CN111458640B - Three-phase current-based alternating current asynchronous motor rotor broken bar fault diagnosis method - Google Patents

Three-phase current-based alternating current asynchronous motor rotor broken bar fault diagnosis method Download PDF

Info

Publication number
CN111458640B
CN111458640B CN202010323423.2A CN202010323423A CN111458640B CN 111458640 B CN111458640 B CN 111458640B CN 202010323423 A CN202010323423 A CN 202010323423A CN 111458640 B CN111458640 B CN 111458640B
Authority
CN
China
Prior art keywords
firefly
algorithm
current
fundamental frequency
asynchronous motor
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010323423.2A
Other languages
Chinese (zh)
Other versions
CN111458640A (en
Inventor
刘飞
李睿彧
梁霖
徐光华
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xian Jiaotong University
Original Assignee
Xian Jiaotong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xian Jiaotong University filed Critical Xian Jiaotong University
Priority to CN202010323423.2A priority Critical patent/CN111458640B/en
Publication of CN111458640A publication Critical patent/CN111458640A/en
Application granted granted Critical
Publication of CN111458640B publication Critical patent/CN111458640B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/34—Testing dynamo-electric machines
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G06F2218/04—Denoising
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction

Abstract

A fault diagnosis method for broken bars of an alternating current asynchronous motor rotor based on three-phase current is characterized in that three-phase current signals of an alternating current asynchronous motor stator are synchronously collected; setting the VMD decomposition number K according to the signal characteristics; changing the fixed moving step length of the firefly in the basic firefly algorithm GSO, initializing the initial position of the firefly based on the chaotic sequence, providing a chaotic variable step length firefly improved algorithm CSVGSO, and determining the fitness function of the analysis current signal; optimizing the VMD decomposition penalty parameter alpha by using CSVGSO; performing VMD decomposition on the three-phase stator current based on the punishment parameter alpha corresponding to the global optimal fitness value and the decomposition number K, and extracting modal components corresponding to the fundamental frequency of the current signal; determining a reference phase of the current signal based on the Park transformation; substituting the obtained phase information into an LMS algorithm, and extracting a fundamental frequency component in the current signal to perform self-adaptive filtering; carrying out spectrum analysis on the filtered current signal to detect the characteristic frequency of the broken bar fault of the motor rotor; the method has the advantages of high accuracy and strong robustness.

Description

Three-phase current-based alternating current asynchronous motor rotor broken bar fault diagnosis method
Technical Field
The invention belongs to the technical field of mechanical fault diagnosis and monitoring, and particularly relates to a three-phase current-based method for diagnosing a broken bar fault of an alternating current asynchronous motor rotor.
Background
At the initial stage of the broken bar fault of the rotor of the alternating current asynchronous motor, the fault characteristic frequency component of the current signal is weak and is easily covered by the fundamental frequency component and difficult to identify, so that the method for inhibiting the current fundamental frequency component is concerned by a plurality of scholars. The LMS adaptive filtering principle is simple, the calculation is convenient, and the LMS adaptive filtering principle can be used for filtering the fundamental frequency component. The filtering of the fundamental frequency component belongs to single-component adaptive filtering, so that the accurate construction of the reference signal is a key technology. Compared with EMD and LMD, VMD decomposition has a solid theoretical basis, and the problems of end effect, mode aliasing and the like in recursive decomposition are avoided, so that VMD can be used for carrying out adaptive decomposition on a current signal. The GSO method is used as a novel meta-heuristic SI optimization algorithm, can simultaneously search a plurality of optimal solutions of multi-modal functions, and is very suitable for processing complex local optimization problems, so that the GSO method can be used for further optimizing VMD decomposition parameters. However, the basic firefly algorithm also has the following defects that the firefly position initialization sampling random mode may generate a large amount of fireflies far away from the optimal solution and at the solution space edge position, and the optimal solution may not be generated. The movement of the firefly in a fixed step length can cause slow convergence and poor optimization of the algorithm at the later stage and oscillation phenomena around the peak value. Therefore, further improvements are also needed for the GSO process.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention aims to provide the alternating current asynchronous motor rotor bar breaking fault diagnosis method based on the three-phase current, and the method has the advantages of high accuracy and strong robustness.
In order to achieve the purpose, the invention adopts the technical scheme that:
a three-phase current-based fault diagnosis method for broken bars of an alternating current asynchronous motor rotor comprises the following steps:
step 1: synchronous acquisition of three-phase current signals i of alternating current asynchronous motor statorU、iV、iW；
Step 2: setting the VMD decomposition number K according to the characteristics of the current signal;
and step 3: changing the fixed moving Step size of the firefly in a basic firefly algorithm GSO (Glowworm Swarm Optimization), initializing the initial position of the firefly based on a chaotic sequence, and providing a Chaos Variable Step size firefly improved algorithm CSVGSO (Chaos Variable Step-size Glowworm Optimization);
and 4, step 4: determining a fitness function;
and 5: optimizing a VMD decomposition penalty parameter alpha by using the CSVGSO algorithm;
step 6: performing VMD decomposition on the stator three-phase current based on the punishment parameter alpha corresponding to the global optimal fitness value and the decomposition number K determined in the step 2, and extracting modal components corresponding to the fundamental frequency;
and 7: determining a reference signal phase based on the Park transformation;
and 8: substituting the phase information obtained in the step 7 into an LMS algorithm, and carrying out self-adaptive filtering on the fundamental frequency component in the extracted modal signal;
and step 9: and carrying out spectrum analysis on the filtered signals to detect the characteristic frequency of the rotor broken bar fault.
And 2, setting the VMD decomposition number K according to the characteristics of the current signal, wherein the specific method comprises the following steps:
because the stator current signal of the alternating current asynchronous motor is expressed in a superposition mode of fundamental frequency and odd harmonic modulation components thereof, and only the frequency component less than half of the sampling frequency F can be accurately collected according to the Shannon sampling theorem, the decomposition number K is set according to the following formula:
Figure BDA0002462289180000031
in the formula fsIs the fundamental frequency of the current; f is the sampling frequency; k is the number of decompositions.
In the step 3, the CSVGSO algorithm proposes a variable step length moving strategy according to the firefly luciferin value and the neighborhood range to replace the fixed step length in the GSO algorithm, and the specific moving strategy is as follows:
Figure BDA0002462289180000032
in the formula, si(T) the ith firefly at the TthStep size of movement in iteration, smax、sminMaximum and minimum moving step, l, respectivelymax(T)、rmin(T) is respectively the maximum value of all firefly luciferin and the minimum value of the neighborhood decision range in the Tth iteration, |i(T)、
Figure BDA0002462289180000033
The distribution is the fluorescein value and the neighborhood range of the ith firefly in the Tth iteration, eta is a proportional adjustment coefficient, and eta is more than 0 and less than 1 to ensure that the moving step length of the firefly is in the range of smin,smax]In the meantime.
And 3, initializing the initial position of the firefly in a feasible solution space of the penalty parameter alpha according to the chaotic sequence by the CSVGSO algorithm to increase the ergodicity, regularity and diversity of the population.
The fitness function determining method in the step 4 comprises the following steps: the VMD theoretical decomposition result is K modal components with center frequencies from low to high, and the current signal is in a superposition form of fundamental frequency and odd harmonic modulation components thereof, so that the center frequency interval of adjacent modal components is theoretically 2-fold frequency of the fundamental frequency, and the interval of the center frequencies of the adjacent modal components is set as fi(i ═ 1, 2.., K-2), and defines a fitness function:
Figure BDA0002462289180000034
the step 5 of optimizing the VMD decomposition penalty parameter alpha by using the CSVGSO algorithm comprises the following specific steps:
step 5.1: initializing relevant parameters of a firefly algorithm;
step 5.2: initializing a firefly initial position in a feasible solution space of a penalty parameter alpha based on the chaotic sequence;
step 5.3: setting maximum and minimum moving step length smax、sminProportional adjustment coefficient eta and maximum number of iterations titerAnd making the current iteration time T equal to 1;
step 5.4: updating the fitness value of each firefly;
step 5.5: updating the moving step length of each firefly according to the variable step length moving strategy;
step 5.6: updating the position and neighborhood range of each firefly;
step 5.7: t is T +1 and T is judged to be less than or equal to TiterIf so, return to step 5.4, otherwise end the loop.
And 6, after VMD decomposition is carried out on the stator three-phase current, K single-component modal signals with the center frequency from low to high are obtained, and the fundamental frequency component corresponds to the first modal component.
The specific determination method of the fundamental frequency component in step 7 is as follows: for the extracted fundamental frequency modal component i of the three-phase current of the statoru、iv、iwCarrying out Park conversion to obtain two-phase current id、iqAnd i isqIs idBased on idAnd iqComponent structure analysis signal iz：
iz＝id+j·iq (4)
For analytic signal izBy finding the amplitude angle, the instantaneous phase information of the reference signal can be determined
Figure BDA0002462289180000041
t represents a time series.
The specific method for performing adaptive filtering by using the LMS algorithm in step 8 is as follows: the filtering process for the fundamental frequency component can be regarded as a single-component interference adaptive filtering process, and thus the following set of orthogonal signals is constructed:
Figure BDA0002462289180000051
where a is the initial amplitude of the set of quadrature signals,
initial tap coefficients and learning step sizes are then set and adaptive filtering is performed based on continuous iteration of the LMS algorithm (Least Mean Square).
The invention has the beneficial effects that:
the CSVGSO algorithm of the invention provides a variable step length moving strategy to replace the fixed step length in the basic firefly algorithm to solve the problems of slow later convergence and low optimization efficiency of the GSO algorithm; the penalty parameter alpha of VMD decomposition can be effectively optimized according to the CSVGSO algorithm, and the decomposition number K is set according to the characteristics of the motor current signal, so that the problem of VMD decomposition parameter selection can be effectively solved; the multi-component current signal can be effectively decomposed into K single-component signals based on VMD decomposition, and the comprehensive phase information of three-phase current can be utilized to calculate the phase of the current signal by combining with Park conversion, so that the problem of frequency estimation error of traditional LMS single-component adaptive filtering is effectively avoided, and the filtering efficiency is further improved. The method can effectively solve the problem that the fault characteristic frequency component is weak and is easily covered by the fundamental frequency component and difficult to identify in the initial stage of the broken bar fault of the rotor of the alternating current asynchronous motor, and further strengthen the application of the three-phase current information of the stator in the fault diagnosis of the motor.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
FIG. 2 is a flow chart of the CSVGSO algorithm of the present invention.
FIG. 3(a) shows the three-phase current i of the stator of the AC asynchronous motor with rotor broken bar fault synchronously collected according to the embodimentU、iV、iWA time domain diagram, (b) is iUAnd (4) a frequency domain graph.
Fig. 4 shows the VMD decomposition results of the embodiment, wherein (a) is a time domain diagram of each modal component, and (b) is a frequency domain diagram of each modal component.
FIG. 5 shows an embodiment of the reference signal i obtained by LMS adaptive filteringrefAnd iUThe contrast map of the fundamental frequency modal components of the signal, wherein (a) is a time domain contrast map, and (b) is a frequency domain contrast map.
FIG. 6 shows an embodiment of the filtered signal iresAnd original iUSignal contrast map, wherein (a) is time domain contrast map, and (b) is iresThe signal spectrum.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples.
Referring to fig. 1, the method for diagnosing the broken bar fault of the rotor of the alternating current asynchronous motor based on the three-phase current comprises the following steps:
step 1: synchronous acquisition of three-phase current signals i of alternating current asynchronous motor statorU、iV、iW；
Step 2: setting the VMD decomposition number K according to the characteristics of the current signal;
and step 3: changing the fixed moving step length of the firefly in the GSO algorithm, initializing the initial position of the firefly based on the chaotic sequence, and providing a CSVGSO algorithm;
and 4, step 4: determining a fitness function;
and 5: optimizing a VMD decomposition penalty parameter alpha by using the CSVGSO algorithm;
step 5.1: initializing relevant parameters of a firefly algorithm;
step 5.2: initializing a firefly initial position in a feasible solution space of a penalty parameter alpha based on the chaotic sequence;
step 5.3: setting the maximum moving step smaxAnd a minimum moving step sminProportional adjustment coefficient eta and maximum number of iterations titerAnd making the current iteration time T equal to 1;
step 5.4: updating the fitness value of each firefly;
step 5.5: updating the moving step length of each firefly according to the variable step length moving strategy;
step 5.6: updating the position and neighborhood range of each firefly;
step 5.7: t is T +1 and T is judged to be less than or equal to TiterIf yes, returning to the step 5.4, otherwise, ending the circulation;
step 6: performing VMD decomposition on the stator three-phase current based on the punishment parameter alpha corresponding to the global optimal fitness value and the decomposition number K determined in the step 2, and extracting modal components corresponding to the fundamental frequency;
and 7: determining a reference signal phase based on the Park transformation;
and 8: substituting the phase information obtained in the step 7 into an LMS algorithm, and carrying out self-adaptive filtering on the fundamental frequency component in the extracted modal signal;
and step 9: and carrying out spectrum analysis on the filtered signals to detect the characteristic frequency of the rotor broken bar fault.
The invention comprises three processes: CSVGSO algorithm, VMD decomposition and LMS single component adaptive filtering.
Referring to fig. 2, the CSVGSO algorithm specifically includes the steps of:
(1) initialization: randomly generating n fireflies with the same fluorescein value and neighborhood range in a feasible solution space; setting the sensing radius rsFluorescein growth factor gamma, fluorescein attenuation factor rho, neighborhood decision range rdFixed moving step length s of firefly, gain constant beta and expected neighbor number nt；
(2) And (3) updating fluorescein: updating the fluorescein value of the firefly according to the change of the position of the firefly along with the iteration times according to the following formula:
li(T+1)＝(1-ρ)li(T)+γJ(xi(T+1))
in the formula Ii(T) represents the fluorescein value, x, of the ith firefly at the Tth iterationi(T +1) represents the location of the ith firefly at iteration T +1, and J represents the objective function value of the firefly at that location;
(3) and (3) determining the neighbor: firefly searches for individuals with a higher fluorescein value than itself in its neighborhood as neighbors according to the following formula:
Figure BDA0002462289180000071
in the formula, Ni(T) and
Figure BDA0002462289180000072
respectively the neighbor number and the neighborhood decision range, dis, of the ith firefly in the Tth iterationij(T) is the Euclidean distance of the ith and jth fireflies at the Tth iteration;
(4) the firefly moves: each firefly moves to neighbor j as follows:
Figure BDA0002462289180000081
in the formula, pij(T) is the probability that the ith firefly moved to neighbor j at the tth iteration;
(5) firefly location tracking: the firefly location is updated as follows:
Figure BDA0002462289180000082
(6) neighborhood update: the neighborhood range of the firefly is updated as follows:
Figure BDA0002462289180000083
(7) and (3) judging the termination of the algorithm: judging T is less than or equal to TiterIf yes, the method returns to the step (2), otherwise, the algorithm is terminated, and the loop is exited.
The CSVGSO algorithm initializes the firefly position according to the chaotic sequence to increase the ergodicity, regularity and diversity of the population, wherein the chaotic sequence is generated according to the following Logistic mapping:
zn+1＝θzn(1-zn)n＝0,1,2...
wherein theta is a control variable, and when theta is 4,0 ≦ z0When the sequence is less than or equal to 1, the sequence is in a complete chaotic state.
The optimization efficiency of the algorithm is increased by replacing the fixed moving step length of the firefly with the variable step length according to the following rule, the problems of slow convergence and poor optimization of the later period of the algorithm are improved,
Figure BDA0002462289180000084
in the formula, si(T) is the moving step length of ith firefly at the Tth iteration, smax、sminMaximum and minimum moving step, l, respectivelymax(t)、rmin(t) all firefly luciferin maxima and neighborhoods at the t-th iteration, respectivelyMinimum value of decision range,/i(T)、
Figure BDA0002462289180000085
The distribution is the fluorescein value and the neighborhood range of the ith firefly in the Tth iteration, eta is a proportional adjustment coefficient, and eta is more than 0 and less than 1 to ensure that the moving step length of the firefly is in the range of smin,smax]To (c) to (d);
the rule automatically adjusts the moving step length according to the fluorescein value and the neighborhood range of the firefly, so that individuals with lower fluorescein values and larger neighborhood ranges have relatively large moving step lengths to improve the global optimizing capability, and individuals with higher fluorescein values and smaller neighborhood ranges have relatively small moving step lengths to enhance the local searching capability.
The VMD decomposition comprises the following specific steps:
(1) for each intrinsic mode function ukAnd obtaining an analytic signal of the signal by using Hilbert transform:
Figure BDA0002462289180000091
(2) adding an exponential term into the analytic signal to adjust the respective estimated center frequency, so as to modulate each modal spectrum to a corresponding baseband:
Figure BDA0002462289180000092
(3) using the above demodulated signal H1To estimate the center frequency omega of each modekThereby obtaining a constrained variation problem:
Figure BDA0002462289180000093
Figure BDA0002462289180000094
wherein u is the original signal;
(4) the problem is converted into an unconstrained variational problem by introducing a secondary penalty factor and Lagrange multipliers:
Figure BDA0002462289180000095
in the formula, alpha is a penalty factor, and lambda (t) is a Lagrange multiplier;
therefore, the complex modulation signals can be subjected to self-adaptive decomposition by continuously updating the center frequency and the bandwidth, and K single-component modulation signals are obtained.
Because the stator current signal of the alternating current asynchronous motor is represented as a superposition form of a current fundamental frequency and an odd harmonic component modulation signal thereof, and only a frequency component less than half of a sampling frequency F can be accurately acquired according to Shannon's sampling theorem, a decomposition number K is set according to the following formula:
Figure BDA0002462289180000101
when the VMD decomposition penalty parameter alpha is optimized based on the CSVGSO algorithm, the fitness function determining method comprises the following steps: the result of the VMD theory decomposition is K modal components with the center frequency from low to high, and the current signal is expressed in a superposition form of fundamental frequency and odd harmonic modulation components thereof. Therefore, the central frequency interval of the adjacent modal components is theoretically 2 times of the fundamental frequency; setting the interval of center frequencies of adjacent modal components as fi(i ═ 1, 2.., K-2), and defines a fitness function:
Figure BDA0002462289180000102
the LMS single-component adaptive filtering comprises the following steps:
xc(t) and xs(t) is a set of mutually orthogonal reference signals, and the specific iterative process of the LMS algorithm is represented as follows:
Figure BDA0002462289180000103
where y (t) is the LMS algorithm output reference signal, ∈ (t) is the filter output signal, d (t) is the input signal, [ mu ] is the learning step length, ω isc(t) and ωs(t) is the current tap coefficient;
the self-adaptive filtering comprises the following specific steps:
(1) respectively extracting single-component modal signals i corresponding to fundamental frequency components after VMD decomposition of stator three-phase currentu、iv、iw；
(2) To iu、iv、iwCarrying out Park conversion to obtain two-phase current id、iq(ii) a Wherein the Park transformation formula is:
Figure BDA0002462289180000111
(3) based on idAnd iqComponent structure analysis signal iz：
iz＝id+j·iq
For analytic signal izAmplitude and angle determination of instantaneous phase information of reference signal
Figure BDA0002462289180000112
Figure BDA0002462289180000113
(4) The following set of orthogonal signals is constructed:
Figure BDA0002462289180000114
(5) then setting initial tap coefficients and learning step sizes and carrying out adaptive filtering based on continuous iteration of the LMS algorithm.
The present invention will be described in further detail with reference to specific examples.
In the embodiment, in order to verify the accuracy and robustness of the method, a squirrel-cage three-phase alternating current asynchronous fault motor with a built-in rotor broken bar is adopted for experimental verification. Selecting a current sensor with the sensitivity of 100mv/A and selecting an 8-channel data acquisition card; setting a fundamental frequency f of a motor supply currentsThe motor is operated at a steady state under a rated load through an electromagnetic clutch at 26 Hz; setting the sampling frequency to 400Hz, the data acquisition length to 4000, and setting the synchronously acquired three-phase current i of the rotor broken bar fault alternating current asynchronous motor stator in the graph of FIG. 3(a)U、iV、iWA time domain diagram, (b) is iUAnd (4) a frequency domain graph.
The fundamental frequency of the current and the 3, 5 and 7 harmonic components thereof are determined according to the VMD decomposition number K determination rule in the invention, K is set to 5, and then
Figure BDA0002462289180000115
Performing 10 sub-optimization on the VMD decomposition penalty parameter alpha for a fitness function based on a GSO algorithm and the CSVGSO provided by the invention, wherein the firefly position, namely the optimization parameter VMD penalty parameter alpha is in [100,3000 ]]And (4) internal random distribution, wherein the GSO algorithm and the CSVGSO algorithm initialization parameter settings are shown in Table 1. Table 2 is a comparison table of the number of iterations performed to obtain the optimal penalty parameter α for these two algorithms. It can be seen that the CSVGSO algorithm requires a significantly smaller number of iterations. Table 3 shows the optimal penalty parameter α and the corresponding fitness value obtained by performing 10 sub-optimizations on the VMD decomposition penalty parameter α based on the CSVGSO algorithm proposed by the present invention.
TABLE 1 GSO Algorithm and CSVGSO Algorithm parameter selection Table
Figure BDA0002462289180000121
TABLE 2 GSO Algorithm and CSVGSO Algorithm iteration number comparison table
Figure BDA0002462289180000122
TABLE 3 CSVGSO Algorithm optimization result Table
Figure BDA0002462289180000123
It can be seen that when the penalty parameter α is 882, the optimal fitness value is minimum, so the number of decompositions K is set to 5, and the penalty parameter α is 882 to perform VMD decomposition on the stator three-phase current. Fig. 4 is the decomposition result, and it can be seen that the VMD decomposes the current signal into 5 single-component modulation signals from low frequency to high frequency. The modal components 1 to 4 correspond to single-component modulation signals of fundamental frequency, 3-order harmonic, 5-order harmonic and 7-order harmonic in the original current signal, respectively, and the modal component 5 is a residual high-frequency component.
Extracting single-component modal signal i corresponding to fundamental frequency component in three-phase currentu、iv、iwCarrying out Park conversion to calculate phase information of fundamental frequency component, substituting the phase information into LMS algorithm to construct reference signal and initialize tap coefficient [0.2,0.2]Learning step length 0.001, reference signal initial amplitude A being 1, carrying out adaptive filtering, and finally obtaining reference signal iref＝ωcxc(t)+ωsxs(t) of (d). FIG. 5 shows irefAnd iuComparing the figures, it can be seen that the reference signal can be effectively constructed and the amplitude and phase of the reference signal can be accurately determined by the present invention. Obtaining a residual signal i after adaptive filteringresTo i, pairresPerforming frequency spectrum analysis as shown in fig. 6, it can be seen that the fundamental frequency component has been completely suppressed, so that the characteristic frequency f of the rotor broken bar fault can be clearly detecteds±kfbr(k＝1,2,3...,fbr＝2sfsAnd s is slip).
The invention improves the basic firefly algorithm and provides the CSVGSO algorithm, and solves the problems of slow later convergence and low optimization efficiency of the firefly algorithm. The penalty parameter alpha of VMD decomposition can be effectively optimized according to the CSVGSO algorithm, and the decomposition number K is set according to the motor current signal characteristics, so that the problem of VMD decomposition parameter selection can be effectively solved. The reference signal of the fundamental frequency component can be accurately constructed based on Park transformation and combined with an LMS single-component adaptive filtering algorithm, and effective filtering is carried out on the reference signal. The method can effectively solve the problem that the fault characteristic frequency component is weak and is easily covered by the fundamental frequency component and difficult to identify in the initial stage of the broken bar fault of the rotor of the alternating current asynchronous motor, and further strengthen the application of the three-phase current information of the stator in the fault diagnosis of the motor.

Claims (5)

1. A three-phase current-based AC asynchronous motor rotor broken bar fault diagnosis method is characterized by comprising the following steps:
step 1: synchronous acquisition of three-phase current signals i of alternating current asynchronous motor statorU、iV、iW；
Step 2: setting the VMD decomposition number K according to the characteristics of the current signal;
and step 3: changing the fixed moving Step size of the firefly in a basic firefly algorithm GSO (Glowworm Swarm Optimization), initializing the initial position of the firefly based on a chaotic sequence, and providing a Chaos Variable Step size firefly improved algorithm CSVGSO (Chaos Variable Step-size Glowworm Optimization);
the CSVGSO algorithm initializes the firefly position according to the chaotic sequence to increase the ergodicity, regularity and diversity of the population, wherein the chaotic sequence is generated according to the following Logistic mapping:
zn+1＝θzn(1-zn) n＝0,1,2...
wherein theta is a control variable, and when theta is 4,0 ≦ z0When the sequence is less than or equal to 1, the sequence is in a complete chaotic state;
the optimization efficiency of the algorithm is increased by replacing the fixed moving step of the firefly with a variable step according to the following rule,
Figure FDA0002997015990000011
in the formula, si(T) is the moving step length of ith firefly at the Tth iteration, smax、sminMaximum and minimum moving step, l, respectivelymax(t)、rmin(t) the maximum values of all firefly luciferin and the minimum value of the neighborhood decision range in the t-th iteration, li(T)、
Figure FDA0002997015990000012
The distribution is the fluorescein value and the neighborhood range of the ith firefly in the Tth iteration, eta is a proportional adjustment coefficient, and eta is more than 0 and less than 1 to ensure that the moving step length of the firefly is in the range of smin,smax]To (c) to (d);
the rule automatically adjusts the moving step length according to the fluorescein value and the neighborhood range of the firefly;
and 4, step 4: determining a fitness function;
the fitness function determining method comprises the following steps: the VMD theoretical decomposition result is K modal components with the center frequency from low to high, and the current signal is in a superposition form of fundamental frequency and odd harmonic modulation components thereof, so that the center frequency interval of the adjacent modal components is 2 frequency multiplication of the fundamental frequency theoretically; therefore, after the VMD decomposition number K is determined, the interval of the center frequencies of the adjacent modal components is set as fi1,2, K-2, and defines a fitness function:
Figure FDA0002997015990000021
and 5: optimizing a VMD decomposition penalty parameter alpha by using the CSVGSO algorithm; the method comprises the following specific steps:
step 5.1: initializing relevant parameters of a firefly algorithm;
step 5.2: initializing the initial position of the firefly in a feasible solution space [100-3000] of a penalty parameter alpha based on the chaotic sequence;
step 5.3: setting maximum and minimum moving step length smax、sminProportional adjustment coefficient eta and maximum number of iterations titerAnd making the current iteration time T equal to 1;
step 5.4: updating the fitness value of each firefly;
step 5.5: updating the moving step length of each firefly according to the variable step length moving strategy;
step 5.6: updating the position and neighborhood range of each firefly;
step 5.7: t is T +1 and T is judged to be less than or equal to TiterIf yes, returning to the step 5.4, otherwise, ending the circulation;
step 6: performing VMD decomposition on the stator three-phase current based on the punishment parameter alpha corresponding to the global optimal fitness value and the decomposition number K determined in the step 2, and extracting modal components corresponding to the fundamental frequency;
and 7: determining a reference signal phase based on the Park transformation;
and 8: substituting the phase information obtained in the step 7 into an LMS algorithm, and carrying out self-adaptive filtering on the fundamental frequency component in the extracted modal signal;
and step 9: and carrying out spectrum analysis on the filtered signals to detect the characteristic frequency of the rotor broken bar fault.
2. The method for diagnosing the rotor broken bar fault of the alternating current asynchronous motor based on the three-phase current as claimed in claim 1, wherein the VMD decomposition number K is set according to the characteristics of the current signal in the step 2, and the specific method comprises the following steps:
because the stator current signal of the alternating current asynchronous motor is expressed in a superposition mode of fundamental frequency and odd harmonic modulation components thereof, and only the frequency component less than half of the sampling frequency F can be accurately collected according to the Shannon sampling theorem, the decomposition number K is set according to the following formula:
Figure FDA0002997015990000031
in the formula fsIs the fundamental frequency of the current; f is the sampling frequency; k is the number of decompositions.
3. The method for diagnosing the rotor bar breakage fault of the alternating current asynchronous motor based on the three-phase current as claimed in claim 1, wherein the step 6 is used for performing VMD decomposition on the three-phase current of the stator to obtain K single-component modal signals with the center frequency from low to high, and the fundamental frequency component corresponds to the first modal component.
4. The method for diagnosing the broken bar fault of the rotor of the alternating current asynchronous motor based on the three-phase current as claimed in claim 1, wherein the specific determination method of the fundamental frequency component in the step 7 is as follows: for the extracted fundamental frequency modal component i of the three-phase current of the statoru、iv、iwCarrying out Park conversion to obtain two-phase current id、iqAnd i isqIs idBased on idAnd iqComponent structure analysis signal iz：
iz＝id+j·iq
For analytic signal izBy finding the amplitude angle, the instantaneous phase information of the reference signal can be determined
Figure FDA0002997015990000032
t represents time.
5. The method for diagnosing the rotor broken bar fault of the alternating current asynchronous motor based on the three-phase current according to claim 1, wherein a specific method for performing adaptive filtering by using an LMS algorithm in the step 8 is as follows: the filtering process for the fundamental frequency component is considered as a single-component interference adaptive filtering process, and therefore the following set of orthogonal signals is constructed:
Figure FDA0002997015990000041
where a is the initial amplitude of the set of quadrature signals,
initial tap coefficients and learning step sizes are then set and adaptive filtering is performed based on continuous iteration of the LMS algorithm (Least Mean Square).
CN202010323423.2A 2020-04-22 2020-04-22 Three-phase current-based alternating current asynchronous motor rotor broken bar fault diagnosis method Active CN111458640B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010323423.2A CN111458640B (en) 2020-04-22 2020-04-22 Three-phase current-based alternating current asynchronous motor rotor broken bar fault diagnosis method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010323423.2A CN111458640B (en) 2020-04-22 2020-04-22 Three-phase current-based alternating current asynchronous motor rotor broken bar fault diagnosis method

Publications (2)

Publication Number Publication Date
CN111458640A CN111458640A (en) 2020-07-28
CN111458640B true CN111458640B (en) 2021-05-28

Family

ID=71686112

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010323423.2A Active CN111458640B (en) 2020-04-22 2020-04-22 Three-phase current-based alternating current asynchronous motor rotor broken bar fault diagnosis method

Country Status (1)

Country Link
CN (1) CN111458640B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113364362B (en) * 2021-06-30 2023-03-24 青岛恒星智能装备有限公司 Firefly optimization-based induction motor rotation speed estimation method
CN113866631B (en) * 2021-09-02 2022-06-10 浙江大学 Cage type asynchronous motor rotor broken bar fault diagnosis method
CN117349735B (en) * 2023-12-05 2024-03-26 国家电投集团云南国际电力投资有限公司 Fault detection method, device and equipment for direct-current micro-grid and storage medium

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
AU2007275451A1 (en) * 2006-07-19 2008-01-24 Edsa Micro Corporation Systems and methods for real-time advanced visualization for predicting the health, reliability and performance of an electrical power system
CN103063419A (en) * 2012-10-30 2013-04-24 西安交通大学 Quality testing device for motorized spindle rotor
CN105277362A (en) * 2015-11-23 2016-01-27 西安交通大学 Gear fault detection method on the basis of multi-position turning angle signals of encoders
CN106651464A (en) * 2016-12-30 2017-05-10 中国地质大学（武汉） Agricultural product price prediction method and apparatus
CN107192951A (en) * 2017-07-31 2017-09-22 湖南科技大学 A kind of micro- method for diagnosing faults of wind-driven generator three-phase rotor current
CN108197648A (en) * 2017-12-28 2018-06-22 华中科技大学 A kind of Fault Diagnosis Method of Hydro-generating Unit and system based on LSTM deep learning models
CN109444740A (en) * 2018-11-14 2019-03-08 湖南大学 A kind of the malfunction intellectual monitoring and diagnostic method of Wind turbines
CN110046562A (en) * 2019-04-01 2019-07-23 湖南大学 A kind of wind power system health monitor method and device
CN110595765A (en) * 2019-08-26 2019-12-20 西安理工大学 Wind turbine generator gearbox fault diagnosis method based on VMD and FA _ PNN

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101025430A (en) * 2007-03-28 2007-08-29 华北电力大学 Cage type asynchronous motor rotor strip-broken failure detecting method
CN108022025B (en) * 2017-12-28 2020-08-18 华中科技大学 Wind speed interval prediction method and system based on artificial neural network

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
AU2007275451A1 (en) * 2006-07-19 2008-01-24 Edsa Micro Corporation Systems and methods for real-time advanced visualization for predicting the health, reliability and performance of an electrical power system
CN103063419A (en) * 2012-10-30 2013-04-24 西安交通大学 Quality testing device for motorized spindle rotor
CN105277362A (en) * 2015-11-23 2016-01-27 西安交通大学 Gear fault detection method on the basis of multi-position turning angle signals of encoders
CN106651464A (en) * 2016-12-30 2017-05-10 中国地质大学（武汉） Agricultural product price prediction method and apparatus
CN107192951A (en) * 2017-07-31 2017-09-22 湖南科技大学 A kind of micro- method for diagnosing faults of wind-driven generator three-phase rotor current
CN108197648A (en) * 2017-12-28 2018-06-22 华中科技大学 A kind of Fault Diagnosis Method of Hydro-generating Unit and system based on LSTM deep learning models
CN109444740A (en) * 2018-11-14 2019-03-08 湖南大学 A kind of the malfunction intellectual monitoring and diagnostic method of Wind turbines
CN110046562A (en) * 2019-04-01 2019-07-23 湖南大学 A kind of wind power system health monitor method and device
CN110595765A (en) * 2019-08-26 2019-12-20 西安理工大学 Wind turbine generator gearbox fault diagnosis method based on VMD and FA _ PNN

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
A Broken Rotor Bar Fault Diagnosis Approach Based on Singular Value Decomposition and Variational Mode Decomposition;Dan Zou et.al;《2019 IEEE Transportation Electrification Conference and Expo, Asia-Pacific》;20191121;第1-6页 *
基于改进萤火虫算法的永磁同步电机PID参数优化研究;胡国祥 等;《变频器世界》;20190630;第81-85页 *
萤火虫—粒子群优化神经网络的异步电机转子断条故障诊断;乔维德;《电机与控制应用》;20171231;第44卷(第1期);第83-88页 *

Also Published As

Publication number Publication date
CN111458640A (en) 2020-07-28

Similar Documents

Publication Publication Date Title
CN111458640B (en) Three-phase current-based alternating current asynchronous motor rotor broken bar fault diagnosis method
CN110850482A (en) Transient electromagnetic signal-noise separation method based on variational modal decomposition principle
CN107192878A (en) A kind of trend of harmonic detection method of power and device based on compressed sensing
CN110138698B (en) Initial phase joint estimation method and device for frequency offset of high-order modulation linear mixed signal
CN110690931B (en) Digital signal adaptive code rate estimation method and device based on multi-wavelet-base combination
CN108414833B (en) Accurate estimation method for signal component frequency
CN110726875B (en) New energy flexible direct-current grid-connected transient harmonic detection method and system
CN110598170A (en) Data prediction method based on FEEMD decomposition time sequence
CN108732440A (en) A kind of transient power quality detection method and system
CN108181617B (en) Filtering method of non-linear frequency modulation system based on tensor product model transformation
CN111082541B (en) Control method and device of mobile wireless power transmission system
CN112751633B (en) Broadband spectrum detection method based on multi-scale window sliding
CN111046485B (en) Method for estimating unmanned ship rolling motion response function by recursive least square method
CN106101039A (en) A kind of based on the data auxiliary adjustable frequency deviation estimating method of precision
CN100574130C (en) The detection method and the device of TD-SCDMA system same frequency sub-district intermediate code
CN115986965A (en) Load and mutual inductance synchronous identification method for multi-frequency multi-load wireless power transmission system
CN111563631A (en) Wind power generation power prediction method and device based on variational modal decomposition
Zheng et al. Time-frequency feature-based underwater target detection with deep neural network in shallow sea
CN113640842B (en) Direct-spread signal capturing method based on intelligent Doppler search
CN104184530B (en) A kind of detection method of uplink interference and device
CN116559584A (en) Self-adaptive voltage sag waveform decomposition method and system
CN104202097B (en) A kind of high-velocity scanning method for digital monitoring receiver
CN116559579A (en) Improved VMD and Teager energy operator fault positioning method
CN107181708B (en) Frequency estimating methods and location receiver
CN115343570A (en) Power grid impedance online identification method and device based on variable forgetting factor recursive least square algorithm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant