A method for extracting vibration signals from AC motor current

Abstract

This application discloses a method for extracting vibration signals from AC motor current, enabling the effective extraction of vibration signals from the current of AC motor equipment during operation, thereby supporting intelligent diagnosis of equipment vibration faults. The method involves: acquiring 1 second of current data from the AC motor equipment; dividing this 1 second current monitoring data into segments according to each sine wave cycle in the time domain; generating a standard sine wave data sequence with a root mean square and the same number of samples as the monitoring data within each sine wave cycle; calculating the relative deviation between each monitoring data point and the generated sine wave data according to the sequence; setting the deviation to 0 if it is less than a certain threshold, such as 10%; retaining the deviation data with positive and negative signs if it is greater than the threshold; and finally, after processing the current monitoring data for all cycles, the resulting deviation data sequence is clean vibration wave data, which can be directly used for intelligent vibration diagnosis of equipment faults.

Description

Technical Field

[0001] This application relates to a method for extracting vibration signals from AC motor current, enabling the effective acquisition of extremely weak but pure and complete vibration signals from current monitoring data during normal operation of AC motor equipment. This supports intelligent diagnosis of equipment vibration faults. The method involves: acquiring 1 second of current data from the AC motor equipment; dividing this 1 second current monitoring data into segments according to each sine wave cycle in the time domain; generating a standard sine wave data sequence with a root mean square and the same number of samples as the monitoring data within each sine wave cycle; calculating the relative deviation between each monitoring data point and the generated sine wave data according to the sequence; setting the deviation to 0 if it is less than a certain threshold, such as 10%; retaining the deviation data with positive and negative signs if it is greater than the threshold; and finally, after processing the current monitoring data for all cycles, the resulting deviation data sequence is clean vibration wave data, which can be directly used for intelligent vibration diagnosis of equipment faults. Background Technology

[0002] During normal operation, AC motor equipment often contains extremely weak current signals, i.e., vibration waves, that reflect the vibration characteristics of the equipment. This signal is of great value for intelligent fault diagnosis. However, because the vibration signal is too weak compared to the AC main frequency signal, it is difficult to extract using traditional spectrum analysis methods. This is mainly because, firstly, the extremely weak signal in the time domain is easily submerged by current frequency harmonic noise in the frequency domain; secondly, Fourier transform in the time-to-frequency domain transformation results in information loss for discrete data sequences, further weakening the effective signal in the time domain. Therefore, it is necessary to extract the vibration signal directly in the time domain. Extracting pure and complete vibration signals from AC motor equipment allows for direct intelligent vibration diagnosis of AC motor equipment faults, eliminating the cost of directly installing vibration monitoring instruments on the equipment to collect vibration data. Therefore, developing data processing methods for extracting vibration signals from AC motor equipment is particularly important. Summary of the Invention

[0003] This application provides a method for extracting vibration signals from AC motor current, enabling the effective acquisition of extremely weak but pure and complete vibration signals from the current monitoring data of the AC motor equipment during normal operation, supporting intelligent diagnosis of equipment vibration faults. The method involves: acquiring 1 second of current data from the AC motor equipment; dividing this 1 second of current monitoring data into segments according to each sine wave cycle in the time domain; generating a standard sine wave data sequence with a root mean square and the same number of samples as the monitoring data within each sine wave cycle; calculating the relative deviation between each monitoring data point and the generated sine wave data according to the sequence; setting the deviation to 0 if it is less than a certain threshold, such as 10%; retaining the deviation data with positive and negative signs if it is greater than the threshold; and finally, after processing the current monitoring data for all cycles, the resulting deviation data sequence is clean vibration wave data, which can be directly used for intelligent vibration diagnosis of equipment faults.

[0004] Firstly, this is a data processing technology that can effectively collect extremely weak but pure and complete vibration signals from equipment current monitoring data, see... Figure 1 This is one of the claims of this invention. Because the vibration signal in conventional motor current signals is extremely weak, ranging from a few thousandths to a few ten-thousandths of the AC current, it has been consistently ignored by mainstream industry applications. This invention achieves the extraction of pure and complete vibration signals from AC motor equipment, which can be directly used for intelligent vibration diagnosis of AC motor equipment faults, saving the cost of directly installing vibration monitoring instruments on the equipment to collect vibration data. During normal operation, the various mechanical parts of the motor equipment generate mechanical vibration signals related to the equipment structure due to the electric rotor's operation. These mechanical vibration signals, in turn, affect the AC current value by influencing the subtle spatial changes in the cutting magnetic lines of force in the motor brushes. Therefore, this signal is very weak and has extremely low resolution, requiring direct extraction from the current time domain signal. Extraction; This technology differs from the current fluctuation signal identification and diagnosis technology generated by motor equipment during power-on and power-off, poor circuit contact, and short circuit accidents. The difference lies in the fact that the current wave signal intensity during motor power-on and power-off, poor circuit contact, and short circuit accidents is very obvious, generally exceeding 5% of the normal current value, and sometimes reaching about 100%. This technology also differs from the current fluctuation signal identification and diagnosis technology generated by motor equipment during large changes in workload. For example, when machine tool tools are operating and idling, the motor equipment will generate large current fluctuation changes. The difference lies in the fact that the current fluctuation signal intensity caused by large changes in the workload of the motor equipment is still very obvious, generally exceeding 5% of the normal current value. The current fluctuation frequency reflects a combination of factors such as load change characteristics and tool material characteristics.

[0005] Secondly, this invention collects current data from an AC motor device for 1 second. This is a standard operation in sampling technology; to complete the technical verification, at least 50 cycles of current data must be collected for AC signals.

[0006] Thirdly, this invention divides the 1-second current monitoring data in the time domain into each sine wave cycle; starting from the first current value of 0, the third current value of 0 is the first sine wave cycle; the third to fifth current values ​​of 0 are the second sine wave cycle, and so on, until the 1-second data is processed. The number of samples n in different cycle segments may vary slightly. This is a routine operation in time domain signal processing technology and a necessary technical route for vibration wave extraction. Since the vibration signal is too weak compared to the AC main frequency signal, it is difficult to extract using traditional spectrum analysis methods. This is mainly because, on the one hand, extremely weak signals in the time domain are easily submerged by current frequency harmonic noise signals in the frequency domain; on the other hand, Fourier transform will cause information loss for discrete data sequences during the transformation from the time domain to the frequency domain, which will further weaken the effective signal in the time domain. Therefore, it is necessary to extract the vibration signal directly in the time domain.

[0007] Fourthly, this invention generates a root mean square within each sine wave period. A standard sine wave data sequence with the same sample size n as the monitoring data. This operation simulates and monitors a 50Hz standard AC signal with the same amplitude as the current.

[0008] Fifthly, the present invention calculates each monitoring data point sequentially. With generating sine wave data relative deviation between Choose a deviation threshold, for example, a threshold of 10%. Less than 10% is considered negligible and is set to 0. A deviation greater than 10% is considered a valid signal, and the deviation data, including both positive and negative signs, is retained. This serves as an effective signal for the vibration wave. The operation involves removing the 50Hz standard AC signal and white noise from the monitored current, leaving only the pure vibration signal.

[0009] Sixthly, after processing the current monitoring data for all cycles, the present invention finally forms a deviation data sequence. This is clean vibration wave data, whose spectral characteristics are completely consistent with the fault spectral characteristics of vibration detection. Therefore, it can be directly used for intelligent vibration diagnosis of equipment faults. Attached Figure Description

[0010] Figure 1 This is a method logic diagram of an embodiment of this application.

Claims

1. A method of extracting a vibration signal from an alternating current motor current, characterized by, The method comprises the following steps: (1) collecting current time domain data of an alternating current motor device during operation; (2) dividing the current time domain data according to the period of a sine wave in the time domain to obtain a plurality of continuous single-period current data segments; (3) for each single-period current data segment, the following operations are performed: (3.1) Calculate the root mean square value of the current of all sampling points in the data segment wherein is the number of sampling points of the data segment, is the current value of the sampling point; (3.2) generating a standard sinusoidal data sequence } wherein ; (3.3) calculating each sample value of the single-cycle current data segment point by point the difference between the generated corresponding standard sine wave value = ;​ (3.4) Select a deviation threshold, if the absolute value is less than the deviation threshold, it is considered negligible, set to 0, if the absolute value is greater than the deviation threshold, it is considered valid signal, keep the deviation data with positive and negative signs as the effective signal of the vibration wave; (4) Calculate the difference sequence obtained from all periods { The data are merged to form a complete deviation data sequence. (5) based on the complete deviation data sequence, vibration fault diagnosis is performed on the alternating current motor device.

Images