DE3332941A1 - Method and device for sound analysis of machines and plants - Google Patents

Abstract

The method for sound analysis of running machines measures the noise generated by the specimen segment by segment, the segments being determined by the operating cycle of the machine. Only one specific measured value, e.g. peak value of the noise is determined inside each segment and subsequently stored in digital form. The stored digital values of a plurality of segments are compared to one another and/or to prescribed threshold values. <IMAGE>

Description

Method and device for sound analysis of machines and systems Description The invention relates to a method according to the preamble of claim 1 and to a device according to the preamble of claim 14th

From DE-OS 29 50 869 a method and a device for Quality testing of materials (test items) known, in which the test item for Acoustic emission analysis is subjected to a load and the resulting therefrom Noise emissions are recorded, in sections depending on the specified Load levels. Since the beginning of crack formation in loaded specimens by sonic makes emissions clearly noticeable, you can determine at which load level the crack formation begins, or at the non-destructive one Material testing examine whether the test specimen has certain stress levels without cracking endures.

The device according to DE-OS 29 50 869 has a sound sensor for this purpose and a measuring device for determining a specific parameter of the test object, for example Load or deflection. The signals from the sound pick-up are consecutive an amplifier, an analyzer, a pulse shaper and a pulse height analyzer fed, the multiple outputs of the pulse height analyzer each to a counter are fed. The counters are controlled by signals from the measuring device for the parameter of the device under test controlled, the outputs of the counters being fed to memories.

DE-PS 30 12 773 discloses a method for monitoring machines and components are known in the operating state by recording their sound vibrations, in the case of fluctuations in the sound oscillation amplitudes via an integrator with a predetermined value Discharge slope can be detected. The time constant for the discharge characteristic is like this chosen that the discharge time at least over the smallest to be expected Period duration of the sound oscillation extends, the integrator of each detected peak value of the amplitude is discharged starting until the during the Discharge time, the amplitude of the sound oscillation changes, the discharge value of the integrator has reached. It then becomes the difference between the discharge value and the respective detected peak value of the amplitude, and the successive difference values are used to detect irregularities in the machine and its components compared.

From an essay by Dunegan and Green, "Factors Affecting Acoustic Emission Response From Materials "(Acoustic emission ASTM STP.505. American Society for Testing and Materials, 1972, pages 100-113) it is also known to use a Acoustic emission analysis on the stress or deformation of a test object relate.

When analyzing vibration signals, regardless of whether they takes place in the time domain or frequency domain, is usually the size of the oscillation as a peak value or as a mean value or the energy of the oscillation as an effective value or integration value determined.

The different measurands result in a simplified form by that the signals are rectified once or twice and a capacitor is charged with coordinated loading and unloading times. By skillful Coordination of the during and discharge times of the capacitor and the coordination of the time constants with the downstream measuring and display system then different results result Information about the waveform of the signals.

With this measuring method, continuous sinusoidal oscillations can be achieved describe exactly. However, the further the signal form deviates from a sine and additionally occurs discontinuously or stochastically, the further the deviations Measured values differ from the actual signal sizes and are no longer sufficient to describe characteristic behavior of a system.

Instead of analog measurement technology, it is now possible to use the Signal with sufficiently fast analog-digital converters with a given sampling rate to digitize and then further process in computers.

Depending on the speed of the A / D converter, one sample can the respective instantaneous value of the oscillation can be taken over and digitized. Ever further examined this Signal deviates from the sinusoidal shape, d. H. the more high-frequency components the signal contains, the more digital measured values must be scanned in order to be able to reproduce the signal shape exactly. at a real-time measurement of relatively high-frequency signals is therefore the storage capacity even larger computers can be used to capacity quickly.

With the above-described DE-PS 30 12 773 you get a compression of information, d. H. the complete signal curve no longer has to be stored and processed on the other hand, so much information goes through this information compression lost that not all the required tests can be carried out.

The object of the invention is therefore the method and the device of the aforementioned type to the effect that they can be used while the machines are running Record and save as little information as possible in order to achieve the desired sound analysis to be able to perform.

This object is achieved by the characterizing features of the claims 1 or 14 solved. Advantageous embodiment and development of the invention are to be found in the subclaims.

The basic idea of the invention is therefore that the signal does not more continuously scanned and buffered, but in larger segments is broken down and only one value is saved for each segment.

The width of the segments is determined by the movement behavior of the test object determined, for example in engines by a predetermined crankshaft angle or in the case of gearboxes by the angle of rotation of a face width. The only reading of one Segment is then saved and evaluated. A combination is particularly favorable of analog and digital signal processing. Within each segment, the The only value to be recorded is determined using analog technology and then digitized and in digital form saves. This combination of analog and digital signal processing solves the problem of coordinating the charging or discharge times of the capacitor avoided. The charging time of the capacitor is selected so quickly that the highest frequencies in the frequency range to be analyzed Signals can still be recorded properly. At the end of each segment there is the one on the capacitor pending value digitized, whereupon the capacitor can be deleted very quickly since the discharge time constant has no influence on the analysis result.

You can adapt the signal shape by choosing the segment width. In this way, the storage capacity can be relieved depending on the required analysis accuracy will.

In the following the invention is based on an exemplary embodiment explained in more detail in connection with the drawing.

They show: FIG. 1 an example of a signal curve for explanation the method according to the invention; 2 shows a basic circuit diagram of a device for sound analysis according to an embodiment of the invention; Fig. 3 is a circuit diagram of another embodiment of an analyzer for use in the block diagram according to FIG. 2; 4 shows the signal profile of individual signals to explain the mode of operation of the circuit diagram of FIG. 2.

In Fig. 1, a sound noise signal is an example of a running Machine shown. Here it is assumed that the machine runs smoothly, so that an exactly linear relationship between time t and a movement quantity the machine, e.g. B. Crankshaft angle exist. The signal to be analyzed becomes now divided into individual segments S1-S10, with segment S1 runs from time to-t1, which in the example of an internal combustion engine is a crankshaft angle of ovo corresponds to.

Only one specific value Am1 -Am1o determined. In the exemplary embodiment in FIG. 1, this is the peak value the amplitude of the signal. According to other variants of the invention, instead of Peak value can also be determined by the integral over the segment, the effective value of the signal or some other value. The determined specific values Am1-Am1o are then stored in digital form at the end of each assigned segment. During the evaluation, the stored digital values from Am1-Am1o are combined with one another and / or compared with threshold values. Fig. 2 shows a circuit diagram of a device, with which the process according to the invention can be carried out. A sound sensor 10 receives the sound signals generated by the test object. The sound sensor 10 can be a microphone, a piezo-electric transducer and can either be direct be coupled to the test object for structure-borne noise measurement, or a little further be arranged away from him for the measurement of airborne noise. For testing Motors is the sound pick-up for a frequency range from approx. 4 Hz to 100 kHz designed. A structure-borne sound sensor and a microphone can also be provided in parallel be.

The electrical output signals of the sound sensor 10 are shown in amplified by a preamplifier 11 and then in a high-pass filter 12 and a downstream one Low pass 13 filters so that a specific frequency band is selected will. The low-pass filter 13 is followed by an amplifier 14, its output signals here a full-wave rectifier 15 are fed. Via a downstream Changeover switch 16, which is controlled by signals on a line 17, are the Output signals of the rectifier 15 optionally on a line 18 or one Line 19 of a 1st measuring channel 20 or a 2nd measuring channel 21 is given. In a practical Embodiment is of course the switch 16 with electronic components built up and thus works contactless.

The two channels 20 and 21 are constructed identically, so that it is sufficient describe the 1st channel 20 in more detail. The analyzer is here as a peak detector formed, which is realized by a capacitor 22 (23) between the Line 18 (19) and ground. Since the amplifier 14 and the rectifier 15 have a have low internal resistance, the capacitor 22 has a very small charging time constant and thus always saves the highest voltage (peak value) of the signal on the Line 18. A controllable switch 24 is parallel to the capacitor 22 (23) (25), which can be actuated via signals on a control line 26 (27). If the switch 24 is closed, the capacitor 22 becomes almost suddenly unload. Oem capacitor 22 (23) is a controllable analog / digital converter 28 (29) connected downstream, which, controlled by signals on a control line 33 (34), converts the analog signals present at its input into digital signals and transmitted to a control circuit 32 via a data line 30 (31). This steering group 32 is available for both channels 20 and 21 together.

The digital signals are transmitted from the control circuit 32 via a data line 35 transmitted to a memory 37 and stored there.

Since the control circuit 32 contains a computer that carries out the evaluation, The stored data can also be transferred from the memory 37 via a data line 36 transferred back to the control circuit.

Furthermore, a measuring device 38 for the specific movement size of the Test object, for example for the crankshaft angle of an engine, to the control circuit 32 connected and finally also a timer 39 in the event that the segments should have a constant duration for certain measurements.

In the following the operation of the device according to FIG. 2 will be explained.Es it is assumed that an internal combustion engine is to be tested and that the Measuring device 38 measures the crankshaft angle. A segment is supposed to be a crankshaft angle of 10. At time t0 (FIG. 1), which corresponds to the crankshaft angle of 0o should correspond, the measuring process is started and channel 20 is selected. By a pulse on the line 17, the switch 16 is in the position shown brought. The voltage on capacitor 22 then represents the peak value of the sensed, filtered. amplified and rectified sound signal. At the time t (Fig. 1), which corresponds to a crankshaft angle of 10, is the first segment 51 finished. The switch 16 is now switched over via a signal on the line 17, so that subsequent signals are routed to channel 21. Appears at the same time an activation signal for the analog / digital converter 28 on the control line 33, which now its input signal, which corresponds to the value Aml (Fig. 1), into a digital one Converts signal and transmitted via line 30 to control circuit 32 from where it is passed on to the memory 37. As soon as the analog / digital conversion is completed, the switch 24 is activated via a signal on the control line 26 closed, so that the capacitor 22 short-circuited and thus very quickly load will. Meanwhile, the channel 21 has already started its work, where now during of segment S2 (t1-t2) is described in a corresponding manner in connection with channel 20 the measuring process is running. At time t2 (crankshaft angle 2 °), now again switched to channel 20, where the process described above for the segment S3 expires. The two channels 20 and 21 thus work alternately. The reason for the arrangement of two channels is that the discharging of the capacitors, the analog-to-digital conversion and the transfer of the digital values to the memory but require a certain period of time during which a measuring channel is not ready to receive is. Measurements of interest could occur during this gap, so it to increase the accuracy it is advantageous to provide the two channels. With relatively However, you can also use low frequencies of the sound signal to be analyzed get by with a measuring channel if the discharge time is constant of the capacitor and the The duration of the analog-digital conversion is very small in relation to the period of the are signals to be analyzed.

When testing the engine, it is sufficient to measure 7200 crankshaft angles perform. In memory 37 there are then camshaft angles with a segment width of 10 A total of 720 peak values are stored in digital form, which are then evaluated will. There are now various options for the evaluation: 1. Specifications for Good / bad detection are set as limit values (threshold values) or limit curves programmed; 2. An adaptive evaluation is carried out in which the system when measuring a larger number of aggregates, enter which aggregate accepted and which one should be discarded. The evaluation circuit can use this then automatically the associated noise patterns or limit values or limit curves identify and save; 3. the limit values and limit patterns are automatically based on statistical functions from a large number of measurements calculated or programmed.

There are now different sub-variants for these three ways. at The individual measured values of each segment with a limit value are a sub-variant compared for this very segment.

Further information is obtained when the measured values of the individual segments are compared with one another, for example the difference or the quotient of the values Aml and Am2 is formed.

Finally, the measured values of several segments can also be grouped together be combined, for example, by the segments S1-S5 a first group and the segments S6-S10 form a second group. From the values of a group can then a special value can be selected, for example the maximum value of the values Am1 Am5 or the minimum value of the values Am1 ~ Am5 the latter then represents with the described Peak value detection represents the minimum of the maxima of the segments S1-S5. The selected Values of the individual groups are then compared with one another, for example by Difference or quotient formation.

Fig. 3 shows a modified analyzer, here in the form of a Integrator is formed. The capacitors 22 and 23 of FIG. 2 are here replaced by an integrator made up of resistor 40 and capacitor 41. the Switches 24 and 25 are here via a transistor 43 with an operating resistor 42 replaced, whereby the base 44 of the transistor is then connected to the control lines 26 resp. 27 are connected. The other parts correspond to those of FIG. 2 and are therefore no longer shown.

4 shows the course of various signals at individual measuring points of the device of FIG. 2. With U38 are pulses of the measuring device 38 denotes, which determine the segment width. This is where the control signals become U17 derived on line 17 for switch 16. The voltage across the capacitor 22 is denoted by U22, while the voltage across the capacitor 23 is denoted by U23 is. It can be seen that the two capacitors act alternately and act as a peak value detector work. With AD1 and AD2 are the control signals for the analog-digital converter 28 and 29 respectively. These signals appear shortly after the end of the respective Segment. After the signals ADl or

AD2, the control signals U26 and U27 are generated, which the switches 24 and 25 close and the capacitors 22 and 23 then discharge. For sound analysis the segment width is of particular importance. If you choose the segment width relatively small, based on the period of the signal to be analyzed, one obtains practically an envelope curve of the signal to be analyzed, but on the other hand also a relatively large number of measured values, which causes problems with memory capacity brings with it.

If, on the other hand, the segment width is chosen to be very large, it is reduced the resolving power, d. H. certain signal changes of interest cannot more can be captured.

It can be seen from the above that a time analysis is carried out with the invention which can even be executed in real time as a rule. As the storage capacity with the invention is relatively lightly loaded, the entire analysis can be carried out in parallel can be carried out in several channels in which the signals are each different be filtered.

In other words, time analyzes can be carried out for several frequency channels can be performed so that a frequency weighted analysis is obtained. Though above the engine test has been referred to several times, it should be emphasized that any running machines or systems can be analyzed with the invention; for example the rolling noise of gears can be analyzed during the gearbox test and by assigning the segment width to a tooth spread the Rolling noises of the individual teeth of the gearwheel can be analyzed.

With the evaluation described above, not only good / bad information can be obtained obtained; rather, very complex noises can be made into very special ones Errors are detected. For example, it is the peak value in the engine test within a very specific segment, in which just a specific cylinder is ignited too high, this indicates a bearing fault in the crankshaft bearing for this cylinder (incorrect bearing play). Have amplitudes in other segments on the valve control, while an excessive background noise level (minima the measured values of all segments) for incorrect tension of the toothed belt for the camshaft drive indicates. If individual maxima are shifted by a segment compared to the target value, then this indicates incorrect ignition timing.

Certain rhythms or time shifts can also be set Recognize the occurrence of certain signal components or also beats, d. H. Overlays with other background noise.

When setting the segment width, the lower limit is still to Note that the times for discharging the capacitor, the analog-to-digital conversion and the transmission of the signal to the memory are shorter than the time for one Segment. If this condition cannot be met, further parallel ones must be used Measuring channels, e.g. B. three or more may be provided. The control with evaluation circuit 32 is implemented in practice with a programmable microcomputer that is used for any measuring tasks can be reprogrammed.

All in the claims, the description and the drawing The technical details shown can be used both on their own and in any Combination be essential to the invention.

List of reference symbols 10 sound pick-up 11 preamplifier 12 high-pass filter 13 Low-pass filter 14 Amplifier 15 Rectifier 16 Changeover switch 17 Control line for 16 18 signal line for channel 20 19 signal line for channel 21 20 1st channel 21 2nd channel 22 capacitor (analyzer) 23 capacitor (analyzer) 24 switch 25 Switch 26 control line for 24 27 control line for 25 28 analog-digital converter 29 Analog-digital converter 30 Data line 31 Data line 32 Control circuit / evaluation circuit 33 control line for 28 34 control line for 29 35 data line to 37 36 data line of 37 37 Memory / Display 38 Meter 39 Timer 40 Resistor 41 Capacitor 42 resistor 43 transistor 44 control line (corresponds to 2627) - blank page-

Claims (20)

Method and device for sound analysis of machines and systems Claims 1. Method for sound analysis of machines and systems (test specimen), in which the sound noises generated by the test object are measured and segment-wise (in sections) can be analyzed, the segments depending on specific Characteristics of the test specimen are determined, that is to say e t, then the segments are determined by the sequence of movements of the Pruflinos Only one specific measured value of the sound noise is determined within each segment and is cached and then in digital form is stored, and that the stored digital values of several segments with one another and / or are compared with predetermined threshold values. 2. The method according to claim 1, characterized in that the specific The measured value of each segment is determined in analog form and then converted from analog to digital, whereupon the temporarily stored analog measured value is deleted. 3. The method according to claim 1 or 2, characterized in that the specific measured value the peak value of the sound noise signal during the segment is. 4. The method according to claim 1 or 2, characterized in that the specific measured value the integral of the sound noise signal integrated over the Segment is. 5. The method according to claim 1 or 2, characterized in that the specific measured value the envelope curve of the sound noise signal during the segment is. 6. The method according to any one of claims 1 to 5, characterized in that that the segment is a divisor of a periodic movement quantity of the test object. 7. The method according to any one of claims 1 to 5, characterized in that that the segment is a divisor of the measurement time. 8. The method according to any one of claims 1 to 7, characterized in that that the minimum width of a segment 1800 (or g)) of the sound noise signal component with the highest frequency to be analyzed. 9. The method according to any one of claims 1 to 8, characterized in that that the digital measured values of several Segments combined in groups a specific group value is determined from the measured values of a group and that specific group value with a specific group value of another Group and / or a threshold value is compared. 10. The method according to claim 9, characterized in that the specific Group value is the maximum value of the measured values in a group. 11. The method according to claim 9, characterized in that the specific Group value is the minimum value of the maxima of the measured values of the segments in a group. 12. The method according to any one of claims 1 to 11, characterized in that that it is carried out for different frequency ranges of the sound noise. 13. The method according to any one of claims 1 to 11, characterized in that that the method in at least two parallel and alternating measuring channels is carried out. 14. Device for sound analysis of machines and systems (test item) with a sound sensor that detects sound noises emitted by the test object and is converted into electrical sound noise signals, with a measuring device, the specific Characteristics of the test object recorded with a downstream of the sound sensor Analyzer, with a control circuit that depends on signals from the measuring device controls an evaluation circuit, and with one of the evaluation circuit connected downstream Memory that shows that the measuring device (38, 39) Fixed segments (S1-S10) of the movement sequence of the test object detected that the Analyzer (22, 23) only one specific value within each segment (Am1 Am10) the sound noise sig- nale detects that the memory (37) per segment only stores this specific value and that the evaluation circuit (32) the im Memory (37) compares stored values with one another and / or with a threshold value. 15. The device according to claim 14, characterized in that the Analyzer (22, 23) is a peak value detector. 16. The device according to claim 15, characterized in that the Peak detector consists of a capacitor (22; 23) to which the sound noise signals can be supplied, as well as from a controllable switch connected in parallel to this (24, 25; 43) to quickly discharge the capacitor. 17. The device according to claim 14, characterized in that the Analyzer is an integrator (40, 41). 18. Device according to one of claims 14 to 17, characterized in that that the analyzer is followed by a controllable analog-digital converter (28, 29) is. 19. Device according to one of claims 14 to 18, characterized in that that two measuring channels (20, 21) connected in parallel and working alternately are provided which can be connected alternately to the sound sensor (10) via a switch (16) are. 20. Device according to one or more of claims 14 to 19, characterized in that several parallel and simultaneously working measuring channels are intended for different frequency ranges.