Method and system for evaluating performance of motor based on electric signal spectrum analysis

Abstract

The application provides a motor performance evaluation method and system based on electric signal spectrum analysis, and relates to the technical field of motor evaluation. The application collects stator current signals under the transient working condition of motor start-stop or load mutation, synchronously acquires vibration acceleration and high-frequency acoustic emission signals through a sound and vibration fusion sensor, and performs amplification and anti-aliasing filtering processing through an analog conditioning circuit. The three are subjected to joint time-frequency transformation, the instantaneous frequency trajectory and energy distribution atlas representing electromagnetic and mechanical coupling characteristics are extracted, and then coupling characteristic data is generated. In a preset motor transient response model, the change process of the characteristic data during the transient period is analyzed to determine the dynamic matching degree of electromagnetic force and mechanical inertia, and the performance stability and dynamic response capability of the motor are quantitatively evaluated according to the dynamic matching degree, so that a transient performance evaluation result is obtained, and the precise evaluation of the electromagnetic and mechanical coupling state of the motor under the transient working condition can be realized.

Description

Technical Field

[0001] This application relates to the field of motor evaluation technology, and in particular to a method and system for evaluating motor performance based on electrical signal spectrum analysis. Background Technology

[0002] As a core power device in the industrial field, the performance of electric motors directly affects production efficiency and energy consumption. Therefore, accurate evaluation of motor performance is of great significance. Under transient operating conditions such as motor start-up, stopping, or sudden load changes, its dynamic response capability and operational stability are key indicators for measuring the overall performance of the motor. Related evaluation technologies are widely used in high-end manufacturing, new energy vehicles, and aerospace fields.

[0003] Existing technologies include a motor performance evaluation method based on electrical signal spectrum analysis. This method acquires the motor's stator current signal, extracts the current spectrum characteristics using techniques such as Fast Fourier Transform, and combines this with indicators such as harmonic distortion rate and three-phase imbalance to evaluate the motor's performance. This type of method primarily focuses on analyzing the motor's operating state from the perspective of electrical parameters, reflecting some characteristics related to stator windings and power supply quality.

[0004] However, the aforementioned evaluation methods relying solely on a single electrical signal are insufficient to comprehensively capture the complex dynamic behavior of a motor during transient processes. This is particularly true when electromagnetic force interacts with mechanical inertia, as the coupling state reflected by physical quantities such as mechanical vibration and acoustic emission cannot be directly obtained through current signals. The lack of simultaneous observation methods for the electromagnetic-mechanical coupling state makes it difficult for existing technologies to accurately assess the degree of electromechanical dynamic matching within the motor under transient operating conditions. Therefore, existing technologies suffer from the technical problem of difficulty in simultaneously acquiring and accurately evaluating the electromagnetic-mechanical coupling state under transient operating conditions. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for evaluating motor performance based on electrical signal spectrum analysis, so as to solve the problem that it is difficult to synchronously acquire and accurately evaluate the electromagnetic and mechanical coupling state under transient operating conditions in the prior art.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a method for evaluating motor performance based on electrical signal spectrum analysis, comprising: During transient operating conditions such as motor start-up, shutdown, or sudden load changes, one stator current signal of the motor is collected. By using an acoustic-vibration fusion sensor installed on the motor bearing, the vibration acceleration signal and high-frequency acoustic emission signal corresponding to the stator current signal are acquired, and the vibration acceleration signal and the high-frequency acoustic emission signal are amplified and anti-aliasing filtered by the analog conditioning circuit in the acoustic-vibration fusion sensor. By performing a joint time-frequency transformation on the stator current signal, the processed vibration acceleration signal, and the processed high-frequency acoustic emission signal, an instantaneous frequency trajectory and energy distribution spectrum characterizing the electromagnetic and mechanical coupling characteristics of the motor are obtained. Based on the instantaneous frequency trajectory and the energy distribution map, coupling feature data is generated, which is used to reflect the electromagnetic and mechanical coupling state of the motor under the transient operating condition. In the preset transient response model of the motor, the change process of the coupling characteristic data during the transient operating condition is analyzed, and the dynamic matching degree of the electromagnetic force and mechanical inertia of the motor under the transient operating condition is determined based on the change process. Based on the dynamic matching degree, the performance stability and dynamic response capability of the motor under the transient operating conditions are quantitatively evaluated to obtain the transient performance evaluation results of the motor.

[0007] Optionally, the step of performing a joint time-frequency transformation on the stator current signal, the processed vibration acceleration signal, and the processed high-frequency acoustic emission signal to obtain the instantaneous frequency trajectory and energy distribution spectrum characterizing the electromagnetic and mechanical coupling characteristics of the motor includes: The stator current signal, the processed vibration acceleration signal, and the processed high-frequency acoustic emission signal are synchronously organized according to the acquisition time axis to form three discrete sequences. Hilbert-Huang transforms are performed on the three sets of discrete sequences respectively to extract the instantaneous frequency ridges of each sequence, and the corresponding time-frequency energy spectra are generated by short-time Fourier transform. In the time-frequency energy spectrum, an energy threshold is set, and continuous regions with energy values ​​higher than the energy threshold are extracted as candidate event segments; The instantaneous frequency ridges of the stator current signal and the processed vibration acceleration signal are cross-correlated on the time axis to calculate the correlation coefficient curve between the two. The continuous time intervals in the correlation coefficient curve that exceed the preset correlation coefficient threshold are logically ANDed with the candidate event segment to obtain the associated event segment. Using the instantaneous frequency ridge of the processed high-frequency acoustic emission signal, the start and end boundaries of the waveform are calibrated within the associated event segment using a peak detection algorithm; Based on the associated event segment and the starting and ending boundaries, the energy centroid trajectories of the three sets of signals in the time-frequency energy spectrum are extracted. The projection of the energy centroid trajectory on the time axis is taken as the instantaneous frequency trajectory, and the energy diffusion pattern of the energy centroid trajectory on the frequency-time plane is taken as the energy distribution map.

[0008] Optionally, the step of using the analog conditioning circuit within the acoustic-vibration fusion sensor to amplify and anti-aliasing the vibration acceleration signal and the high-frequency acoustic emission signal includes: The analog conditioning circuit within the acoustic-vibration fusion sensor receives the vibration acceleration signal from the vibration-sensitive element and the high-frequency acoustic emission signal from the acoustic emission-sensitive element, respectively. The vibration acceleration signal and the high-frequency acoustic emission signal are respectively input to the AC coupling path in the analog conditioning circuit to filter out the DC component in their respective signals; The vibration acceleration signal and the high-frequency acoustic emission signal after AC coupling are respectively sent to the variable gain amplifier in the analog conditioning circuit. According to the original amplitude range of the vibration acceleration signal and the high-frequency acoustic emission signal, the amplification factor of the variable gain amplifier is adjusted so that the amplitude of the two amplified signals falls within the preset amplitude window range. The amplified two signals are respectively input to the low-pass filter in the analog conditioning circuit. The frequency components in the two signals that are higher than the cutoff frequency are filtered out by the cutoff frequency set by the low-pass filter. The two signals processed by the low-pass filter are output separately as the processed vibration acceleration signal and the processed high-frequency acoustic emission signal.

[0009] Optionally, generating coupled feature data based on the instantaneous frequency trajectory and the energy distribution map includes: The instantaneous frequency trajectory is subjected to a first-order difference operation to identify the rising segment and the falling segment where the rate of change of the frequency value exceeds a preset rate of change threshold, which are respectively used as the electromagnetic force action range and the mechanical inertia response range. In the energy distribution spectrum, the energy centroid frequency at each time point is calculated, and an energy centroid frequency curve is generated that varies with time. The first moment centroid time of the energy centroid frequency curve within the electromagnetic force action range is extracted as the characteristic moment of the electromagnetic force, and the second moment centroid time of the energy centroid frequency curve within the mechanical inertia response range is extracted as the characteristic moment of the mechanical response. Calculate the time difference sequence between the characteristic time of the electromagnetic force and the characteristic time of the mechanical response, and calculate the proportional relationship between the time difference sequence and the length of the electromagnetic force action interval and the length of the mechanical inertia response interval; The time difference sequence and the proportional relationship are vectorized and combined to generate a multidimensional time deviation vector, and the multidimensional time deviation vector is used as the coupling feature data. The multidimensional time deviation vector is used to characterize the dynamic matching relationship between the electromagnetic force action time sequence and the mechanical inertia response time sequence at the energy centroid level.

[0010] Optionally, the step of analyzing the change process of the coupling characteristic data during the transient operating condition in a preset motor transient response model, and determining the dynamic matching degree of the motor electromagnetic force and mechanical inertia under the transient operating condition based on the change process, includes: The coupling characteristic data is input into the preset motor transient response model, which includes a first reference time interval characterizing the electromagnetic force change trend and a second reference time interval characterizing the mechanical inertia response trend. In the motor transient response model, the multidimensional timing deviation vector in the coupled feature data is compared with the first reference timing interval and the second reference timing interval respectively, and the first matching state of the timing deviation falling into the first reference timing interval, the second matching state falling into the second reference timing interval, and the third matching state that deviates from both the first reference timing interval and the second reference timing interval are identified. The timing deviation is statistically analyzed to determine the first cumulative number of times the first matching state occurs, the second cumulative number of times the second matching state occurs, and the third cumulative number of times the third matching state occurs during the transient operating condition. Based on the numerical relationship between the first cumulative count, the second cumulative count, and the third cumulative count, the synchronization degree between the electromagnetic force action sequence and the mechanical inertia response sequence under the transient operating condition is determined, and the synchronization degree is used as the dynamic matching degree.

[0011] Optionally, the step of quantitatively evaluating the performance stability and dynamic response capability of the motor under the transient operating condition based on the dynamic matching degree to obtain the transient performance evaluation result of the motor includes: Obtain the synchronization degree value in the dynamic matching degree, and compare the synchronization degree value with a preset stability threshold and a preset response threshold respectively; When the synchronization degree value is greater than or equal to the stability threshold, the performance stability of the motor is rated as the first stability level; when the synchronization degree value is less than the stability threshold, the performance stability is rated as the second stability level. When the synchronization degree value is greater than or equal to the response threshold, the dynamic response capability of the motor is rated as the first response level; when the synchronization degree value is less than the response threshold, the dynamic response capability is rated as the second response level. The first stability level or the second stability level is combined with the first response level or the second response level to form a combined identifier that includes stability level information and response level information, and the combined identifier is output as the transient performance evaluation result.

[0012] Optionally, before performing joint time-frequency transformation on the stator current signal, the processed vibration acceleration signal, and the processed high-frequency acoustic emission signal, the method further includes: The waveform correspondence between the processed vibration acceleration signal and the processed high-frequency acoustic emission signal in the time domain is obtained. Based on the waveform correspondence, the transient segment affected by external forces is identified from the processed vibration acceleration signal, and the accompanying segment that coincides with the transient segment in time is identified from the processed high-frequency acoustic emission signal. The accompanying segment is distinguished from the processed high-frequency acoustic emission signal to obtain the distinguished high-frequency acoustic emission signal; The amplitude variation profile of the differentiated high-frequency acoustic emission signal in the non-transient segment outside the transient segment is extracted, and the correlation profile of the amplitude variation is analyzed with the waveform of the processed vibration acceleration signal in the non-transient segment to obtain the correlation value; When the correlation value is lower than a preset threshold, the waveforms corresponding to the non-transient segment in the processed vibration acceleration signal and the distinguished high-frequency acoustic emission signal are jointly marked as ineffective segments, and the stator current signal, the processed vibration acceleration signal and the distinguished high-frequency acoustic emission signal in the time interval corresponding to the ineffective segment are removed in the joint time-frequency transformation.

[0013] Secondly, this application provides a motor performance evaluation system based on electrical signal spectrum analysis, comprising: The acquisition module is used to acquire one stator current signal of the motor during transient operating conditions such as motor start-up, shutdown, or sudden load changes. The processing module is used to acquire the vibration acceleration signal and high-frequency acoustic emission signal corresponding to the stator current signal through the acoustic vibration fusion sensor installed on the motor bearing, and to amplify and anti-aliasing filter the vibration acceleration signal and the high-frequency acoustic emission signal using the analog conditioning circuit in the acoustic vibration fusion sensor. The transformation module is used to perform joint time-frequency transformation on the stator current signal, the processed vibration acceleration signal, and the processed high-frequency acoustic emission signal to obtain the instantaneous frequency trajectory and energy distribution spectrum characterizing the electromagnetic and mechanical coupling characteristics of the motor. The generation module is used to generate coupling feature data based on the instantaneous frequency trajectory and the energy distribution map. The coupling feature data is used to reflect the electromagnetic and mechanical coupling state of the motor under the transient operating condition. The analysis module is used to analyze the change process of the coupling characteristic data during the transient operating condition in a preset motor transient response model, and determine the dynamic matching degree of the motor electromagnetic force and mechanical inertia under the transient operating condition based on the change process. The evaluation module is used to quantitatively evaluate the performance stability and dynamic response capability of the motor under the transient operating conditions based on the dynamic matching degree, and obtain the transient performance evaluation results of the motor.

[0014] Thirdly, this application provides an electronic device, comprising: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the motor performance evaluation method based on electrical signal spectrum analysis as described in the first aspect above.

[0015] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the motor performance evaluation method based on electrical signal spectrum analysis as described in the first aspect above.

[0016] The motor performance evaluation method based on electrical signal spectrum analysis provided in this application simultaneously acquires stator current, vibration acceleration, and high-frequency acoustic emission signals under transient operating conditions, and improves signal quality through simulation conditioning, laying a reliable data foundation for subsequent multi-source information fusion. By performing joint time-frequency transformation on the three signals, the instantaneous frequency trajectory and energy distribution of electromagnetic and mechanical coupling can be intuitively presented, effectively revealing the coupling characteristics in the transient process. Based on the analysis of the changes in coupling characteristic data during the transient period using a preset model, the dynamic matching degree of electromagnetic force and mechanical inertia can be accurately reflected, thereby realizing a quantitative evaluation of the transient performance stability and dynamic response capability of the motor.

[0017] Furthermore, the current, vibration acceleration, and high-frequency acoustic emission signals are synchronously organized into three discrete sequences along the time axis. Instantaneous frequency ridges are extracted using Hilbert-Huang transform, and combined with short-time Fourier transform to generate time-frequency energy spectra. Candidate event segments are extracted by energy thresholding. Cross-correlation analysis is performed on the instantaneous frequency ridges of current and vibration acceleration. Logical AND operation is performed between the continuous intervals with correlation coefficients exceeding the threshold and the candidate event segments to obtain associated event segments. The start and end boundaries of the waveform are then marked within these segments using the instantaneous frequency ridge of the high-frequency acoustic emission signal. Finally, the energy centroid trajectories of the three signals in the time-frequency energy spectrum are extracted, and their time-axis projection is used as the instantaneous frequency trajectory, and the energy diffusion pattern is used as the energy distribution map. This step, through multi-signal synchronous organization and multi-transform fusion, can accurately separate key event segments of electromagnetic and mechanical coupling in transient processes. It uses cross-correlation analysis and logical operations to suppress non-correlated interference, and combines the high time-domain resolution of high-frequency acoustic emission signals to accurately calibrate event boundaries. This significantly improves the extraction accuracy and noise resistance of instantaneous frequency trajectories and energy distribution maps, providing more reliable time-frequency feature support for subsequent coupling feature analysis. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating a motor performance evaluation method based on electrical signal spectrum analysis provided in this application embodiment; Figure 2 A flowchart illustrating another method for evaluating motor performance based on electrical signal spectrum analysis provided in this application embodiment; Figure 3 This is a schematic diagram of a motor performance evaluation system based on electrical signal spectrum analysis, provided as an embodiment of this application. Detailed Implementation

[0020] Existing motor performance evaluation methods mostly rely on a single stator current signal to extract electrical characteristics through spectrum analysis. However, under transient conditions such as start-up, shutdown, and sudden load changes, there is a complex dynamic coupling between electromagnetic force and mechanical inertia. Current signals alone cannot simultaneously obtain mechanical response information such as vibration and acoustic emission, resulting in significant information loss in the evaluation of electromechanical coupling states during transient processes.

[0021] To address the aforementioned issues, this application proposes a method for evaluating the transient performance of a motor based on the joint analysis of electrical and acoustic / vibration signals. This method simultaneously acquires stator current signals, as well as vibration acceleration and high-frequency acoustic emission signals obtained through an acoustic / vibration fusion sensor. A joint time-frequency transformation is performed on these three types of signals to generate instantaneous frequency trajectories and energy distribution maps reflecting the electromagnetic-mechanical coupling characteristics. This allows for the analysis of the dynamic matching degree between electromagnetic force and mechanical inertia, thereby achieving a quantitative assessment of the motor's transient performance. By introducing simultaneous sensing and joint analysis of multi-source signals, this method overcomes the limitation of observing mechanical dynamic behavior with a single electrical signal, achieving comprehensive capture and accurate evaluation of the electromechanical coupling state under transient operating conditions.

[0022] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] The core of this application is to provide a method for evaluating motor performance based on electrical signal spectrum analysis, and a flowchart of one specific implementation is shown below. Figure 1 As shown, the method includes: S101. Under transient operating conditions such as motor start-up / stop or sudden load changes, acquire one stator current signal of the motor.

[0024] Transient operating conditions refer to the rapid changes in the motor's operating state, including the starting process of the motor accelerating from a standstill to its rated speed, the stopping process of decelerating from its rated speed to a stop, and the load change process of sudden increase or decrease in load torque. The stator current signal refers to the current waveform flowing through the motor's stator windings; this signal reflects the dynamic changes in the motor's electromagnetic field.

[0025] In this embodiment, a stator current signal is collected from the motor power supply line using a current sensor. The sampling frequency must satisfy the Nyquist sampling theorem to ensure that the key frequency components in the signal are completely captured, thereby obtaining the electrical response information of the motor during the transient process and providing a basic data source for subsequent electromagnetic and mechanical coupling analysis.

[0026] As an example, taking a three-phase asynchronous motor with a rated power of 5 kW as an example, during the motor startup process, the stator current signal is collected on one phase of the motor power line through a current transformer. The sampling frequency is set to 10 kHz and the sampling time is 5 seconds, covering the complete startup process of the motor from standstill to stable operation, resulting in a current timing sequence with a length of 50,000 sampling points.

[0027] S102. By using an acoustic-vibration fusion sensor installed on the motor bearing, the vibration acceleration signal and high-frequency acoustic emission signal corresponding to the stator current signal are acquired, and the vibration acceleration signal and high-frequency acoustic emission signal are amplified and anti-aliasing filtered by the analog conditioning circuit in the acoustic-vibration fusion sensor.

[0028] Among them, the acoustic-vibration fusion sensor refers to a composite sensor that integrates vibration-sensitive elements and acoustic emission-sensitive elements. Installed in the bearing area of ​​a motor, it can simultaneously sense mechanical vibration and high-frequency elastic wave signals in the bearing region. The vibration acceleration signal reflects the vibration intensity changes of the motor's mechanical structure in the low-to-mid-frequency range, while the high-frequency acoustic emission signal reflects the instantaneous elastic waves generated within the material due to friction, impact, or crack propagation. Together, they constitute the physical characterization of the mechanical state. The analog conditioning circuit refers to the circuit module within the sensor that performs analog domain processing on the raw signal, including functions such as amplification and filtering.

[0029] In this embodiment, the vibration acceleration signal and high-frequency acoustic emission signal corresponding to the stator current signal time are acquired synchronously by an acoustic-vibration fusion sensor. The two signals are amplified and anti-aliasing filtered by the analog conditioning circuit inside the sensor, thereby improving the signal quality and providing a synchronous and clean data foundation for subsequent multi-source signal fusion.

[0030] As an example, the acoustic-vibration fusion sensor is mounted on the bearing housing of the motor drive end. When the motor starts, mechanical vibration occurs in the bearing area, and the vibration-sensitive element outputs a vibration acceleration signal with an amplitude range of 0 mV to 500 mV. Simultaneously, the contact between the bearing rolling elements and the raceway generates high-frequency stress waves, and the acoustic emission-sensitive element outputs a high-frequency acoustic emission signal with an amplitude range of 0 mV to 100 mV. The two signals are transmitted through internal wires of the sensor to two independent input channels of the analog conditioning circuit for further processing.

[0031] In S102, the analog conditioning circuit within the acoustic-vibration fusion sensor is used to amplify and anti-aliasing the vibration acceleration signal and the high-frequency acoustic emission signal, specifically including: S1021. The analog conditioning circuit in the acoustic-vibration fusion sensor receives the vibration acceleration signal from the vibration-sensitive element and the high-frequency acoustic emission signal from the acoustic emission-sensitive element, respectively.

[0032] In this embodiment, the analog conditioning circuit inside the acoustic-vibration fusion sensor simultaneously receives two input signals: one from the raw vibration acceleration signal output by the vibration-sensitive element, and the other from the raw high-frequency acoustic emission signal output by the acoustic emission-sensitive element. The two signals enter the conditioning circuit in parallel, laying the foundation for subsequent synchronous processing.

[0033] As an example, when the motor starts, the vibration sensing element outputs a vibration acceleration signal with an amplitude range of 0 mV to 500 mV, and the acoustic emission sensing element outputs a high-frequency acoustic emission signal with an amplitude range of 0 mV to 100 mV. The two signals are transmitted to two independent input channels of the analog conditioning circuit through the internal wires of the sensor.

[0034] S1022. The vibration acceleration signal and the high-frequency acoustic emission signal are respectively input to the AC coupling path in the analog conditioning circuit to filter out the DC component in their respective signals.

[0035] In this context, the AC coupling path refers to a RC coupling circuit consisting of a series capacitor and a resistor. Its function is to block the DC component of the signal, allowing only the AC component to pass through. The DC component refers to the part of the signal with a frequency of zero or extremely slow change, usually caused by sensor bias voltage or temperature drift.

[0036] In this embodiment, the vibration acceleration signal is input to the first AC coupling path of the analog conditioning circuit, and the high-frequency acoustic emission signal is input to the second AC coupling path. The two signals pass through the DC blocking capacitors in their respective paths to filter out the DC bias in the signals and retain the AC components that reflect dynamic changes, thus obtaining two AC signals after removing DC.

[0037] As an example, the original vibration acceleration signal contains a 50 mV DC bias. After passing through the AC coupling path, this 50 mV DC component is filtered out, and the output signal fluctuates around 0 V. The original high-frequency acoustic emission signal contains a 20 mV DC bias, which is also filtered out. This processing allows the dynamic range of both signals to fully utilize the effective range of the subsequent amplification circuit.

[0038] S1023. The vibration acceleration signal and the high-frequency acoustic emission signal after AC coupling are respectively sent to the variable gain amplifier in the analog conditioning circuit. According to the original amplitude range of the vibration acceleration signal and the high-frequency acoustic emission signal, the amplification factor of the variable gain amplifier is adjusted so that the amplitude of the two amplified signals falls within the preset amplitude window range.

[0039] A variable gain amplifier is an amplifier circuit whose gain can be adjusted dynamically according to the amplitude of the input signal. The preset amplitude window range refers to the optimal input voltage range of the analog-to-digital converter, which is usually set to 50% to 80% of the converter's range to balance signal resolution and clipping margin.

[0040] In this embodiment, the vibration acceleration signal after AC coupling is sent to a first variable gain amplifier, and the high-frequency acoustic emission signal after AC coupling is sent to a second variable gain amplifier. The peak amplitude of the two signals is detected respectively, and the gain coefficients of the two amplifiers are independently adjusted according to the detection results, so that the maximum amplitude of the amplified vibration acceleration signal and the amplified high-frequency acoustic emission signal both fall within the preset amplitude window range, thereby obtaining the best quantization accuracy in the subsequent analog-to-digital conversion.

[0041] As an example, the preset amplitude window range is 2V to 3V, and the analog-to-digital converter range is 5V. The peak value of the vibration acceleration signal after AC coupling is approximately 0.4V. The amplification factor required to adjust its peak value to 2.5V is 6.25x, therefore the gain of the first variable gain amplifier is set to 6.25x. The peak value of the high-frequency acoustic emission signal after AC coupling is approximately 0.1V. The amplification factor required to adjust its peak value to 2.5V is 25x, therefore the gain of the second variable gain amplifier is set to 25x. After amplification, the peak values ​​of both signals are approximately 2.5V, falling within the preset window range.

[0042] S1024. The amplified two signals are input to the low-pass filter in the analog conditioning circuit respectively. The frequency components in the two signals that are higher than the cutoff frequency are filtered out by the cutoff frequency set by the low-pass filter.

[0043] The cutoff frequency is the frequency value at which the filter's frequency response drops to -3 dB, and it is set according to the effective frequency band range of the signal.

[0044] In this embodiment, the amplified vibration acceleration signal is input to a first low-pass filter, and the amplified high-frequency acoustic emission signal is input to a second low-pass filter. A first cutoff frequency is set according to the effective frequency range of the vibration acceleration signal to filter out high-frequency noise above that frequency. A second cutoff frequency is set according to the effective frequency range of the high-frequency acoustic emission signal to filter out noise components above that frequency, thereby achieving anti-aliasing filtering and preventing high-frequency noise from folding into the effective frequency band in subsequent sampling.

[0045] As an example, the effective frequency range of the vibration acceleration signal is 0 Hz to 2 kHz, so the cutoff frequency of the first low-pass filter is set to 2.5 kHz to filter out high-frequency noise above 2.5 kHz; the effective frequency range of the high-frequency acoustic emission signal is 20 kHz to 100 kHz, so the cutoff frequency of the second low-pass filter is set to 100 kHz to filter out noise components above 100 kHz.

[0046] S1025. The two signals processed by the low-pass filter are output separately as the processed vibration acceleration signal and the processed high-frequency acoustic emission signal.

[0047] In this embodiment, the vibration acceleration signal after passing through the low-pass filter is output from the first output terminal of the analog conditioning circuit as the processed vibration acceleration signal; the high-frequency acoustic emission signal after passing through the low-pass filter is output from the second output terminal of the analog conditioning circuit as the processed high-frequency acoustic emission signal.

[0048] As an example, the analog conditioning circuit outputs a processed vibration acceleration signal with a peak value of approximately 2.5 volts and a frequency range of 0 Hz to 2.5 kHz; the processed high-frequency acoustic emission signal also outputs a processed signal with a peak value of approximately 2.5 volts and a frequency range of 20 kHz to 100 kHz. Both signals are fed into an analog-to-digital converter for synchronous sampling, with sampling rates set to 5 kHz and 200 kHz respectively to meet the requirements of their respective frequency bands.

[0049] This application, through the above steps, utilizes an acoustic-vibration fusion sensor to synchronously acquire vibration acceleration and high-frequency acoustic emission signals, and combines an analog conditioning circuit to complete signal amplification and anti-aliasing filtering, thereby achieving high-quality synchronous acquisition of motor mechanical state signals and providing a reliable data foundation for subsequent multi-source signal fusion.

[0050] Prior to S103, it also includes: The waveform correspondence between the processed vibration acceleration signal and the processed high-frequency acoustic emission signal in the time domain is obtained. Based on the waveform correspondence, transient segments affected by external forces are identified from the processed vibration acceleration signal, and accompanying segments that coincide with the transient segments in time are identified from the processed high-frequency acoustic emission signal. The accompanying segments are distinguished from the processed high-frequency acoustic emission signal to obtain the distinguished high-frequency acoustic emission signal. The amplitude change contours of the distinguished high-frequency acoustic emission signal in the non-transient segments outside the transient segments are extracted, and the correlation degree between the amplitude change contours and the waveforms of the processed vibration acceleration signal in the non-transient segments is analyzed to obtain the correlation degree value. When the correlation degree value is lower than a preset threshold, the waveforms of the corresponding non-transient segments in the processed vibration acceleration signal and the distinguished high-frequency acoustic emission signal are jointly marked as ineffective segments. The stator current signal, the processed vibration acceleration signal, and the distinguished high-frequency acoustic emission signal within the time interval corresponding to the ineffective segments are removed in the joint time-frequency transformation.

[0051] The transient segment refers to the brief fluctuation range in the vibration acceleration signal caused by external impact or sudden interference. The signal characteristics within this segment are independent of the motor's mechanical state. The accompanying segment refers to the corresponding interval in the high-frequency acoustic emission signal that overlaps with the transient segment in time. The high-frequency signal within this interval may be mixed with external interference components. The non-transient segment refers to the time interval other than the transient segment. Correlation analysis is a quantitative method for calculating the similarity of two signal waveforms, which can employ algorithms such as Pearson correlation coefficient or dynamic time warping distance. The preset threshold is a pre-set lower limit for the correlation degree. Values ​​below this threshold indicate low correlation between the two signals in the non-transient segment, suggesting that the signal may be affected by interference or poor sensor contact during this period, and thus belong to an ineffective segment.

[0052] In this embodiment, the time-domain waveforms of the processed vibration acceleration signal and the processed high-frequency acoustic emission signal are first obtained, and the two are aligned and compared on the time axis. Based on the characteristics such as amplitude abrupt change and waveform distortion in the vibration acceleration signal, the transient segment affected by external action is identified. At the same time, the accompanying segment that coincides with the transient segment in the high-frequency acoustic emission signal is found, and the accompanying segment is distinguished from the high-frequency acoustic emission signal to obtain the distinguished high-frequency acoustic emission signal.

[0053] Then, the amplitude change profile of the differentiated high-frequency acoustic emission signal in the non-transient segment outside the transient segment is extracted. The correlation between this amplitude change profile and the waveform of the processed vibration acceleration signal in the same non-transient segment is calculated to obtain the correlation value.

[0054] When the correlation value is lower than the preset threshold, it indicates that the two signals in the non-transient segment lack consistency. The waveforms of the non-transient segment in the processed vibration acceleration signal and the distinguished high-frequency acoustic emission signal are marked as non-effective segments. In the subsequent joint time-frequency transformation, the stator current signal, the processed vibration acceleration signal and the distinguished high-frequency acoustic emission signal in the time interval corresponding to the non-effective segment are removed together, and only the effective segment signal is retained for subsequent analysis.

[0055] As an example, the processed vibration acceleration signal exhibits an abnormal increase in amplitude between 0.2 and 0.25 seconds, identified as a transient segment caused by external impact. The processed high-frequency acoustic emission signal also shows a sudden increase in amplitude within the same time period; this segment is identified as an accompanying segment and separated from the high-frequency acoustic emission signal. Within the non-transient segment between 0.3 and 0.5 seconds, the amplitude variation profile of the separated high-frequency acoustic emission signal is extracted, yielding a sequence [0.1, 0.12, 0.11, 0.13, 0.12] volts. The waveform amplitude sequence of the processed vibration acceleration signal within the same time period is extracted as [0.2, 0.21, 0.19, 0.22, 0.20] volts. The Pearson correlation coefficient between the two sequences is calculated, yielding a correlation value of 0.95, higher than the preset threshold of 0.8. Therefore, this non-transient segment is determined to be a valid segment and retained. If the correlation value of a non-transient segment is less than 0.8, all three signals in that segment will be removed and will not participate in the subsequent joint time-frequency transformation.

[0056] S103. Perform a joint time-frequency transformation on the stator current signal, the processed vibration acceleration signal, and the processed high-frequency acoustic emission signal to obtain the instantaneous frequency trajectory and energy distribution spectrum characterizing the electromagnetic and mechanical coupling characteristics of the motor.

[0057] Joint time-frequency transformation refers to the synchronous time-frequency domain analysis of multiple signals to extract the pattern of signal frequency change over time. Instantaneous frequency trajectory is the curve showing the instantaneous frequency of a signal changing over time, reflecting the dynamic evolution of the signal frequency. Energy distribution spectrum refers to the distribution of signal energy in the time and frequency planes, usually presented as a time-frequency spectrum.

[0058] like Figure 2 As shown, S103 specifically includes: S1031. The stator current signal, the processed vibration acceleration signal, and the processed high-frequency acoustic emission signal are synchronously organized according to the acquisition time axis to form three discrete sequences.

[0059] In this embodiment, the stator current signal, the processed vibration acceleration signal, and the processed high-frequency acoustic emission signal are aligned along a unified time coordinate axis to ensure that the three sets of signals correspond precisely in time. Since the sampling rates of the three signals may be different, the three sets of signals are converted into the same time point sequence through interpolation or resampling methods, forming three discrete sequences of the same length, with each sampling point of each sequence corresponding to the same physical moment.

[0060] As an example, the stator current signal is sampled at 10 kHz, the processed vibration acceleration signal at 5 kHz, and the processed high-frequency acoustic emission signal at 200 kHz. Resampling all three to 10 kHz yields three discrete sequences, each 50,000 bytes long, covering a startup process from 0 to 5 seconds.

[0061] S1032. Perform Hilbert-Huang transform on the three sets of discrete sequences respectively, extract the instantaneous frequency ridges of each sequence, and use short-time Fourier transform to generate the corresponding time-frequency energy spectrum.

[0062] Hilbert-Huang transform is an adaptive time-frequency analysis method that includes two steps: empirical mode decomposition and Hilbert transform. It can extract the nonlinear and non-stationary instantaneous frequency characteristics of a signal. The instantaneous frequency ridge refers to the curve showing how the instantaneous frequency value of a signal changes over time. Short-time Fourier transform is a classic time-frequency analysis method that performs a Fourier transform on the signal using a sliding time window to obtain the time-frequency energy distribution.

[0063] In this embodiment, empirical mode decomposition (EMD) is first performed on the three discrete sequences, decomposing each sequence into several intrinsic mode function (IMF) components. The instantaneous phase of each component's analytic signal is calculated using Hilbert transform, and the instantaneous frequency is obtained by differentiating the instantaneous phase. The IMF component with the largest energy proportion is selected as the principal component, and the curve showing the instantaneous frequency change of this component over time is extracted as the instantaneous frequency ridge. Simultaneously, short-time Fourier transform is performed on the three discrete sequences, and an appropriate window function and window length are selected to calculate the power spectral density within each time window. The results are then arranged in chronological order to form three sets of time-frequency energy spectra.

[0064] As an example, a Hilbert-Huang transform is performed on the discrete sequence of the stator current signal to extract its instantaneous frequency ridge. During the motor startup process, this ridge gradually rises from 0 Hz to 50 Hz. At the same time, a Hanning window with a window length of 256 sampling points and a step size of 64 sampling points is used to perform a short-time Fourier transform on the current signal to obtain the time-frequency energy spectrum of the current signal during startup. The horizontal axis represents time from 0 seconds to 5 seconds, the vertical axis represents frequency from 0 Hz to 100 Hz, and the color depth represents energy intensity.

[0065] S1033. In the time-frequency energy spectrum, set an energy threshold and extract continuous regions with energy values ​​higher than the energy threshold as candidate event segments.

[0066] The energy threshold refers to a pre-set lower limit for energy intensity, used to distinguish active signal regions from noisy background regions. The candidate event segment refers to a spatiotemporal region in the time-frequency energy spectrum where the energy value is continuously higher than the threshold, representing time periods and frequency bands where significant changes occur in the signal.

[0067] In this embodiment, energy thresholds are set for the three sets of time-frequency energy spectra, and each time-frequency point is traversed to identify points with energy values ​​higher than the threshold. High-energy points that are continuous in the time domain and adjacent in the frequency domain are marked as connected components, and the time span and frequency span of each connected component are extracted. Connected components with continuous time spans are selected as candidate event segments.

[0068] As an example, the energy threshold of the time-frequency energy spectrum of the current signal is set to 10% of the maximum energy value, and a continuous high-energy region between 0.5 seconds and 1.2 seconds and within the frequency range of 0 Hz to 60 Hz is extracted as candidate event segment A; the energy threshold of the time-frequency energy spectrum of the vibration acceleration signal is set to 15% of the maximum energy value, and a continuous high-energy region between 0.6 seconds and 1.1 seconds and within the frequency range of 10 Hz to 40 Hz is extracted as candidate event segment B; the energy threshold of the time-frequency energy spectrum of the high-frequency acoustic emission signal is set to 8% of the maximum energy value, and a continuous high-energy region between 0.55 seconds and 1.15 seconds and within the frequency range of 30 kHz to 80 kHz is extracted as candidate event segment C.

[0069] S1034. Perform cross-correlation analysis on the time axis of the instantaneous frequency ridge of the stator current signal and the processed vibration acceleration signal, calculate the correlation coefficient curve of the two, and perform a logical AND operation between the continuous time intervals in the correlation coefficient curve that exceed the preset correlation coefficient threshold and the candidate event segment to obtain the associated event segment.

[0070] Cross-correlation analysis calculates the similarity between two signals at different time offsets. The correlation coefficient curve reflects the synchronous changes of the two signals under the original time alignment condition. A preset correlation coefficient threshold is a pre-defined lower limit for the correlation coefficient; values ​​above this threshold indicate a strong correlation between the two signals within that time period. Logical AND operation refers to taking the intersection of two sets.

[0071] In this embodiment, the instantaneous frequency ridge of the stator current signal and the instantaneous frequency ridge of the processed vibration acceleration signal are aligned on the time axis, and the sliding window correlation coefficient between the two is calculated to obtain a curve showing the correlation coefficient changing over time. All continuous time intervals exceeding a preset correlation coefficient threshold are marked on the correlation coefficient curve to obtain strongly correlated time intervals. Then, a logical AND operation is performed between the strongly correlated time intervals and the time intervals of the candidate event segments extracted in step S1033, i.e., the intersection of the two in time is taken to obtain the associated event segments.

[0072] As an example, the correlation coefficient curve between the instantaneous frequency ridge of the stator current and the instantaneous frequency ridge of the vibration acceleration is calculated. A preset correlation coefficient threshold of 0.7 is set, resulting in a strong correlation time interval of 0.55 seconds to 1.25 seconds. This interval is then logically ANDed with the time intervals of candidate event segment A (0.5 seconds to 1.2 seconds) and candidate event segment B (0.6 seconds to 1.1 seconds). The intersection of these three time intervals is taken to obtain the associated event segment of 0.6 seconds to 1.1 seconds.

[0073] S1035. Using the instantaneous frequency ridge of the processed high-frequency acoustic emission signal, the start and end boundaries of the waveform are calibrated within the associated event segment using a peak detection algorithm.

[0074] Peak detection algorithms are algorithms that identify local maxima in signal waveforms and can be used to determine the start and end times of transient events. The start boundary refers to the time point at which the event begins, and the end boundary refers to the time point at which the event ends.

[0075] In this embodiment, within the time range of the associated event segment, the instantaneous frequency ridge waveform of the processed high-frequency acoustic emission signal is acquired; a peak detection algorithm is applied to the waveform to identify the time point when the waveform amplitude first exceeds a preset starting threshold as the starting boundary, and the time point when the waveform amplitude last falls back below a preset ending threshold as the ending boundary, thereby accurately calibrating the precise interval of the event on the time axis.

[0076] As an example, within the associated event segment of 0.6 seconds to 1.1 seconds, the amplitude of the instantaneous frequency ridge waveform of the high-frequency acoustic emission signal first exceeds the starting threshold at 0.62 seconds, marked as the starting boundary; and finally falls back below the ending threshold at 1.05 seconds, marked as the ending boundary. Therefore, the calibrated starting boundary is 0.62 seconds, and the ending boundary is 1.05 seconds.

[0077] S1036. Based on the associated event segments and the starting and ending boundaries, extract the energy centroid trajectories of the three sets of signals in the time-frequency energy spectrum, take the projection of the energy centroid trajectory on the time axis as the instantaneous frequency trajectory, and take the energy diffusion pattern of the energy centroid trajectory on the frequency time plane as the energy distribution map.

[0078] Here, the energy centroid trajectory refers to the curve showing the change in the position of the energy distribution center on the frequency axis at each time point in the time-frequency energy spectrum over time. The energy centroid frequency refers to the central frequency value of the energy distribution at each time point. The energy diffusion pattern refers to the width of the energy distribution in the frequency direction and its evolution over time.

[0079] In this embodiment, the time window to be analyzed is precisely determined based on the time range of the associated event segment and the start and end boundaries defined in step S1035. Within this time window, the energy centroid frequency at each time point is calculated for the time-frequency energy spectra of the three sets of signals. The energy centroid frequencies at each time point are connected in chronological order to obtain the energy centroid trajectory. The projection of this trajectory onto the time axis is taken as the instantaneous frequency trajectory, thus obtaining the instantaneous frequency trajectories corresponding to each of the three sets of signals. Simultaneously, the energy diffusion pattern of the energy centroid trajectory on the frequency-time plane is extracted, including statistical characteristics such as the standard deviation and skewness of the energy distribution at each time point, forming an energy distribution map.

[0080] As an example, within a time window of 0.62 seconds to 1.05 seconds, the energy centroid frequency of the stator current signal time-frequency energy spectrum at each time point is calculated, and the instantaneous frequency trajectory is obtained as rising from 48 Hz to 52 Hz and then falling back to 50 Hz. The full width at half maximum (FWHM) of the energy distribution at each time point is calculated, and the width sequence is obtained as expanding from 5 Hz to 12 Hz and then contracting back to 6 Hz. This width sequence is used as part of the energy distribution spectrum.

[0081] Through the above steps, this application performs joint time-frequency transformation of stator current, vibration acceleration and high-frequency acoustic emission signal, realizing deep fusion and feature extraction of multi-source signals, significantly improving the accuracy and anti-interference ability of instantaneous frequency trajectory and energy distribution spectrum, and providing reliable feature input for subsequent coupling analysis.

[0082] S104. Generate coupled feature data based on instantaneous frequency trajectory and energy distribution map.

[0083] Among them, coupling feature data refers to multidimensional feature vectors extracted from instantaneous frequency trajectories and energy distribution maps, which can quantitatively describe the temporal relationship between electromagnetic force and mechanical inertia response.

[0084] S104 specifically includes: S1041. Perform first-order difference operation on the instantaneous frequency trajectory to identify the rising segment and falling segment where the rate of change of frequency value exceeds the preset rate of change threshold, and use them as the electromagnetic force action range and mechanical inertia response range, respectively.

[0085] The preset rate of change threshold is a pre-defined upper limit for the absolute value of the frequency change rate; exceeding this value indicates a significant frequency change. The electromagnetic force action range refers to the period during which the frequency rises rapidly, corresponding to the active electromagnetic force action phase. The mechanical inertia response range refers to the period during which the frequency falls rapidly, corresponding to the mechanical inertia-dominated response phase.

[0086] In this embodiment of the application, a first-order difference operation is performed on the instantaneous frequency trajectory to obtain a curve of frequency change rate changing with time; the continuous time interval where the change rate exceeds a preset positive threshold is identified as the rising segment, which is used as the electromagnetic force action interval; the continuous time interval where the change rate is lower than a preset negative threshold is identified as the falling segment, which is used as the mechanical inertia response interval.

[0087] As an example, the instantaneous frequency trajectory of the stator current signal is subjected to first-order difference, with a preset rate-of-change threshold of 10 Hz per second. Frequency changes exceeding 10 Hz per second within the interval of 0.62 to 0.85 seconds are identified and marked as electromagnetic force application intervals; frequency changes below negative 10 Hz per second within the interval of 0.85 to 1.05 seconds are identified and marked as mechanical inertia response intervals.

[0088] S1042. In the energy distribution map, calculate the energy centroid frequency at each time point and generate an energy centroid frequency curve that varies with time.

[0089] In this embodiment of the application, the energy centroid frequency at each time point is extracted from the energy distribution map, and the energy centroid frequency values ​​at each time point are connected in chronological order to form an energy centroid frequency curve.

[0090] As an example, the energy centroid frequency at each time point from 0.62 seconds to 1.05 seconds was calculated from the energy distribution spectrum of the stator current signal. The energy centroid frequency curve was found to be 48.5 Hz at 0.62 seconds, rising to 52.3 Hz at 0.85 seconds, and falling to 49.8 Hz at 1.05 seconds.

[0091] S1043. Extract the first moment centroid time of the energy centroid frequency curve within the electromagnetic force action range as the characteristic moment of the electromagnetic force, and extract the second moment centroid time of the energy centroid frequency curve within the mechanical inertia response range as the characteristic moment of the mechanical response.

[0092] The first-order moment centroid time refers to the time point corresponding to the centroid of the first-order statistical moment of the curve within a specified interval, reflecting the average position and time of the energy centroid frequency within that interval. The second-order moment centroid time refers to the time point corresponding to the centroid of the second-order statistical moment of the curve, reflecting the center of the temporal dispersion of the energy distribution.

[0093] In this embodiment of the application, within the electromagnetic force action range, the first moment of the energy centroid frequency curve is calculated to obtain the time centroid of the energy distribution within the range, which is used as the characteristic moment of the electromagnetic force; within the mechanical inertia response range, the second moment of the energy centroid frequency curve is calculated to obtain the time discrete centroid of the energy distribution within the range, which is used as the characteristic moment of the mechanical response.

[0094] As an example, within the electromagnetic force application range of 0.62 seconds to 0.85 seconds, the moment of the first moment of the energy centroid frequency curve is calculated to be 0.74 seconds, which is taken as the characteristic moment of the electromagnetic force; within the mechanical inertia response range of 0.85 seconds to 1.05 seconds, the moment of the second moment of the energy centroid frequency curve is calculated to be 0.94 seconds, which is taken as the characteristic moment of the mechanical response.

[0095] S1044. Calculate the time difference sequence between the characteristic time of electromagnetic force and the characteristic time of mechanical response, and calculate the proportional relationship between the time difference sequence and the length of the electromagnetic force action interval and the length of the mechanical inertia response interval.

[0096] Here, the time difference sequence refers to the sequence formed by the differences between the characteristic moments of electromagnetic force and the characteristic moments of mechanical response in multiple events. The proportional relationship refers to the ratio between the time difference and the length of each interval.

[0097] In this embodiment, the time difference is obtained by subtracting the characteristic time of the electromagnetic force from the characteristic time of the mechanical response. Since there may be multiple repeated electromagnetic force action and mechanical response events in the transient operating condition, a time difference sequence is formed. The ratio of each time difference to the length of the corresponding electromagnetic force action interval and the ratio to the length of the corresponding mechanical inertia response interval are calculated to obtain a proportional relationship sequence.

[0098] As an example, the calculated time difference is 0.94 seconds minus 0.74 seconds, which equals 0.20 seconds. The electromagnetic force's effective interval is 0.23 seconds, and the ratio of the time difference to the electromagnetic force's effective interval is approximately 0.87 (0.20 divided by 0.23). The mechanical inertia response interval is also 0.20 seconds, and the ratio of the time difference to the mechanical inertia response interval is 1.00 (0.20 divided by 0.20).

[0099] S1045. The time difference sequence and the proportional relationship are vectorized and combined to generate a multi-dimensional time series deviation vector, and the multi-dimensional time series deviation vector is used as the coupling feature data.

[0100] Among them, the multidimensional time-series deviation vector is used to characterize the dynamic matching relationship between the electromagnetic force action time sequence and the mechanical inertia response time sequence at the energy centroid level.

[0101] In this embodiment, the calculated time difference, the ratio of the time difference to the length of the electromagnetic force action interval, and the ratio of the time difference to the length of the mechanical inertia response interval are combined into a multidimensional vector, which is the multidimensional timing deviation vector. If there are multiple events in the transient condition, each event corresponds to a vector, and multiple vectors form a sequence.

[0102] As an example, a time difference of 0.20 seconds, a scale of 0.87, and a scale of 1.00 are combined to form a multidimensional timing deviation vector [0.20, 0.87, 1.00]. During the motor startup process, three similar electromagnetic force action and mechanical response events were identified, forming a vector sequence {[0.20, 0.87, 1.00], [0.18, 0.82, 0.95], [0.22, 0.91, 1.05]}, which serves as the coupling feature data.

[0103] Through the above steps, this application transforms time-frequency characteristics into quantized coupling characteristic data, achieving an accurate characterization of the temporal matching relationship between electromagnetic force and mechanical inertia response, improving the quantifiability of the coupling state, and providing accurate data support for dynamic matching analysis.

[0104] S105. In the preset transient response model of the motor, analyze the change process of the coupling characteristic data during the transient operating condition, and determine the dynamic matching degree of the electromagnetic force and mechanical inertia of the motor under the transient operating condition based on the change process.

[0105] The preset motor transient response model refers to a pre-constructed reference model that describes the ideal temporal relationship between the electromagnetic force and mechanical inertia response of the motor under transient operating conditions, including the theoretical standard time interval. The dynamic matching degree refers to the degree of consistency between the actual coupling characteristics and the model reference value, reflecting the synergy between electromagnetic force and mechanical inertia in the transient process.

[0106] S105 specifically includes: S1051. Input the coupling characteristic data into the preset motor transient response model. The motor transient response model includes a first reference time interval characterizing the trend of electromagnetic force change and a second reference time interval characterizing the trend of mechanical inertia response.

[0107] The first reference time interval refers to the standard range of time difference during the electromagnetic force action phase in the transient response model of the motor, while the second reference time interval refers to the standard range of proportion during the mechanical inertia response phase.

[0108] In this embodiment of the application, a multidimensional time-series deviation vector sequence is input into a preset motor transient response model. The model pre-sets a first reference time interval corresponding to the electromagnetic force and a second reference time interval corresponding to the mechanical inertia response based on the motor design parameters and theoretical dynamic response characteristics.

[0109] As an example, in the motor transient response model, the first reference timing interval is set to a time difference between 0.15 seconds and 0.25 seconds, and the second reference timing interval is a ratio between 0.85 and 1.05.

[0110] S1052. In the transient response model of the motor, the multidimensional timing deviation vector in the coupled feature data is compared with the first reference timing interval and the second reference timing interval respectively, and the timing deviation is identified as the first matching state falling into the first reference timing interval, the second matching state falling into the second reference timing interval, and the third matching state that deviates from both the first reference timing interval and the second reference timing interval.

[0111] The first matching state refers to the state when the time difference component falls into the first reference time interval, the second matching state refers to the state when the ratio component falls into the second reference time interval, and the third matching state refers to the state when neither the time difference nor the ratio component falls into their respective reference intervals.

[0112] In this embodiment, each component of the multidimensional time-series deviation vector is compared with its corresponding reference time-series interval. If the time difference component falls into the first reference time-series interval, it is determined to be a first matching state; if the ratio component falls into the second reference time-series interval, it is determined to be a second matching state; if neither the time difference nor the ratio component falls into their respective reference intervals, it is determined to be a third matching state.

[0113] As an example, when matching and identifying the vector [0.20, 0.87, 1.00], a time difference of 0.20 seconds falls within the first reference time interval of 0.15 seconds to 0.25 seconds, which is determined to be a first matching state; the proportions 0.87 and 1.00 both fall within the second reference time interval of 0.85 to 1.05 seconds, which is determined to be a second matching state. For the vector [0.30, 1.20, 1.10], a time difference of 0.30 seconds does not fall within the first reference time interval, and the proportions 1.20 and 1.10 do not fall within the second reference time interval, so it is determined to be a third matching state.

[0114] S1053. Statistical time-series deviations include the first cumulative number of times the first matching state occurs, the second cumulative number of times the second matching state occurs, and the third cumulative number of times the third matching state occurs during transient operating conditions.

[0115] The first cumulative count refers to the total number of times the first matching state occurs, the second cumulative count refers to the total number of times the second matching state occurs, and the third cumulative count refers to the total number of times the third matching state occurs.

[0116] In this embodiment of the application, during the duration of the transient operating condition, the matching state of each multidimensional time-series deviation vector is identified and counted, and the total number of occurrences of the first matching state, the total number of occurrences of the second matching state, and the total number of occurrences of the third matching state are counted respectively.

[0117] As an example, within 5 seconds of motor startup, a total of 15 multi-dimensional timing deviation vectors were extracted. After matching and identification, it was found that the first matching state appeared 12 times, the second matching state appeared 11 times, and the third matching state appeared 2 times.

[0118] S1054. Based on the numerical relationship between the first cumulative count, the second cumulative count, and the third cumulative count, determine the synchronization degree between the electromagnetic force action sequence and the mechanical inertia response sequence under transient operating conditions, and use the synchronization degree value as the dynamic matching degree.

[0119] The synchronization degree value refers to the quantitative value calculated based on the ratio of the first cumulative count, the second cumulative count, and the third cumulative count, which is used to represent the temporal matching degree between electromagnetic force and mechanical inertia.

[0120] In this embodiment, the synchronization degree value is calculated based on the proportional relationship between the first cumulative count, the second cumulative count, and the third cumulative count. The synchronization degree value is positively correlated with the first and second cumulative counts and negatively correlated with the third cumulative count. The specific calculation method can be weighted summation or ratio operation.

[0121] As an example, when calculating the synchronization level, the first cumulative count (12) and the second cumulative count (11) are added together, and the third cumulative count (2) is subtracted, resulting in 21. Dividing 21 by the total count (15) yields a synchronization level of 1.4. This value is greater than 1, indicating a high degree of temporal matching between the electromagnetic force and the mechanical inertia.

[0122] Through the above steps, this application performs matching analysis between coupling characteristic data and a preset motor transient response model, thereby achieving a quantitative assessment of the dynamic matching degree between electromagnetic force and mechanical inertia. This improves the accuracy and interpretability of identifying transient coupling relationships and provides a reliable quantitative basis for performance evaluation.

[0123] S106. Based on the degree of dynamic matching, the performance stability and dynamic response capability of the motor under transient operating conditions are quantitatively evaluated to obtain the transient performance evaluation results of the motor.

[0124] Among them, performance stability refers to the motor's ability to maintain stable operation during transient processes, while dynamic response capability refers to the motor's ability to quickly follow changes in electromagnetic force. The transient performance evaluation result is the final evaluation index that comprehensively reflects the electromagnetic-mechanical matching state of the motor under transient operating conditions.

[0125] S106 specifically includes: S1061. Obtain the synchronization degree value in the dynamic matching degree, and compare the synchronization degree value with the preset stability threshold and the preset response threshold respectively.

[0126] The preset stability threshold is the boundary value used to divide stability levels, and the preset response threshold is the boundary value used to divide response capability levels.

[0127] In this embodiment of the application, the synchronization degree value calculated in step S1054 is obtained and compared with a preset stability threshold and a preset response threshold respectively to determine the relationship between the synchronization degree value and the two thresholds.

[0128] As an example, the synchronization level is 1.4, the preset stability threshold is 1.2, and the preset response threshold is 1.1. The synchronization level of 1.4 is greater than both the stability threshold and the response threshold.

[0129] S1062. When the synchronization degree value is greater than or equal to the stability threshold, the performance stability of the motor is rated as the first stability level; when the synchronization degree value is less than the stability threshold, the performance stability is rated as the second stability level.

[0130] The first stability level indicates good stability, while the second stability level indicates poor stability.

[0131] In this embodiment of the application, the performance stability is divided into two levels based on the comparison between the synchronization degree value and the stability threshold.

[0132] As an example, the synchronization level of 1.4 is greater than the stability threshold of 1.2, so the performance stability is rated as the first stability level.

[0133] S1063. When the synchronization degree value is greater than or equal to the response threshold, the dynamic response capability of the motor is rated as the first response level; when the synchronization degree value is less than the response threshold, the dynamic response capability is rated as the second response level.

[0134] The first response level indicates a good response capability, while the second response level indicates a poor response capability.

[0135] In this embodiment of the application, the dynamic response capability is divided into two levels based on the comparison result between the synchronization degree value and the response threshold.

[0136] As an example, the synchronization level of 1.4 is greater than the response threshold of 1.1, so the dynamic response capability is rated as the first response level.

[0137] S1064. Combine the first stability level or the second stability level with the first response level or the second response level to form a combined identifier that includes stability level information and response level information, and output the combined identifier as the transient performance evaluation result.

[0138] Among them, the combined identifier refers to the comprehensive identifier formed by combining the stability level and the response level.

[0139] In this embodiment of the application, the assessed stability level and response level are combined to generate a combined identifier, which contains both performance stability information and dynamic response capability information, and is used as the final transient performance evaluation result output.

[0140] As an example, the first stability level and the first response level are combined to generate a combined identifier "stability level A-response level A", and this combined identifier is output as the result of the motor transient performance evaluation.

[0141] Through the above steps, this application transforms the dynamic matching degree into a combined identifier that includes stability level and response level, thereby enabling intuitive output of transient performance evaluation results, improving the readability and engineering applicability of the evaluation results, and facilitating rapid judgment of the transient operating status of the motor.

[0142] Figure 3 This is a schematic diagram of a specific implementation of a motor performance evaluation system based on electrical signal spectrum analysis provided in this application. (Refer to...) Figure 3 The system may include: The acquisition module 31 is used to acquire one stator current signal of the motor under transient conditions such as motor start-up, shutdown or sudden load change. Processing module 32 is used to acquire the vibration acceleration signal and high-frequency acoustic emission signal corresponding to the stator current signal through an acoustic-vibration fusion sensor installed on the motor bearing, and to amplify and anti-aliasing filter the vibration acceleration signal and the high-frequency acoustic emission signal using the analog conditioning circuit in the acoustic-vibration fusion sensor. The transformation module 33 is used to perform joint time-frequency transformation on the stator current signal, the processed vibration acceleration signal and the processed high-frequency acoustic emission signal to obtain the instantaneous frequency trajectory and energy distribution spectrum characterizing the electromagnetic and mechanical coupling characteristics of the motor. The generation module 34 is used to generate coupling feature data based on the instantaneous frequency trajectory and the energy distribution map. The coupling feature data is used to reflect the electromagnetic and mechanical coupling state of the motor under the transient operating condition. Analysis module 35 is used to analyze the change process of the coupling characteristic data during the transient operating condition in a preset motor transient response model, and determine the dynamic matching degree of the motor electromagnetic force and mechanical inertia under the transient operating condition based on the change process. Evaluation module 36 is used to quantitatively evaluate the performance stability and dynamic response capability of the motor under the transient operating conditions based on the dynamic matching degree, and obtain the transient performance evaluation result of the motor.

[0143] The motor performance evaluation system based on electrical signal spectrum analysis in this application is used to implement the aforementioned motor performance evaluation method based on electrical signal spectrum analysis. Therefore, the specific implementation of the motor performance evaluation system based on electrical signal spectrum analysis can be found in the embodiment section of the motor performance evaluation method based on electrical signal spectrum analysis above. The specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0144] This application also provides an electronic device, including: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of any of the above-described motor performance evaluation methods based on electrical signal spectrum analysis.

[0145] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the above-described methods for evaluating motor performance based on electrical signal spectrum analysis.

[0146] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0147] Embodiments of the present invention also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the embodiments of the motor performance evaluation method based on electrical signal spectrum analysis.

[0148] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0149] The foregoing has provided a detailed description of a motor performance evaluation method and system based on electrical signal spectrum analysis provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for evaluating motor performance based on electrical signal spectrum analysis, characterized in that, include: During transient operating conditions such as motor start-up, shutdown, or sudden load changes, one stator current signal of the motor is collected. By using an acoustic-vibration fusion sensor installed on the motor bearing, the vibration acceleration signal and high-frequency acoustic emission signal corresponding to the stator current signal are acquired, and the vibration acceleration signal and the high-frequency acoustic emission signal are amplified and anti-aliasing filtered by the analog conditioning circuit in the acoustic-vibration fusion sensor. By performing a joint time-frequency transformation on the stator current signal, the processed vibration acceleration signal, and the processed high-frequency acoustic emission signal, an instantaneous frequency trajectory and energy distribution spectrum characterizing the electromagnetic and mechanical coupling characteristics of the motor are obtained. Based on the instantaneous frequency trajectory and the energy distribution map, coupling feature data is generated, which is used to reflect the electromagnetic and mechanical coupling state of the motor under the transient operating condition. In the preset transient response model of the motor, the change process of the coupling characteristic data during the transient operating condition is analyzed, and the dynamic matching degree of the electromagnetic force and mechanical inertia of the motor under the transient operating condition is determined based on the change process. Based on the dynamic matching degree, the performance stability and dynamic response capability of the motor under the transient operating conditions are quantitatively evaluated to obtain the transient performance evaluation results of the motor.

2. The method according to claim 1, characterized in that, The process of performing a joint time-frequency transformation on the stator current signal, the processed vibration acceleration signal, and the processed high-frequency acoustic emission signal to obtain the instantaneous frequency trajectory and energy distribution spectrum characterizing the electromagnetic and mechanical coupling characteristics of the motor includes: The stator current signal, the processed vibration acceleration signal, and the processed high-frequency acoustic emission signal are synchronously organized according to the acquisition time axis to form three discrete sequences. Hilbert-Huang transforms are performed on the three sets of discrete sequences respectively to extract the instantaneous frequency ridges of each sequence, and the corresponding time-frequency energy spectra are generated by short-time Fourier transform. In the time-frequency energy spectrum, an energy threshold is set, and continuous regions with energy values ​​higher than the energy threshold are extracted as candidate event segments; The instantaneous frequency ridges of the stator current signal and the processed vibration acceleration signal are cross-correlated on the time axis to calculate the correlation coefficient curve between the two. The continuous time intervals in the correlation coefficient curve that exceed the preset correlation coefficient threshold are logically ANDed with the candidate event segment to obtain the associated event segment. Using the instantaneous frequency ridge of the processed high-frequency acoustic emission signal, the start and end boundaries of the waveform are calibrated within the associated event segment using a peak detection algorithm; Based on the associated event segment and the starting and ending boundaries, the energy centroid trajectories of the three sets of signals in the time-frequency energy spectrum are extracted. The projection of the energy centroid trajectory on the time axis is taken as the instantaneous frequency trajectory, and the energy diffusion pattern of the energy centroid trajectory on the frequency-time plane is taken as the energy distribution map.

3. The method according to claim 1, characterized in that, The process of amplifying and anti-aliasing filtering the vibration acceleration signal and the high-frequency acoustic emission signal using the analog conditioning circuit within the acoustic-vibration fusion sensor includes: The analog conditioning circuit within the acoustic-vibration fusion sensor receives the vibration acceleration signal from the vibration-sensitive element and the high-frequency acoustic emission signal from the acoustic emission-sensitive element, respectively. The vibration acceleration signal and the high-frequency acoustic emission signal are respectively input to the AC coupling path in the analog conditioning circuit to filter out the DC component in their respective signals; The vibration acceleration signal and the high-frequency acoustic emission signal after AC coupling are respectively sent to the variable gain amplifier in the analog conditioning circuit. According to the original amplitude range of the vibration acceleration signal and the high-frequency acoustic emission signal, the amplification factor of the variable gain amplifier is adjusted so that the amplitude of the two amplified signals falls within the preset amplitude window range. The amplified two signals are respectively input to the low-pass filter in the analog conditioning circuit. The frequency components in the two signals that are higher than the cutoff frequency are filtered out by the cutoff frequency set by the low-pass filter. The two signals processed by the low-pass filter are output separately as the processed vibration acceleration signal and the processed high-frequency acoustic emission signal.

4. The method according to claim 1, characterized in that, The process of generating coupled feature data based on the instantaneous frequency trajectory and the energy distribution map includes: The instantaneous frequency trajectory is subjected to a first-order difference operation to identify the rising segment and the falling segment where the rate of change of the frequency value exceeds a preset rate of change threshold, which are respectively used as the electromagnetic force action range and the mechanical inertia response range. In the energy distribution map, the energy centroid frequency at each time point is calculated, and an energy centroid frequency curve is generated that varies with time. The first moment centroid time of the energy centroid frequency curve within the electromagnetic force action range is extracted as the characteristic moment of the electromagnetic force, and the second moment centroid time of the energy centroid frequency curve within the mechanical inertia response range is extracted as the characteristic moment of the mechanical response. Calculate the time difference sequence between the characteristic time of the electromagnetic force and the characteristic time of the mechanical response, and calculate the proportional relationship between the time difference sequence and the length of the electromagnetic force action interval and the length of the mechanical inertia response interval; The time difference sequence and the proportional relationship are vectorized and combined to generate a multidimensional time deviation vector. The multidimensional time deviation vector includes a time difference component and a proportional component. The multidimensional time deviation vector is used as the coupling feature data. The multidimensional time deviation vector is used to characterize the dynamic matching relationship between the electromagnetic force action time sequence and the mechanical inertia response time sequence at the energy centroid level.

5. The method according to claim 4, characterized in that, The step involves analyzing the changes in the coupling characteristic data during the transient operating condition within a preset motor transient response model, and determining the dynamic matching degree between the motor's electromagnetic force and mechanical inertia under the transient operating condition based on the changes. This includes: The coupling characteristic data is input into the preset motor transient response model, which includes a first reference time interval characterizing the electromagnetic force change trend and a second reference time interval characterizing the mechanical inertia response trend. In the motor transient response model, the time difference component and the proportional component in the multidimensional time deviation vector of the coupled feature data are taken as time deviation pairs. The time deviation pairs are compared with the first reference time interval and the second reference time interval respectively. The first matching state is identified in which the time difference component in the time deviation pair falls into the first reference time interval, the second matching state in which the proportional component in the time deviation pair falls into the second reference time interval, and the third matching state in which the time difference component and the proportional component deviate from the first reference time interval and the second reference time interval at the same time. The timing deviation is statistically analyzed to determine the first cumulative number of times the first matching state occurs, the second cumulative number of times the second matching state occurs, and the third cumulative number of times the third matching state occurs during the transient operating condition. Based on the numerical relationship between the first cumulative count, the second cumulative count, and the third cumulative count, the synchronization degree between the electromagnetic force action sequence and the mechanical inertia response sequence under the transient operating condition is determined, and the synchronization degree is used as the dynamic matching degree.

6. The method according to claim 1, characterized in that, The step of quantitatively evaluating the performance stability and dynamic response capability of the motor under the transient operating condition based on the dynamic matching degree, and obtaining the transient performance evaluation result of the motor, includes: Obtain the synchronization degree value in the dynamic matching degree, and compare the synchronization degree value with a preset stability threshold and a preset response threshold respectively; When the synchronization degree value is greater than or equal to the stability threshold, the performance stability of the motor is rated as the first stability level; when the synchronization degree value is less than the stability threshold, the performance stability is rated as the second stability level. When the synchronization degree value is greater than or equal to the response threshold, the dynamic response capability of the motor is rated as the first response level; when the synchronization degree value is less than the response threshold, the dynamic response capability is rated as the second response level. The first stability level or the second stability level is combined with the first response level or the second response level to form a combined identifier that includes stability level information and response level information, and the combined identifier is output as the transient performance evaluation result.

7. The method according to claim 1, characterized in that, Before performing a joint time-frequency transformation on the stator current signal, the processed vibration acceleration signal, and the processed high-frequency acoustic emission signal, the method further includes: The waveform correspondence between the processed vibration acceleration signal and the processed high-frequency acoustic emission signal in the time domain is obtained. Based on the waveform correspondence, the transient segment affected by external forces is identified from the processed vibration acceleration signal, and the accompanying segment that coincides with the transient segment in time is identified from the processed high-frequency acoustic emission signal. The accompanying segment is distinguished from the processed high-frequency acoustic emission signal to obtain the distinguished high-frequency acoustic emission signal; The amplitude variation profile of the differentiated high-frequency acoustic emission signal in the non-transient segment outside the transient segment is extracted, and the correlation profile of the amplitude variation is analyzed with the waveform of the processed vibration acceleration signal in the non-transient segment to obtain the correlation value; When the correlation value is lower than a preset threshold, the waveforms corresponding to the non-transient segment in the processed vibration acceleration signal and the distinguished high-frequency acoustic emission signal are jointly marked as ineffective segments, and the stator current signal, the processed vibration acceleration signal and the distinguished high-frequency acoustic emission signal in the time interval corresponding to the ineffective segment are removed in the joint time-frequency transformation.

8. A motor performance evaluation system based on electrical signal spectrum analysis, characterized in that, include: The acquisition module is used to acquire one stator current signal of the motor during transient operating conditions such as motor start-up, shutdown, or sudden load changes. The processing module is used to acquire the vibration acceleration signal and high-frequency acoustic emission signal corresponding to the stator current signal through the acoustic vibration fusion sensor installed on the motor bearing, and to amplify and anti-aliasing filter the vibration acceleration signal and the high-frequency acoustic emission signal using the analog conditioning circuit in the acoustic vibration fusion sensor. The transformation module is used to perform joint time-frequency transformation on the stator current signal, the processed vibration acceleration signal, and the processed high-frequency acoustic emission signal to obtain the instantaneous frequency trajectory and energy distribution spectrum characterizing the electromagnetic and mechanical coupling characteristics of the motor. The generation module is used to generate coupling feature data based on the instantaneous frequency trajectory and the energy distribution map. The coupling feature data is used to reflect the electromagnetic and mechanical coupling state of the motor under the transient operating condition. The analysis module is used to analyze the change process of the coupling characteristic data during the transient operating condition in a preset motor transient response model, and determine the dynamic matching degree of the motor electromagnetic force and mechanical inertia under the transient operating condition based on the change process. The evaluation module is used to quantitatively evaluate the performance stability and dynamic response capability of the motor under the transient operating conditions based on the dynamic matching degree, and obtain the transient performance evaluation results of the motor.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the motor performance evaluation method based on electrical signal spectrum analysis as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, enables the implementation of the motor performance evaluation method based on electrical signal spectrum analysis as described in any one of claims 1 to 7.

Images