Motor vibration velocity double-channel frequency domain integral calibration method, device and system

Abstract

The application discloses a motor vibration speed double-channel frequency domain integral calibration method, device and system, and a method thereof, and the method comprises the following steps: obtaining an original vibration speed effective value corresponding to a to-be-tested motor sent by an edge end; periodically obtaining an acceleration original signal sequence cached by the edge end, reconstructing a time domain speed reconstruction signal through conjugate symmetric frequency domain integral, and converting to obtain a standard vibration speed effective value; based on the standard vibration speed effective value and the original vibration speed effective value, obtaining a correction coefficient and delivering the correction coefficient to the edge end, so that the edge end dynamically calibrates the real-time output vibration speed based on the correction coefficient. According to the embodiment of the application, the measurement drift and attenuation caused by the performance degradation of the sensor can be effectively offset, in-situ online calibration of the sensor is realized, the problems of fault misjudgment and omission can be avoided, and the accuracy and stability of the rotating machinery vibration monitoring are significantly improved.

Description

Technical Field

[0001] This application relates to the field of signal processing technology, and in particular to a method, apparatus and system for dual-channel frequency domain integration calibration of motor vibration velocity. Background Technology

[0002] In industrial predictive maintenance systems for rotating machinery such as electric motors and oil pumping units, surface vibration velocity is a core monitoring parameter for evaluating the machine's operating condition and identifying early faults. The root mean square (VRMS) vibration velocity is a key quantitative indicator in the ISO 10816 vibration intensity rating standard, effectively supporting the diagnosis and identification of typical early faults such as bearing wear, rotor imbalance, and shaft misalignment. Currently, industrial sites commonly use accelerometers deployed on the surface of rotating machinery to collect vibration signals and perform first-order numerical integration on the discrete time-domain acceleration sequence to obtain vibration velocity indicators that conform to international standards, thus achieving routine monitoring of equipment vibration status.

[0003] However, existing solutions for solving vibration velocities in rotating machinery and calibrating sensors still suffer from numerous technical shortcomings and application bottlenecks in low-power edge industrial applications, making it difficult to simultaneously address the constraints of low computing power at the edge, signal reconstruction accuracy, and long-term online calibration requirements. Traditional low-power microprocessors often employ a single-sided positive frequency interval integration operator when performing frequency domain integration of acceleration signals. This approach disrupts the strict conjugate symmetry of the positive and negative frequency intervals of the discrete spectrum, resulting in a residual imaginary part in the time-domain velocity signal reconstructed by the inverse Fourier transform after frequency domain integration. Directly discarding this imaginary part leads to phase shifts and amplitude distortion in the time-domain velocity signal, severely reducing the accuracy of vibration velocity reconstruction.

[0004] Meanwhile, in order to suppress the spectral leakage problem caused by the truncation of finite-length signals, the existing frequency domain analysis process generally performs windowing preprocessing on the original discrete acceleration sequence. However, while the windowing operation optimizes the spectral characteristics, it also causes the overall energy of the signal to decay, resulting in a lower effective value amplitude of the reconstructed vibration velocity, which directly affects the accuracy and reliability of the equipment vibration fault alarm judgment.

[0005] In addition, the complex operating conditions in industrial settings, such as high temperature, high humidity, and long-term high-frequency vibration, cause the voltage sensitivity of accelerometers to gradually degrade over time, leading to drift in sensor measurement characteristics. Limited by the low power consumption and low computing power of edge microprocessors, existing edge nodes can only achieve low-precision time-domain integration and coarse feature value estimation, unable to perform high-precision signal processing. This results in monotonic shifts and numerical divergence in the vibration velocity feature values ​​continuously reported by the edge as the sensor ages. Furthermore, existing sensor calibration schemes are mostly offline static calibration modes, requiring a fixed excitation signal applied to a standard vibration table for calibration, which cannot meet the requirements of long-term continuous operation of industrial equipment and makes in-situ online dynamic calibration difficult. On the other hand, continuously transmitting high-frequency raw acceleration waveform data from the edge to the central platform for high-precision calculations would significantly consume wireless transmission bandwidth, increasing the system communication load and making it unsuitable for large-scale, routine monitoring applications in industrial settings.

[0006] In summary, there is an urgent need for an adaptive in-situ online calibration and monitoring technology that is compatible with low-power edge devices, can achieve high-precision, distortion-free speed signal reconstruction, lossless energy compensation, and can suppress sensor performance degradation. Summary of the Invention

[0007] Therefore, it is necessary to provide a dual-channel frequency domain integration calibration method, apparatus, and system for motor vibration velocity to address the above-mentioned technical problems and solve at least one of the problems existing in the prior art.

[0008] Firstly, a dual-channel frequency domain integration calibration method for motor vibration velocity is provided, applied to a central server. The method includes: The effective value of the original vibration velocity corresponding to the motor under test sent by the edge terminal is obtained. The effective value of the original vibration velocity is obtained by the edge terminal collecting and buffering the original acceleration signal of the motor under test according to a preset sampling frequency, and then extracting it from the corresponding original acceleration signal according to a preset time window. The original acceleration signal sequence of the edge buffer is periodically acquired, and the time-domain velocity reconstruction signal is obtained by conjugate symmetric frequency domain integration, and the effective value of standard vibration velocity is calculated. Based on the standard effective value of vibration velocity and the original effective value of vibration velocity, a correction coefficient is obtained; The correction coefficient is sent to the edge end so that the edge end can dynamically calibrate the real-time output vibration velocity based on the correction coefficient.

[0009] In one possible implementation, the original effective value of the vibration velocity is obtained by second-level correction of the integral zero-point drift at the edge end based on a non-iterative algebraic dot product analytical algorithm, wherein the second-level correction includes: Determine whether the motor under test is currently in a stopped state by using time-domain zero-speed update dead-zone gating; If so, then force the output time-domain velocity reconstruction signal to be zero; If not, then perform a Tikhonov-regularized algebraic dot product operation on the original acceleration signal sequence in the frequency domain to reconstruct the velocity spectrum; When the energy percentage in the low-frequency cutoff section is not lower than the preset ratio, the integral hard rejection logic is triggered to avoid integral saturation. When the energy proportion in the low-frequency cutoff section is lower than the preset ratio, the Hanning window is used in the time domain to incrementally compensate the reconstructed time-domain velocity signal and output the original effective value of vibration velocity.

[0010] In one possible implementation, obtaining the time-domain velocity reconstruction signal through conjugate symmetric frequency domain integration includes: Frequency domain features are extracted from the original acceleration signal sequence acquired in the current period to obtain the acceleration frequency domain complex spectrum; Based on the conjugate symmetry relationship between positive and negative frequencies, the complex spectrum of the acceleration frequency domain is completed to obtain the complex spectrum of the velocity frequency domain; The inverse Fourier transform of the velocity frequency domain complex spectrum is performed to obtain the time-domain velocity reconstruction signal.

[0011] In one possible implementation, the step of completing the complex spectrum of the acceleration frequency domain based on the conjugate symmetry relationship of positive and negative frequencies to obtain the complex spectrum of the velocity frequency domain includes: Set the spectral amplitudes corresponding to the DC component and Nyquist frequency component in the complex spectrum of the acceleration frequency domain to zero; For the acceleration spectrum components in the entire positive frequency range, frequency domain integration is performed point by point in combination with the corresponding angular frequency to obtain the velocity spectrum components corresponding to each positive frequency point. For all negative frequency ranges, based on the conjugate symmetry of positive and negative frequencies in the real signal spectrum, the conjugate value of the corresponding positive frequency velocity spectrum is taken and filled to complete all negative frequency velocity spectrum components, forming the velocity frequency domain complex spectrum.

[0012] In one possible implementation, the step of extracting frequency domain features from the original acceleration signal sequence acquired within the current period to obtain the acceleration frequency domain complex spectrum includes: The original acceleration signal sequence acquired in the current period is subjected to Hanning windowing processing to obtain a windowed and smoothed time-domain acceleration sequence; A discrete Fourier transform is performed on the windowed and smoothed time-domain acceleration sequence to generate a complex frequency-domain acceleration spectrum.

[0013] In one possible implementation, the conversion to obtain the standard vibration velocity effective value includes: Calculate the root mean square time of the reconstructed velocity signal in the time domain; The time-domain root mean square is energy equivalently corrected based on a preset energy recovery compensation coefficient to obtain the effective value of the standard vibration velocity.

[0014] In one possible implementation, obtaining the correction coefficient based on the standard effective value of vibration velocity and the original effective value of vibration velocity includes: Based on the ratio between the standard effective value of vibration velocity and the original effective value of vibration velocity, the instantaneous deviation coefficient corresponding to the current cycle is obtained; Median filtering is performed on multiple sets of instantaneous deviation coefficients within a preset sliding window to obtain stable correction coefficients.

[0015] In one possible implementation, after the correction coefficient is sent to the edge terminal, the edge terminal uses the effective value of the motor vibration velocity calibrated by the correction coefficient to make an over-limit judgment with a preset vibration alarm threshold. If the effective value of the calibrated motor vibration velocity exceeds the preset vibration alarm threshold, an abnormal motor vibration alarm is output.

[0016] Secondly, a dual-channel frequency domain integration calibration system for motor vibration velocity is provided, applied to a central server. The system includes: The original vibration velocity effective value acquisition unit is used to acquire the original vibration velocity effective value corresponding to the motor under test sent by the edge end. The original vibration velocity effective value is obtained by the edge end collecting and buffering the original acceleration signal of the motor under test according to a preset sampling frequency, and then extracting it from the corresponding original acceleration signal according to a preset time window. The standard vibration velocity effective value acquisition unit is used to periodically acquire the original acceleration signal sequence buffered at the edge end, reconstruct the time-domain velocity signal through conjugate symmetric frequency domain integration, and convert it to obtain the standard vibration velocity effective value. The correction coefficient generation unit is used to obtain the correction coefficient based on the standard effective value of vibration velocity and the original effective value of vibration velocity; The calibration unit is used to send the correction coefficient to the edge end so that the edge end can dynamically calibrate the real-time output vibration velocity based on the correction coefficient.

[0017] In one possible implementation, the original effective value of the vibration velocity is extracted by the edge end after performing second-level correction of the integral zero-point drift based on a non-iterative algebraic dot product analytical algorithm. The second-level correction includes: When the time-domain zero-speed update dead-zone gating determines that the motor under test is not in a stopped state, the frequency domain performs an algebraic dot product operation based on Tikhonov regularization on the original acceleration signal sequence to reconstruct the velocity spectrum. When the energy proportion in the low-frequency cutoff section is not less than a preset ratio, an integral hard rejection is triggered to discard the current calculation. When the energy proportion in the low-frequency cutoff section is less than a preset ratio, the time-domain correction speed is output based on Hanning window incremental compensation.

[0018] Thirdly, a dual-channel frequency domain integration calibration system for motor vibration velocity is provided, comprising: the system includes a central server and at least one edge terminal; each edge terminal is rigidly mounted on the bearing housing of the motor under test and establishes a communication connection with the central server. The edge terminal is used to acquire and locally cache the original acceleration signal of the motor under test at a preset sampling frequency, perform lightweight operation on the cached signal according to a preset time window to extract the original vibration velocity effective value, and upload the original vibration velocity effective value to the central server. When performing lightweight operation on the cached signal, the edge terminal determines whether the state of the motor under test is stopped based on the zero-speed update dead zone gate in the time domain. If it is, it forces the output of the time domain velocity reconstruction signal to be zero. Otherwise, it performs frequency domain Tikhonov regularized algebraic dot product reconstruction on the original acceleration signal. When the proportion of low-frequency cutoff energy is not less than a preset ratio, it triggers integral hard rejection to discard the current calculation. When the proportion of low-frequency cutoff energy is less than a preset ratio, it outputs the original vibration velocity effective value based on Hanning window incremental compensation. The central server includes the aforementioned dual-channel frequency domain integration calibration device for motor vibration velocity, which is used to periodically acquire the original acceleration signal sequence cached at the edge end, reconstruct the time-domain velocity signal through a conjugate symmetric frequency domain integration algorithm and convert it into a standard vibration velocity effective value; then, it calculates a correction coefficient by combining the standard vibration velocity effective value with the original vibration velocity effective value uploaded by the edge end, and sends the correction coefficient to the corresponding edge end, so that the edge end can perform dynamic calibration on the real-time output vibration velocity measurement value according to the correction coefficient.

[0019] The aforementioned dual-channel frequency domain integral calibration method, apparatus, and system for motor vibration velocity includes the following steps: acquiring the original effective value of the vibration velocity of the motor under test sent by the edge terminal, wherein the original effective value of the vibration velocity is obtained by the edge terminal acquiring and buffering the original acceleration signal of the motor under test according to a preset sampling frequency, and then extracting it from the corresponding original acceleration signal according to a preset time window; periodically acquiring the original acceleration signal sequence buffered by the edge terminal, reconstructing the time-domain velocity reconstructed signal through conjugate symmetric frequency domain integral, and converting it into a standard effective value of vibration velocity; obtaining a correction coefficient based on the standard effective value of vibration velocity and the original effective value of vibration velocity; and sending the correction coefficient to the edge terminal so that the edge terminal can dynamically calibrate the real-time output vibration velocity based on the correction coefficient. In this embodiment, the system acquires and buffers raw acceleration signals locally at the edge in real time, extracts effective vibration velocity values ​​in seconds, and periodically acquires raw signal sequences at the server, then performs high-precision velocity reconstruction using conjugate symmetric frequency domain integration. This constructs a dual-channel collaborative monitoring mechanism with coarse edge measurement and fine center calibration, effectively balancing the low-power operation requirements of the edge with the accuracy requirements of equipment vibration monitoring. Dynamic correction coefficients are generated by comparing the deviation between the standard reconstructed effective vibration velocity values ​​and the effective values ​​output locally at the edge. Real-time calibration and correction of the edge measurement results are achieved through a parameter-download closed-loop method, continuously offsetting the impact of complex industrial processes. Under the condition of accelerometer performance degradation leading to monotonic drift and attenuation errors, this method achieves in-situ online self-healing correction of sensor measurement accuracy from an operational mechanism perspective. This eliminates the need for downtime calibration and additional standard excitation equipment, significantly reducing on-site maintenance costs and network transmission bandwidth pressure. At the same time, the vibration velocity values ​​calibrated by this method have higher accuracy and stronger stability, and can accurately match vibration intensity assessment standards for fault threshold determination. This effectively avoids the long-term accuracy deviation, fault misjudgment, and missed judgment problems existing in traditional fixed parameter measurement methods, significantly improving the accuracy, reliability, and long-term stability of vibration condition monitoring and predictive maintenance of rotating machinery. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application 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.

[0021] Figure 1 This is a schematic diagram of a monitoring system architecture for a dual-channel frequency domain integration calibration method for motor vibration velocity according to an embodiment of this application; Figure 2 This is a flowchart illustrating a dual-channel frequency domain integration calibration method for motor vibration velocity in one embodiment of this application; Figure 3 This is a schematic diagram of a module of the Hanning window ENBW energy recovery correction link and root mean square calculation framework in one embodiment of this application; Figure 4 This is a schematic diagram of the principle and spectrum allocation of the conjugate symmetric two-sided frequency domain velocity integral reconstruction algorithm in one embodiment of this application; Figure 5 This is a schematic diagram of a dual-channel frequency domain integration calibration device for motor vibration velocity according to one embodiment of this application; Figure 6 This is a schematic diagram of a computer device according to one embodiment of this application. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of 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 dual-channel frequency domain integration calibration method for motor vibration velocity provided in this embodiment can be applied to... Figure 1 In the monitoring system architecture scenario shown, the specific deployment method is as follows: a wireless monitoring node equipped with a three-axis accelerometer and a low-power MCU chip is used as an edge terminal (also known as a second-level low-power sentinel terminal), which is rigidly installed on the bearing housing of the motor under test to ensure the synchronization and authenticity of vibration signal acquisition; an algorithm service layer with high computing power data processing capabilities is deployed to a central server that can be connected to an industrial network, serving as a minute-level high-computing power benchmark terminal, thus forming a layered monitoring architecture of "edge low-power acquisition and calculation + central high-precision calibration and calculation".

[0024] During operation, the edge device acquires raw vibration acceleration signals of the motor in real time (seconds) using a triaxial accelerometer and caches them locally. Lightweight signal processing is achieved using a low-power MCU, calculating and outputting the local effective vibration velocity value in real time according to a fixed time window. The central server intermittently acquires the raw acceleration signal sequence cached at the edge device via the industrial network (either actively retrieved by the central server or actively uploaded by the edge device). A conjugate symmetric frequency domain integral algorithm is used to reconstruct the high-precision vibration velocity signal, obtaining the standard effective vibration velocity value. The central server generates stable correction coefficients by comparing the deviation characteristics between the standard effective value and the local effective value at the edge device and sends them to the edge device. The edge device dynamically calibrates and corrects the real-time output vibration velocity parameters based on the received correction coefficients, achieving online self-healing correction of sensor measurement deviations. Finally, using the calibrated vibration velocity value as a benchmark, and referring to the ISO 10816 standard, the motor vibration status assessment and fault alarm determination are completed, achieving long-term, high-precision, and high-stability online monitoring and predictive maintenance of the motor vibration status.

[0025] Example 1 In one embodiment, such as Figure 2 As shown, a dual-channel frequency domain integration calibration method for motor vibration velocity is provided, which is applied to... Figure 1 Taking the central server as an example, the explanation includes the following steps: In step S110, the effective value of the original vibration velocity corresponding to the motor under test sent by the edge terminal is obtained. The effective value of the original vibration velocity is obtained by the edge terminal collecting and buffering the original acceleration signal of the motor under test according to a preset sampling frequency, and then extracting it from the corresponding original acceleration signal according to a preset time window. Optionally, an edge-end acquisition device, such as an accelerometer, deployed on the surface of the motor under test, can continuously and in real-time acquire the raw triaxial acceleration signals in the time domain during motor operation at a preset sampling rate fs. The acquired raw acceleration time-domain data is cached in a local FIFO buffer to avoid bandwidth consumption pressure caused by real-time uploading of high-frequency raw data. The edge-end microprocessor can then segment the cached signal into preset second-level time windows, use a lightweight time-domain integration algorithm to complete low-precision numerical conversion, and simultaneously extract the basic features of the effective vibration velocity value. The time-domain integral uses a simple accumulation operation, which has low computing power overhead and fast running speed. It can meet the needs of edge nodes for low power consumption and continuous output of monitoring features in seconds. Finally, it outputs the original vibration velocity effective value corresponding to the current time window and uploads it to the central server.

[0026] In a preferred embodiment of this application, when the edge device performs lightweight operations on the buffered signal to extract the effective value of the original vibration velocity, it performs second-level online self-calibration correction of the integral zero-point drift based on a non-iterative algebraic dot product analytical algorithm. Specifically, this second-level correction includes the following steps: First, the operating state of the motor under test is determined by dead-zone gating in the zero-rate update time domain (ZUPT): the root mean square value of the time-domain acceleration signal is extracted in real time at the edge. and compared with the set zero-speed judgment threshold. Perform a comparison; if If the motor under test is stopped, the system will determine that the motor under test is in a stopped state. In order to prevent the time-domain accumulation integral from generating a serious zero-point drift under zero-bias disturbance, the system will force the output time-domain speed reconstruction signal to be zero. like If the motor under test is determined to be in operation, then frequency domain correction and speed reconstruction are initiated: the time-domain acceleration sequence is transformed to the frequency domain at the edge, and the speed spectrum is calculated using the Tikhonov regularized algebraic dot product reconstruction method to avoid the problem of denominator division by zero or numerical divergence when the traditional frequency domain integral is divided by a very small angular frequency at ultra-low frequencies. The speed spectrum reconstruction calculation formula is expressed as: ; in, Represents the acceleration spectrum components, For the first k The discrete angular frequencies corresponding to the spectral lines λ The parameters for the regularization term are typically set to extremely small positive real constants (e.g., ...). ).

[0027] Secondly, hard rejection detection with integral saturation is performed at the edge: real-time statistical acceleration spectrum is used at a set low-frequency cutoff frequency. The following section represents the percentage of accumulated energy to the total signal energy. If this percentage is not less than a preset proportion, such as 50%, it indicates that the signal is mainly dominated by sensor temperature drift or basic vibration zero drift, which can easily cause integrator saturation distortion. In this case, the system will initiate hard rejection logic, directly discarding the current calculation and refusing to update the correction coefficient. k If the ratio is lower than the preset ratio, such as 50%, it indicates that the signal energy distribution is reasonable and output is allowed.

[0028] Finally, to compensate for the boundary energy attenuation caused by the truncation of the frequency domain window function, Hanning window incremental compensation is performed on the time-domain velocity signal obtained by the inverse transform at the edge. Specifically, time-domain weighted superposition is performed by an incremental factor that is complementary to the energy of the window function to eliminate boundary waveform distortion and output a high-fidelity original vibration velocity effective value. .

[0029] In step S120, the original acceleration signal sequence of the edge buffer is periodically acquired, and the time-domain velocity reconstruction signal is obtained by conjugate symmetric frequency domain integration, and the standard vibration velocity effective value is calculated. Optionally, the central server actively retrieves the raw acceleration signal sequence cached in the local FIFO buffer of the edge terminal according to a preset period (at minute intervals), or receives the raw acceleration signal sequence uploaded by the edge terminal according to a preset period, and records it as a[n] (n=0, ..., N). 1) This allows for the acquisition of complete, unprocessed, raw vibration time-domain data. Then, leveraging its high computing power, the central server abandons the low-precision time-domain accumulation integration method used at the edge and employs a conjugate symmetric frequency-domain integration algorithm to perform high-precision signal reconstruction on the acquired raw acceleration signal sequence. This achieves distortion-free frequency-domain integration by fully preserving the conjugate symmetry of the positive and negative frequency spectra, and the inverse transformation yields a high-fidelity reconstructed time-domain velocity signal. Based on this, the central server performs effective value calculation and conversion on the reconstructed high-precision time-domain velocity signal to obtain a standard vibration velocity effective value with higher accuracy, smaller error, and conformity to the actual physical characteristics of vibration. .

[0030] In step S130, a correction coefficient is obtained based on the standard effective value of vibration velocity and the original effective value of vibration velocity; Optionally, the central server obtains the raw vibration velocity RMS value from the low-precision calculation output of the edge terminal. The standard vibration velocity effective value obtained by high-precision reconstruction through conjugate symmetric frequency domain integration, and itself. The instantaneous deviation coefficient is calculated by comparing the deviations of two sets of effective vibration velocity values ​​from the same acquisition period and the same measuring point. .

[0031] Furthermore, to mitigate abnormal coefficient disturbances caused by random vibrations and instantaneous acquisition noise on-site and to improve the long-term stability of the correction parameters, the system is set with a fixed-length sliding time window. L The instantaneous deviation coefficients generated over multiple periods are stored in a window queue. A median filtering algorithm is then used to sort all instantaneous deviation coefficients within the window and take the median value, removing outliers that deviate from the normal range. The final output is a robust correction coefficient with stronger anti-interference capabilities and smoother numerical changes. This ensures that the correction parameters sent to the edge will not change drastically due to a single accidental error, thus maintaining the stable and reliable accuracy of edge vibration measurement calibration. .

[0032] In step S140, the correction coefficient is sent to the edge end so that the edge end can dynamically calibrate the real-time output vibration velocity based on the correction coefficient.

[0033] Optionally, the central server will calculate the correction coefficients. The data is transmitted to the corresponding edge devices via the industrial network. After receiving and storing the correction coefficients transmitted from the center, the edge devices can dynamically calibrate the real-time output vibration velocity to obtain the calibrated effective value of the vibration velocity. , In subsequent real-time vibration monitoring, this calibrated effective value will be used. The system compares the vibration alarm threshold with the pre-configured vibration intensity standard conforming to ISO 10816 in real time and executes the over-limit judgment logic. When the effective value of the calibrated vibration velocity is greater than the preset alarm threshold, the system immediately generates and pushes an abnormal motor vibration alarm signal, which makes it easier for maintenance personnel to identify equipment faults in a timely manner and realize early warning of rotating machinery faults.

[0034] In this embodiment of the application, a dual-channel frequency domain integration calibration method for motor vibration velocity is provided, including: acquiring the original effective value of vibration velocity of the motor under test transmitted from the edge end. Wherein, the effective value of the original vibration velocity After the edge end acquires and buffers the raw acceleration signal of the motor under test at a preset sampling frequency, it extracts the acceleration signal from the corresponding raw acceleration signal according to a preset time window; the raw acceleration signal sequence buffered by the edge end is periodically acquired, and the time-domain velocity reconstruction signal is obtained through conjugate symmetric frequency domain integration, and the standard vibration velocity effective value is calculated. Based on the aforementioned standard vibration velocity effective value Compared with the original effective value of vibration velocity The correction coefficients are obtained. The correction coefficient The data is sent to the edge device so that the edge device can apply the correction coefficient. Dynamic calibration is performed on the real-time output vibration velocity. In this embodiment, the original acceleration signal is locally acquired and buffered in real time at the edge, and the effective value of vibration velocity is extracted in seconds. This is combined with the server periodically acquiring the original signal sequence and completing high-precision velocity reconstruction based on conjugate symmetric frequency domain integration. This constructs a dual-channel collaborative monitoring mechanism of coarse edge measurement and fine center calibration, effectively balancing the low-power operation requirements of the edge and the accuracy requirements of equipment vibration monitoring. Dynamic correction coefficients are generated by comparing the deviation between the standard reconstructed effective value of vibration velocity and the effective value output locally at the edge. Real-time calibration and correction of the edge measurement results are completed through a parameter-download closed-loop method, which can continuously offset the complexities of industrial processes. Under the condition of accelerometer performance degradation leading to monotonic drift and attenuation errors, this method achieves in-situ online self-healing correction of sensor measurement accuracy from an operational mechanism perspective. This eliminates the need for downtime calibration and additional standard excitation equipment, significantly reducing on-site maintenance costs and network transmission bandwidth pressure. At the same time, the vibration velocity values ​​calibrated by this method have higher accuracy and stronger stability, and can accurately match vibration intensity assessment standards for fault threshold determination. This effectively avoids the long-term accuracy deviation, fault misjudgment, and missed judgment problems existing in traditional fixed parameter measurement methods, significantly improving the accuracy, reliability, and long-term stability of vibration condition monitoring and predictive maintenance of rotating machinery.

[0035] like Figure 3 As shown, in one embodiment of this application, obtaining the time-domain velocity reconstruction signal through conjugate symmetric frequency domain integration includes: Frequency domain features are extracted from the original acceleration signal sequence acquired in the current period to obtain the acceleration frequency domain complex spectrum; Based on the conjugate symmetry relationship between positive and negative frequencies, the complex spectrum of the acceleration frequency domain is completed to obtain the complex spectrum of the velocity frequency domain; The inverse Fourier transform of the velocity frequency domain complex spectrum is performed to obtain the time-domain velocity reconstruction signal.

[0036] Optionally, frequency domain feature extraction processing is performed on the original acceleration signal sequence retrieved in the current period. This is achieved through windowing preprocessing (such as Hanning windowing) and Discrete Fourier Transform (FFT) to convert the time-domain acceleration signal into a complex frequency-domain acceleration spectrum A[k] (k=0, ..., N). 1) The frequency components and amplitude characteristics of the original vibration signal are fully preserved. Based on this, and strictly adhering to the conjugate symmetry characteristics of the positive and negative frequencies in the spectrum, the complex spectrum of the unilateral acceleration in the frequency domain is symmetrically completed and corrected. This is then combined with a frequency domain integral operator to complete the computational transformation, resulting in a complete complex spectrum of the velocity in the frequency domain that satisfies the conjugate symmetry condition. (k=0, ..., N) 1) This approach avoids the problems of residual imaginary part and phase distortion caused by unilateral spectrum processing at the source. Finally, it reconstructs the symmetrical velocity frequency domain complex spectrum. (k=0, ..., N) 1) Perform an inverse Fourier transform (IFFT) to restore the frequency domain signal to the time domain signal, ultimately obtaining a high-fidelity time-domain velocity reconstruction signal without amplitude distortion or phase deviation. [n], where, [n]=IFFT( V [ k ]).

[0037] In one embodiment of this application, the step of completing the complex spectrum of the acceleration frequency domain based on the conjugate symmetry relationship of positive and negative frequencies to obtain the complex spectrum of the velocity frequency domain includes: Set the spectral amplitudes corresponding to the DC component and Nyquist frequency component in the complex spectrum of the acceleration frequency domain to zero; For the acceleration spectrum components in the entire positive frequency range, frequency domain integration is performed point by point in combination with the corresponding angular frequency to obtain the velocity spectrum components corresponding to each positive frequency point. For all negative frequency ranges, based on the conjugate symmetry of positive and negative frequencies in the real signal spectrum, the conjugate value of the corresponding positive frequency velocity spectrum is taken and filled to complete all negative frequency velocity spectrum components, forming the velocity frequency domain complex spectrum.

[0038] Optionally, such as Figure 4 As shown, the spectral values ​​corresponding to the DC component and the Nyquist frequency component in the complex spectrum of acceleration in the frequency domain are uniformly set to zero: V[0]=0, V[ ]=0; Eliminate the influence of low-frequency DC offset and high-frequency aliasing interference on the integration results, and ensure the effectiveness of vibration velocity solution.

[0039] For the acceleration spectrum components across the entire positive frequency range, a point-by-point frequency domain integration is performed using the corresponding angular frequency parameters to accurately obtain the velocity spectrum components at each positive frequency position. For example, the integral of the positive frequency components (k=1,2, ..., ...) 1) Corresponding angular frequency =2πk Then its velocity spectrum is: V[k] = Where j represents the imaginary unit, This represents the sampling frequency, where N is the number of sampling points; Then, utilizing the inherent conjugate symmetry property of positive and negative frequencies in the spectrum of real physical time-domain signals, spectral completion is performed on all negative frequency intervals. This involves sequentially taking the conjugate value of the corresponding positive frequency velocity spectrum and filling it into the corresponding negative frequency position, thus completely completing the negative frequency spectral components. For example, the conjugate construction of negative frequency components (k= +1, ..., N 1) Assignment based on the conjugate symmetry theorem: To ensure that the frequency domain velocity spectrum meets the requirements This avoids the residual imaginary part in the inverse transform time-domain velocity signal caused by traditional one-sided integration.

[0040] like Figure 3 As shown, in one embodiment of this application, the step of extracting frequency domain features from the original acceleration signal sequence acquired in the current period to obtain the acceleration frequency domain complex spectrum includes: The original acceleration signal sequence acquired in the current period is subjected to Hanning windowing processing to obtain a windowed and smoothed time-domain acceleration sequence; A discrete Fourier transform is performed on the windowed and smoothed time-domain acceleration sequence to generate a complex frequency-domain acceleration spectrum.

[0041] Optionally, a Hanning window function is selected to window the original acceleration signal sequence acquired within the current period. The gradually decaying weights at both ends of the Hanning window smooth the original signal, effectively suppressing the spectral leakage problem caused by finite-length signal truncation, and obtaining a time-domain acceleration sequence with smooth waveform transition and faithful frequency characteristics. The windowed and smoothed time-domain acceleration sequence... [n] can be specifically represented by the following formula: [n] = a[n] × w[n]; Among them, the window function The formula is expressed as: .

[0042] After completing the windowing preprocessing, the windowed and smoothed time-domain acceleration sequence [n] Perform an N-point Fast Fourier Transform (FFT) to convert the time-domain vibration signal into a complex frequency-domain acceleration spectrum A[k] (k=0, ..., N) containing complete amplitude, frequency, and phase information. 1) to accurately analyze the frequency components of the motor vibration signal.

[0043] In one embodiment of this application, the conversion to obtain the effective value of the standard vibration velocity includes: Calculate the root mean square time of the reconstructed velocity signal in the time domain; The time-domain root mean square is energy equivalently corrected based on a preset energy recovery compensation coefficient to obtain the effective value of the standard vibration velocity.

[0044] Optionally, since the velocity frequency domain complex spectrum strictly satisfies the conjugate symmetry characteristic of positive and negative frequencies, the time-domain velocity reconstruction signal obtained by inverse Fourier transform... The imaginary part of [n] is always zero due to strict symmetry, and the output signal is a pure real number waveform. This is the time-domain velocity reconstruction signal. [n] Perform root mean square (RMS) calculation to obtain the original time-domain velocity RMS value. Since the initial signal windowing process inevitably causes overall signal energy attenuation, an energy recovery compensation constant can be introduced to eliminate this inherent error and achieve lossless energy solution. Equivalent energy correction is applied to the initial root mean square value in the time domain to accurately compensate for the amplitude and energy loss caused by windowing, ultimately outputting a high-precision, standard-compliant corrected effective value of vibration velocity. Specifically, it can be expressed by the following formula: = × = × .

[0045] In one embodiment of this application, obtaining the correction coefficient based on the standard effective value of vibration velocity and the original effective value of vibration velocity includes: Based on the ratio between the standard effective value of vibration velocity and the original effective value of vibration velocity, the instantaneous deviation coefficient corresponding to the current cycle is obtained; Median filtering is performed on multiple sets of instantaneous deviation coefficients within a preset sliding window to obtain stable correction coefficients.

[0046] Optionally, the original effective value of vibration velocity from the low-precision calculation output at the edge can be obtained. The standard vibration velocity effective value obtained by high-precision reconstruction through conjugate symmetric frequency domain integration, and itself. The instantaneous deviation coefficient is obtained by comparing and calculating the deviation of two sets of effective vibration velocity values ​​from the same acquisition period and the same equipment measurement point. Specifically, it is expressed as: = ; To mitigate abnormal coefficient disturbances caused by random vibrations and instantaneous acquisition noise, and to improve the long-term stability of the correction parameters, the system sets a fixed-length sliding time window, storing the instantaneous deviation coefficients generated over multiple periods into a window queue. A median filtering algorithm is then used to sort all instantaneous deviation coefficients within the window and take the median value, eliminating outliers that deviate from the normal range. The final output is a robust correction coefficient with stronger anti-interference capabilities and smoother numerical changes. This ensures that the correction parameters sent to the edge will not change drastically due to a single accidental error, thus maintaining the stable and reliable accuracy of edge vibration measurement calibration. Among these, the robust correction coefficient... Specifically, this can be expressed by the following formula: = median({ L+1, ..., }).

[0047] The central server receives the correction coefficient. After that, you can Send it back to the edge so that the edge can use it for subsequent calculations. Perform gain calibration alignment to prevent monotonic attenuation of measurement values ​​caused by sensor degradation. = × ; Furthermore, during subsequent real-time vibration monitoring, this calibrated effective value will be used. The system compares the vibration alarm threshold with the pre-configured vibration intensity standard conforming to ISO 10816 in real time and executes the over-limit judgment logic. When the effective value of the calibrated vibration velocity is greater than the preset alarm threshold, the system immediately generates and pushes an abnormal motor vibration alarm signal, which makes it easier for maintenance personnel to identify early equipment failures such as bearing wear and rotor imbalance in a timely manner, and realize early warning of rotating machinery failures.

[0048] In this embodiment, a strict positive and negative frequency conjugate symmetric joint frequency domain integration mechanism is adopted to fully preserve the discrete spectrum symmetry characteristics, enabling the reconstructed time-domain velocity signal to possess pure real-valued physical properties. This fundamentally avoids the phase delay, waveform distortion, and amplitude distortion problems caused by discarding the imaginary part in traditional schemes, effectively improving the fidelity of the vibration velocity signal reconstruction. Simultaneously, to address the signal energy attenuation and velocity RMS calculation deviation caused by Hanning window preprocessing, a dedicated energy recovery constant is introduced to complete lossless energy correction across the entire process. This accurately offsets the amplitude attenuation error caused by window truncation, corrects the accumulated deviation throughout the signal processing flow, significantly improves the accuracy of vibration intensity calculation, and ensures that the equipment vibration state rating and fault threshold determination fully conform to ISO standards. Adhering to the 10816 international standard, this application effectively reduces the probability of false alarms and missed alarms. Furthermore, it innovatively constructs a low-bandwidth closed-loop self-calibration architecture of "second-level real-time monitoring at the edge + minute-level high-precision benchmark calibration at the center." Relying on the high-precision algorithm at the center, robust correction coefficients are generated and lightweightly distributed to the edge nodes. This achieves dynamic self-healing calibration of sensor measurement parameters without requiring equipment shutdown or external standard vibration table for physical excitation, greatly reducing network bandwidth consumption. It effectively offsets the monotonic decay and numerical drift of measurement values ​​caused by sensor performance degradation under complex industrial conditions, ensuring the stability and accuracy of long-term vibration monitoring data of the equipment. This perfectly adapts to industrial-scale, unmanned predictive maintenance application scenarios.

[0049] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0050] Example 2 In one embodiment, a dual-channel frequency domain integration calibration device for motor vibration velocity is provided, which corresponds one-to-one with the dual-channel frequency domain integration calibration method for motor vibration velocity described in the above embodiments. For example... Figure 5 As shown, the dual-channel frequency domain integration calibration device for motor vibration velocity includes a raw vibration velocity RMS value acquisition unit 10, a standard vibration velocity RMS value acquisition unit 20, a correction coefficient generation unit 30, and a calibration unit 40. Detailed descriptions of each functional module are as follows: The original vibration velocity effective value acquisition unit 10 is used to acquire the original vibration velocity effective value corresponding to the motor under test sent by the edge end. The original vibration velocity effective value is obtained by the edge end collecting and buffering the original acceleration signal of the motor under test according to a preset sampling frequency, and then extracting it from the corresponding original acceleration signal according to a preset time window. The standard vibration velocity effective value acquisition unit 20 is used to periodically acquire the original acceleration signal sequence buffered at the edge end, obtain the time-domain velocity reconstruction signal through conjugate symmetric frequency domain integration, and convert it to obtain the standard vibration velocity effective value. The correction coefficient generation unit 30 is used to obtain a correction coefficient based on the standard effective value of vibration velocity and the original effective value of vibration velocity; The calibration unit 40 is used to send the correction coefficient to the edge end so that the edge end can dynamically calibrate the real-time output vibration velocity based on the correction coefficient.

[0051] In one embodiment of this application, the standard vibration velocity effective value acquisition unit 20 is further configured to: Frequency domain features are extracted from the original acceleration signal sequence acquired in the current period to obtain the acceleration frequency domain complex spectrum; Based on the conjugate symmetry relationship between positive and negative frequencies, the complex spectrum of the acceleration frequency domain is completed to obtain the complex spectrum of the velocity frequency domain; The inverse Fourier transform of the velocity frequency domain complex spectrum is performed to obtain the time-domain velocity reconstruction signal.

[0052] In one embodiment of this application, the standard vibration velocity effective value acquisition unit 20 is further configured to: Set the spectral amplitudes corresponding to the DC component and Nyquist frequency component in the complex spectrum of the acceleration frequency domain to zero; For the acceleration spectrum components in the entire positive frequency range, frequency domain integration is performed point by point in combination with the corresponding angular frequency to obtain the velocity spectrum components corresponding to each positive frequency point. For all negative frequency ranges, based on the conjugate symmetry of positive and negative frequencies in the real signal spectrum, the conjugate value of the corresponding positive frequency velocity spectrum is taken and filled to complete all negative frequency velocity spectrum components, forming the velocity frequency domain complex spectrum.

[0053] In one embodiment of this application, the standard vibration velocity effective value acquisition unit 20 is further configured to: The original acceleration signal sequence acquired in the current period is subjected to Hanning windowing processing to obtain a windowed and smoothed time-domain acceleration sequence; A discrete Fourier transform is performed on the windowed and smoothed time-domain acceleration sequence to generate a complex frequency-domain acceleration spectrum.

[0054] In one embodiment of this application, the standard vibration velocity effective value acquisition unit 20 is further configured to: Calculate the root mean square time of the reconstructed velocity signal in the time domain; The time-domain root mean square is energy equivalently corrected based on a preset energy recovery compensation coefficient to obtain the effective value of the standard vibration velocity.

[0055] In one embodiment of this application, the correction coefficient generation unit 30 is further configured to: Based on the ratio between the standard effective value of vibration velocity and the original effective value of vibration velocity, the instantaneous deviation coefficient corresponding to the current cycle is obtained; Median filtering is performed on multiple sets of instantaneous deviation coefficients within a preset sliding window to obtain stable correction coefficients.

[0056] In one embodiment of this application, the effective value of the original vibration velocity is extracted by the edge end after performing second-level correction of the integral zero-point drift based on a non-iterative algebraic dot product analytical algorithm. The second-level correction includes: When the current motor under test is not stopped, the dead zone gate is updated in the time domain. Then, in the frequency domain, algebraic dot multiplication based on Tikhonov regularization is performed on the original acceleration signal sequence to reconstruct the velocity spectrum. If the energy ratio in the low-frequency cutoff section is not less than a preset ratio, a hard rejection is performed to discard the current calculation. If the energy ratio in the low-frequency cutoff section is less than a preset ratio, the time-domain correction speed is output based on Hanning window incremental compensation.

[0057] In this embodiment, a strict positive and negative frequency conjugate symmetric joint frequency domain integration mechanism is adopted to fully preserve the discrete spectrum symmetry characteristics, enabling the reconstructed time-domain velocity signal to possess pure real-valued physical properties. This fundamentally avoids the phase delay, waveform distortion, and amplitude distortion problems caused by discarding the imaginary part in traditional schemes, effectively improving the fidelity of the vibration velocity signal reconstruction. Simultaneously, to address the signal energy attenuation and velocity RMS calculation deviation caused by Hanning window preprocessing, a dedicated energy recovery constant is introduced to complete lossless energy correction across the entire process. This accurately offsets the amplitude attenuation error caused by window truncation, corrects the accumulated deviation throughout the signal processing flow, significantly improves the accuracy of vibration intensity calculation, and ensures that the equipment vibration state rating and fault threshold determination fully conform to ISO standards. Adhering to the 10816 international standard, this application effectively reduces the probability of false alarms and missed alarms. Furthermore, it innovatively constructs a low-bandwidth closed-loop self-calibration architecture of "second-level real-time monitoring at the edge + minute-level high-precision benchmark calibration at the center." Relying on the high-precision algorithm at the center, robust correction coefficients are generated and lightweightly distributed to the edge nodes. This achieves dynamic self-healing calibration of sensor measurement parameters without requiring equipment shutdown or external standard vibration table for physical excitation, greatly reducing network bandwidth consumption. It effectively offsets the monotonic decay and numerical drift of measurement values ​​caused by sensor performance degradation under complex industrial conditions, ensuring the stability and accuracy of long-term vibration monitoring data of the equipment. This perfectly adapts to industrial-scale, unmanned predictive maintenance application scenarios.

[0058] Specific limitations regarding the dual-channel frequency domain integration calibration device for motor vibration velocity can be found in the limitations of the dual-channel frequency domain integration calibration method for motor vibration velocity described above, and will not be repeated here. Each module in the aforementioned dual-channel frequency domain integration calibration device for motor vibration velocity can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0059] Example 3 In one embodiment, a dual-channel frequency domain integration calibration system for motor vibration velocity is provided, such as... Figure 1 As shown, the dual-channel frequency domain integration calibration system for motor vibration velocity includes: the system includes a central server and at least one edge terminal; each edge terminal is rigidly mounted on the bearing housing of the motor under test and establishes a communication connection with the central server; The edge terminal is used to acquire and locally cache the original acceleration signal of the motor under test at a preset sampling frequency, perform lightweight operation on the cached signal according to a preset time window to extract the original vibration velocity effective value, and upload the original vibration velocity effective value to the central server. When performing lightweight operation on the cached signal, the edge terminal determines whether the state of the motor under test is stopped based on the zero-speed update dead zone gate in the time domain. If it is, it forces the output of the time domain velocity reconstruction signal to be zero. Otherwise, it performs frequency domain Tikhonov regularized algebraic dot product reconstruction on the original acceleration signal. When the proportion of low-frequency cutoff energy is not less than a preset ratio, it triggers integral hard rejection to discard the current calculation. When the proportion of low-frequency cutoff energy is less than a preset ratio, it outputs the original vibration velocity effective value based on Hanning window incremental compensation. The central server includes the dual-channel frequency domain integration calibration device for motor vibration velocity described in Embodiment 2 above. It is used to periodically acquire the original acceleration signal sequence buffered at the edge end, reconstruct the time domain velocity signal through the conjugate symmetric frequency domain integration algorithm, and convert it into the standard vibration velocity effective value. Then, it calculates the correction coefficient by combining the standard vibration velocity effective value with the original vibration velocity effective value uploaded by the edge end, and sends the correction coefficient to the corresponding edge end, so that the edge end can perform dynamic calibration on the real-time output vibration velocity measurement value according to the correction coefficient.

[0060] It should be noted that the edge end includes an edge sentinel module, which includes a data acquisition device (such as an accelerometer) and a processor. The data acquisition device acquires the raw time-domain triaxial acceleration signal of the motor under test at a preset acquisition frequency, and the processor performs lightweight operations on the buffered signal according to a preset time window to extract the effective value of the raw vibration velocity. The original vibration velocity effective value is transmitted via a wireless transmission module, such as industrial wired Ethernet or industrial wireless communication networks (LoRa / NB-IoT / industrial 5G). The data is uploaded to the central server. This central server includes a high-performance benchmark module and a self-calibration module. The high-performance benchmark module periodically acquires the raw acceleration signal sequence cached at the edge, reconstructs the time-domain velocity signal using a conjugate symmetric frequency domain integration algorithm, and then calculates the effective value of the standard vibration velocity. The self-calibration calculation module is used to combine the standard vibration velocity RMS value. Compared with the original vibration velocity RMS value uploaded from the edge end Calculate the correction factor .

[0061] In one embodiment, a computer device is provided, which may be a terminal device, and its internal structure diagram may be as follows: Figure 6As shown, the computer device includes a processor, memory, and network interface connected via a system bus. The processor provides computational and control capabilities. The memory includes a readable storage medium storing computer-readable instructions. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer-readable instructions implement a dual-channel frequency domain integration calibration method for motor vibration velocity. The readable storage medium provided in this embodiment includes both non-volatile and volatile readable storage media.

[0062] In this application embodiment, a computer device is provided, including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor. When the processor executes the computer-readable instructions, it implements the steps of the dual-channel frequency domain integration calibration method for motor vibration velocity described above.

[0063] In this embodiment of the application, a readable storage medium is provided, which stores computer-readable instructions. When the computer-readable instructions are executed by a processor, they implement the steps of the dual-channel frequency domain integration calibration method for motor vibration velocity described above.

[0064] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing related hardware with computer-readable instructions. These computer-readable instructions can be stored in a non-volatile readable storage medium or a volatile readable storage medium. When executed, these computer-readable instructions can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0065] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0066] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A dual-channel frequency domain integration calibration method for motor vibration velocity, characterized in that, Applied to a central server, the method includes: The effective value of the original vibration velocity corresponding to the motor under test sent by the edge terminal is obtained. The effective value of the original vibration velocity is obtained by the edge terminal collecting and buffering the original acceleration signal of the motor under test according to a preset sampling frequency, and then extracting it from the corresponding original acceleration signal according to a preset time window. The original acceleration signal sequence of the edge buffer is periodically acquired, and the time-domain velocity reconstruction signal is obtained by conjugate symmetric frequency domain integration, and the effective value of standard vibration velocity is calculated. Based on the standard effective value of vibration velocity and the original effective value of vibration velocity, a correction coefficient is obtained; The correction coefficient is sent to the edge end so that the edge end can dynamically calibrate the real-time output vibration velocity based on the correction coefficient.

2. The dual-channel frequency domain integration calibration method for motor vibration velocity as described in claim 1, characterized in that, The original effective value of vibration velocity is obtained by second-level correction of integral zero-point drift at the edge end based on a non-iterative algebraic dot product analytical algorithm. The second-level correction includes: Determine whether the motor under test is currently in a stopped state by using time-domain zero-speed update dead-zone gating; If so, then force the output time-domain velocity reconstruction signal to be zero; If not, then perform a Tikhonov-regularized algebraic dot product operation on the original acceleration signal sequence in the frequency domain to reconstruct the velocity spectrum; When the energy percentage in the low-frequency cutoff section is not lower than the preset ratio, the integral hard rejection logic is triggered to avoid integral saturation. When the energy proportion in the low-frequency cutoff section is lower than the preset ratio, the Hanning window is used in the time domain to incrementally compensate the reconstructed time-domain velocity signal and output the original effective value of vibration velocity.

3. The dual-channel frequency domain integration calibration method for motor vibration velocity as described in claim 1, characterized in that, The process of obtaining the time-domain velocity reconstruction signal through conjugate symmetric frequency domain integration includes: Frequency domain features are extracted from the original acceleration signal sequence acquired in the current period to obtain the acceleration frequency domain complex spectrum; Based on the conjugate symmetry relationship between positive and negative frequencies, the complex spectrum of the acceleration frequency domain is completed to obtain the complex spectrum of the velocity frequency domain; The inverse Fourier transform of the velocity frequency domain complex spectrum is performed to obtain the time-domain velocity reconstruction signal.

4. The dual-channel frequency domain integration calibration method for motor vibration velocity as described in claim 3, characterized in that, The process of completing the complex spectrum of acceleration in the frequency domain based on the conjugate symmetry relationship of positive and negative frequencies to obtain the complex spectrum of velocity in the frequency domain includes: Set the spectral amplitudes corresponding to the DC component and Nyquist frequency component in the complex spectrum of the acceleration frequency domain to zero; For the acceleration spectrum components in the entire positive frequency range, frequency domain integration is performed point by point in combination with the corresponding angular frequency to obtain the velocity spectrum components corresponding to each positive frequency point. For all negative frequency intervals, based on the conjugate symmetry characteristic of positive and negative frequencies in the spectrum of real numbers, the conjugate value of the corresponding positive frequency velocity spectrum is taken and filled to complete all negative frequency velocity spectrum components, forming the velocity frequency domain complex spectrum.

5. The dual-channel frequency domain integration calibration method for motor vibration velocity as described in claim 3, characterized in that, The step of extracting frequency domain features from the original acceleration signal sequence acquired in the current period to obtain the complex frequency spectrum of acceleration includes: The original acceleration signal sequence acquired in the current period is subjected to Hanning windowing processing to obtain a windowed and smoothed time-domain acceleration sequence; A discrete Fourier transform is performed on the windowed and smoothed time-domain acceleration sequence to generate the acceleration frequency domain complex spectrum.

6. The dual-channel frequency domain integration calibration method for motor vibration velocity as described in claim 1, characterized in that, The conversion to obtain the standard vibration velocity effective value includes: Calculate the root mean square time of the reconstructed velocity signal in the time domain; The time-domain root mean square is energy equivalently corrected based on a preset energy recovery compensation coefficient to obtain the effective value of the standard vibration velocity.

7. The dual-channel frequency domain integration calibration method for motor vibration velocity as described in claim 1, characterized in that, The correction coefficient, obtained based on the standard effective value of vibration velocity and the original effective value of vibration velocity, includes: Based on the ratio between the standard effective value of vibration velocity and the original effective value of vibration velocity, the instantaneous deviation coefficient corresponding to the current cycle is obtained; Median filtering is performed on multiple sets of instantaneous deviation coefficients within a preset sliding window to obtain stable correction coefficients.

8. A dual-channel frequency domain integration calibration device for motor vibration velocity, characterized in that, The device, applied to a central server, includes: The original vibration velocity effective value acquisition unit is used to acquire the original vibration velocity effective value corresponding to the motor under test sent by the edge end. The original vibration velocity effective value is obtained by the edge end collecting and buffering the original acceleration signal of the motor under test according to a preset sampling frequency, and then extracting it from the corresponding original acceleration signal according to a preset time window. The standard vibration velocity effective value acquisition unit is used to periodically acquire the original acceleration signal sequence buffered at the edge end, reconstruct the time-domain velocity signal through conjugate symmetric frequency domain integration, and convert it to obtain the standard vibration velocity effective value. The correction coefficient generation unit is used to obtain the correction coefficient based on the standard effective value of vibration velocity and the original effective value of vibration velocity; A calibration unit is used to send the correction coefficient to the edge end so that the edge end can dynamically calibrate the real-time output vibration velocity based on the correction coefficient.

9. The dual-channel frequency domain integration calibration device for motor vibration velocity as described in claim 8, characterized in that, The original effective value of vibration velocity is extracted by the edge end after performing second-level correction of the integral zero-point drift based on a non-iterative algebraic dot product analytical algorithm. The second-level correction includes: When the time-domain zero-speed update dead-zone gating determines that the motor under test is not in a stopped state, the frequency domain performs an algebraic dot product operation based on Tikhonov regularization on the original acceleration signal sequence to reconstruct the velocity spectrum. When the energy proportion in the low-frequency cutoff section is not less than a preset ratio, an integral hard rejection is triggered to discard the current calculation. When the energy proportion in the low-frequency cutoff section is less than a preset ratio, the time-domain correction speed is output based on Hanning window incremental compensation.

10. A dual-channel frequency domain integration calibration system for motor vibration velocity, characterized in that, The system includes: a central server and at least one edge terminal; each edge terminal is rigidly mounted on the bearing housing of the motor under test and establishes a communication connection with the central server; The edge terminal is used to acquire and locally cache the original acceleration signal of the motor under test at a preset sampling frequency, perform lightweight operation on the cached signal according to a preset time window to extract the original vibration velocity effective value, and upload the original vibration velocity effective value to the central server. When performing lightweight operation on the cached signal, the edge terminal determines whether the state of the motor under test is stopped based on the zero-speed update dead zone gate in the time domain. If it is, it forces the output of the time domain velocity reconstruction signal to be zero. Otherwise, it performs frequency domain Tikhonov regularized algebraic dot product reconstruction on the original acceleration signal. When the proportion of low-frequency cutoff energy is not less than a preset ratio, it triggers integral hard rejection to discard the current calculation. When the proportion of low-frequency cutoff energy is less than a preset ratio, it outputs the original vibration velocity effective value based on Hanning window incremental compensation. The central server includes a dual-channel frequency domain integration calibration device for motor vibration velocity as described in claim 8 or 9. This device is used to periodically acquire the original acceleration signal sequence cached at the edge end, reconstruct the time-domain velocity signal using a conjugate symmetric frequency domain integration algorithm, and convert it into a standard effective value of vibration velocity. Then, it calculates a correction coefficient by combining the standard effective value of vibration velocity with the original effective value of vibration velocity uploaded by the edge end, and sends the correction coefficient to the corresponding edge end, so that the edge end can perform dynamic calibration on the real-time output vibration velocity measurement value based on the correction coefficient.

Images