Vibration signal detection system and method for motor state diagnosis

Abstract

The invention relates to a vibration signal detection system for motor state diagnosis and a method thereof, the system comprises a piezoelectric acceleration sensor, a charge transfer module, a range selection module, an amplification module, an oscilloscope acquisition module and a data analysis module which are connected in sequence, and the piezoelectric acceleration sensor is used for acquiring charges generated by vibration; the charge transfer module is used for determining charge change in the vibration process; the range selection module is used for adjusting the gain of the amplification module; the amplification module is used for amplifying the charge change signal; the oscilloscope acquisition module is used for acquiring the amplified charge change signal; and the data analysis module performs fast Fourier transform and frequency sweep sampling calculation processing on the signals acquired by the oscilloscope acquisition module to obtain a vibration frequency value. Compared with the prior art, the method detects the dynamic process of vibration, is rich in perceivable characteristic parameters, effectively improves the measurement precision of motor vibration signals, and is suitable for motor state diagnosis application based on technologies such as deep learning.

Description

Technical Field

[0001] The present invention relates to the technical field of motor vibration detection, and in particular to a vibration signal detection system and method for motor state diagnosis. Background Art

[0002] The vibration parameters of the motor are one of the most important parameters that reflect the mechanical failure of the motor. Although directly testing the vibration parameters of the motor is an effective method, it is difficult to sample most motors on site or it is impossible to install them (such as rail car motors). Piezoelectric accelerometers are one of the most widely used methods for vibration signal acquisition. They have the characteristics of high sensitivity, wide frequency response range and high resonant frequency. They are mainly composed of piezoelectric sensor elements, measurement conversion circuits, elastic elements, protection elements, etc. After the vibration is detected, the internal piezoelectric sensor element will generate an electric charge when subjected to tensile / compressive stress. The amount of charge is proportional to the magnitude of the tensile / compressive stress. The charge is output through the measurement conversion circuit to obtain an electrical signal representing the vibration signal.

[0003] The piezoelectric accelerometers commonly used in the market can be divided into charge output type and voltage output type (IEPE type sensors). The charge output type does not process the charge signal and needs to be connected to a charge-voltage conversion circuit for signal acquisition. The voltage output type has a charge amplifier or voltage amplifier connected inside the sensor. It only needs to provide the sensor with an excitation source (and an amplifier circuit) to acquire signals, which is easy to use.

[0004] However, the voltage signal output by the existing IEPE sensors only represents an acceleration result value within the range, and the obtained signal (non-real signal) cannot be analyzed by spectrum.

[0005] The voltage output by the sensor connected to the charge amplifier is approximately equal to the ratio of the charge to the integral capacitor. Since the charge amplifier has memory, the output signal is the result of the integration of the charge on the integral capacitor and the signal of the charging and discharging process of the capacitor. It does not truly reflect the changes in the vibration process and is only an approximate measurement of the vibration signal.

[0006] In addition, the sensor connected to a voltage amplifier amplifies the weak voltage generated by the charge into a voltage signal that can be collected, but the accuracy of the output signal will be affected by the connecting cable. Summary of the invention

[0007] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a vibration signal detection system and method for motor status diagnosis, which can effectively improve the measurement accuracy of the motor vibration signal.

[0008] The object of the present invention can be achieved by the following technical scheme: A vibration signal detection system for motor state diagnosis, comprising a piezoelectric acceleration sensor, a charge transfer module, a range selection module, an amplification module, an oscilloscope acquisition module and a data analysis module connected in sequence, wherein the piezoelectric acceleration sensor is used to collect the charge generated by vibration;

[0009] The charge transfer module is used to determine the charge change during the vibration process;

[0010] The range selection module is used to adjust the gain of the amplification module;

[0011] The amplification module is used to amplify the charge change signal;

[0012] The oscilloscope acquisition module is used to collect the amplified charge change signal;

[0013] The data analysis module performs fast Fourier transform (FFT) and frequency sweep sampling calculation processing on the signal collected by the oscilloscope acquisition module, and outputs a vibration frequency value.

[0014] Furthermore, the piezoelectric acceleration sensor is specifically a charge output type piezoelectric sensor.

[0015] Furthermore, the charge transfer module adopts a pseudo-differential structure, one end of which is connected to the signal line of the piezoelectric acceleration sensor, and the other end of which is connected to the signal line shielding layer of the piezoelectric acceleration sensor.

[0016] Furthermore, the charge transfer module and the housing of the piezoelectric acceleration sensor share a common ground.

[0017] Furthermore, the range selection module includes a T-type resistor network and an electronic switch, and the electronic switch is adjusted to connect the tap of the T-type resistor network node to adjust the gain of the amplification module accordingly.

[0018] Furthermore, the amplification module is specifically a differential amplifier circuit.

[0019] Furthermore, the amplification factor of the amplification module is:

[0020] K=K 1 *K 2

[0021] Among them, K 1 To adjust the amplification factor after the electronic switch is connected to the tap of the T-type resistor network node, K 2 It is a fixed multiple of the differential amplifier circuit.

[0022] Further, the data analysis module includes an FFT spectrum analysis unit, a frequency sweep sampling unit and a comparison unit, wherein the FFT spectrum analysis unit is used to perform a fast Fourier transform on the signal collected by the oscilloscope acquisition module to obtain an approximate value of a target frequency in the signal;

[0023] The sweep frequency sampling unit is used to determine the sweep frequency sampling frequency range and perform sweep frequency sampling within the range according to the target frequency approximation value;

[0024] The comparison unit is used to compare the target frequency amplitude at each sampling frequency, screen out the maximum amplitude, and use the frequency corresponding to the maximum amplitude as the output vibration frequency value.

[0025] A vibration signal detection method for motor state diagnosis comprises the following steps:

[0026] S1. Use a piezoelectric acceleration sensor to collect the charge generated by vibration in real time, and use a charge transfer module to capture the charge change signal during the vibration process;

[0027] S2. The gain of the amplification module is adjusted by the range selection module, and the amplification module amplifies the charge change signal;

[0028] S3. Use the oscilloscope acquisition module to collect the amplified charge change signal, and transmit it to the data analysis module for fast Fourier transform and sweep frequency sampling calculation processing, and output the vibration frequency value.

[0029] Furthermore, the specific process of step S3 is as follows:

[0030] S31, using an oscilloscope acquisition module to collect the amplified charge change signal, and transmit it to a data analysis module;

[0031] S32, the data analysis module first performs a fast Fourier transform on the received signal to determine an approximate value of the target characteristic frequency;

[0032] Then, the frequency range of the frequency sweep sampling is determined based on the approximate value of the target characteristic frequency and combined with the Nyquist principle, and the frequency sweep sampling is performed within the frequency range;

[0033] Finally, by comparing the target frequency amplitude at each sampling frequency, the maximum amplitude is screened out, and the frequency corresponding to the maximum amplitude is output as the vibration frequency value.

[0034] Compared with the prior art, the present invention has the following advantages:

[0035] The present invention designs a piezoelectric acceleration sensor, a charge transfer module, a range selection module, an amplification module, an oscilloscope acquisition module and a data analysis module connected in sequence, and uses a piezoelectric acceleration sensor to collect the charge generated by vibration; uses a charge transfer module to determine the charge change in the vibration process; uses a range selection module to adjust the gain of the amplification module; uses the amplification module to amplify the charge change signal; uses an oscilloscope acquisition module to collect the amplified charge change signal; uses a data analysis module to perform fast Fourier transform and sweep frequency sampling calculation processing on the signal collected by the oscilloscope acquisition module to output the vibration frequency value. Therefore, for the application of piezoelectric acceleration sensor in vibration measurement, a charge transfer module is designed to capture the real charge change in the vibration process, rather than relying solely on the processed acceleration result value. Then, the range selection module and the amplification module are combined to ensure the accuracy of signal measurement, and automatic gain adjustment is realized to provide more detailed information for the vibration process. In addition, combined with the method of digital signal processing, on the basis of roughly estimating the target frequency range by FFT, a scheme combining sweep frequency sampling with amplitude comparison is proposed, which can more accurately determine the characteristic frequency in the vibration signal.

[0036] In the present invention, the charge transfer module adopts a pseudo-differential structure, that is, the signal line and the signal line shielding layer of the sensor are respectively connected to the input end of the differential circuit. This structure can enhance the anti-interference ability to a certain extent, while reducing the influence of the length of the sensor signal line (capacitance to ground) on the sensor sensitivity, thereby accurately capturing the charge change signal.

[0037] The present invention combines a T-type resistor network, an electronic switch and a differential amplifier circuit, wherein the T-type resistor network and the input end of the differential amplifier circuit are directly connected, and the electronic switch is adjusted to connect the tap of the T-type resistor network node to adjust the amplification gain accordingly, thereby avoiding a large error between the actual amplification factor and the ideal amplification factor, and adopting a differential structure can eliminate common mode interference and minimize EMI electromagnetic interference.

[0038] The present invention proposes a scheme for roughly estimating the target frequency range based on FFT, and performing frequency sweep sampling and amplitude comparison on this basis, that is, firstly, performing FFT spectrum analysis to determine the target characteristic frequency estimation value; then combining the estimation value, determining the frequency range of frequency sweep sampling according to the Nyquist principle; finally, performing frequency sweep sampling within the determined frequency range, and comparing the target frequency amplitude at each sampling frequency until the maximum amplitude is obtained, and the frequency corresponding to the maximum amplitude is the closest target characteristic frequency. This not only improves the accuracy of frequency measurement, but also enhances the understanding of the dynamic characteristics of the vibration system. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a schematic diagram of the system structure of the present invention;

[0040] Figure 2 Schematic diagram of the charge transfer circuit in the present invention;

[0041] Figure 3 It is a schematic diagram of a T-type resistor network in the present invention;

[0042] Figure 4 Schematic diagram of the differential amplifier circuit in the present invention;

[0043] Figure 5 It is a schematic diagram of the method flow of the present invention;

[0044] Figure 6 It is a schematic diagram of the numerical analysis process in the present invention;

[0045] Figure 7 This is a spectrum diagram generated by 10 Hz square wave excitation in the embodiment;

[0046] Figure 8 This is a spectrum diagram generated by 20 Hz square wave excitation in the embodiment;

[0047] Fig. 9 This is a spectrum diagram generated by 30 Hz square wave excitation in the embodiment.

[0048] Explanation of the marks in the figure: 1. Piezoelectric acceleration sensor, 2. Charge transfer module, 3. Range selection module, 4. Amplification module, 5. Oscilloscope acquisition module, 6. Data analysis module. DETAILED DESCRIPTION

[0049] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0050] Example

[0051] like Figure 1 As shown, a vibration signal detection system for motor status diagnosis includes a piezoelectric acceleration sensor 1, a charge transfer module 2, a range selection module 3, an amplification module 4, an oscilloscope acquisition module 5 and a data analysis module 6 connected in sequence, wherein the piezoelectric acceleration sensor 1 is used to collect the charge generated by vibration;

[0052] The charge transfer module 2 is used to determine the charge change during the vibration process;

[0053] The range selection module 3 is used to adjust the gain of the amplification module;

[0054] The amplification module 4 is used to amplify the charge change signal;

[0055] The oscilloscope acquisition module 5 is used to collect the amplified charge change signal;

[0056] The data analysis module 6 performs fast Fourier transform and frequency sweep sampling calculation processing on the signal collected by the oscilloscope acquisition module 5, and outputs a vibration frequency value.

[0057] The measurement method of piezoelectric acceleration sensor is studied based on piezoelectric theory, and the motor state is analyzed and evaluated from the perspective of vibration signal. Piezoelectric sensor has the characteristics of strong anti-interference ability, high detection accuracy, wide frequency band range, etc., and is often used in vibration signal detection. Since the commonly used piezoelectric sensor now has a charge amplifier or voltage amplifier integrated inside, its output signal represents the result value of acceleration rather than the actual vibration process signal. Therefore, this embodiment designs a charge transfer circuit based on the charge output type piezoelectric sensor in order to obtain the actual charge change process that can reveal the vibration process. In order to ensure the accuracy of signal measurement and automatic gain adjustment, a method of combining T-type resistor network and differential amplifier circuit is adopted, combined with a digital oscilloscope to collect signals, and finally uploaded to the LabVIEW host computer for data analysis. At the same time, based on the rough estimation of the target frequency range based on FFT, a method of sweep frequency sampling and amplitude comparison is proposed to improve the accuracy of the measured target frequency.

[0058] The specific contents are:

[0059] (i) Sensor selection: In order to obtain the charge change that can characterize the vibration process, this embodiment selects a charge output type piezoelectric sensor, and at the same time requires the sensor to have the characteristics of high sensitivity, wide bandwidth, and strong anti-interference ability.

[0060] In this embodiment, the selection index of the piezoelectric acceleration sensor is:

[0061] Signal output type: charge signal;

[0062] Rated output range: ±3000g;

[0063] Frequency bandwidth: 1Hz-10kHz;

[0064] Sensitivity: 16.57pc / g;

[0065] Resonant frequency: 40kHz.

[0066] (ii) Charge transfer circuit design: Figure 2As shown. For vibration signals in the medium and high frequency bands, the piezoelectric sensor collects vibration signals to generate charges, and the current generated by charging the capacitor through the charges is amplified into a voltage signal output; for low-frequency signals, due to the low-frequency resistance of the capacitor, the low-frequency response is poor. In order to improve the low-frequency response characteristics, a small resistor is connected in parallel at both ends of the capacitor; the charge transfer circuit adopts a pseudo-differential structure, that is, the signal line and the signal line shielding layer of the sensor are respectively connected to the input end of the differential circuit (one end of the charge transfer circuit is connected to the sensor signal output, and the other end of the charge transfer circuit is connected to the sensor signal line shielding layer, and the charge transfer circuit and the sensor housing are grounded). This structure can enhance the anti-interference ability to a certain extent, while reducing the influence of the length of the sensor signal line (capacitance to ground) on the sensor sensitivity.

[0067] In this embodiment, the frequency bandwidth of the charge transfer circuit is 1 Hz-10 kHz.

[0068] (III) Amplification circuit design: A method combining a T-type resistor network, an electronic switch and a differential amplifier circuit is used, wherein the T-type resistor network is as follows: Figure 3 As shown, the differential amplifier circuit is as follows Figure 4 This method changes the amplification factor K by adjusting the tap of the T-type resistor network node connected by the electronic switch. 1 , and then through the fixed multiple K of the differential amplifier circuit 2 , we get the gain adjustable amplification factor K = K 1 *K 2 The circuit has the following two advantages:

[0069] 1) If the gain resistor of the differential amplifier circuit is directly selected by the electronic switch, the amplification factor will be affected by the resistance of the electronic switch itself, that is, there will be a large error between the actual amplification factor and the ideal amplification factor; while by directly connecting the T-type resistor network and the input end of the differential amplifier circuit, the error can be avoided due to the virtual disconnection characteristics of the operational amplifier (infinite input impedance).

[0070] 2) The differential structure can eliminate common mode interference and minimize EMI. Since the amplified signal is the voltage difference between the two ends, the same interference has the same effect on the two ports of the differential signal, and the interference caused by this part cancels each other out.

[0071] In this embodiment, the design index of the amplifier circuit is:

[0072] Circuit structure: Take the voltage signal from the T-type resistor network and connect it to the differential circuit;

[0073] Frequency bandwidth: 1Hz-10kHz;

[0074] Magnification: 1.56, 6.25, 25, 100.

[0075] (iv) Oscilloscope acquisition: Connect the oscilloscope probe to the output end of the differential circuit, and then connect it to the host computer through the LAN interface or USB interface, and upload the signal collected by the oscilloscope to LabVIEW for analysis.

[0076] In this embodiment, the oscilloscope sampling index is:

[0077] Sampling rate: maximum 1.25Gsa / s;

[0078] Precision: 12 bits;

[0079] Frequency bandwidth: 0-70MHz.

[0080] (V) LabVIEW signal analysis: First, perform FFT spectrum analysis to determine the estimated value of the target characteristic frequency; then, combine the estimated value and determine the frequency range of the swept frequency sampling according to the Nyquist principle; finally, perform swept frequency sampling within the determined frequency range, and compare the target frequency amplitude at each sampling frequency (that is, compare the amplitudes corresponding to the target frequencies in the previous and next FFT spectrum diagrams in sequence) until the maximum amplitude is obtained. The frequency corresponding to the maximum amplitude is the closest target characteristic frequency.

[0081] Based on the above hardware system, a vibration signal detection method for motor status diagnosis is implemented, such as Figure 5 As shown, the following steps are included:

[0082] S1. Use a piezoelectric acceleration sensor to collect the charge generated by vibration in real time, and use a charge transfer module to capture the charge change signal during the vibration process;

[0083] S2. The gain of the amplification module is adjusted by the range selection module, and the amplification module amplifies the charge change signal;

[0084] S3. Use the oscilloscope acquisition module to collect the amplified charge change signal, and transmit it to the data analysis module for fast Fourier transform and sweep frequency sampling calculation processing, and output the vibration frequency value.

[0085] Considering that traditional frequency measurement methods rely on a fixed sampling rate, which may lead to inaccurate target frequency identification, especially under complex working conditions, the motor vibration frequency range is wide and variable, and a single sampling strategy is difficult to meet the needs.

[0086] Therefore, this solution proposes a method in step S3 to roughly estimate the target frequency range based on FFT, and perform frequency sweep sampling and amplitude comparison on this basis. This method not only improves the accuracy of frequency measurement, but also enhances the understanding of the dynamic characteristics of the vibration system. The vibration signal is synthesized by computer and numerically analyzed. The numerical analysis process is as follows: Figure 6As shown. Combined Figure 7 , Figure 8 , Fig. 9 The in-loop simulation results show that (the vibration spectrum analysis measured in the experiment is as follows Figure 7 , Figure 8 and Fig. 9 As shown, the vibration table excitation signal is a square wave with an amplitude of 10V, frequencies of 10Hz, 20Hz and 30Hz respectively, and the oscilloscope sampling rate is 100k). This method can significantly improve the accuracy of frequency measurement.

[0087] In summary, the charge transfer circuit designed in this scheme can characterize the charge change of the actual vibration signal process, combined with the differential amplifier circuit with automatic gain adjustment, to achieve accurate vibration signal measurement, and has the characteristics of strong anti-interference ability. In data processing, on the rough estimate of the target frequency of FFT spectrum analysis, the swept frequency sampling and amplitude comparison method are used to improve the measurement accuracy of the target characteristic frequency. This scheme can be applied in practice and can be used for intelligent diagnosis of motor testing, effectively improving the accuracy of motor vibration measurement.

Claims

1. A vibration signal detection system for motor status diagnosis, characterized in that: It comprises a piezoelectric acceleration sensor (1), a charge transfer module (2), a range selection module (3), an amplification module (4), an oscilloscope acquisition module (5) and a data analysis module (6) which are connected in sequence, wherein the piezoelectric acceleration sensor (1) is used to collect charges generated by vibration; The charge transfer module (2) is used to determine the charge change during the vibration process; The range selection module (3) is used to adjust the gain of the amplification module (4); The amplification module (4) is used to amplify the charge change signal; The oscilloscope acquisition module (5) is used to acquire the amplified charge change signal; The data analysis module (6) performs fast Fourier transform (FFT) and frequency sweep sampling calculation processing on the signal collected by the oscilloscope collection module (5), and outputs a vibration frequency value.

2. A vibration signal detection system for motor status diagnosis according to claim 1, characterized in that: The piezoelectric acceleration sensor (1) is specifically a charge output type piezoelectric sensor.

3. A vibration signal detection system for motor status diagnosis according to claim 1, characterized in that: The charge transfer module (2) adopts a pseudo-differential structure, one end of which is connected to the signal line of the piezoelectric acceleration sensor (1), and the other end of which is connected to the signal line shielding layer of the piezoelectric acceleration sensor (1).

4. A vibration signal detection system for motor status diagnosis according to claim 3, characterized in that: The shells of the charge transfer module (2) and the piezoelectric acceleration sensor (1) share a common ground.

5. A vibration signal detection system for motor status diagnosis according to claim 1, characterized in that: The range selection module (3) comprises a T-type resistor network and an electronic switch, and the electronic switch is adjusted to connect the tap of the T-type resistor network node to adjust the gain of the amplification module (4) accordingly.

6. A vibration signal detection system for motor status diagnosis according to claim 5, characterized in that: The amplification module (4) is specifically a differential amplification circuit.

7. A vibration signal detection system for motor status diagnosis according to claim 5, characterized in that: The magnification factor of the amplification module (4) is: K=K1*K2 Among them, K1 is the gain factor that changes after the tap of the T-type resistor network node is turned on by adjusting the electronic switch, and K2 is the fixed factor of the differential amplifier circuit.

8. A vibration signal detection system for motor status diagnosis according to claim 1, characterized in that: The data analysis module (6) comprises an FFT spectrum analysis unit, a frequency sweep sampling unit and a comparison unit, wherein the FFT spectrum analysis unit is used to perform a fast Fourier transform on the signal collected by the oscilloscope collection module (5) to obtain an approximate value of a target frequency in the signal; The sweep frequency sampling unit is used to determine the sweep frequency sampling frequency range and perform sweep frequency sampling within the range according to the target frequency approximation value; The comparison unit is used to compare the target frequency amplitude at each sampling frequency, screen out the maximum amplitude, and use the frequency corresponding to the maximum amplitude as the output vibration frequency value.

9. A vibration signal detection method for motor state diagnosis, applied to a vibration signal detection system for motor state diagnosis as claimed in claim 1, characterized in that: The following steps are involved: S1, using a piezoelectric acceleration sensor (1) to collect the charge generated by vibration in real time, and using a charge transfer module (2) to capture the charge change signal during the vibration process; S2, adjusting the gain of the amplification module (4) through the range selection module (3), and the amplification module (4) amplifies the charge change signal; S3, using the oscilloscope acquisition module (5) to collect the amplified charge change signal, and transmit it to the data analysis module (6) for fast Fourier transform and sweep frequency sampling calculation processing, and output to obtain the vibration frequency value.

10. A vibration signal detection method for motor status diagnosis according to claim 9, characterized in that: The specific process of step S3 is: S31, using the oscilloscope acquisition module (5) to collect the amplified charge change signal, and transmit it to the data analysis module (6); S32, the data analysis module (6) first performs a fast Fourier transform on the received signal to determine an approximate value of the target characteristic frequency; Then, the frequency range of the frequency sweep sampling is determined based on the approximate value of the target characteristic frequency and combined with the Nyquist principle, and the frequency sweep sampling is performed within the frequency range; Finally, by comparing the target frequency amplitude at each sampling frequency, the maximum amplitude is screened out, and the frequency corresponding to the maximum amplitude is output as the vibration frequency value.

Images