FR2577323A1 - Method of selecting an acoustic signal and corresponding selection device - Google Patents

Abstract

The amplitudes of each alternation of a rectified electrical signal corresponding to the acoustic signal are compared with a low and a high threshold. A time interval of predetermined fixed duration T1 starting when the signal first overshoots the low threshold is determined. The method involves seeing whether the signal overshoots the low threshold and then seeing whether there is an overshoot of the high threshold during the time interval T1. If there is, the signal is accepted; if not, it is rejected. The selection device comprises amplifiers 12, 13, monostable flip-flops 14, 14', and multivibrators 15, 16. Application particularly to an installation for monitoring impacts in the primary circuit of a pressurised water nuclear reactor.

Description

 The invention relates to a method for selecting an acoustic signal picked up in an industrial installation to determine this signal can be used for the detection of a phenomenon resulting in an acoustic emission in the installation.

 It is known to provide monitoring of a complex industrial installation such as a pressurized water nuclear reactor, by an acoustic method. It has been proposed, for example, to detect the acoustic waves accompanying the shocks of a migrating body on certain parts of the primary circuit of a pressurized water nuclear reactor, thanks to a set of sensors arranged in certain parts of the circuit. primary.

 The detection of migrant bodies must be carried out in an extremely safe and extremely rapid manner, in order to avoid deterioration of certain parts of the primary circuit which may possibly jeopardize the safety of the reactor while avoiding untimely shutdowns of this reactor.

 Such acoustic monitoring methods can also make it possible to detect the beginning of a part rupture and to intervene on the installation before this part is detached and has become a migrant body.

 It is also known to look for the initiations of rupture in installatians or complex industrial components. during the tests preceding their entry into service, by recording the acoustic signals emitted by these structures under the effect of the stresses undergone during the tests.

 The impacts and propagation of the cracks are in fact Mechanical phenomena which result in the emission of acoustic waves, the arrival time of which can be determined with very good precision on a sensor placed on the installation. monitoring or testing course, or in its immediate vicinity. It is thus possible to very precisely determine the zone or zones in which the acoustic waves have been emitted, using a set of sensors placed at precise locations in the installation, this localization of the origin of the acoustic waves allows. more effective monitoring or control of the installation.

 However. the acoustic sensors receive not only the useful waves directly translating the phenomena for which detection is carried out, but also parasitic waves such as the echoes of the useful waves or waves due to phenomena different from the impacts or propagation of the cracks. such as friction or hydraulic phenomena (flow noises, cavitation).

 Multiple echoes of a useful signal can lead to interpretation errors or uncertainties in the measurement of arrival times and therefore in the location of the origin of waves. Furthermore. when storing information on magnetic tape or disc. these signals unnecessarily fill the storage media.

 Acoustic waves of mechanical origin reflecting impact or cracking are characterized by a very rapid rise in amplitude, which makes it possible to determine with great precision their moment of arrival on a sensor. It is not the same for echoes of these direct waves or frictions or other mechanical parasites which have a time of rise in amplitude much longer. The acoustic signals received by piezoelectric type sensors are transformed by these sensors into electrical signals whose amplitude directly reflects the amplitude of the acoustic waves received. For their processing, these signals are first rectified and are in the form of a series of alternations of variable amplitudes. The electrical signals representative of acoustic signals directly translating a mechanical phenomenon have a time of rise in amplitude very short while the rectified electrical signals translating the multiple echoes of these direct waves or frictions, for example, have a time of rise in amplitude much longer.

 Until now, no process has been known which makes it possible to analyze these signals very easily and very quickly to determine whether they are usable for the detection of a phenomenon of mechanical origin in the installation on which the monitoring or the control.

 The object of the invention is therefore to propose a method for selecting an acoustic signal picked up in an industrial installation to determine whether this signal can be used for the detection of a mechanical phenomenon resulting in an acoustic emission, each signal emitted and sensed in the installation being put into the form of a rectified electrical signal, this process should allow rapid selection, by simple means, without having to analyze the signal in detail.

 For this purpose - the amplitude of each of the alternations of the rectified electrical signal corresponding to the acoustic signal is compared to a first fixed - predetermined amplitude, called the low threshold, - we compare the amplitude of each of the alternations of the rectified electrical signal to one second predetermined fixed amplitude. greater than the first and called the high threshold, - a time interval of predetermined fixed duration Ti is determined, originating from the start of the first crossing of the low threshold by alternating signal.

- a search is made if a breach of the high threshold takes place during the time interval of duration TI, - and we accept the acoustic signal considered to be usable if at least a breach of the high threshold takes place during the time interval of duration T1 .

 The invention also relates to a device allowing the selection of acoustic signals, by the method according to the invention.

 In order to clearly understand the invention.

We will now describe, by way of nonlimiting example, with reference to the figures appended.

 an embodiment of a device for implementing the method according to the invention and the functioning of this device, in the case of a set of acoustic sensors placed in the vicinity of a zone of the primary circuit of a reactor pressurized water nuclear power plant in which impact monitoring is carried out.

In these figures
- Fig. t is a very schematic representation of the device for selecting acoustic signals associated with four sensors constituting the input elements of four measurement lines.

 - Fig. 2 is a detailed diagram of the constitution of the signal selection device as concerned with a measurement line.

 - Figs. 3A and 3B are representations of acoustic signals picked up in the primary circuit of a nuclear reactor.

 - Figs. 4A, 4B, 4-C and 4D are representations of electrical signals at different stages of their processing in the selection device represented in FIGS. 1 and 2.

 - Fig. 5 is a diagram showing the different stages for obtaining an electrical validation signal, in the case of a direct acoustic signal translating an impact.

 - Fig. 6 is a diagram showing the different stages corresponding to those of FIG. 5, in the case of a parasitic acoustic signal picked up by an impact monitoring device.

 In Fig. 1, we can see four sensors ta, lb, lc and id which are piezoelectric acoustic sensors arranged at determined locations in the vicinity of an area of the primary circuit of a pressurized water nuclear reactor, for the detection of pacts in this primary circuit. Each of the sensors 1 constitutes the input of a measurement line, the initial part of which is shown in FIGS. 1 and 2 is a signal selection device. The signals selected at the output of this device are processed in a part of the impact monitoring installation of the nuclear power plant, not shown; this subsequent processing includes in particular the determination of the orders and arrival times of the acoustic waves on the various sensors.

 The selection device successively comprises a preamplification stage 2, a signal rectification stage 3, a stage 4 for comparing the rectified signal with thresholds and for developing discrimination signals and finally a stage 5 for developing an initial acoustic signal validation signal.

 The various validation signals are transmitted to an OR gate 7 with four inputs enabling a validation signal to be delivered in the event that at least one of the four inputs itself receives a validation signal.

 The various intermediate signals being processed can be viewed as shown in Figs. 3 to 6.

 In Fig. 2, it can be seen that the output of a preamplifier 2 is connected, via a capacitor and a decoupling resistor designated by the reference 10, to the stage 3 for rectifying the signal constituted by two operational amplifiers 12 and 13 associated with different elements according to a usual assembly. A potentiometer 11 makes it possible to adjust the minimum value of the signal to the value OV.

 The signal comparison and processing stage 5 consists of two monostable flip-flops 14 and 14 'receiving on the one hand the rectified signal at the output of stage 3 and on the other hand a threshold value which can be adjusted with a potentiometer 17 (or you '). The flip-flops 14 and 15 'emit an all-or-nothing signal triggered when an alternation of the rectified signal exceeds the corresponding amplitude threshold and interrupted when the alternation falls below the threshold. The threshold value entered in the monostable rocker 14 and adjusted by the potentiometer 17 is lower than the threshold value entered in the monostable rocker 14 'and adjusted by the potentiometer 17'. The first threshold value is called low threshold and the second high threshold value.

 Stage 5 of processing a validation signal is constituted by two monostable multivibrators 15 and 16 arranged in series with which potentiometers 21 and 22 are associated respectively, making it possible to adjust the times T1 and T2 of transmission of signals d constant amplitude, by multivibrators 15 and 16 triggered by an input signal.

 The second multivibrator 16 is retriggerable.

 The signal at the output of the second multivibrator 16 of duration T2 is combined by a gate 20 with the discrimination signal coming from the monostable flip-flop 14 ′ and resulting from the comparison of the rectified signal with the high threshold. Gate 20 is a NOR gate and outputs the validation signal Y.

 The form of this validation signal makes it possible to retain or on the contrary to eliminate the acoustic signal received in the selection device. It allows, for example. perform digital processing of the acoustic signal at the origin of the validation signal.

 We will now describe the different signals obtained in the selection device and the operation of this device.

 In Figs. 3A and 3B, there are two types of signals received by a sensor 1 from an impact monitoring installation of a pressurized water nuclear reactor.

 The signal shown in FIG. 3A corresponds to the direct acoustic waves received by the sensor following a shock. This signal has a very steep front, with a very rapid and large amplitude increase. Such a signal can be easily used to determine its arrival time on the sensor.

 Fig. 3B represents an unwanted signal received by the sensor. This signal is difficult to interpret and may correspond to the echo of the shock signal or to a phenomenon different from a shock, such as friction in the primary circuit or hydraulic turbulence. On the other hand, it would be excrementally difficult to exploit this signal, for example to determine its instant of arrival on the sensor 1. In fact, the rise in amplitude is slow and of small magnitude, so that the signal is distinguished little background noise for a relatively long period of time.

 The processing of the signals received by the sensor 1 in a selection device as shown in FIGS. 1 and 2 makes it possible to accept an acoustic signal as shown in 3A and to reject a spurious signal as shown in 36.

 In Fig. 4a, we see an electrical signal supplied by the piezoelectric sensor 1, after the preamplification stage 2. In FIG. iB. the same signal is seen at the output of the rectification stage 3. The alternations of the signal on its entry into the comparison stage 4 are then all positive. In the comparison stage 4, these positive half-waves are compared in amplitude on the one hand to the value of the low threshold 25 and on the other hand to the value of the high threshold 26.

 In the case shown in FIG. 4B, the six mid-wave alternations 28 exceed in amplitude the low threshold 25. On the other hand, only the two central half-waves 28a and 28b exceed the high threshold 26.

 In Fig. 4C, we see the low threshold discrimination signal emitted by the monostable flip-flop 14. This discrimination signal comprises six successive pulses 29 initiated by the crossing of the low threshold 25 by an alternation 28 of the signal and interrupted by the fallout of the corresponding alternation below the low threshold 25. The pulses 29 therefore have a duration corresponding to the duration of exceeding the low threshold by the alternations of the rectified electrical signal.

In the same way, we have represented on the
Fig. 4D, the two pulses 30 corresponding to the amplitude exceeding of the high threshold 26, by the alternation 28a and 28b of the rectified electrical signal.

 The low threshold discrimination signal A and the high threshold discrimination signal O are sent to the output of the comparison stage 4, on the first multivibrator 15 and on the NOR gate 20 respectively.

In Fig. 5, a low threshold discrimination signal A has been shown from top to bottom, a signal T1 at the output of the first multivibrator 15, a signal
T2 at the output of the second multivibrator 16, a high threshold discrimination signal 9 and finally the discrimination signal Y obtained from these signals at the output of the gate 20. These signals correspond to the processing of the acoustic signal represented in FIG. 3A.

In the multivibrator 15, the initiation of the first pulse 29a of the discrimination signal
A triggers a constant voltage signal of duration T1.

At the end of the duration TI, the voltage of the signal from 15 is canceled and a second signal is initiated by the next pulse 29b of the low threshold discrimination signal. This second signal also has a constant voltage and a duration T1. The signals of duration T1 are received by the second multivibrator 16 which is triggered when the first signal T1 falls. The multivibrator 16 provides a signal of constant voltage amplitude and duration T2. This duration signal T2 and the high threshold discrimination signal B are transmitted to gate 20 which outputs the validation signal Y.

 The signal T2 is an inhibition signal which cancels all the pulses of the high threshold discrimination signal B. We therefore find in the validation signal Y only the pulses 30a and 30b of the high threshold discrimination signal B which were sent before the initiation of the signal T2, that is to say during the duration T1 following the first pulse 29a of the low threshold discrimination signal. These pulses are positive and allow to validate an acceptable acoustic signal for the subsequent treatment. The signal T2 makes it possible to inhibit the selection system for a few tens of microseconds after the reception of a signal. We thus obtain a separation of the different selection operations and an elimination of multiple echoes of a wave received by the sensors. The pulses of the validation signal allow the triggering of the signal processing system, for example to count the differences in signal arrival time on the various sensors used for monitoring an area of the primary circuit of the nuclear reactors or triggering the digital acquisition of corresponding analog signals.

 In Fig. 6, the processing of a second signal is shown, analogous to the processing of the first signal which is the subject of FIG. 5 and corresponding to the signal of FIG. 38.

 The first pulse 29a of the low threshold discrimination signal A triggers a signal of constant voltage and duration TI thanks to the multivibrator 15. This first signal of duration T1 is followed by a second signal of the same duration triggered by pulse 29b of the low threshold discrimination signal A.

 The fallout of the signal T1 triggers the second multivibrator 16 which emits a signal of constant amplitude and duration T2. This duration signal T2 and the high threshold discrimination signal B are combined by means of the terminal 20 which supplies the validation signal Y as an output.

The high threshold discrimination signal 8 only presents pulses 30 after the signal T1 has fallen, that is to say during the transmission of the signal.
T2. This signal T2 causes the inhibition of these pulses, so that the validation signal Y has no pulse. This validation signal translates an acoustic signal whose amplitude rise time is too long for the signal to be acceptable for further processing. The validation signal is then perfectly flat and the acoustic signal is not accepted by the monitoring device.

 The method and the device according to the invention therefore provides a criterion for discriminating acoustic signals which is very easy to implement with very great speed and without having to analyze the signal received by the sensor in detail.

 On the other hand, by modifying the duration of the signal T1. we can modify this criterion. This adjustment can be obtained by a simple potentiometer as described.

 It is also possible to modify the criteria leading to the preparation of the discrimination signals, that is to say the low threshold and the high threshold taken into account by the comparison device.

 The inhibition signal T2, the duration of which can be adjusted, allows the processing of the different acoustic waves received by the sensor to be perfectly separated. The multivibrator 16 can be retriggered during a following selection operation, its role being limited to inhibiting the validation signal for an adjustable duration corresponding to the echoes.

 The invention is not limited to the embodiment which has been described. It is thus possible to imagine carrying out the selection electronics in a different manner which has been described in detail with reference to FIG. 2.

 One can imagine other modes of obtaining a validation signal by directly verifying the presence of an overshoot of the high threshold for a period of time of a predetermined duration after the first overshoot of the low threshold by the signal.

 One can imagine the use of the method and the selection device according to the invention in the case of the processing of acoustic signals of mechanical origin, in any industrial installation. This selection method can be used both for monitoring an industrial installation and for checking a complex structure with large dimensions.

Claims (9)

 1.- Method for selecting an acoustic signal picked up in an industrial installation to determine whether this signal can be used for the detection of a mechanical phenomenon resulting in an acoustic emission, each acoustic signal emitted and picked up in the installation being in the form of a rectified electrical signal, characterized by the fact - that the amplitude of each of the half-waves (28) of the rectified electrical signal corresponding to the acoustic signal is compared to a first predetermined fixed amplitude (25), called low threshold, - that the amplitude of each of the half-waves (28) of the rectified electrical signal is compared to a second predetermined fixed amplitude (26), greater than the first low and called the high threshold, - that a time interval of duration is determined predetermined fixed Tî originating from the beginning of the first overshoot of the low threshold (25) by an alternation (28) of the signal, - that we are looking for if an overshoot of the high threshold (26) takes place during the interval of time duration TI, - and that the acoustic signal considered to be usable is accepted if at least one crossing of the high threshold (26) takes place during the time interval duration T1.  2.- selection method according to claim 1, characterized in that - a low threshold discrimination signal A is produced comprising a pulse 29 initiated by each crossing of the low threshold (25) by an alternation (28) of the signal rectified, of a duration corresponding to the duration of the exceeding of the low threshold (25) by the alternation (28) of the rectified signal, - that a high threshold discrimination signal B is produced comprising a pulse initiated by each of the exceedances of the high threshold (26) by an alternation (28), of a duration corresponding to the duration of exceeding the high threshold (26) by the alternation (28), - that a first amplitude signal is produced fixed and of predetermined duration T1 initiated by the first pulse (29a) of the low threshold discrimination signal A, - that a second signal of fixed amplitude and of predetermined duration T2 is initiated initiated by the end of the first signal of duration T1, - and that a validation signal Y is produced by combining the signal B of high threshold discrimination and the duration signal T2 called the inhibition signal, so that the validation signal Y comprises pulses corresponding only to the pulses of the high threshold discrimination signal B emitted outside the period of activation of the inhibition signal T2, that is to say during the activation period of the first signal of duration T1.  3. A selection method according to any one of claims 1 and 2. characterized in that the values of the low threshold 25 and the high threshold 26 can be adjusted.  4.- selection method according to any one of claims 1, 2 and 3, characterized in that the duration Tt of the first signal and the duration T2 inhibition signal can be set.  5. Selection method according to claim 2, characterized in that the duration T2 of the inhibition signal is a few tens of microseconds.  6.- Device for selecting an acoustic signal picked up in an industrial installation and put in the form of an electric signal rectified by means of a preamplifier (2) and a rectifier (3) characterized in that 'it comprises a device for comparing the signal to a predetermined low threshold (25) and to a predetermined high threshold (26), consisting of two monostable flip-flops 114. 14') each receiving the rectified electrical signal (28) and a threshold value predetermined (25. 26) and a stage for preparing a validation signal constituted by two monostable multivibrators (15. 16) in series, the first of which (15) receives the signal at the output of the monostable rocker (t4) performing the comparison of the signal with the low threshold for the preparation of a signal of predetermined fixed duration T1 initiated by the comparison signal at the low threshold and the second (16) of which receives the signal of duration T1 emitted by the first multivibrator ( 15>, for the preparation of a signal of predetermined duration rminée T2 initiated by the end of the duration signal T1 as well as a gate (20) receiving on the one hand the duration signal T2 or inhibition signal and on the other hand the signal at the output of the monostable rocker (14 ' ) of comparison with the high threshold, the gate (20) developing a validation signal Y resulting from the inhibition of the comparison signal at the high threshold at the output of (14 ') by the signal of duration T2.  7.- Selection device according to claim 6, characterized in that each of the monostable flip-flops (14, 14 ') is associated with a potentiometer (17, 17') for adjusting the threshold value supplied to the flip-flop monostable (14. 14 ') corresponding.  8.- Selection device according to claim 6, characterized in that each of the multivibrators (15, 16) is associated with a potentiometer (21, 22) for adjusting the value of the duration TI, T2 of the transmitted signal by the multivibrator (15.  16).  9. Application of the selection method according to any one of claims 1 to 6, to the selection of acoustic waves in an installation for detecting impacts in the primary circuit of a pressurized water nuclear reactor.