Electric motor diagnostic equipment

Abstract

Provided is an electric motor diagnosis device which is capable of diagnosing the presence of an abnormality in an electric motor, by detecting sidebands which form peaks at both sides in the vicinity of a power supply frequency, even in an electric motor having a load torque which varies. The current of an electric motor 5 is detected by a current detector 4 and inputted from a current input unit 10. Power spectrum analysis results, which are obtained by performing frequency analysis a plurality of times on the current waveform when the current is stable, are subjected to averaging processing by a logical operation unit 11, sidebands are detected from the power spectrum analysis results which have been subjected to the averaging processing, to determine the presence of an abnormality in the electric motor 5, and an alarm is outputted from an alarm output unit 21 if it is determined that an abnormality has occurred.

Description

Electric motor diagnostic device This invention relates to a motor diagnostic device used in control centers, such as closed-circuit distribution panels, to diagnose whether or not there is an abnormality in an induction motor. Conventionally, a method for diagnosing abnormalities in induction motors and equipment driven by induction motors has been proposed, which involves measuring the load current of an induction motor, performing frequency analysis, and focusing on the sidebands generated on both sides of the operating frequency. Based on the state of short-period vertical waveform disturbances and long-period vertical waveform oscillations (swells), abnormalities in induction motors and equipment driven by induction motors have been diagnosed. (For example, Patent Document 1) Patent No. 4782218 Conventional equipment anomaly diagnosis methods have a problem in that when load torque fluctuations occur in induction motors, the spectral intensity on both sides near the power supply frequency (operating frequency) increases, becoming greater than the vibration intensity of the sidebands that occur in a peak shape on both sides of the power supply frequency, making it difficult to detect the sidebands. This invention was made to solve the above-mentioned problems, and aims to provide a motor diagnostic device that can diagnose the presence or absence of abnormalities in a motor by detecting sidebands that occur in a peak shape on both sides near the power supply frequency, even in a motor where the load torque fluctuates. The motor diagnostic device according to this invention is characterized by comprising: a current input unit that detects and inputs the current of the motor; an FFT analysis unit that analyzes the power spectrum of the current when the current from the current input unit is in a stable state; a peak detection calculation unit that detects the peak location of the power spectrum obtained by the FFT analysis unit; an averaging calculation unit that averages multiple power spectra analyzed by the FFT analysis unit; a sideband extraction unit that extracts sidebands from the power spectrum averaged by the averaging calculation unit; and an alarm output unit that outputs an alarm when a sideband with a signal strength exceeding a set value is extracted by the sideband extraction unit. According to the present invention, there are provided a current input unit for detecting and inputting the current of an electric motor, an FFT analysis unit for analyzing the power spectrum of the current when the current from the current input unit is in a stable state, a peak detection arithmetic unit for detecting peak points of the power spectrum obtained by the FFT analysis unit, an averaging arithmetic unit for averaging a plurality of power spectra analyzed by the FFT analysis unit, a sideband extraction unit for extracting sidebands of the power spectrum averaged by the averaging arithmetic unit, and an alarm output unit for performing an alarm output when a sideband having a signal intensity equal to or higher than a set value is extracted by the sideband extraction unit. Therefore, by performing the analysis of the power spectrum by the FFT analysis unit when the current is in a stable state, peak points generated on both sides of the power frequency can be surely detected. Further, by averaging a plurality of power spectra by the averaging arithmetic unit, for example, the signal intensity of peak points mixed in the power spectrum due to noise or the like is reduced, so that sidebands can be more surely extracted by the sideband extraction unit. Even in an electric motor in which the load torque fluctuates, by detecting sidebands that peak on both sides near the power frequency, it is possible to obtain an effect of obtaining a diagnostic device for an electric motor that can diagnose the presence or absence of an abnormality in the electric motor. It is a schematic configuration diagram showing the installation situation of the diagnostic device for an electric motor in Embodiment 1 of the present invention. It is a block diagram showing the configuration of the logical operation unit of the diagnostic device for an electric motor in Embodiment 1 of the present invention. It is an explanatory diagram explaining the frequency analysis result when the load fluctuation of the electric motor in the diagnostic device for an electric motor in Embodiment 1 of the present invention is large. It is an explanatory diagram explaining the conversion of the frequency axis of the diagnostic device for an electric motor in Embodiment 1 of the present invention. It is a flowchart explaining the operation of the diagnostic device for an electric motor in Embodiment 1 of the present invention. It is a schematic configuration diagram showing the installation situation of the diagnostic device for an electric motor in Embodiment 2 of the present invention. It is a block diagram showing the configuration of the logical operation unit of the diagnostic device for an electric motor in Embodiment 2 of the present invention. It is an explanatory diagram explaining the trend analysis of the diagnostic device for an electric motor in Embodiment 2 of the present invention. It is an explanatory diagram explaining the setting of the threshold value of the diagnostic device for an electric motor in Embodiment 1 of the present invention. Hereinafter, embodiments of the present invention will be described, and the same or corresponding parts in each figure will be denoted by the same reference numerals and described. Embodiment 1. Figure 1 is a schematic diagram showing the installation of the motor diagnostic device in Embodiment 1 of this invention; Figure 2 is a block diagram showing the configuration of the logic unit of the motor diagnostic device in Embodiment 1 of this invention; Figure 3 is an explanatory diagram illustrating the frequency analysis results of the motor diagnostic device in Embodiment 1 of this invention when the motor load fluctuation is large; Figure 4 is an explanatory diagram illustrating the conversion of the frequency axis of the motor diagnostic device in Embodiment 1 of this invention; Figure 5 is a flowchart illustrating the operation of the motor diagnostic device in Embodiment 1 of this invention; and Figure 9 is an explanatory diagram illustrating the setting of the threshold value of the motor diagnostic device in Embodiment 1 of this invention. In Figure 1, the main circuit 1, which is drawn in from the power grid, is equipped with a circuit breaker 2, an electromagnetic contactor 3, and a current detector 4 such as an instrument transformer that detects the load current of one phase of the three-phase main circuit 1. An electric motor 5, such as a three-phase induction motor, which is the load, is connected to the main circuit 1, and the machinery 6 is driven by the electric motor 5. The electric motor diagnostic device 7 is equipped with a rated information input unit 8 in which the power frequency, rated output, rated voltage, rated current, number of poles, rated rotational speed, etc. of the electric motor 5 are pre-inputted, and a rated information storage unit 9 in which the rated information input from the rated information input unit 8 is stored. The rated information is easily obtainable by looking at the manufacturer's catalog of the electric motor 5 or the nameplate attached to the electric motor 5. If there are multiple electric motors 5 to be diagnosed, the rated information of all electric motors 5 to be diagnosed is pre-inputted, but in the following explanation, only one electric motor 5 will be described. Furthermore, the motor diagnostic device 7 includes a current input unit 10 for inputting the current detected by the current detector 4, a logic operation unit 11 for diagnosing whether or not there is an abnormality in the motor 5 using the current input from the current input unit 10, and an alarm output unit 21 that outputs an alarm, such as by lighting an alarm or an abnormality lamp, when an abnormality is detected by the logic operation unit 11. The configuration of the logic operation unit 11 will be explained based on Figure 2. The logic operation unit 11 includes a current fluctuation calculation unit 12 that determines whether or not there is a fluctuation in the current input from the current input unit 10, an FFT analysis interval determination unit 13 that uses the results obtained by the current fluctuation calculation unit 12 to extract a stable current interval and determine a power spectrum analysis interval, an FFT analysis unit 14 that performs power spectrum analysis using the current in the interval determined by the FFT analysis interval determination unit 13, a peak detection calculation unit 15 that detects peak locations included in the power spectrum analyzed by the FFT analysis unit 14, and a peak location attributable to the rotation frequency that is obtained from the peak locations detected by the peak detection calculation unit 15. The system consists of a rotation frequency band determination unit 16, a frequency axis conversion calculation unit 17 that matches the rotation frequency band frequencies of multiple power spectra, an averaging calculation unit 18 that averages the multiple power spectra whose frequency axes have been converted by the frequency axis conversion calculation unit 17, a sideband extraction unit 19 that uses the power spectra averaged by the averaging calculation unit 18 to extract whether there are peak locations on both sides of the power supply frequency other than the rotation frequency band (hereinafter, these peak locations will be referred to as sidebands), and a sideband determination unit 20 that determines whether the signal strength of the sidebands is above a set value when sidebands are extracted by the sideband extraction unit 19. The current fluctuation calculation unit 12 calculates the statistical variation of the current value based on the current from the current input unit 10. Methods for calculating variation include, for example, standard deviation and Mahalanobis distance. The FFT analysis interval determination unit 13 determines the power spectrum analysis interval by extracting only the current intervals where the current value is stable and the variation is below a threshold, based on the statistical variation of the current value obtained by the current fluctuation calculation unit 12. Generally, when the load torque of the electric motor 5 fluctuates, the current value varies, and if power spectrum analysis is performed on a current waveform with large variation, the signal intensity on both sides near the power supply frequency increases, as shown in Figure 3, and peaks such as sidebands disappear. A threshold is set in the FFT analysis interval determination unit 13 to prevent this. The FFT analysis unit 14 calculates the current power spectrum intensity by performing frequency analysis using the current waveform input to the section determined by the FFT analysis section determination unit 13. By performing power spectrum analysis with a current waveform where the current value is stable, the power spectrum intensity does not increase on both sides near the power supply frequency, and any peaks will appear reliably. The peak detection calculation unit 15 detects peaks due to power supply frequency, rotation frequency, sidebands, and other peaks from the analysis results of the current power spectrum intensity. Peaks can be detected by extracting the portion where the steep slope of the result calculated by first, second, and third-order differential calculations reverses. By performing differential calculations up to the third order, it becomes possible to detect peaks with smaller signal intensities. Peaks due to power supply frequency occur at the power supply frequency (generally 50 Hz or 60 Hz) stored in the rated information storage unit 9, so they can be easily confirmed. The rotation frequency band determination unit 16 determines the rotation frequency from the rated rotation speed stored in the rated information storage unit 9, and extracts peak locations with similar signal strength located near positions shifted by the rotation frequency on both sides of the power supply frequency. Generally, the motor 5 experiences slippage depending on the load torque, causing a shift in rotation speed, so the peak locations caused by the rotation frequency also appear shifted accordingly. The rotation frequency band determination unit 16 extracts peak locations within the frequency band that takes this shift into account and determines them as the rotation frequency band. The frequency axis conversion calculation unit 17 is necessary to correctly perform the averaging calculation performed by the averaging calculation unit 18. Generally, the location of sidebands generated by an abnormality in the motor 5 is closely related to the rotation frequency band, and the frequency band of the sidebands is often a multiple of the rotation frequency band. Furthermore, as explained above, the rotation frequency band shifts depending on the load torque of the motor 5. For this reason, it is necessary to align the frequency axes of multiple power spectrum analysis results to be averaged using a peak-tracking method. Specifically, as shown in Figure 4, if the rotation frequency band is at a position fr away from the power supply frequency and the sideband frequency is at a position fb away from the power supply frequency, and the rotation frequency band when the motor 5 is unloaded is at a position fr' away from the power supply frequency, then the conversion rate α is α = fr' / fr, and the position fb' of the sidebands under no load can be obtained as fb' = α・fb. In this way, the frequency axes of all peak locations are converted by multiplying the rotation frequency band by the conversion rate α. Although the above explanation described the case where the frequency axis is aligned with the no-load state, the frequency axis conversion calculation unit 17 can be configured to align the frequency axis of multiple power spectrum analysis results to be averaged with a predetermined load, for example, by aligning the frequency axis with the rated load state. The averaging calculation unit 18 averages the results of multiple power spectrum analyses whose frequency axes have been aligned by the frequency axis conversion calculation unit 17. By averaging, the base noise can be reduced and the signal-to-noise ratio at peak locations can be improved. Specifically, if the results of 10 power spectrum analyses are averaged, the signal strength of peak locations caused by noise, etc., which occurred in only one analysis, will be reduced to one-tenth. On the other hand, in the case of rotation frequency bands or sidebands, peak locations occur in all 10 analyses, and since the frequency axis has been converted using the peak-following method and the frequencies have been aligned, the signal strength at the peak locations will not change even after averaging. Note that the above explanation describes the case of averaging 10 power spectrum analysis results, but it is not limited to 10 times; multiple analyses can be averaged. The sideband extraction unit 19 extracts sidebands from the power spectrum analysis results averaged by the averaging calculation unit 18, specifically peak locations that occur on both sides of the power supply frequency at the same frequency shift. Candidate sidebands are selected from the peak locations obtained by the peak detection calculation unit 15. If a peak location occurs on only one side of the power supply frequency, it is determined not to be a sideband and is not extracted. The sideband detection unit 20 determines whether the motor 5 is malfunctioning based on the number of sidebands and signal strength extracted by the sideband extraction unit 19. If it determines that the motor 5 is malfunctioning, it outputs an alarm from the alarm output unit 21. Next, the operation will be explained based on Figure 5. The motor diagnostic device 7 is started at predetermined time intervals and performs the following processes. In step 101, the current of the motor 5 detected by the current detector 4 is input to the current input unit 10. In step 102, the current fluctuation calculation unit 12 calculates the variation in the effective value (hereinafter referred to as the current value) of the current input from the current input unit 10, and the FFT analysis interval determination unit 13 uses the calculation result to determine whether the current is in a stable state. If the determination result indicates an unstable state (NO) where the variation in the current value is above a preset threshold, the process returns to step 101 and is repeated until the current is in a stable state. If the current is in a stable state (YES), the process proceeds to step 103. Regarding the threshold, for example, field data of multiple motors is acquired in advance, and a range with a small variation value is selected from the current variation value (standard deviation) of that data, and the selected value is used as the threshold. As a concrete example of the calculation, as shown in Figure 9, the variation value is calculated 50 times and sorted in ascending order. The fifth smallest variation value, 0.8, is determined as the threshold value. Alternatively, instead of using prior field data, a certain learning period may be set in the motor 5, and the same calculation may be performed using the current variation value (standard deviation) acquired during the learning period. In step 103, the FFT analysis unit 14 uses the current waveform in the stable current value range to perform frequency analysis between 0 Hz and 120 Hz (twice the power supply frequency of 60 Hz), and passes the power spectrum analysis result to the peak detection calculation unit 15. In step 104, the peak detection calculation unit 15 detects all peak locations included in the power spectrum analysis result. In step 105, the rotation frequency band determination unit 16 extracts the peak locations that are in the rotation frequency band from the detected peak locations and determines the rotation frequency band. In step 106, the frequency axis conversion calculation unit 17 converts the frequency axis of all peak locations so that the detected rotation frequency band becomes the rotation frequency band under no load. In step 107, the operations from step 101 to step 106 are repeated 10 times to collect 10 power spectrum analysis results with converted frequency axes. In step 108, the averaging calculation unit 18 averages the 10 power spectrum analysis results that have been collected. In step 109, the sideband extraction unit 19 extracts sidebands by focusing on the peak locations of the averaged power spectrum analysis results. In step 110, if the sideband extraction unit 19 does not extract sidebands, or if sidebands are extracted but the signal intensity is smaller than the set value, the sideband determination unit 20 determines that there is no abnormality in the motor 5 (NO) and terminates the diagnostic process. On the other hand, if the signal intensity of the sideband extracted by the sideband extraction unit 19 is greater than the set value (YES), the unit determines that there is an abnormality in the motor 5 and sends a signal to the alarm output unit 21, and in step 111, the alarm output unit 21 outputs an alarm and terminates the diagnostic process. The set value for the sideband determination unit 20 is determined by learning the signal intensity A of the sideband under normal conditions, calculating the standard deviation σ, and setting the range in which 99.7% of the data exists where the detected sideband peak value is A + 3σ. Alternatively, a safety factor α (for example, 2 or more) can be applied, resulting in A + 3σ × α. Another way to determine the above setting value is to use data from past failures of similar motors. As the number of failure cases increases, the sidebands will allow for more accurate determination of the failure location and severity. As explained above, by performing power spectrum analysis on the current waveform when the current value is stable, peaks such as sidebands can be reliably identified. Furthermore, by performing averaging, noise and other unwanted signals are reduced, enabling more accurate fault diagnosis. Embodiment 2. Figure 6 is a schematic diagram showing the installation of the motor diagnostic device in Embodiment 2 of this invention, Figure 7 is a block diagram showing the configuration of the logic unit of the motor diagnostic device in Embodiment 2 of this invention, and Figure 8 is an explanatory diagram illustrating the trend analysis of the motor diagnostic device in Embodiment 2 of this invention. Embodiment 1 described a case in which an abnormality of the motor 5 is diagnosed using the averaged power spectrum analysis results, but Embodiment 2 describes a case in which the averaged power spectrum analysis results are saved in time series and trend monitoring is performed. In Figure 6, the motor diagnostic device 7 is equipped with an FFT information storage unit 22 that stores the averaged power spectrum analysis results in a time series, and as shown in Figure 7, a trend analysis unit 23 is provided in the logical operation unit 11. Next, the operation of the trend analysis unit 23 will be explained. The trend analysis unit 23 focuses on sidebands of specific frequencies in the power spectrum analysis results stored in time series in the FFT information storage unit 22, and calculates and displays the signal intensity of those sidebands in time series as shown in Figure 8. The set values ​​in Figure 8 are the set values ​​used in the sideband determination unit 20. By performing trend analysis in this way, for example, if the sidebands were caused by wear of the bearings of the electric motor 5, a small signal intensity of the sidebands indicates that the degree of bearing wear is small and does not constitute an abnormality. However, since bearing wear gradually increases, the signal intensity of the sidebands also gradually increases in time series as shown in Figure 8. Therefore, the time when the signal intensity of the sidebands reaches the set value and bearing replacement becomes necessary can be determined by performing trend analysis. Note that the other parts are the same as in Embodiment 1, so their explanation will be omitted. Furthermore, within the scope of this invention, it is possible to freely combine each embodiment, or to modify or omit embodiments as appropriate. 1. Main circuit, 2. Circuit breaker, 3. Electromagnetic contactor, 4. Current detector, 5. Electric motor, 6. Mechanical equipment, 7. Electric motor diagnostic device, 8. Rated information input unit, 9. Rated information storage unit, 10. Current input unit, 11. Logic operation unit, 12. Current fluctuation calculation unit, 13. FFT analysis interval determination unit, 14. FFT analysis unit, 15. Peak detection calculation unit, 16. Rotation frequency band determination unit, 17. Frequency axis conversion calculation unit, 18. Averaging calculation unit, 19. Sideband extraction unit, 20. Sideband determination unit, 21. Alarm output unit, 22. FFT information storage unit, 23. Trend analysis unit

Claims

------July 26, 2016------(OCR) Page 1 of 1 page Claims 1. Integrated electric motor diagnostic device comprising: a current input unit for detecting and inputting electric motor current; an FFT analysis unit for, when the current from the current input unit is in a stable state, analyzing the power spectrum of the current; a peak detection calculation unit for detecting peaks in the power spectrum obtained by the FFT analysis unit; an averaging calculation unit for averaging a number of power spectra analyzed by the FFT analysis unit; a sideband extraction unit for extracting various sideband waves from the power spectrum averaged by the averaging calculation unit;And an alarm output unit for outputting alarm signals when sideband waves with different signal strengths equal to or greater than the set value are extracted by the sideband extraction unit.

2. Electric motor diagnostic device according to claim 1, where the averaging unit performs averaging in a state in which the various frequency axes of a number of power spectrums have been transformed in that state in order to align the peaks due to the various rotating bands.

3. Electric motor diagnostic device according to claim 1 or 2, where information on the power spectrum averaged by the averaging unit is stored and the sideband waves are based on trend observation. ------------Revised 27 / 11 / 2018 1. Electric motor diagnostic device comprising: Current input unit for the detection and current input of electric motors;The FFT analysis timing unit calculates the variation of the pre-generated current variation data, based on the standard deviation or Mahalanobis distance, and determines that a stable state is achieved when the current variation from the input current unit within such a time period does not exceed a threshold value chosen to enable the detection of sideband waves in the power spectrum; the FFT analysis unit is for analyzing the power spectrum that has been set to a stable state by the FFT analysis timing unit; the peak detection calculation unit is for detecting peaks in the power spectrum generated by the FFT analysis unit; the averaging calculation unit is for averaging a number of power spectra analyzed by the FFT analysis unit; and the sideband extraction unit is for extracting sideband waves from the power spectrum averaged by the averaging calculation unit.And the alarm output unit for outputting alarm signals when sideband waves with different signal strengths equal to or greater than the set value are intercepted by the sideband interception unit-----------------------------------------------;