detected in the presence of cavitation C: - on the one hand the first arrival times TOAn corresponding to the reflection of the first emitted ultrasonic signal e(t) on the blades 3, as illustrated by [Fig. 11 A], these first arrival times TOAn being noted TOAn(3) in [Fig. 11C]; - on the other hand the second arrival times TOAn corresponding to the reflection of the first emitted ultrasonic signal e(t) on the clouds C of cavitation bubbles, as well as illustrated by [Fig. llB], these second arrival times TOAn being noted TOAn(C) in [Fig. llC]. These second arrival times TOAn TOAn(C) are lower than the first arrival times TOAn being noted TOAn(3) corresponding to the emission at the blades 3, because the cloud C of cavitation bubbles is located between the wall 13 and the blades 3. The duration during which the arrival times TOAn remain low and fluctuating corresponds to the lifetime DVC of the cavitation cloud C.

[0146] Figure 11D shows that when cavitation C is present, the average mTOA of the arrival TOAn times decreases to a second average mTOA(o of the arrival TOAn times, lower than the first average mT0A(3) of the arrival TOAn times of absence of cavitation C.

[0147] According to an embodiment of the invention, the calculation unit 40 determines that a cavitation C is present, when the average mT0A of the arrival times TOAn is less than a first threshold Sm (which may be equal to the first average mT0A(3) of the arrival times TOAn of absence of cavitation C).

[0148] According to one embodiment of the invention, the calculation unit 40 determines a position Pc of the cavitation C from the average mTOA of the arrival times TOAn. For example, the average mTOA of the arrival times TOAn makes it possible to locate the cavitation C in the region defined by the product (c.mToA / 2) between the propagation speed c of the signals e and x and the average mTOA of the arrival times TOAn, divided by 2.

[0149] Figure 11D shows that when cavitation C is present, the variance of the arrival TOAn times increases to a second variance Vc of the arrival TOAn times, greater than the first variance V3 of the arrival TOAn times in the absence of cavitation C.

[0150] According to an embodiment of the invention, the calculation unit 40 determines that a cavitation C is present, when the variance of the arrival times TOAn is greater than a determined threshold V (which may be equal to the first variance V3 of the arrival times TOAn of absence of cavitation C).

[0151] According to one embodiment of the invention, the calculation unit 40 determines a lifetime DVC of the cavitation C from the variance of the arrival times TOAn and / or from the average mT0A of the arrival times TOAn.

[0152] The DVC lifetime of the cavitation C can be defined by the duration during which the average mT0A of the arrival times TOAn is different by at least 50% (or by more than a percentage X greater than 50% and less than 100%) compared to the first average mT0A(3) of the arrival times TOAn of absence of cavitation C.

[0153] The DVC lifetime of cavitation C can be defined by the duration during which the variance of the arrival times TOAn is different by at least 50% (or more of a percentage Y greater than 50% and less than 100%) compared to the first variance V3 of the TOAn times of arrival of absence of cavitation C.

[0154] According to one embodiment of the invention, the calculation unit 40 determines a direction Dc of displacement of the cavitation C from the instantaneous band BWk.

[0155] [Fig. 11E] shows that during cavitation C the spectral indicator (instantaneous band BWk) varies with respect to the first instantaneous band BW3 of absence of cavitation, because the movement of cavitation C modifies the emitted spectrum, which constitutes a broadband Doppler effect.

[0156] The variation of the instantaneous band BWk is proportional to the angle 0 between on the one hand the axis AE of the path of the signal received x(t) by the transmitting transducer 20 and the direction Dc of movement of the cloud C of cavitation bubbles: if there is no variation of the instantaneous band BWk, in accordance with the theory of the Doppler effect, the cavitation cloud moves in a direction Dc perpendicular to the axis AE of reception. If the variation of the instantaneous band BWk is large, the cloud C of cavitation bubbles moves towards the wall 13 and its erosive character is very probable.

[0157] Thus, if the instantaneous band BWk increases relative to the first instantaneous band BW3 of absence of cavitation, this corresponds to the angle 0 between 0 and jt / 2, which indicates that the cloud C of cavitation bubbles is moving towards the transducers 20, 30: the larger the instantaneous band BWk, the more the angle 0 converges towards 0. Conversely, if the instantaneous band BWk decreases, this corresponds to the angle 0 between ir / 2 and ji, which indicates that the cloud C of cavitation bubbles is moving away from the transducers 20, 30.

[0158] This thus makes it possible to determine from the instantaneous band BWk the orientation of the cavitation cloud C, without needing several ultrasonic strings, a characteristic which stands out from the state of the art.

[0159] In the absence of cavitation, the emitted waves e(t) are reflected by the blades 3 if the transducers 20, 30 are positioned at the level of the wheel casing, or by the wall 17 if they are positioned downstream of the wheel. As soon as cavitation occurs, the emitted waves e(t) are reflected well before, because the cloud of cavitation air bubbles is positioned between the transmitter 20 / receiver 30 and the turbine components - blades or the walls of the aspirator. Then the propagation path is shorter and the estimation of the arrival times of the received signals x(t) makes it possible to determine the position of the cavitation cloud C relative to the point of emission. Since broadband e(t) emitted signals are used (characterized by a spectral band equivalent to the center frequency of the spectrum), it is possible to detect a wide range of cavitation characterized by bubbles of arbitrary diameters.More precisely, the diameters of cavitation bubbles are related to the reflection indices of acoustic waves, which in turn depend on the frequencies. The larger the diameter of the cavitation bubbles, the greater the . minimum frequency of reflected acoustic waves decreases. A cavitation bubble with a large diameter (around 5 mm for example) will reflect all frequencies above 500 kHz, while a cavitation bubble diameter of 2.5 mm will reflect all frequencies above 1 MHz. It can be considered that the denser the cloud of cavitation bubbles, the greater the energy of the reflected signal x(t) will be, because the range of reflected frequencies will be wider.

[0160] Figure 12 shows on the ordinate the values ​​of the arrival times TOAn, having been calculated by the method according to the invention, for 2 pairs of ultrasonic transducers 20, 30 (corresponding to the 2 curves) arranged on the outlet wall 17, for a turbine group 1 of a hydraulic power station, for different dates shown on the abscissa, as well as the variances (J^OA of the arrival times TOAn and an example of the threshold V determined for these variances.

[0161] Figure 13 shows on the ordinate the values ​​of the arrival times TOAn, having been calculated by the method according to the invention, for 2 pairs of ultrasonic transducers 20, 30 (corresponding to the 2 curves) arranged on the outlet wall 17, for another turbine group 1 of a hydraulic power station, for different dates shown on the abscissa, as well as the variances of the arrival times TOAn and an example of the threshold V determined for these variances.

[0162] Figure 14 shows, as a function of time t on the abscissa, values ​​obtained from the angle ANG3 of orientation of the blades 3 and the angle ANG9 of orientation of the guide vanes 9 on the ordinate, these angles ANG3 and ANG9 being elements contributing to the overall efficiency of the machine 1. It can be seen that the variation of these parameters ANG3 and ANG9 is accompanied by variations in the arrival times TOAn in terms of mean mT0A and variance, thus testifying to the presence of cavitation C, identified by the ellipses.

[0163] Another detailed confirming analysis consisted of comparing the level of cavitation C, estimated by acoustic methods, with the efficiencies of machine 1. It was thus demonstrated (see Figure 15) that, for certain propellers (defined by the points corresponding to the flow rates, on the abscissa and the efficiencies, on the ordinate), when moving away from the optimal operating point, cavitation C is indeed present for the group studied. Figure 15 shows at the top the efficiency, expressed in % on the ordinate, as a function of a flow rate of machine 1 expressed in m3 / s on the abscissa. Figure 15 shows at the bottom the arrival time TOAn, expressed in seconds, as a function of a flow rate of machine 1 expressed in m3 / s on the abscissa. Figure 15 shows the correlation between the operating points and the presence of cavitation indicated by the acoustic parameters.Referring to Figure 15, we observe that when the yield is maximum (along the vertical line Rmax) the variance ny^^dcs arrival time TOAn is . minimal, which corresponds to the absence of cavitation. When the operation is no longer optimal (according to the FC arrow) the variance rr^^becomes very large.

[0164] According to one embodiment of the invention, each emitting ultrasonic transducer 20 comprises an active mode, which can be controlled by the calculation unit 40, and a passive mode which can be controlled by the calculation unit 40.

[0165] When the emitting ultrasonic transducer 20 is in the active mode, the first step E1, the second step E2 and the third step E3 are carried out.

[0166] When the emitting ultrasonic transducer 20 is in the passive mode, the first step E1, the second step E2 and the third step E3 are not carried out and the emitting ultrasonic transducer 20 does not emit a first ultrasonic signal e or e(t). The calculation unit 40 controls by default the emitting ultrasonic transducer 20 in the passive mode. When the emitting ultrasonic transducer 20 is in the passive mode, the receiving ultrasonic transducer(s) 30 receives a third signal xy. This third signal xy corresponds to the vibration of the system, present in the liquid water passage conduit 12.

[0167] The calculation unit 40 calculates the power spectral density DSP^f) of the third signal xy having been received in each of several time slots k-1, k mentioned above, following one another and each having an identical time slot duration D.

[0168] The calculation unit 40 controls the emitting ultrasonic transducer 20 in the active mode, when the power spectral density DSPy(k)(f) of the third signal xy having been received in a current time slot k (or its representative value DSP'y(k)(f), mentioned below) differs (in modulus) by more than a second non-zero percentage B (threshold) given with respect to the power spectral density DSPy^^f) of the third signal xy having been received in a time slot prior to the current time slot k (which may for example be the time slot k-1 preceding the current time slot k). This second percentage B may for example be equal to 50% or may be greater than 50%.Indeed, when the power spectral density DSPy(k)(f) or DSP' y(k)(f) of the current time slot k differs by more than the percentage B compared to the previous time slot, this is a change in hydraulic regime, likely to generate cavitation in the broad sense (erosive or not). This active mode makes it possible not to use the device according to the invention in its entirety and thus avoids recovering an excess of data of no interest for the study of cavitation.

[0169] The power spectral density DSPy(k)(f) is calculated in the same way as above, over a reception time window F; of a prescribed duration W, the reception time windows F; being temporally shifted one after the other by the prescribed period Tr, for i ranging from 1 to M prescribed, namely according to the following equation: ,,. 1 । fi-w ।2 The representative value DSPy=w|J(n).w+1MO -^p( -jlTTft) dt^ DSP'y(k)(f) of the power spectral density DSP;(k)(f) of the second ultrasonic signal xk(t) having been received is equal to the average of the power spectral density DSPy(k)(f) of the third signal xy over the time slot k, according to the following equation: — 1 ( f ) During the active regime the trans emission drivers begin to periodically emit broadband e(t) signals.

[0170] The invention can be applied to a hydraulic machine 1 (for example hydraulic turbine 1) of a hydraulic power plant for producing electricity, but also of a nuclear power plant for producing electricity or of a thermal power plant for producing electricity.

[0171] Of course, values ​​other than the numerical examples mentioned above may be provided. Of course, the embodiments, features, possibilities and examples described above may be combined with each other or selected independently of each other.

Claims

Claims

1. Method for measuring erosive cavitation of a hydraulic machine (1) in operation, the hydraulic machine (1) comprising walls (13, 16, 17) delimiting a conduit (12) for the passage of liquid water and at least one movable member (11, 3, 9), which is located in the conduit (12) for the passage of liquid water and which is set in motion during the operation of the hydraulic machine (1), characterized in that the method comprises the following steps: during a first step (El), emission of at least one first wideband pulsed ultrasonic signal (e) by at least one emitting ultrasonic transducer (20), which is positioned on a first side (131, 171) of at least one first (13, 17) of the walls (13, 16, 17) and which is oriented through the conduit (12) for the passage of liquid water towards another side (172) of the first wall (17) or towards the at least one movable member (11, 3, 9), during a second step (E2), reception,by at least one receiving ultrasonic transducer (30), which is positioned on the first side (131, 171) of the first wall (13, 17), of at least one second ultrasonic signal (x) received in response to the first ultrasonic signal (e) having been emitted, during a third step (E3), calculation, by a calculation unit (40) connected to the at least one transmitting ultrasonic transducer (20) and to the at least one receiving ultrasonic transducer (30), from the first ultrasonic signal (e) having been emitted and from the second ultrasonic signal (x) having been received, of at least one indicator of erosive cavitation of the hydraulic machine (1), from the calculation by the calculation unit (40) of at least one arrival time (TOAn) of the second ultrasonic signal (x) having been received relative to the first ultrasonic signal (e) having been emitted and / or from the calculation by the unit (40) calculation of an instantaneous spectral band (BWk) defined by a frequency range,for which a representative value (DSP(k)(f)) of the power spectral density (DSP^f)) of the second ultrasonic signal (x) having been received is greater than a first given non-zero percentage (A) of the calculated average value (DSPmean) of the representative value (DSP(k)(f)) of the power spectral density (DSP;(k)(f)) over time.,

2. Method according to claim 1, characterized in that the first broadband pulsed ultrasonic signal (e) comprises a plurality of pulses (n) emitted with a prescribed period (Tr) of pulse repetition, non-zero.

3. Method according to claim 2, characterized in that each pulse (n) has, during a modulation duration (Lp), less than the prescribed period (Tr) of repetition of the pulses, a frequency modulation greater than or equal to a low value (fj of frequency modulation, non-zero, and less than or equal to a high value (f2) of frequency modulation, non-zero.

4. Method according to claim 3, characterized in that the frequency modulation is linear or cubic and goes from the low value (fi) of frequency modulation to the high value (f2) of frequency modulation over the duration (Lp) of modulation or from the high value (f2) of frequency modulation to the low value (fj) of frequency modulation over the duration (Lp) of modulation.

5. Method according to any one of claims 2 to 4, characterized in that during the third step (E3), the calculation unit (40) calculates for each pulse (n) of the first ultrasonic signal e(t) having been emitted the arrival time TOAn of the second ultrasonic signal x(t) having been received relative to the first ultrasonic signal e(t) having been emitted according to the following equations: TOAn = argmaxt [ G ( t ) ]2 £0 _ J" x(r j. n is the index of each pulse.

6. Method according to any one of claims 2 to 4, characterized in that during the third step (E3), the calculation unit (40) calculates for each pulse (n) of the first ultrasonic signal e(t) having been emitted the arrival time TOAn of the second ultrasonic signal x(t) having been received relative to the first ultrasonic signal e(t) having been emitted according to the following equations: TOAn = argmaxt [ G ( t ) ]2 G0 = FFT'^ FFT{x] *FFT{ <?}]]OÛ n est l’indice de chaque impulsion, FFT{x} est la transformée de Fourier rapide du deuxième signal ultrasonore x(t) ayant été reçu, FFT{e] est le conjugué complexe de la transformée de Fourier rapide du premier signal ultrasonore e(t), FFT^ est la transformée de Fourier rapide inverse.

7. Method according to claim 5 or 6, characterized in that during the third step (E3), the calculation unit (40) calculates the at least one indicator of erosive cavitation of the hydraulic machine (1) as comprising the average ^toA) of the arrival times TOAn for the pulses (n) of the first ultrasonic signal e(t) having been emitted and / or the variance (fT?^) of the arrival times TOAn for the pulses (n) of the first ultrasonic signal e(t) having been emitted and / or the standard deviation (^toA) of the arrival times TOAn for the pulses (n) of the first ultrasonic signal e(t) having been emitted.

8. Method according to claim 7, characterized in that during the third step (E3), the calculation unit (40) determines that a cavitation (C) is present, when the average (mT0A) of the arrival times TOAn is lower than a first determined threshold (Sm).

9. Method according to claim 7 or 8, characterized in that during the third step (E3), the calculation unit (40) determines that a cavitation (C ) is present, when the variance (¢7^^) of the arrival times TOAn is greater than a determined threshold (V, Vc).

10. Method according to any one of claims 7 to 9, characterized in that during the third step (E3), when the calculation unit (40) has determined that a cavitation (C ) is present, the calculation unit (40) determines a position (Pc) of the cavitation (C ) in the region defined by the product (c.mToA / 2) between the speed (c ) of propagation of the ultrasonic signals (e, x) and the average ^toA) of the arrival times TOAn, divided by 2.

11. Method according to any one of claims 7 to 10, characterized in that during the third step (E3), when the calculation unit (40) has determined that a cavitation (C ) is present, the calculation unit (40) determines a lifetime (DVC) of the cavitation (C ) from the variance of the arrival times TOAn and / or from the average (mT0A) of the arrival times TOAn.

12. Method according to any one of the preceding claims, when they depend at least on claim 3, characterized in that during the third step (E3), the calculation unit (40) calculates: the power spectral density DSP;(k)(f) of the second ultrasonic signal xk(t) having been received for each of several reception time windows F; having a prescribed duration W covering the pulse n, the reception time windows F; being shifted temporally one after the other for i ranging from 1 to M, according to the following equation: ,, । pw > 2where M is a DSP-k(f) ■ exp( -j27rft)-dt\ prescribed number of reception time windows F;, the representative value DSP^^f) of the power spectral density DSP^f) of the second ultrasonic signal xk(t) having been received being equal to the average power spectral density DSP(k)(f) for the power spectral density DSP^f) over the M time windows F; reception, according to the following equation: DSP^k' ( / ) = 77 DSPÏk) ( / ) the average value DSPmean calculated from the representative value DSP(k)(f) of the power spectral density DSP;(k) over time according to the following equation: r» votf / 2 edUo , x , , the spectral band BWk ins- DSPmem - p^^XxDSP \J)df tantanized defined by the frequency range suppf, for which the representative value DSP(k)(f) of the power spectral density DSP;(k)(f) of the second ultrasonic signal xk(t) having been received is greater than the first given non-zero percentage A of the calculated average value DSPmean of the representative value DSP(k)(f) of the power spectral density DSP;(k)(f) over time according to the following equation: BWk = suppf\DSP{k} DSPmean\ where fi is the low frequency modulation value, f2 is the high frequency modulation value, f is the frequency, t is the time.;

13. Method according to claim 12, characterized in that the reception time windows F; overlap one after the other.

14. Method according to claim 12 or 13, characterized in that the prescribed duration W of the reception time window F; is greater than or equal to 1 time or 2 times or 3 times the prescribed period (Tr) of repetition of the pulses.

15. Method according to any one of claims 12 to 14, characterized in that during the third step (E3), when the calculation unit (40) has determined that a cavitation (C ) is present, the calculation unit (40) determines a direction (Dc) of movement of the cavitation (C ) from the instantaneous spectral band (BWk).

16. Method according to any one of the preceding claims, characterized in that the first given non-zero percentage A is greater than or equal to 50% or 70% or 90% or 95%.

17. Method according to any one of the preceding claims, characterized in that the at least one transmitting ultrasonic transducer (20) comprises an active mode, controllable by the computing unit (40) and in which the first step (E1), the second step (E2) and the third step (E3) are carried out, and a passive mode, controllable by the computing unit (40) and in which the first step (E1), the second step (E2) and the third step (E3) are not carried out and the at least one transmitting ultrasonic transducer (20) does not transmit, the computing unit (40) controlling by default in the third step (E3) the at least one transmitting ultrasonic transducer (20) in the passive mode, in which the at least one receiving ultrasonic transducer (30) receives a third signal (xy), the computing unit (40) calculating in the third step (E3) the power spectral density (DSPy(k)(f),DSP'y(k)(f)) of the third signal (xy) having been received in each of several time slots (k-1, k) following one another and each having an identical time slot duration (D), the calculation unit (40) controlling during the third step (E3) the at least one ultrasonic transducer (20) emitting in the active mode, when the power spectral density (DSP^f), DSP'y(k)(f)) of the third signal (xy) having been received in a current time slot (k) differs by more than a second given non-zero percentage (B) with respect to the power spectral density (DSP^f), DSP'y(k)(f)) of the third signal (xy) having been received in a time slot (k-1) prior to the current time slot (k).,

18. Method according to claim 17, characterized in that the time slot (k-1) prior to the current time slot (k) is the time slot (k-1) preceding the current time slot (k).

19. Device (100) for measuring erosive cavitation of a hydraulic machine (1) in operation, the hydraulic machine (1) comprising walls (13, 16, 17) delimiting a conduit (12) for the passage of liquid water and at least one movable member (11, 3, 9), which is located in the conduit (12) for the passage of liquid water and which is set in motion during operation of the hydraulic machine (1), characterized in that the measuring device (100) comprises: at least one emitting ultrasonic transducer (20), which is positioned on a first side (131, 171) of at least one first (13, 17) of the walls (13, 16, 17) and which is oriented through the liquid water passage conduit (12) towards another side (172) of the first wall (17) or towards the at least one movable member (11, 3, 9), the at least one emitting ultrasonic transducer (20) being capable of emitting at least one first broadband pulsed ultrasonic signal (e), at least one receiving ultrasonic transducer (30), which is positioned on the first side (131, 171) of the first wall (13, 17), and which is capable of receiving at least one second ultrasonic signal (x) received in response to the first ultrasonic signal (e) having been emitted, a computing unit (40) connected to the at least one transducer ultrasonic (20) transmitter and at least one ultrasonic transducer (30) receiver,the calculation unit (40) being capable of calculating from the first ultrasonic signal (e) having been emitted and from the second ultrasonic signal (x) having been received, at least one indicator of erosive cavitation of the hydraulic machine (1), from the calculation by the calculation unit (40) of at least one arrival time (TOAn) of the second ultrasonic signal (x) having been received relative to the first ultrasonic signal (e) having been emitted and / or from the calculation by the calculation unit (40) of an instantaneous spectral band (BWk) defined by a frequency range, for which a representative value (DSP(k)(f)) of the power spectral density (DSP;(k)(f)) of the second ultrasonic signal (x) having been received is greater than a first given non-zero percentage (A) of the calculated average value (DSPmean) of the representative value (DSP(k)(f)) of the power spectral density (DSP;(k)(f)) over time.,

20. Computer program, comprising code instructions for implementing the third step (E3) of the method for measuring erosive cavitation of a hydraulic machine (1) according to any one of claims 1 to 18, when implemented on a computer (40).

Images