Method and apparatus for suppressing fundamental current component based on particle filter

Abstract

The application discloses a kind of based on particle filter's method and device for inhibiting current fundamental component, belong to motor fault diagnosis technical field, method includes: based on the stator line current signal of acquisition, construct for estimating particle swarm of fundamental component, each particle in particle swarm has two parameters of amplitude and initial phase, two parameters meet preset normal distribution and uniform distribution respectively;Pass through the distribution of propagation equation and measurement value to particle swarm is propagated and weight is updated, estimate the state value of stator line current fundamental component at different sampling time, obtain fundamental component estimate value;The stator line current signal of acquisition is subtracted fundamental component estimate value, obtains the current signal after inhibiting fundamental component.The application realizes accurate estimation of fundamental component based on particle filter by constructing suitable particle swarm, effectively inhibits the fundamental component in stator line current, improves the identifiable nature of fault characteristic current harmonic, improves the reliability of asynchronous motor fault diagnosis.

Description

Technical Field

[0001] This invention relates to a method and apparatus for suppressing the fundamental frequency component of current based on a particle filter, belonging to the field of motor fault diagnosis technology. Background Technology

[0002] Asynchronous motors may experience various faults during operation, such as rotor bar breakage and bearing damage. These faults can lead to three-phase imbalance and air gap changes, introducing fault-related characteristic current harmonics into the stator line current. By monitoring the presence and amplitude of these fault characteristic current harmonics in the stator line current, the type and severity of the asynchronous motor fault can be determined. However, in actual spectrum analysis, due to spectral leakage of the fundamental frequency component contained in the stator line current, the spectral peaks of the fault characteristic current harmonics are often covered by the broadened spectral peaks of the fundamental frequency component, making it impossible to effectively identify the peaks of the fault characteristic current harmonics. Therefore, suppressing the fundamental frequency component contained in the stator line current is of great significance for improving the reliability of motor current signal analysis in diagnosing asynchronous motor faults.

[0003] In actual operation, asynchronous motors experience fluctuations in parameters such as grid voltage, internal motor temperature, and operating conditions, leading to strong nonlinear characteristics in the stator current signal. Particle filters, a filtering technique that estimates the target state by directly fitting the probability distribution of the target state, possess unique advantages in processing nonlinear signals and have been widely used in fields such as robot navigation and computer vision. However, their application in estimating fundamental frequency current signals is limited, primarily due to the difficulty in constructing a particle swarm capable of fitting the probability distribution of a sinusoidal signal state. Therefore, new solutions are needed to suppress the fundamental frequency component in the stator current signal using particle filters. Summary of the Invention

[0004] To address the aforementioned problems, this invention proposes a method and apparatus for suppressing the fundamental frequency component of current based on a particle filter, which can improve the reliability of asynchronous motor fault diagnosis.

[0005] The technical solution adopted by this invention to solve its technical problem is as follows: In a first aspect, embodiments of the present invention provide a method for suppressing the fundamental frequency component of current based on a particle filter, comprising the following steps: Step S1: Based on the acquired stator line current signal, construct a particle swarm for estimating the fundamental frequency component. The particle swarm includes multiple particles, each particle having two parameters: amplitude and initial phase. The amplitude of the fundamental frequency component satisfies a preset normal distribution, and the initial phase of the fundamental frequency component satisfies a preset uniform distribution. Step S2: Propagate and update the particle swarm using the propagation equation and the distribution of measured values, estimate the state value of the fundamental frequency component of the stator line current at different sampling times, and obtain the estimated value of the fundamental frequency component. Step S3: Subtract the estimated value of the fundamental frequency component from the acquired stator line current signal to obtain the current signal after suppressing the fundamental frequency component.

[0006] As one possible implementation of this embodiment, step S1 includes the following steps: Step S11: Based on the stator line current signal collected within a preset time period, obtain the maximum value of the stator line current signal within the preset time period. and minimum value Calculate the mean estimate of the fundamental frequency component amplitude: ; Step S12, based on the mean estimate Standard deviation of the preset fundamental frequency component amplitude : ; Step S13, based on the mean estimate and standard deviation Randomly selected from a normal distribution Amplitude samples of each fundamental frequency component , ,in, It is a preset positive integer. ; Step S14, from the interval is Randomly selected from a uniform distribution Initial phase sample of each fundamental frequency component , ,in, It is a preset positive integer. ; Step S15, based on amplitude samples and initial phase sample Constructing particle swarms; Step S16: Assign initial weights to each particle in the particle swarm. : ,in, The total number of particles, .

[0007] As one possible implementation of this embodiment, the construction of the particle swarm includes: Based on amplitude samples and initial phase sample Through formula Constructing particles , , , .

[0008] As one possible implementation of this embodiment, step S2 includes the following steps: Step S21: Establish a particle propagation model and calculate the prior estimate of each particle at the next moment. Step S22: Update the weights of each particle based on the measured values ​​and the measurement noise model; Step S23: Calculate the posterior estimate of the fundamental frequency component based on the updated weights.

[0009] As one possible implementation of this embodiment, the particle propagation model is as follows: , in, Let be the prior estimate of the fundamental frequency component of the stator current obtained based on the k-th particle at the m-th sampling time. and These represent the amplitude and phase angle of the fundamental frequency component corresponding to the k-th particle, respectively. The fundamental frequency of the motor current is . Sampling frequency, This is the sequence number of the sampled sequence (initially 1). This is the noise term.

[0010] As one possible implementation of this embodiment, the noise item The mean is 0 and the standard deviation is 0. The normal distribution, presuppose .

[0011] As one possible implementation of this embodiment, the formula for updating the weights of each particle is: , in, , In the formula, Let be the weight of the k-th particle at the m-th sampling time. This represents the measured value of the fundamental frequency component of the stator line current at the m-th sampling time; for the state of the k-th particle at the m-th sampling time, Satisfying the mean is The standard deviation is normal distribution (preset) ); This represents the weight of the k-th particle at the (m-1)-th sampling time.

[0012] As one possible implementation of this embodiment, the calculation of the posterior estimate of the fundamental frequency component includes: The posterior estimate of the fundamental frequency component of the stator line current at the m-th sampling time is calculated by weighted averaging. : , in, The total number of particles, Let be the weight of the k-th particle at the m-th sampling time.

[0013] As one possible implementation of this embodiment, the method for suppressing the fundamental frequency component of current based on a particle filter further includes the following steps: Step S4: Extract fault features from the current signal after suppressing the fundamental frequency component to achieve fault diagnosis of the asynchronous motor.

[0014] Secondly, embodiments of the present invention provide a device for suppressing the fundamental frequency component of a current based on a particle filter, comprising: The particle swarm construction module is used to construct a particle swarm for estimating the fundamental frequency component based on the acquired stator line current signal. The particle swarm includes multiple particles, each of which has two parameters: amplitude and initial phase. The amplitude of the fundamental frequency component satisfies a preset normal distribution, and the initial phase of the fundamental frequency component satisfies a preset uniform distribution. The particle swarm propagation module is used to propagate the particle swarm and update its weights through the propagation equation and the distribution of measured values, estimate the state value of the fundamental frequency component of the stator line current at different sampling times, and obtain the estimated value of the fundamental frequency component. The fundamental frequency component suppression module is used to subtract the estimated value of the fundamental frequency component from the acquired stator line current signal to obtain the current signal after suppressing the fundamental frequency component.

[0015] As one possible implementation of this embodiment, the device for suppressing the fundamental frequency component of the current based on a particle filter further includes: The fault feature extraction module is used to extract fault features from the current signal after suppressing the fundamental frequency component, so as to realize the fault diagnosis of asynchronous motor.

[0016] The beneficial effects of the technical solutions of the embodiments of the present invention are as follows: This invention achieves accurate estimation of the fundamental frequency component based on a particle filter by constructing a suitable particle swarm. This effectively suppresses the fundamental frequency component in the stator line current, thereby improving the identifiability of fault characteristic current harmonics and enhancing the reliability of asynchronous motor fault diagnosis. This method is suitable for signal environments with strong nonlinearity and has high practical value and promising prospects for wider application. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a method for suppressing the fundamental frequency component of current based on a particle filter, according to an exemplary embodiment. Figure 2 This is a schematic diagram of a device for suppressing the fundamental frequency component of a current based on a particle filter, according to an exemplary embodiment. Detailed Implementation

[0018] To more clearly illustrate the technical features of the present invention, the present invention will be described in detail below through specific embodiments and in conjunction with the accompanying drawings.

[0019] like Figure 1 As shown in the figure, an embodiment of the present invention provides a method for suppressing the fundamental frequency component of current based on a particle filter, comprising the following steps: Step S1: Based on the acquired stator line current signal, construct a particle swarm for estimating the fundamental frequency component. The particle swarm includes multiple particles, each particle having two parameters: amplitude and initial phase. The amplitude of the fundamental frequency component satisfies a preset normal distribution, and the initial phase of the fundamental frequency component satisfies a preset uniform distribution. Step S2: Propagate and update the particle swarm using the propagation equation and the distribution of measured values, estimate the state value of the fundamental frequency component of the stator line current at different sampling times, and obtain the estimated value of the fundamental frequency component. Step S3: Subtract the estimated value of the fundamental frequency component from the acquired stator line current signal to obtain the current signal after suppressing the fundamental frequency component.

[0020] As one possible implementation of this embodiment, step S1 includes the following steps: Step S11: Based on the stator line current signal collected within a preset time period, obtain the maximum value of the stator line current signal within the preset time period. and minimum value Calculate the mean estimate of the fundamental frequency component amplitude: ; Step S12, based on the mean estimate Standard deviation of the preset fundamental frequency component amplitude : ; Step S13, based on the mean estimate and standard deviation Randomly selected from a normal distribution Amplitude samples of each fundamental frequency component , ,in, It is a preset positive integer. ; Step S14, from the interval is Randomly selected from a uniform distribution Initial phase sample of each fundamental frequency component , ,in, It is a preset positive integer. ; Step S15, based on amplitude samples and initial phase sample Constructing particle swarms; Step S16: Assign initial weights to each particle in the particle swarm. : ,in, The total number of particles, .

[0021] As one possible implementation of this embodiment, the construction of the particle swarm includes: Based on amplitude samples and initial phase sample Through formula Constructing particles , , , .

[0022] As one possible implementation of this embodiment, step S2 includes the following steps: Step S21: Establish a particle propagation model and calculate the prior estimate of each particle at the next moment. Step S22: Update the weights of each particle based on the measured values ​​and the measurement noise model; Step S23: Calculate the posterior estimate of the fundamental frequency component based on the updated weights.

[0023] As one possible implementation of this embodiment, the particle propagation model is as follows: , in, Let be the prior estimate of the fundamental frequency component of the stator current obtained based on the k-th particle at the m-th sampling time. and These represent the amplitude and phase angle of the fundamental frequency component corresponding to the k-th particle, respectively. The fundamental frequency of the motor current is . Sampling frequency, This is the sequence number of the sampled sequence (initially 1). This is the noise term.

[0024] As one possible implementation of this embodiment, the noise item The mean is 0 and the standard deviation is 0. The normal distribution, presuppose .

[0025] As one possible implementation of this embodiment, the formula for updating the weights of each particle is: , in, , In the formula, Let be the weight of the k-th particle at the m-th sampling time. This represents the measured value of the fundamental frequency component of the stator line current at the m-th sampling time; for the state of the k-th particle at the m-th sampling time, Satisfying the mean is The standard deviation is normal distribution (preset) ); This represents the weight of the k-th particle at the (m-1)-th sampling time.

[0026] As one possible implementation of this embodiment, the calculation of the posterior estimate of the fundamental frequency component includes: The posterior estimate of the fundamental frequency component of the stator line current at the m-th sampling time is calculated by weighted averaging. : , in, The total number of particles, Let be the weight of the k-th particle at the m-th sampling time.

[0027] As one possible implementation of this embodiment, the method for suppressing the fundamental frequency component of current based on a particle filter further includes the following steps: Step S4 involves extracting fault features from the current signal after suppressing the fundamental frequency component to achieve fault diagnosis of the asynchronous motor. Specifically, first, the current signal after suppressing the fundamental frequency component is subjected to spectral analysis. Then, characteristic frequencies related to specific faults are identified. Next, the current amplitude at the characteristic frequencies is extracted. Finally, based on the comparison between the characteristic amplitude and a preset threshold, the fault type and severity are determined, thereby achieving fault diagnosis of the asynchronous motor.

[0028] like Figure 2 As shown in the figure, an embodiment of the present invention provides a device for suppressing the fundamental frequency component of current based on a particle filter, comprising: The particle swarm construction module is used to construct a particle swarm for estimating the fundamental frequency component based on the acquired stator line current signal. The particle swarm includes multiple particles, each of which has two parameters: amplitude and initial phase. The amplitude of the fundamental frequency component satisfies a preset normal distribution, and the initial phase of the fundamental frequency component satisfies a preset uniform distribution. The particle swarm propagation module is used to propagate the particle swarm and update its weights through the propagation equation and the distribution of measured values, estimate the state value of the fundamental frequency component of the stator line current at different sampling times, and obtain the estimated value of the fundamental frequency component. The fundamental frequency component suppression module is used to subtract the estimated value of the fundamental frequency component from the acquired stator line current signal to obtain the current signal after suppressing the fundamental frequency component.

[0029] As one possible implementation of this embodiment, the device for suppressing the fundamental frequency component of the current based on a particle filter further includes: The fault feature extraction module is used to extract fault features from the current signal after suppressing the fundamental frequency component, so as to realize the fault diagnosis of asynchronous motor.

[0030] Before suppressing the fundamental frequency component of the stator line current, this invention first determines the core sampling parameter of the stator line current signal: sampling frequency. The sampling frequency is 12.8kHz, and the sampling duration is... The sampling time is 10.24 seconds, and the total number of sampling points is: The specific implementation process of this invention, based on a particle filter, for suppressing the fundamental frequency component of current mainly includes the following key steps.

[0031] Step 1: Construction of the particle swarm: The particles in the particle swarm constructed in this invention have two parameters: amplitude and initial phase.

[0032] 1. Set the amplitude of the fundamental frequency component to satisfy a normal distribution.

[0033] 2. Based on the stator line current signal collected over a period of time (e.g., 2 seconds), obtain an estimate of the mean amplitude of the fundamental frequency component signal. The calculation method is as follows: 1) Find the maximum value of the stator line current signal within 2 seconds. 2) Find the minimum value of the stator current signal within 2 seconds. 3) Perform calculations to obtain .

[0034] 3. Based on Standard deviation of the preset fundamental frequency component amplitude In this embodiment, the preset value is based on empirical values. .

[0035] 4. Based on and Randomly selected Amplitude samples of each fundamental frequency component , ,in, It is a preset positive integer. The present invention is pre-set during implementation. .

[0036] 5. Set the initial phase of the fundamental frequency component to satisfy the interval as follows: The uniform distribution.

[0037] 6. Based on interval Uniform distribution, random sampling Initial phase sample of each fundamental frequency component , ,in, It is a preset positive integer. The present invention is pre-set during implementation. .

[0038] 7. Based on amplitude samples and initial phase sample and preset and Constructing 2500 particles , Its construction method is as follows , , , .

[0039] 8. Set particles initial weights : ,in, The total number of particles, ,Right now .

[0040] Step two, particle swarm propagation: The propagation model of particle swarm optimization is as follows: This represents the m-th sampling time (sampling period is ). ,Right now, (seconds), the state value of the fundamental frequency component of the stator line current.

[0041] For the Particles ( ), The prior estimate is ,in, Let be the prior estimate of the fundamental frequency component of the stator current obtained based on the k-th particle at the m-th sampling time. and These represent the amplitude and phase angle of the fundamental frequency component corresponding to the k-th particle, respectively. The fundamental frequency of the motor current is . Sampling frequency, The sequence number is the number of the sampled sequence (initially 1). The noise term has a mean of 0 and a standard deviation of . The normal distribution, presuppose .

[0042] This represents the measured value of the fundamental frequency component of the stator line current at the m-th sampling time; for the state of the k-th particle at the m-th sampling time, Satisfying the mean is The standard deviation is The normal distribution, presuppose .

[0043] During the propagation of the particle swarm, the weight of each particle changes, and the method for updating the particle weight is as follows: At the (m-1)th sampling time, the kth particle weight To the m-th sampling time, the k-th particle weight For example, the update: , , Calculate the posterior estimate of the fundamental frequency component of the stator line current at the m-th sampling time. : , Thus, the estimation of the fundamental frequency component of the stator line current at the m-th sampling time is completed.

[0044] Step 3, suppression of the fundamental frequency component: By subtracting the estimated value of the fundamental frequency component from the acquired stator current signal, the current signal after suppressing the fundamental frequency component is obtained, thus achieving the suppression (or filtering) of the fundamental frequency component contained in the stator current.

[0045] Step 4, Fault Diagnosis: The current signal after suppressing the fundamental frequency component is obtained. First, the spectrum of the current signal after suppressing the fundamental frequency component is analyzed. Then, the characteristic frequency related to the specific fault is identified. Next, the current amplitude at the characteristic frequency is extracted. Finally, the fault type and severity are determined based on the comparison between the characteristic amplitude and the preset threshold, thereby realizing the fault diagnosis of the asynchronous motor.

[0046] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for suppressing the fundamental frequency component of current based on a particle filter, characterized in that, Includes the following steps: Step S1: Based on the acquired stator line current signal, construct a particle swarm for estimating the fundamental frequency component. The particle swarm includes multiple particles, each particle having two parameters: amplitude and initial phase. The amplitude of the fundamental frequency component satisfies a preset normal distribution, and the initial phase of the fundamental frequency component satisfies a preset uniform distribution. Step S2: Propagate and update the particle swarm using the propagation equation and the distribution of measured values, estimate the state value of the fundamental frequency component of the stator line current at different sampling times, and obtain the estimated value of the fundamental frequency component. Step S3: Subtract the estimated value of the fundamental frequency component from the acquired stator line current signal to obtain the current signal after suppressing the fundamental frequency component; Step S2 includes the following steps: Step S21: Establish a particle propagation model and calculate the prior estimate of each particle at the next moment. Step S22: Update the weights of each particle based on the measured values ​​and the measurement noise model; Step S23: Calculate the posterior estimate of the fundamental frequency component based on the updated weights; The particle propagation model is as follows: ， in, Let be the prior estimate of the fundamental frequency component of the stator current obtained based on the k-th particle at the m-th sampling time. and These represent the amplitude and phase angle of the fundamental frequency component corresponding to the k-th particle, respectively. The fundamental frequency of the motor current is . Where is the sampling frequency, and m is the sequence number of the sampling sequence. V Noise term; The formula for updating the weights of each particle is: ， in, The total number of particles, ， In the formula, Let be the weight of the k-th particle at the m-th sampling time. This represents the measured value of the fundamental frequency component of the stator line current at the m-th sampling time; for the state of the k-th particle at the m-th sampling time, Satisfying the mean is The standard deviation is u The normal distribution; This represents the weight of the k-th particle at the (m-1)-th sampling time.

2. The method for suppressing the fundamental frequency component of current based on a particle filter according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Based on the stator line current signal collected within a preset time period, obtain the maximum value of the stator line current signal within the preset time period. and minimum value Calculate the mean estimate of the fundamental frequency component amplitude: ; Step S12, based on the mean estimate Standard deviation of the preset fundamental frequency component amplitude : ; Step S13, based on the mean estimate and standard deviation Randomly selected from a normal distribution Each fundamental frequency component amplitude sample , ,in, It is a preset positive integer. ; Step S14, from the interval is Randomly selected from a uniform distribution Initial phase sample of each fundamental frequency component , ,in, It is a preset positive integer. ; Step S15, based on amplitude samples and initial phase sample Constructing particle swarms; Step S16: Assign initial weights to each particle in the particle swarm. : ,in, The total number of particles, .

3. The method for suppressing the fundamental frequency component of current based on a particle filter according to claim 2, characterized in that, The constructed particle swarm includes: Based on amplitude samples and initial phase sample Through formula Constructing particles , ; ; .

4. The method for suppressing the fundamental frequency component of current based on a particle filter according to claim 1, characterized in that, The posterior estimate of the fundamental frequency component includes: The posterior estimate of the fundamental frequency component of the stator line current at the m-th sampling time is calculated by weighted averaging. : ， in, The total number of particles, Let be the weight of the k-th particle at the m-th sampling time.

5. The method for suppressing the fundamental frequency component of current based on a particle filter according to any one of claims 1-4, characterized in that, It also includes the following steps: Step S4: Extract fault features from the current signal after suppressing the fundamental frequency component to achieve fault diagnosis of the asynchronous motor.

6. An apparatus for suppressing the fundamental frequency component of current based on a particle filter, used to implement the method for suppressing the fundamental frequency component of current based on a particle filter as described in any one of claims 1 to 5, characterized in that, include: The particle swarm construction module is used to construct a particle swarm for estimating the fundamental frequency component based on the acquired stator line current signal. The particle swarm includes multiple particles, each of which has two parameters: amplitude and initial phase. The amplitude of the fundamental frequency component satisfies a preset normal distribution, and the initial phase of the fundamental frequency component satisfies a preset uniform distribution. The particle swarm propagation module is used to propagate the particle swarm and update its weights through the propagation equation and the distribution of measured values, estimate the state value of the fundamental frequency component of the stator line current at different sampling times, and obtain the estimated value of the fundamental frequency component. The fundamental frequency component suppression module is used to subtract the estimated value of the fundamental frequency component from the acquired stator line current signal to obtain the current signal after suppressing the fundamental frequency component.

7. The device for suppressing the fundamental frequency component of current based on a particle filter according to claim 6, characterized in that, Also includes: The fault feature extraction module is used to extract fault features from the current signal after suppressing the fundamental frequency component, so as to realize the fault diagnosis of asynchronous motor.

Images